Enhanced Hosting Capacity Analysis - MN Solar...

Enhanced Hosting Capacity AnalysisFINAL REPORT OCTOBER 15, 2018

PREPARED FOR MINNESOTA DEPARTMENT OF COMMERCE AND THE MINNESOTA SOLAR PATHWAYS PROJECT

PREPARED BY SMARTER GRID SOLUTIONS (SGS) NEW YORK, NY

2018: Smarter Grid Solutions Inc.

MN Solar Pathways: Enhanced Hosting Capacity Analysis | October 15, 2018 | Page 1

Document name: MN Solar Pathways: Enhanced Hosting Capacity Analysis Document number: 200491 03E

Version Issue Date Authors Reviewed by

Approved by

Description DUR Number

A 03/31/2018

Martin Paul

Rachael Taljaard

Paige Medley

Chad Abbey Initial issue 2950

B 04/14/2018

Martin Paul

Rachael Taljaard

Paige Medley

Rachael Taljaard

Paige Medley

Chad Abbey

Updated following customer comment

2972

C 07/19/2018

Paige Medley Rachael Taljaard

Paige Medley

Chad Abbey Final Version 3159

D 08/02/2018 Paige Medley

Rachael Taljaard

Chad Abbey

Final Version – Utility Comments

Incorporated 3189

E 08/31/2018 Paige Medley

Chad Abbey

Chad Abbey

Final Version – MN COMM Comments

Incorporated

3232

Contact: Name: Paige Medley Job Title: Consultant Email: [email protected] Smarter Grid Solutions Inc. 335 Madison Avenue, 4th Floor New York, NY 10017


1 Acknowledgements This project was made possible by a grant from the U.S. Department of Energy Solar Energy Technologies Office and the Minnesota Department of Commerce. Stacy Miller, Project Manager (651) 539-1859 [email protected] Smarter Grid Solutions (SGS) would like to acknowledge the contributions of Shyamal Patel and Daniel Cohen for their work on this PV Hosting Capacity Report. Additionally, SGS would like to thank Morgan Putnam from Clean Power research for his input and guidance throughout the project; Kelsey Horowitz from NREL for providing the NREL cost databases and Smart Inverter cost analysis; Patrick Dalton, Edmund Shannon, and John Harlander from Xcel Energy; Nathan Jensen, Brian Scheer, and Jeff McKeever from Otter Tail Power (OTP); Josh Quinnell from Center for Energy and Environment; and the entire Technical Committee guiding the Minnesota Solar Pathways project for the opportunity to participate in such an exciting study.

mailto:[email protected]


2 Disclaimer

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.


3 Introduction To better understand how technology can increase the grid’s ability to integrate solar energy from photovoltaics (PV) in alignment with Minnesota’s goal of producing 10% of its electricity from solar by 2030, the Minnesota Department of Commerce contracted with Smarter Grid Solutions (SGS) as part of the Solar Pathways project. SGS investigated how advanced technologies can be used to increase PV hosting capacity without the need for traditional network reinforcements. Two Minnesota utilities, Xcel Energy (Xcel) and Otter Tail Power (OTP), provided data for four feeders for use as test cases in both rural and urban networks. SGS first conducted a static hosting capacity analysis to determine the existing available hosting capacity for PV, using typical utility network interconnection screens. SGS then conducted an enhanced hosting capacity analysis on the feeders, investigating the following advanced technology strategies to determine their impact on available hosting capacity:

• Smart inverters (distributed and centrally-coordinated control); • Energy storage; • Thermal load shifting; • Curtailment; • Combined control of the above four technologies.

This report outlines the background information of the Minnesota context, the data received for use in the analysis, the analysis methodology, results of the analysis, a cost analysis, and conclusions that can be drawn from the project.

3.1 Hosting Capacity Analysis

Distributed energy resource (DER) hosting capacity is defined by EPRI as “the amount of DER that can be accommodated without adversely impacting quality or reliability under existing control configurations and without requiring infrastructure upgrades”1. Hosting capacity can be affected by a variety of factors, some of which are listed below:

• Configuration of the distribution system: o Method of voltage control; o Topology; o Load type and location; o Circuit phasing; o Local weather patterns.

• DER characteristics: o Type (technology); o Control functions (real/reactive power, voltage etc.);

1 EPRI, “Impact Factors, Methods, and Considerations for Calculating and Applying Hosting Capacity”, 2018


o Manufacturer and design; o Single of multiple sites/types; o Efficiency.

The algorithms that have been developed to calculate hosting capacity are evolving, the four most prominent in use today are: stochastic, streamlined, iterative, and hybrid/DRIVE (used by Xcel Energy). Choosing between these algorithms is not straightforward with the main consideration being the intended use of the results. The different methods can provide similar answers but this is an evolving area so any comparison between the results is premature, particularly as convergence to a standard approach is likely as each adopts best practices from the others leading to lessen distinction between them.

Whichever method is selected, the results are highly dependent on the input data and assumptions; therefore, care should be taken to make them as realistic as possible. The hosting capacity gives an indication of the potential for DER on a feeder to help with system planning and to inform the public, guiding DER development towards the most suitable locations. The results are highly dependent on the accuracy of models and assumptions.

A hosting capacity calculation gives a MW value that implies a high level of precision which may be misleading; this value could vary significantly with alterations in the inputs, assumptions or model parameters. When analyzing the results of a hosting capacity calculation they should only be used as an indication of whether the feeder is congested or not, with little meaning attached to the absolute value.

3.2 Solar Development in Minnesota

Solar PV developments vary in size from residential scale systems of a few kilowatts to commercial, community, and utility-scale sites of multiple megawatts. One of the most prominent types of solar PV in Minnesota are centrally located PV sites of 1 to 5 MW providing electricity to local subscribers, known as community solar gardens, which have driven much of the state’s growth in solar capacity. They represent approximately 53% of the 467 MW of solar capacity installed during 20172, with most of the rest from utility-scale (44%), and just 3% from rooftop-scale systems. This shows a clear trend towards larger centralized PV sites. This project analyzed the hosting capacity for a single large development on each feeder, representing a community solar garden or utility-scale development. The potential for many distributed sites are not considered in this report. This also aligns with Xcel’s 2017 hosting capacity analysis which used the centralized, single-site scenario rather than the distributed, multi-site scenario. Xcel justified this decision on the basis that it more accurately reflects the characteristics of DER deployment programs in Minnesota.

2 Minnesota achieves 714 MW of solar capacity in 2017, nearly triples cumulative capacity from 2016. http://mn.gov/commerce-stat/pdfs/MN714MW.pdf

http://mn.gov/commerce-stat/pdfs/MN714MW.pdf


3.3 Smart Inverters

Smart inverters offer a variety of functions that can be used to enhance network operation by regulating the real and reactive power output from solar PV and other DERs. To understand how smart inverters could increase PV hosting capacity, we must first understand the characteristics of a smart inverter. EPRI has provided a thorough, well defined set of standard smart inverter functions covering a range of requirements from basic event logging and status monitoring to autonomous grid support of voltage and frequency3. These functions are given in Figure 1.

Monitoring and Scheduling

Frequency Support Real Power Support Power Factor Support Voltage Support

Basic Device Settings and Limits

Frequency-watt Function

Limit DER Power Output Function

Fixed Power Factor Function

Dynamic Volt-watt Function

Connect/Disconnect Function

Low/High Frequency Ride-Through Requirements

Dynamic Real-Power Support

Volt-Var Function Dynamic Reactive Current Support Function

DER Settings to Manage Multiple Grid Configurations (Including Islanding)

Peak Power Limiting Function

Watt-Power Factor Function

Volt-Watt Function

Status Monitoring Points

Load and Generation Following Function

Low/High Voltage Ride-Through Requirements

Event Logging and Reporting

Watt-Var Function

Time Adjustment Function

Battery Storage: Price-based Charge/Discharge Function

Battery Storage: Direct Charge/Discharge

Management Function Battery Storage:

Coordinated Charge/Discharge

Management Function *The function called “Price or Temperature Driven Functions” is not specific to any of the fields above.

Figure 1: EPRI smart inverter functions3

From this broad range of functions, the most potential for increasing PV hosting capacity is offered by real power and power factor support which can be used to mitigate voltage and thermal constraints that often limit the hosting capacity identified on distribution systems. The functions that are investigated in this project are given in Table 1 where they are split into two configurations: stand-alone where the smart inverter is only taking local measurements, and coordinated where the smart inverter is communicating with a wider control system to manage remote constraints.

3 EPRI, Common Functions for Smart Inverters: 4th Edition, https://www.epri.com/#/pages/product/000000003002008217/

https://www.epri.com/#/pages/product/000000003002008217/


Table 1: Project smart inverter functions

Configuration Function Constraint Managed Description Stand-Alone Volt-Watt Voltage Reduce real power when

measured voltage exceeds code

Volt-Var Voltage Provide reactive support based on the measured

voltage Co-ordinated Limit DER Power Output

Function Thermal (Remote to DER) Limit power output when

instructed based on external signal (Co-

ordinated by real-time control system)

3.4 Increasing Hosting Capacity & Real-Time Control

If a utility were to interconnect PV beyond a feeder’s hosting capacity there is the risk that network limits could affect supply security, power quality and the safety of utility staff and customers. The traditional approach is to upgrade overhead lines, cables and substations to accommodate the additional generation, however this is expensive and time consuming. An alternative approach is to use real-time control to monitor the network and take actions when approaching a limit to mitigate the situation, adjusting generation, load or storage for example. A real-time control system is the foundation on which multiple DERs such as storage, generation and load control can be integrated most efficiently into the network. Network security is protected with the option of taking preventative actions when required. (A generator that does not reduce its output might be disconnected, for example). A proven application of this type of real-time control system is to manage thermal constraints during periods of excess generation. For example, a circuit that is overloaded when generation is at full export during periods of low demand. In this situation, a control system can monitor the constrained circuit and reduce generation when it exceeds the limit, known as curtailing generation. From the generator’s perspective, it would like to generate energy but the network capacity is not available to export it. This energy is known as curtailed energy, or curtailment, is usually given in MWh, and represents wasted energy – energy that the PV could generate if network capacity were available. A real-time control system could also manage voltage constraints. Voltage measurements are taken on the network and the generation output, load, or storage is adjusted to keep the voltage within its limits. A utility could therefore interconnect PV on a feeder beyond its static hosting capacity by installing a real-time control system, but the PV site would be subject to periods where it cannot generate at its full potential. This effect is quantified as curtailment, shown conceptually in Figure 2. As PV increases beyond the static hosting capacity, the amount of energy curtailed increases.


Figure 2: Diagram showing increasing curtailment as installed PV is increased


4 Received Data Data has been provided for four feeders from OTP and Xcel. This section outlines the data received and any processing that has been undertaken prior to the hosting capacity analysis.

4.1 Demand Data Measured load data has been provided by OTP and Xcel with the details given in Table 2. The data uses differing timeframes due to availability of data for each of the individual substations but in all cases covers at least one year.

Table 2: Measured load data for each feeder

Feeder Type Time From Time To Feeder A Rural 09/12/2016 00:15 09/12/2017 10:45

Feeder B Mixed Suburban-Rural 10/01/2016 00:15 10/01/2017 00:00

Feeder C Rural 03/01/2015 00:00 05/01/2016 00:00

Feeder D Mixed Suburban-Rural 01/01/2017 00:00 12/31/2017 00:00

4.1.1 Demand Data Processing

The demand data contained periods of zero demand and missing data possibly due to outage events. These periods are summarized in Table 3. These are periods of abnormal operation that are not considered in the analysis and have been replaced with synthesized data entries.

Table 3: Periods of unsuitable data in load profiles

Feeder From To

Feeder A 12/26/2016 14:45 12/26/2016 15:45 01/26/2017 20:45 01/26/2017 21:00

Feeder B 05/11/2017 09:45 05/11/2017 14:30

Feeder C

27/04/2016 08:00 27/04/2016 23:00 28/04/2015 09:00 29/04/2016 20:00 30/04/2015 00:00 21/05/2015 12:00 07/06/2015 10:00 28/12/2015 14:00

Feeder D

03/01/2017 11:00 03/01/2017 13:30 03/20/2017 10:30 03/20/2017 11:00 03/27/2017 08:00 03/31/2017 14:00 04/10/2017 07:30 04/10/2017 15:30 05/04/2017 23:30 05/05/2017 23:00

The data was processed to remove abnormal events from the dataset so there was a continuous load profile for the entire period of analysis. If a zero load time-step appears in isolation then the average of the time-step before and after is used. In the event of a prolonged outage the preceding 24 hours are used to replace the zero values. Figure 3 shows an example of this process.


Figure 3: Diagram of load data cleansing

4.1.2 Solar Irradiation Data

Solar irradiance data was provided by CPR from its SolarAnywhere platform, with individual datasets for each feeder based on its geographical location. Annual irradiance data was provided in hourly time resolution for 2014, 2015, 2016, and 2017. The irradiance data was converted to a normalized profile for each year and then averaged across the four years to give a single normalized profile for each feeder. The resulting generation profiles are shown in Figure 4, Figure 5, Figure 6 and Figure 7. These profiles will be referred to as normalized PV profiles.


Figure 4: Normalized PV profile, Feeder A

Figure 5: Normalized PV profile, Feeder B


Figure 6: Normalized PV profile, Feeder C

Figure 7: Normalized PV profile, Feeder D

4.2 Feeder Models The data for the four feeders in this study was provided as network models in Synergi Electric. The feeder models were validated through meetings with OTP and Xcel, and it was decided with the utilities that reverse power flow would not be the limiting constraint for hosting capacity. The focus is on thermal and voltage constraints, with the rapid voltage change constraint being


the absolute threshold that cannot be breached. The following sections will describe each feeder’s characteristics upon which its hosting capacity is based.

4.3 Feeder A

The Feeder A Substation has a 41.6/12.5 kV substation transformer rated at 3.75 MVA, connecting a single distribution feeder in a rural area. The feeder has a radial configuration with 1,960 kVA of connected industrial load (a large industrial customer) and 81.6 kVA of connected distributed load. Reactive power support is provided by a 300 kVAr manually operated capacitor, installed in the middle of the feeder. According to OTP, the capacitor is switched on during peak load and switched off during off-peak load, which is observed in the measured demand data where there is a distinct seasonal variation in power factor during the year, with lower power factors during the peak winter period. There is no generation connected on the feeder. The feeder details are summarized in Table 4. The breaker at the feeder head has a lower continuous rating than the transformer to which it is connected (99 A compared to 173 A). The overload on this breaker is not regarded in the analysis.

Table 4: Feeder A key details

4.4 Feeder B

The Feeder B Substation is a 115/12.5 kV substation with four distribution feeders located in a mixed suburban rural area. A 115/12.5 kV transformer rated at 37.4 MVA is installed at the substation. The Synergi model has a full representation of this feeder with adjacent feeders represented as spot loads at the substation. Reactive power support is provided by a 450 kVAr manually operated capacitor located along the feeder. OTP have stated that the capacitor is kept on irrespective of the demand. There is no generation currently connected on the feeder. The key feeder details are summarized in Table 5. The demand data for this feeder was measured at the substation. To get the individual feeder demand, the substation demand was scaled by the proportion of connected transformer capacity on the feeder (29% of the substation).

Parameter Value Unit Rated Voltage 12.5 kV Total Length 6.9 miles

Connected Capacity 3730 kVA Peak Metered Demand 1940 kW

Minimum Metered Demand 82 kW

Worst Case Operating Point Time 5/14/2017 13:45

Solar Irradiance 95.6 % Demand 106 kW


Table 5: Feeder B key details

Parameter Value Unit Rated Voltage 12.5 kV Total Length 13.7 miles

Connected Transformer Capacity 10.3 MVA Peak Metered Demand 5.9 MW

Minimum Metered Demand 1.1 MW

Worst Case Operating Point Time 07/05/2017 13:45

Solar Irradiance 95.8 % Demand 2.4 MW

4.5 Feeder C

The Feeder C Substation has a 115/34.5 kV substation transformer rated at 46.7 MVA, connecting a single distribution feeder located in a rural area. Feeder C has a radial configuration with two main branches, one heading north and one heading south of the substation. There are transformers further along these branches to step down the voltage to 12.5 kV for the loads towards the end of the feeder, with the voltage supported by regulators and field capacitors. There is existing PV generation on the feeder at nine large customer sites, nine individual generator sites and with a small amount distributed along five feeder sections, giving the total PV capacity shown in Table 6. This generation is mostly connected to the 34.5 kV sections of the feeder and concentrated closer to the substation. The limiting feeder section is the 336 Al overhead line going into the substation with a rating of 560 A (33.5 MVA).

Table 6: Feeder C key details

Parameter Value Unit Rated Voltage 34.5/12.5 kV Total Length 252 Miles

Connected Capacity 42.2 MVA Peak Metered Demand 10.2 MW

Minimum Metered Demand 2.2 MW

Worst Case Operating Point Time 5/24/15 13:00:00

Solar Irradiance 97.58 % Demand 3.9 MW

Existing PV Generation Large Customer 15.5 MW

Generators 9 MW Distributed 0.06 MW


4.6 Feeder D

The Feeder D substation contains a 115/13.8 kV transformer rated at 50 MVA, connecting eight 13.8 kV distribution feeders located in a mixed suburban rural area. There is existing PV generation on the feeder with most generation located at four large PV sites: one 4 MW, two 5 MW, and a smaller 0.25 MW. There is also a small amount of distributed PV on some single phase sections totaling 24kW. The voltage along the feeder is supported by regulators and capacitors. The limiting feeder section is the 336 Aluminum overhead line going into the substation with a rating of 560 A (13.4 MVA).

Table 7: Feeder D key details

Parameter Value Unit Rated Voltage 13.8 kV Total Length 71.5 Miles

Connected Capacity 27525 kVA Peak Metered Demand 9101 kW

Minimum Metered Demand 1400 kW

Worst Case Operating Point Time 03-May-17 13:00:00

Solar Irradiance 97.00 % Demand 1500 kW


5 Static Hosting Capacity Methodology The analysis for this stage of the project was undertaken in Synergi Electric (version 6.0.0). This software has extensive functionality for distribution network planning, including a tool for calculating the PV hosting capacity using traditional approaches. Using these Synergi tools, each feeder is analyzed individually to answer the following questions:

i. Where is the best location for a community scale PV site that would represent a Solar Garden?

ii. What is the current PV hosting capacity of the feeder? iii. What constraints are limiting the hosting capacity?

5.1 Synergi PV Hosting Capacity Tool Synergi provides five approaches for computing the PV hosting capacity. Each method was assessed to determine its relevance to the project requirements. Table 8 presents a summary of these options.

Table 8: Summary of PV hosting capacity analysis approaches

Method Description

Stochastic analysis Grows PV at random locations around the model

Provides hosting capacity analysis for multiple PV interconnections on

the feeder

Feeder rating Calculates the hosting capacity

based on feeder rating and minimum daytime load

Does not consider the PV location, line loading or voltage limits

Feeder maximum demand Calculates the hosting capacity based on feeder demand and

minimum daytime load

Does not consider the PV location, line loading or voltage limits

Grow PV on each feeder to exception

Gradually increments the existing PV generation until voltage or loading exception is reached

Does not consider new PV locations.

Sectional incremental Grow PV on one feeder section at a

time until voltage or loading exception is reached.

Provides hosting capacity analysis for single PV interconnection on

the feeder

The sectional incremental method was chosen as the most appropriate for this project, as it provides an indication of the hosting capacity of a single solar garden development at different locations on the system. Synergi’s hosting capacity tool only considers voltage and feeder loading for this calculation, and not the broader range of assessment criteria that is commonly used, such as the effect on protection systems and the risk of rapid voltage change. Smarter Grid Solutions therefore implemented scripts in Python to give a more detailed assessment of the feeders by considering these additional constraints in order to align this analysis with the EPRI DRIVE tool. The additional constraints are labelled as “SGS Script” in Table 9.


The SGS script follows the same procedure as that used in Synergi’s sectional incremental capacity method. A PV generator is added at each section of the feeder and its capacity is increased until a limit is breached, where the maximum allowable PV size for that constraint is defined. The hosting capacity is then the lowest value across all the constraints that are considered.

Table 9: Constraints considered for static hosting capacity

Screen SGS Limit (OTP)

SGS Limit (XCEL)

Implementation

Primary Over Voltage 105% 105% Synergi Primary Voltage Deviation 3% 3% SGS Script

Regulator Voltage Deviation 1.5% 50% of regulator bandwidth

SGS Script

Thermal for Discharging DER 100% 100% Synergi Additional Element Fault Constraint 10% 10% SGS Script Breakers Relay Reduction of Reach 5% 10% SGS Script

Reverse Power Flow4 0% 0% Synergi

5.2 Identification of Feeder Worst Case Operating Point To calculate the hosting capacity it is necessary to select a fixed operating point at which to run the calculations. This is typically set to be maximum generation and minimum demand, representing the worst-case boundary condition for distributed generation. Although this general case can be applied to all DER technology, it is unrealistically pessimistic when considering only PV, which has a very predictable generation characteristic based on the patterns of solar irradiance, with peak generation during the day unlikely to coincide with periods of minimum demand which typically occur during the night. This project therefore calculates the hosting capacity at the worst case operating point, based on the load data from the utilities and the solar irradiance profile from CPR, which is calculated as follows:

1. Normalize input data to get two datasets: % load, and % solar irradiance throughout the year;

2. Summarize for each month by hour of the day: a. The minimum % load; b. The maximum % radiation; c. The difference between a. and b. (the worst case operating point for that time).

3. Get the maximum worst case operating point over the year.

The load for the worst case operating point is allocated on the feeder based on the existing power factor that is set in the Synergi model.

4 This screen was included for information only; it was not used in determination of hosting capacity.


6 Enhanced Hosting Capacity Methodology The enhanced hosting capacity analysis process is shown in Figure 8. Phase 1 quantifies the effect of increasing PV on the feeder through annual load flow simulations. A series of load flow calculations are used to generate an annual curtailment profile for increasing PV sizes on each feeder. An annual curtailment profile is a time-series representation of the PV’s curtailed power for each period during the year (in MW). The annual curtailment profiles are used to investigate how curtailment can be reduced with storage or thermal load shifting in Phase 2. These technologies are modelled in Excel. A storage solution is specified based on the results to derive costs in the next phase. The information derived in the earlier phases is used in Phase 3 to demonstrate how the solutions can be combined and to cost and compare the different solutions.

Figure 8: Analysis process

6.1 Phase 1: Load Flow Studies A series of load flow calculations will be used to simulate the annual export for a range of PV sites, with actions taken when necessary to mitigate constraints.


The size of the PV site is chosen so that it is larger than the hosting capacity, so there are periods during the year where there are violations of thermal or voltage limits at constraint locations. When a violation is observed, the PV export is reduced until the issue is resolved, to bring the feeder loading or voltage back within limits, mimicking the response of a real-time distribution control system. The PV curtailment is summed to give an estimate of annual curtailment.

6.1.1 Scenario

A community scale PV generator, or Solar Garden, (The PV Site) is added to the feeder and a series of load flow calculations is used to model the network over a year. This is repeated for all the feeders, with different sizes of PV. A unique combination of feeder, PV size and inverter settings is referred to as a scenario for the purposes of this report. The PV site is modelled as a balanced three-phase inverter connected generator using Synergi’s PV model.

6.1.2 Smart Inverter Settings

The IEEE 1547-2018 standard defines minimum requirements and a number of operating modes for smart inverters, some of which can be employed to increase the hosting capacity of a given circuit. Figure 9 presents both the IEEE 1547-2018 default settings and those used in the present study. The constant fixed leading 0.95 power factor was considered as part of the static hosting capacity results. The smart inverter functionalities considered as part of the enhanced hosting capacity analysis included the following characteristics:

• Operates with a variable power factor between 0.9 lagging to 0.9 leading based on the Volt-VAr characteristic given in Figure 9. This corresponds to an inverter MVA rating of 110% of the PV site’s MW rating. More aggressive settings—those that reach the maximum MVAr settings for a tighter voltage range—were used in an effort to maximize hosting capacity for this technology option. Inverter control compatibility with other grid voltage control systems must be examined when using more aggressive settings.

• Real power is reduced at higher voltages according to the Volt-Watt curve shown in Figure 10. More aggressive settings were used in an effort to enable higher hosting capacity5. The Volt-Watt curve used is outside of the range of allowable settings in IEEE 1547-2018. The standard intentionally sequences the response of Volt-Watt to come after exhaustion of much of the capability from Volt-VAr.

5 It should be noted that the initial voltage at which power curtails (1.04 pu) is lower than the minimum range (1.05 pu) in the standard and hence is not representative of inverters employed in North America. Therefore, the curtailment results should be taken to be greater than would be expected using inverters with the normal range.


Figure 9: Smart inverter Volt-VAR characteristic (IEEE default curve shown in black and curve used in study in blue)

Figure 10: Smart inverter Volt-Watt characteristic (IEEE default curve shown in black and curve used in study in blue)


6.1.3 Constraints

The constraints considered in the enhanced hosting capacity analysis are limited to primary over voltage, primary voltage deviation, regulator voltage deviation, and thermal overload. It was decided in discussion with CPR, OTP and Xcel that reverse power flow and fault level constraints will not apply in this stage of the analysis. The enhanced hosting capacity analysis considers the unity power factor case as the baseline for hosting capacity and a result, the constraints identified for that case are those managed in this part of the study. It should be noted that the utility partners’ preference is to first address voltage constraints using fixed power factor settings. Given that the leading power factor results for the static hosting capacity study change the nature of the limiting constraints, since the unity power factor case provides a fuller picture of how different advanced technologies can be used to resolve different constraints, it was decided to use this approach for the enhanced hosting capacity analysis. Scenarios have been selected to investigate how increasing PV affects the network and whether these constraints can be managed using the advanced technologies. Thermal constraints are managed to the continuous rating of the circuit given in the Synergi model. The rapid voltage change constraint is taken as the upper limit of PV size for the scenarios, as the technologies being investigated offer no potential to manage this type of constraint in the dynamic perspective, at least without detailed discussions with the utility partner to agree on an approach for modelling a control approach that could offer mitigation. Modulating power factor as a function of PV output power and potentially of feeder loading represents a potential route to address the risk of rapid voltage change, however, the project did not have the time or appropriate high resolution data to study technology solutions for this constraint without distracting from the core results of the study. Therefore, the rapid voltage change constraint is taken as the binding constraint that limits the enhanced hosting capacity.

6.1.4 Simulation of Annual Curtailment Profile

This section explains the time-series calculations that are carried out for each scenario to produce an annual curtailment profile. The process is as follows:

1. The PV Site is added to the load flow model with the scenario’s rating and with smart inverter settings to manage the local voltage.

2. For each period in the year: a. The load is set by scaling the entire feeder load (the load is balanced where

possible). b. The PV Site’s output is set based according to the normalized PV profile. c. Existing PV generators and generation at large customers is set according to the

normalized PV profile. d. The load flow is calculated. e. Circuit loading at constraint location is checked for thermal violations.


f. If there is a violation, the PV Site’s output is adjusted to remove the violation, and the load flow solution recalculated. This process is repeated if necessary until the violation is removed.

g. Results for the period are saved (see Section 5.1.5). h. Proceed to next time-step.

6.1.5 Results for Each Time Period

At each time-step in the time-series calculation, the following is saved for later analysis. 1. Feeder voltages at constraint locations; 2. Feeder loading at constraint locations; 3. PV generation export including curtailment (P,Q); 4. Estimated headroom (𝐻𝐻) – discussed in the following section.

Pre- and post-curtailment data is saved for all the required parameters, representing the PV export with and without managing the constraints.

6.1.6 Estimated Headroom

The estimated headroom (𝐻𝐻) is the amount of capacity available for PV export at any instant, considering the constraints on the feeder. Headroom is used to indicate those periods where storage is allowed to discharge or thermal load can be shifted, and by how much. The process to determine the headroom is outlined in Table 10.

Table 10: Process to estimate headroom

Description Calculation

𝑺𝑺𝒍𝒍𝒍𝒍𝒍𝒍𝒙𝒙 Power limit at constraint 𝑥𝑥 considering the

constraints of the feeder from the static hosting capacity analysis and the load flow model.

Taken from load flow model

𝑺𝑺𝒙𝒙 Apparent power at constraint location 𝑥𝑥 Power calculated in load flow

𝑯𝑯 The resultant headroom of the PV site considering all

the constraints. This is the constraint that will be violated first if the PV export is increased.

𝐻𝐻 = 𝑺𝑺𝒍𝒍𝒍𝒍𝒍𝒍𝒙𝒙 − 𝑺𝑺𝒙𝒙

6.1.7 Phase 2: Investigate Solutions

The annual curtailment profiles derived in Phase 1 assume that PV generation is reduced during curtailed periods and therefore the energy is lost. The estimated curtailed energy represents an excess of generation on the feeder that cannot be exported on the network. Different methods and technologies will be investigated to manage this excess generation:

• Reduce the PV output so the energy is not generated in the first place (curtailment – calculated above);

• Store the energy in a battery (storage); • Increase thermal load to absorb the energy (thermal load shifting).


Each of these solutions is analyzed to define the high level requirements of a solution for each feeder with estimated costs. This analysis is used to make conclusions and recommendations about leading ways to economically maximize the installed PV capacity on the feeders.

6.1.8 Storage

To investigate the use of battery storage, an ideal battery is simulated for each configuration to derive an ideal battery profile, which gives the energy stored in an ideal battery over an annual period based on the annual curtailment profile. The objective of the ideal battery is to completely remove the curtailment of PV generation. This profile is then used to specify the requirements of an actual battery in terms of power and energy rating. A battery will be specified for two different scenarios:

1. Store all curtailed energy (so that 100% of PV energy is utilised). 2. Store a proportion of the curtailed energy so that the total curtailed energy is equal to 5%

of total PV generation.

6.1.9 Ideal Battery Model

Prior to calculating the real rating for the battery, an ideal battery model is first considered. To generate the ideal battery profile, a model of an ideal battery is used based on the following assumptions and the description given in Table 11 and Table 12:

• The battery is located at the PV site. • The battery is only used for managing network constraints (not considering other market

opportunities such as arbitrage, frequency support etc.). • The battery has infinite capacity and 100% efficiency. • The battery can charge during constrained periods. • The battery discharges whenever it has stored energy at a rate equivalent to the

available headroom. Table 11: Ideal battery model nomenclature

Symbol Description 𝑻𝑻 Time period of the simulation (from the annual curtailment profile) 𝑬𝑬𝒕𝒕 Energy stored in the battery at time 𝑡𝑡 𝑬𝑬𝒕𝒕−𝟏𝟏 Energy stored in the battery from the previous time step (at time 𝑡𝑡 − 𝑇𝑇) 𝑪𝑪𝒕𝒕 Curtailed power at time 𝑡𝑡 from the annual curtailment profile 𝑯𝑯𝒕𝒕 Available headroom at time 𝑡𝑡, this is the rate at which the storage can be discharged before

violating network limits


Table 12: Calculations for ideal battery profile

Description Battery State Condition Calculation The battery will initially be fully discharged Initial state 𝑡𝑡 = 1 𝐸𝐸𝑡𝑡 = 0

When PV is curtailed, the curtailed energy is stored in the battery Charging 𝐶𝐶𝑡𝑡 > 0 → 𝐻𝐻𝑡𝑡 < 0 𝐸𝐸𝑡𝑡 = 𝐸𝐸𝑡𝑡−1 + 𝐶𝐶𝑡𝑡𝑇𝑇

When the PV is not curtailed, the battery will discharge at the maximum rate available

according to the available network capacity (defined by the headroom)

Discharging 𝐻𝐻𝑡𝑡 > 0 → 𝐶𝐶𝑡𝑡 = 0 𝐸𝐸𝑡𝑡 = 𝐸𝐸𝑡𝑡−1 − 𝐻𝐻𝑡𝑡𝑇𝑇

6.1.10 Battery Rating

The ideal battery profile is used to specify the rating required for a real battery, described in Table 13.

Table 13: Calculation of storage ratings

Rating Unit Calculation Power (MW) MW The maximum rate of charge/discharge observed in the ideal battery profile.

Energy (MWh) MWh The maximum stored energy from the ideal battery profile.

Inverter Rating MVA The power rating of the battery plus a reactive range of 0.9 lagging to 0.9 leading power factor.

6.1.11 Limitations

It should be noted that this methodology is not intended to design a battery that will guarantee network security for all conditions. The battery is rated for a single annual load profile, which may not be the worst case or even typical. Variations in load and generation during a day may result in situations where there is no battery storage available to store excess energy and therefore the network limits would be violated without some other method of control. Guaranteeing network security with a battery operated in isolation would require a specification that could be unfeasibly large and expensive. It is therefore recommended that any storage solution is part of a wider network control scheme. The results from the analysis are only intended to demonstrate the approximate requirements for a battery system to understand whether a battery would be feasible for that particular application.

6.2 Thermal Load Shifting 6.2.1 Source Data

CPR provided electric hot water demand information for a typical day. There are four scenarios: • 2025 Low; • 2025 High; • 2050 Low; • 2050 High.


This project is considering the 2025 Low scenario as an accurate representation of the current situation, as agreed with CPR. This scenario assumes that there is no new electric hot water demand to consider in the network, and on average the controllable hot water load is 20% of the total hot water demand. This typical daily information is scaled down to a “per house” equivalent, based on the 140,000 houses/tanks used to generate the profile. This is then scaled based on the customer information received for each feeder. Key assumptions are:

• It is assumed that the hot water demand generated for each feeder in this manner is already included in the total feeder demand received from the utilities.

• In the event that the scaled hot water demand is greater than the feeder residential demand, it is assumed that the feeder residential demand solely represents hot water demand, and the hot water demand profile is adjusted accordingly.

6.2.2 Feeder Customer Data

OTP and Xcel provided information on the customer diversity in the feeders selected for analysis. This information is displayed in Table 14 along with the average controllable demand obtained using the CPR thermal load data. It is assumed that only residential hot water demand is controllable.

Table 14: Feeder customer data and average controllable demand

Feeder Number of Customers % Residential Average Controllable Demand (kWh)

Feeder A 78 74 29.0 Feeder B 231 37 40.7 Feeder C 3175 90 886.1 Feeder D 1678 93 471.3

6.2.3 Thermal Load Shifting Model

Thermal load shifting is modelled as an increase in load during curtailed periods (during peak solar generation) with a corresponding decrease in load later in the day. A model of thermal load shifting is developed from the annual curtailment profile to generate a shifted load profile and a corresponding shifted curtailment profile. The shifted load profile is a load profile with reduced load during periods of curtailment. The shifted curtailment profile is the annual curtailment profile including the effect of thermal load shifting. The model of thermal load shifting is based on the following assumptions:

• The CPR data is used to derive a data series giving the percentage of controllable thermal load for each hour of the year. This represents load which could be turned down in response to a control signal. It is assumed that this load could also be turned up or shifted to a different time of the day.

• Data is only available for a typical day; it is assumed that this applies for the entire year. In general domestic hot water demand does not vary much from day to day.


• Thermal load shifting is only used to mitigate thermal and reverse power flow constraints upstream of its location.

• It is assumed that the customer will always respond to signals to shift their load.

Table 15 and Table 16 describe the process to generate the shifted load and shifted curtailment profiles.

Table 15: Thermal load shifting nomenclature

Symbol Description 𝑳𝑳𝒕𝒕 The load profile at time 𝑡𝑡 𝑳𝑳𝒕𝒕𝝈𝝈 The shifted load profile at time 𝑡𝑡 𝜹𝜹𝒕𝒕 The control action at time 𝑡𝑡, where positive is an increase in load and negative a decrease 𝑪𝑪𝒕𝒕 The annual curtailment profile at time 𝑡𝑡 𝑪𝑪𝒕𝒕𝝈𝝈 The shifted curtailment profile at time 𝑡𝑡 𝜽𝜽𝒕𝒕 Thermal load available for control at time 𝑡𝑡 (downstream of constraint)

𝝈𝝈𝒕𝒕−𝟏𝟏 Total shifted load that has been supplied prior to time 𝑡𝑡 (resulting in a reduction in load at 𝑡𝑡 if there is headroom available)

𝝈𝝈𝒕𝒕 Cumulative sum of shifted load up to time 𝑡𝑡

𝑯𝑯𝒕𝒕 Available headroom at time 𝑡𝑡 (this is the amount that load can be reduced without violating network

limits) Table 16: Thermal load shifting model

Step Description Condition Calculation

Calculate control action at each time

step

No Control Load is unchanged when there are is no

curtailment and no load was shifted to earlier in the day.

𝐶𝐶𝑡𝑡 = 0 𝜎𝜎𝑡𝑡−1 = 0 𝛿𝛿𝑡𝑡 = 0

Load Increase When PV is curtailed, load is increased to

compensate for the curtailed power, or at the full amount of thermal load available.

𝐶𝐶𝑡𝑡 > 0 𝛿𝛿𝑡𝑡 = min(𝐶𝐶𝑡𝑡 ,𝜃𝜃𝑡𝑡)

Load Decrease Load that was shifted to earlier in the day is

supplied later when there is headroom available.

𝐶𝐶𝑡𝑡 = 0 𝜎𝜎𝑡𝑡−1 > 0 𝛿𝛿𝑡𝑡 = −min(𝜎𝜎𝑡𝑡−1,𝐻𝐻𝑡𝑡)

Calculate shifted profiles

The shifted load and curtailment profiles are calculated at every time step from the

calculated control actions. -

𝐿𝐿𝑡𝑡𝜎𝜎 = 𝐿𝐿𝑡𝑡 + 𝛿𝛿𝑡𝑡 𝐶𝐶𝑡𝑡𝜎𝜎 = 𝐶𝐶𝑡𝑡 − 𝛿𝛿𝑡𝑡 𝜎𝜎𝑡𝑡 = 𝜎𝜎𝑡𝑡−1 + 𝛿𝛿𝑡𝑡

6.2.4 Phase 3: Modelling of Co-ordinated Solution

A final scenario is used to demonstrate how an advanced distribution control system could be used to accommodate numerous different technologies to manage constraints. For this study, the following assumptions are made based on the results from the earlier analysis:


• The combined advanced technologies approach is applied to all sizes of PV modelled in the enhanced hosting capacity analysis.

• The shifted load profile is used to represent the load on the feeder. • The advanced distribution control system monitors the network constraints and assign

actions in the following order of priority: 1. Load shifting; 2. Storage; 3. Curtailment.

For this study the shifted load profile is used to model the response from thermal load shifting and then a battery is sized to ensure the residual curtailment is no more than 3% of the potential annual PV export. A battery charging profile is derived to represent a storage system that charges only when curtailment goes above 3% of annual PV export. Storage is sized according to the same methodology presented in Section 5.1.8.


7 Results and Discussion This section contains the static and enhanced hosting capacity results for the four feeders. The static hosting capacity analysis provides results for two fixed power factors. Both 0.95 leading and unity are considered where 0.95 leading is the preferred utility power factor for modelling, and unity provides a more conservative view of hosting capacity. The enhanced hosting capacity analysis builds on the results of the static hosting capacity analysis, which serves as baseline for the analysis. The unity power factor results are used as basis for the enhanced hosting capacity analysis to assess the potential impact of advanced technologies, and any curtailment that may be experienced. The 0.95 leading power factor serves as the reference point for comparison for advanced technologies.

7.1 Feeder A 7.1.1 Static Hosting Capacity

The worst case scenario occurs on May 13th, 2017 at 13:00. 7.1.2 Unity Power Factor

The results for the individual constraints for a sample of five sections on the feeder are presented in Table 17. The PV Site will be situated 1.05 miles from the substation for the enhanced hosting capacity analysis. For this feeder section the binding constraint is primary over voltage, with a static hosting capacity of 1.14 MW.

Table 17: Feeder A section hosting capacity (MW) results by constraint for unity PF

Distance from

Substation (Miles)

Maximum PV Hosting Capacity (MW)

Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

1.05 1.14 5.00 13.7 4.03 2.70 1.52 0.11 1.61 0.79 5.00 13.0 4.03 2.46 1.52 0.11 2.61 0.50 2.93 10.7 4.03 1.99 1.52 0.11 3.24 0.34 1.99 9.96 4.03 1.76 1.52 0.11 3.24 0.34 1.91 9.96 4.03 1.76 1.52 0.11

Figure 11 shows that the static hosting capacity varies along the feeder according to the different factors that become important. The primary voltage deviation, thermal overload, additional element fault current, breaker relay reduction of reach and reverse power flow all remain relatively static along the length of the feeder. Primary voltage deviation and regulator voltage deviation both decrease as distance along the feeder increases.


Figure 11: PV limit for each constraint along Feeder A for unity PF

7.1.3 0.95 Power Factor

The results for the individual constraints for a sample of five sections on the feeder are presented in Table 18. The PV Site will be situated 1.05 miles from the substation for the enhanced hosting capacity analysis. For this feeder section, with 0.95 PF, the binding constraint is the thermal overload with a static hosting capacity of 4.23 MW.

Table 18: Feeder A section hosting capacity results by constraint type for 0.95 leading PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

1.05 >5.00 >5.00 >15.0 4.23 3.40 7.15 0.110 1.61 >5.00 >5.00 >15.0 4.23 2.93 7.15 0.110 2.61 >5.00 >5.00 >15.0 4.23 1.99 8.79 0.110 3.24 >5.00 >5.00 >15.0 4.23 1.76 8.79 0.110 3.24 >5.00 >5.00 >15.0 4.23 1.76 8.79 0.110

Figure 12 shows that the static hosting capacity varies along the feeder according to the different factors. The available static hosting capacity for PV decreases for Additional Element Fault Current as distance along the feeder increases; whereas, the Breaker Relay Reduction of Reach increases between 1.51 miles from the substation and 2.61 miles from the substation.


The other constraints remain consistent along the length of the feeder, which is likely due to the fact that the feeder is less than 4 miles long.

Figure 12: PV limit for each constraint along Feeder A for 0.95 PF

7.1.4 Enhanced Hosting Capacity

The connection point for the PV Site is 1.05 miles along the feeder. For the unity power factor case, this location is subject to a voltage constraint. There is also a rapid voltage change constraint due to the primary voltage deviation limit, the PV size is limited to this level and sizes are listed below.

Table 19: Feeder A PV site location characteristics for enhanced hosting capacity study

Distance (miles)

Static Hosting Capacity PV Sizes Primary Over

Voltage Primary Voltage

Deviation Thermal Overload

1.05 1.14 5.00 4.03 0.78, 1.56, 2.34, 2.73, 3.12, 3.51, 3.9

The key constraint in Feeder A is voltage and the smart inverter works to resolve the over voltage constraint. The Volt-Watt curve of the inverter starts to take action at 1.04 pu and brings the output to zero if voltage exceeds 1.06 pu, Figure 13. Analysis of the results revealed that the default voltage regulator settings changes the secondary voltage set-point to 1.04 from 1.02 for reverse power flow, which in part explains the high levels of curtailment observed as the voltage becomes bias around 1.04 when PV is producing, as one can note in Figure 13. The second factor is the effect of the V-W curve which operates frequently between 1.04 and 1.05.


The power flow is always within limits (Figure 14), and the generator power factor is low during times of power reduction due to the Volt-Watt curve (Figure 15).

Figure 13: Bus voltage for PV site of 3.90 MW. The voltage frequently goes above the limit.

Figure 14: Power flow at thermal constraint location for PV site of 3.90 MW. The power flow is within the

thermal limit throughout the year.


Figure 15: Power factor for PV site of 3.90 MW. The power factor is low during periods of curtailment due to

Volt-Watt curve.

Table 21 presents the results of the advanced technologies’ ability to maximize PV export. The smart inverters are able to resolve the voltage issues seen on the feeder; however, this results in significantly reduced export of up to 38% for a 3.9 MW capacity site. As in Feeder A, thermal load shifting has a limited benefit, and the battery required to fully maximize PV export is 4 MW/24 MWh.

Table 20: Feeder A: Advanced technology impact on hosting capacity

Advanced technology Ability to increase hosting capacity Smart Inverters (Stand Alone) Yes Smart Inverters (Co-ordinated) No

Thermal Load Shifting No Storage Yes


Table 21: Feeder A: Advanced technology results

Smart Inverters (Stand Alone)

Smart Inverters (Co-ordinated) Thermal Load Shifting Energy Storage

PV [MW]

Voltage Driven Curtailment

Thermal Driven Curtailment6

Curtailment Reduction

Curtailment After Shifting Rating

MWh % MWh % MWh % MWh % MW MWh 0.78 103 9.40% N/A N/A 62.0 60.2% 41.0 3.80% 0.20 1.20 1.56 272 12.5% N/A N/A 82.0 30.1% 190 8.70% 0.60 3.40 2.34 587 17.9% N/A N/A 95.0 16.2% 492 15.0% 1.50 7.20 2.73 804 21.0% N/A N/A 99.0 12.3% 705 18.4% 2.50 11.5 3.12 1160 26.5% N/A N/A 103 8.90% 1060 24.2% 2.90 15.8 3.51 1590 32.2% N/A N/A 105 6.60% 1480 30.1% 3.30 22.4 3.90 2070 37.8% N/A N/A 107 5.20% 1960 35.9% 3.60 24.3

Using an approach that combines the advanced technologies via a real-time, part-distributed control system can exploit the benefits of each and reduce the curtailment experienced by PV connecting beyond the static limit in Feeder A. The results in Table 22 show the impact of thermal load shifting and energy storage on curtailment. Thermal load shifting is the first advanced technology used to maximize PV export, if further action is required then energy storage is used. Energy storage is sized in order to ensure a maximum reduction of export to 3% of potential annual output.

The results show that curtailment is 3%, with a battery sized at 3.5 MW/22.7 MWh for a PV size of 3.9 MW. The implication of installing energy storage and utilizing an inverter with Volt-Watt control capability results in a requirement to divert PV export energy to storage when the smart inverter detects high voltages, rather than reducing the PV generation.

Table 22: Feeder A: Results of combining advanced technologies

Combined Technology PV Annual Thermal Load

Shifted Storage Rating Curtailment

MW MWh MW MWh MWh % 0.78 62.0 0.10 0.50 33.0 3.00% 1.56 82.0 0.50 2.60 65.0 3.00% 2.34 95.0 1.40 6.30 99.0 3.00% 2.73 99.0 2.50 10.6 117 3.10% 3.12 103 2.80 14.8 131 3.00% 3.51 105 3.20 21.4 148 3.00% 3.90 107 3.50 22.7 164 3.00%

6 Curtailment using centrally co-ordinated smart inverters are listed as N/A as the voltage issues are addressed by smart inverters alone, so there is no added value in adding coordination.


The graph in Figure 16 illustrates the change in curtailment experienced by the PV site as rated capacity increases. Thermal load shifting slightly reduces curtailment from that experienced with smart inverters, however combining the benefits of all advanced technologies significantly reduces the curtailment.

Figure 16: Feeder A curtailment with increasing PV size

7.2 Feeder B

7.2.1 Static Hosting Capacity

The worst case scenario for this feeder occurs on May 5th, 2017 at 13:00. 7.2.1.1 Unity Power Factor

The results for the individual constraints for a sample of five sections on the feeder are presented in Table 23 (variations in hosting capacity are due to the location on the feeder). The PV Site will be situated 0.68 miles from the substation for the enhanced hosting capacity analysis. For this section the binding constraint is the thermal overload, with a static hosting capacity of 12.5 MW.


Table 23: Feeder B section results for unity PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.680 20.3 24.1 >40 12.5 14.1 12.2 2.4 1.21 13.2 15.3 >40 12.5 9.1 12.2 2.4 1.74 8.0 9.1 >40 4.0 7.8 12.2 2.4 2.47 4.3 4.7 >40 4.0 5.3 12.8 2.4 3.02 2.0 2.2 25.3 5.1 2.2 12.8 2.4

Figure 17 shows that the static hosting capacity varies along the feeder according to the different factors that become important. The reverse power flow constraint is constant along the feeder; however, OTP decided that reverse power flow should not be the binding constraint for the static hosting capacity. The thermal overload constraint depends on where the PV is located along the feeder, and is lower towards the end of the feeder.

The constraints concerning voltage show a decaying exponential characteristic, with lower PV limits further along the feeder. This is related to the network being weaker further away from the substation; therefore, any injection of power will have a greater effect on the voltage. The fault level injection from the PV decreases further along the feeder, and the impact of fault level on the breaker remains relatively static.

Figure 17: PV limit for each constraint along Feeder B for unity PF


7.2.1.2 0.95 Power Factor

The results for the individual constraints for a sample of five sections on the feeder are presented in Table 24 (variations in hosting capacity are due to the location on the feeder). The PV Site will be situated 0.68 miles from the substation for the enhanced hosting capacity analysis. For this section the binding constraint is the thermal overload, with a static hosting capacity of 13.1 MW.

Table 24: Feeder B section results for 0.95 PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.680 40.0 40.0 24.7 13.1 12.8 10.9 2.42 1.21 40.0 37.8 22.2 13.1 8.44 10.9 2.42 1.74 40.0 40.0 21.6 4.15 7.19 11.6 2.42 2.47 17.8 8.44 20.3 4.24 5.31 11.6 2.42 3.02 40.0 19.1 14.1 5.34 2.19 11.6 2.42

Figure 18 shows that the static hosting capacity varies along the feeder according to the different factors that become important. The thermal overload constraint depends on where the PV is located along the feeder, and is lower towards the end of the feeder.

The regulator voltage deviation decreases as distance increases, however primary over voltage and primary voltage deviation remain relatively static until further along the feeder. This is related to the network being weaker further away from the substation therefore, any injection of power will have a greater effect on the voltage. The fault level injection from the PV decreases further along the feeder, and the impact of fault level on the breaker remains relatively static.


Figure 18: PV limit for each constraint along Feeder B for 0.95 PF


The connection point for the PV site is 0.68 miles along the feeder. This location is limited by thermal, voltage and the rapid voltage change constraint as listed in Table 25. The binding constraint is a thermal overload constraint on the underground cable into Feeder B’s substation. This constraint is monitored in the simulation and the generation reduced when the limit is violated. The PV sizes that are investigated are limited to below the limit of the primary voltage deviation constraint of 24.1MW

Table 25: Feeder B PV site location characteristics for enhanced hosting capacity study

Distance (miles) Static Hosting Capacity

PV Sizes Primary Over Voltage

Regulator Voltage Deviation

Thermal Overload

0.680 12.5 20.3 24.1 4.8, 9.6, 14.4, 16.8, 19.2, 21.5, 24.1

The results show that a PV site of 24 MW capacity can be incorporated without violating any voltage limits, as shown in Figure 19, where both pre- and post- curtailment voltages are less than 1.05 pu throughout the year. The smart inverters maintain the voltages on the feeder, and the impact on the PV site’s power factor is shown in Figure 21. The voltage deviations are not extreme, as the power factor range for the PV site is only altered between unity and 0.97. The smart inverter settings implemented allow for a range of power factor down to 0.9. Maximizing hosting capacity in this scenario would require remote monitoring of the thermal constraint as the power flow is frequently above the limit, as shown in Figure 20, with the power as high as 160% of the circuit rating.


Figure 19: Bus Voltage for PV Site of 24 MW. The voltage is below the limit throughout the

year.

Figure 20: Power Flow at Thermal Constraint Location for PV Site of 24 MW. The power flow frequently goes

above the limit.


Figure 21: Power Factor for PV Site of 24 MW. The power factor varies as a function of voltage due to the Volt-

Var characteristic.

The results in Table 26 show that thermal load shifting has a limited benefit for increasing the size of the PV site. This is due to a combination of factors, the residential population on the feeder is very low and no control is exercised over commercial or industrial customers. This significantly restricts the impact thermal load shifting can have in Feeder B. Energy storage can be used to eliminate curtailment, however a size of 9 MW/48 MWh is required to connect 24 MW of PV.

Table 26: Feeder B: Advanced technology impact on hosting capacity

Advanced Technology Ability to Increase Hosting Capacity Smart Inverters (Stand Alone) No Smart Inverters (Co-ordinated) Yes

Thermal Load Shifting Yes Storage Yes


Table 27: Feeder B: Advanced technology results



PV [MW]

Voltage Driven Curtailment7

Thermal Driven Curtailment



MWh % MWh % MWh % MWh % MW MWh 4.8 0 0 0 0.0% 0 0.0% 0 0.0% 0 0 9.6 0 0 0 0.0% 0 0.0% 0 0.0% 0 0

14.4 N/A N/A 3 0.0% 0 0.0% 2 0.0% 1 2 16.8 N/A N/A 40 0.2% 2 5.0% 38 0.2% 3 9 19.2 N/A N/A 297 1.1% 11 3.7% 286 1.1% 5 20 21.5 N/A N/A 921 3.1% 22 2.4% 899 3.0% 7 33 24 N/A N/A 1990 5.9% 33 1.7% 1957 5.8% 9 48

Using an approach that combines the advanced technologies, via a real-time, part-distributed control system can exploit the benefits of each and reduce the curtailment experienced by PV connecting beyond the static limit in Feeder B. The results in Table 28 show the impact of thermal load shifting and energy storage on curtailment.

The results show that curtailment is less than 3%, without a requirement for energy storage, up until a PV size of 24 MW. A battery sized at 7.7 MW/35.1 MWh is required to ensure curtailment is less than 3% for this size of PV.

Table 28: Feeder B: Results of combining advanced technologies


Shifted Storage Rating Curtailment

MW MWh MW MWh MWh % 4.8 0 0 0 0 0.0% 9.6 0 0 0 0 0.0%

14.4 0 0 0 2 0.0% 16.8 2 0 0 30 0.2% 19.2 11 0 0 168 1.1% 21.5 22 0 0 400 3.0% 24 33 7.7 35.1 999 3.0%

The graph in Figure 22 illustrates the change in curtailment experienced by the PV site as rated capacity increases. Thermal load shifting reduces curtailment slightly from that experienced with smart inverters, however combining the benefits of all advanced technologies significantly reduces the curtailment.

7 These column are listed as N/A beyond the static hosting capacity as smart inverters alone cannot resolve the thermal constraint and therefore, cannot increase hosting capacity beyond the static level.


Figure 22: Feeder B Curtailment with increasing PV Size

7.3 Feeder C


The worst case scenario occurs on May 24th, 2015 at 13:00.

7.3.1.1 Unity Power Factor

The worst case scenario occurs on May 24th, 2015 at 13:00. The results for the individual constraints for 5 locations along the length of the feeder are presented in Table 29. The PV Site will be situated 4 miles from the substation for the enhanced hosting capacity analysis. For this feeder section the binding constraint is primary over voltage, with a static hosting capacity of 12.2 MW.


Table 29: Feeder C section results for unity PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.04 20.0 20.0 20.0 14.6 20.0 12.3 0.00 4.00 12.2 20.0 12.7 14.3 17.0 12.3 0.00 7.50 0.46 5.99 6.09 5.78 9.84 13.6 0.00 11.0 0.78 5.47 20.0 3.79 6.09 13.0 0.00 15.0 1.65 2.34 0.47 0.00 4.22 12.7 0.00

Figure 23 shows that the static hosting capacity varies along the feeder according to the different factors that become important. Primary over voltage decreases further along the feeder. This is a trend common with primary and regulator voltage deviation. In general, thermal overload remains fairly static along the length of the feeder.

Figure 23: PV limit for each constraint along Feeder C for unity PF

7.3.1.2 0.95 Power Factor

The results for the individual constraints for 5 locations along the length of the feeder are presented in Table 29. The PV Site will be situated 4 miles from the substation for the enhanced hosting capacity analysis. For this feeder section, the binding constraint now becomes thermal overload, with a static hosting capacity of 15.0 MW.


Table 30: Feeder C section results for 0.95 PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.04 20.0 20.0 20.0 15.3 20.0 11.4 0.00 4.00 20.0 20.0 20.0 15.0 15.8 11.4 0.00 7.50 20.0 8.91 3.3 0.00 8.91 12.0 0.00 11.0 20.0 20.0 14.8 3.98 5.16 11.4 0.00 15.0 20.0 20.0 1.1 0.00 3.91 11.1 0.00

Figure 23 shows that the static hosting capacity varies along the feeder according to the different factors that become important. Although the feeder section selected for modelling the PV site connection does not experience a primary overvoltage constraint, other feeder sections at the same distance from the substation do experience overvoltage issues. Primary voltage deviation tends to decrease as the distance from the substation decreases. In general the thermal overload, additional element fault current and breaker relay reduction of reach remain fairly static along the feeder. The thermal overload sections do not have a consistent downward trend as the distance increases from the substation because the feeder has different branches that have varying conductor sizes.

Figure 24: PV limit for each constraint along Feeder C for 0.95 PF



The connection point for the PV Site is 4 miles along the feeder. This location is subject to a voltage constraint. There is also a rapid voltage change constraint at 12.7 MW due to the regulator voltage deviation limit.

Table 31: Feeder C PV site location characteristics for enhanced hosting capacity study

Distance (miles)

Static Hosting Capacity

PV Sizes Primary Over Voltage

Regulator Voltage

Deviation

Thermal Overload

4.00 12.2 12.7 14.3 2.54, 5.08, 7.62, 8.89, 10.2, 11.4, 12.7

The results show that a PV site of 12.7 MW capacity can be incorporated without violating any voltage limits, as shown in Figure 25, where both pre- and post- curtailment voltages are less than 1.05 pu throughout the year. The smart inverter works to maintain voltage below the limit, and the impact on generator power factor is shown in Figure 27. The power flow at the thermal constraint location does not go above its limit for any of the scenarios (see Figure 26) therefore there is no requirement to control the real power of the PV site. The binding constraint in this scenario is the rapid voltage change constraint limit, which is hit at 12.7 MW, and the advanced technologies cannot mitigate this constraint under the assumptions of this analysis, as previously discussed.

Figure 25: Bus voltage for PV site of 12.7 MW. The voltage is below the limit throughout the year.


Figure 26: Power Flow at thermal constraint location for PV site of 12.7 MW.

Figure 27: Power factor for PV site of 12.7 MW. The power factor is reduced due to the smart inverter Volt-VAr

curve.

Smart inverters are able to increase hosting capacity beyond the static limit; however, curtailment of real power is not required. The voltage constraint in Feeder C is not extreme as the power factor range for the PV site is only altered between unity and 0.97. The smart inverter settings allow for a power factor up to 0.9. The results in Table 33 show that although hosting capacity has been increased beyond the static limit, it cannot exceed the rapid voltage change constraint. However, we can note in the 0.95 leading power factor case for static hosting capacity, there is a positive impact on this constraint and thus, the utility may allow for


capacities beyond this limit, under the requirement that the leading power factor mode is configured as part of the interconnection agreement. The impact of the additional PV does not breach voltage or thermal limits, resulting in zero curtailment for all PV capacities up to 12.7 MW. As there is no real power curtailment, thermal load shifting and energy storage, or any combination of the advanced technologies, provide no additional benefit for Feeder C.

Table 32: Feeder C: Advanced technology impact on hosting capacity

Advanced technology Ability to increase hosting capacity Smart Inverters (Stand Alone) Yes Smart Inverters (Co-ordinated) No

Thermal Load Shifting No Storage No Table 33: Feeder C: Advanced technology results



PV [MW]

Voltage Driven Curtailment




MWh % MWh % MWh % MWh % MW MWh

2.54 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

5.08 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

7.62 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

8.89 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

10.16 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

11.43 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

12.7 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0

Table 34: Feeder C: Results of combining advanced technologies


Shifted Storage Rating8 Curtailment

MW MWh MW MWh MWh %

2.54 0.00 N/A N/A 0.00 0.00% 5.08 0.00 N/A N/A 00.0 0.00% 7.62 0.00 N/A N/A 0.00 0.00% 8.89 0.00 N/A N/A 0.00 0.00%

10.16 0.00 N/A N/A 0.00 0.00% 11.43 0.00 N/A N/A 0.00 0.00% 12.7 0.00 N/A N/A 0.00 0.00%

8 No curtailment was required in any scenarios, so storage could not be justified.


7.4 Feeder D


The worst case scenario occurs on May 3rd, 2017 at 13:00. 7.4.1.1 Unity Power Factor

The results for the individual constraints for 5 locations along the length of the feeder are presented in Table 35. The PV Site will be situated 2.4 miles from the substation for the enhanced hosting capacity analysis. For this feeder section the binding constraint is thermal overload, with a static hosting capacity of 0 MW, and regulator voltage deviation restricting hosting capacity to 2.97 MW.

Table 35: Feeder D section results for unity PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.02 20.0 20.0 20.0 0.08 20.0 16.7 0.00 2.40 3.57 20.0 2.97 0.00 6.41 16.7 0.00 4.75 1.32 20.0 3.28 0.00 4.22 14.8 0.00 7.60 1.85 15.5 0.78 0.00 3.60 13.3 0.00 9.42 1.23 2.03 0.78 0.00 3.28 13.6 0.00

Figure 28 shows that the static hosting capacity varies along the feeder according to the different factors that become important. Primary over voltage, regulator voltage deviation, and additional element fault current decrease significantly along the length of the feeder. Breaker relay reduction of reach also shows a slight downward trend. Primary voltage deviation, although static for a considerable length of the feeder, does shown a decrease at certain feeder sections, and further along the length.


Figure 28: PV limit for each constraint along Feeder D for unity PF

7.4.2 0.95 Power Factor

The results for the individual constraints for 5 locations along the length of the feeder are presented in Table 36. The PV Site will be situated 2.4 miles from the substation for the enhanced hosting capacity analysis. For this feeder section the binding constraint is thermal overload, with a static hosting capacity of 0 MW, and regulator voltage deviation restricting hosting capacity to 8.91 MW, again a significant improvement over the unity case.

Table 36: Feeder D section results for 0.95 PF

Distance from

Substation (Miles)


Primary Over

Voltage

Primary Voltage

Deviation

Regulator Voltage

Deviation

Thermal Overload

Additional Element

Fault Current

Breaker Relay

Reduction of Reach

Reverse Power Flow

0.02 20.0 20.0 20.0 0.08 20.0 15.8 0.00 2.40 20.0 16.4 8.91 0.00 5.78 14.5 0.00 4.75 20.0 7.34 3.91 0.00 3.59 11.7 0.00 7.60 17.3 7.66 4.53 0.00 3.28 10.2 0.00 9.42 20.0 9.22 4.22 0.00 2.97 10.2 0.00

Figure 29 shows that the static hosting capacity varies along the feeder according to the different factors that become important. Primary voltage deviation, regulator voltage deviation, and additional element fault current decrease along the length of the feeder. Breaker relay reduction of reach also shows a slight downward trend. Primary over voltage, although static for a considerable length of the feeder, does show a decrease at certain feeder sections, and


further along the length. This indicates that a combined smart inverter approach with thermal constraint management leads to significant increases in hosting capacity.

Figure 29: PV limit for each constraint along Feeder D for 0.95 PF (thermal overload is behind the reverse power

flow constraint)


The connection point for the PV site is 2.40 miles along the feeder. This location is subject to a thermal constraint on the 336 AL conductor overhead line that makes up the main branch of the feeder back to the substation. There is also a rapid voltage change constraint due to the regulator voltage deviation limit; the PV sizes are limited to this level and are listed in Table 37.

Table 37: Feeder D: PV site location characteristics for enhanced hosting capacity study

Distance (miles)

Static Hosting Capacity (MW)

PV Sizes Primary Over

Voltage

Primary Voltage Deviation

Thermal Overload

2.4 3.57 2.97 0.00 0.59, 1.19, 1.78, 2.10, 2.38, 2.67, 2.97

The voltage on the feeder never exceeds 1.04 therefore the Volt-Watt control from the smart inverter never comes into play (Figure 30). The smart inverter uses reactive power control and the impact on the PV site’s power factor is shown in Figure 32. The smart inverter power factor varies in a small range between unity and 0.98 indicating that there is not a major voltage issue. The smart inverter settings implemented allow for a range power factor up to 0.9. There is a thermal constraint, and power flow does exceed the limit set. This is shown in Figure 31.


Figure 30: Bus voltage for PV site of 2.97 MW. The voltage is below the limit throughout the year.

Figure 31: Power flow at thermal constraint location for PV site of 2.97 MW. The power flow goes above the

limit during periods of peak generation, requiring curtailment.


Figure 32: Power factor for PV site of 2.97 MW. The PF is reduced due to the smart inverter Volt-VAr curve.

Table 39 shows that the advanced technologies can increase the hosting capacity beyond the static limit with relatively low levels of curtailment; however, rapid voltage change is the binding constraint for this feeder, based on the unity power factor analysis for static hosting capacity. As previously discussed, if the utility bases the interconnection analysis on the assumption of a leading non-unity power factor, there is an opportunity to extend hosting capacity further through the use of managed curtailment.

Table 38: Feeder D: Advanced technology impact on hosting capacity

Advanced technology Ability to increase hosting capacity Smart Inverters (Stand Alone) No Smart Inverters (Co-ordinated) Yes

Thermal Load Shifting Yes Storage Yes


Table 39: Feeder D: Advanced technology results



PV [MW]

Voltage Driven Curtailment9




MWh % MWh % MWh % MWh % MW MWh

0.59 N/A N/A 0.00 0.00% 0.00 0.00% 0.00 0.00% 0.00 0.00

1.19 N/A N/A 0.00 0.00% 0.00 0.00% 0.00 0.00% 0.00 0.00

1.78 N/A N/A 1.00 0.00% 1.00 100% 0.00 0.00% 0.00 0.00

2.10 N/A N/A 2.00 0.10% 2.00 100% 0.00 0.00% 1.00 1.00

2.38 N/A N/A 4.00 0.10% 2.00 50.0% 1.00 0.00% 1.00 2.00

2.67 N/A N/A 7.00 0.20% 4.00 57.1% 3.00 0.10% 1.00 2.00

2.97 N/A N/A 11.0 0.30% 5.00 45.5% 5.00 0.10% 1.00 3.00

There is no additional benefit from combining advanced technologies as thermal load shifting can mitigate the very low levels of curtailment. As a result of this conclusion, the combined technologies are not presented in Figure 33.

Table 40: Feeder D: Results of combining advanced technologies


Shifted Storage Rating10 Curtailment

MW MWh MW MWh MWh %

0.59 0.00 N/A N/A 0.00 0.00% 1.19 0.00 N/A N/A 0.00 0.00% 1.78 1.00 N/A N/A 0.00 0.00% 2.10 2.00 N/A N/A 0.00 0.00% 2.38 2.00 N/A N/A 1.00 0.00% 2.67 4.00 N/A N/A 3.00 0.10% 2.97 5.00 N/A N/A 5.00 0.00%

9 Uncoordinated smart inverters are not able to address the thermal issue and so cannot enhance hosting capacity. 10 Due to the low levels of curtailment, storage is not required to maintain below the 3% curtailment limit.


Figure 33: Feeder D curtailment with increasing PV size


7.5 Summary of Results This section summarizes the Enhanced Hosting Capacity results across each of the studied feeders at Unity Power Factor, to compare the ability of each technology option to enhance the hosting capacity. The results from the enhanced Hosting Capacity Analysis inform the Cost Analysis presented in Section 7. The use of reactive power generally has a positive effect on all voltage related constraints (voltage limitation and rapid voltage fluctuations) and some of the other constraints. However, these should be taken with some caution since the operation of the smart inverter settings need to be considered together with other voltage regulating equipment to avoid undesirable effects, such as under-voltages should the inverter draw in too many MVArs under specific generation-load conditions. This report does not advocate for a specific reactive power control mode; the utility should perform the analysis given the specificities of their system to validate the approach does not lead to unforeseen consequences.

Table 41: Hosting capacity summary for Feeder A

Scenario Hosting Capacity Limit

Unity power factor (MW)

Limiting Constraint Curtailment %

Baseline (Static Hosting Capacity) 1.14 Voltage 0.00

Smart Inverter (no coordination) 3.90 Rapid Voltage Change 37.8

Smart Inverter (centrally coordinated) 3.90 Rapid Voltage Change 37.8

Thermal Load Shift 1.14 Voltage 0.00 Energy Storage 3.90 Rapid Voltage Change 0.00

Coordinated control of all technologies 3.90 Rapid Voltage Change 3.00

Table 42: Hosting capacity summary for Feeder B


Unity power factor (MW)


Baseline (Static Hosting Capacity) 12.5 Thermal 0.00

Smart Inverter (no coordination) 12.5 Thermal 0.00


Thermal Load Shift Between 12.5 and 14.4 Thermal 0.00 Energy Storage 24.0 Rapid Voltage Change 0.00

Coordinated control of all technologies 24.0 Rapid Voltage Change 2.90


Table 43: Hosting capacity summary for Feeder C


Unity Power Factor (MW)


Baseline (Static Hosting Capacity) 12.2 Voltage 0

Smart Inverter (no coordination) 12.7 Rapid Voltage Change 0

Smart Inverter (centrally coordinated) 12.7 Rapid Voltage Change 0

Thermal Load Shift 12.7 Rapid Voltage Change 0 Energy Storage 12.7 Rapid Voltage Change 0

Coordinated control of all technologies 12.7 Rapid Voltage Change 0

Table 44: Hosting capacity summary for Feeder D


Unity Power Factor (MW)


Baseline (Static Hosting Capacity) 0 Thermal 0

Smart Inverter (no coordination) 1.19 Rapid Voltage Change 0


Thermal Load Shift Between 1.19 and 1.78 Rapid Voltage Change 0 Energy Storage 2.97 Rapid Voltage Change 0

Coordinated control of all technologies 2.97 Rapid Voltage Change 0


8 Cost Analysis

8.1 Introduction A high-level cost analysis was conducted to compare the costs of each advanced technology scenario studied in the enhanced hosting capacity phase of the Minnesota Solar Pathways project. The costs of the advanced technologies are compared against the PV Capacity (MW) scenarios for each feeder. The goal of the cost analysis is to determine what the most cost effective advanced technology scenario is in order to achieve the highest PV capacity on each feeder. In order to compare the costs of the advanced technology scenarios, the baseline metric is the PV size to determine how much each advanced technology would cost to achieve that level of PV capacity. Table 45 provides the PV (MW) capacities that were used to compare the costs for each advanced technology scenario. Additional PV hosting capacities were used to provide more granularity of data for the cost analysis.

Table 45: PV capacities used for cost comparison of advanced technology scenarios

Feeder Distance from Substation (Miles)

Capacity Option (MW) 1 2 3 4 5 6 7

Feeder A 1.05 0.78 1.56 2.34 2.73 3.12 3.51 3.90 Feeder B 0.68 4.80 9.60 14.4 16.8 19.2 21.5 24.0 Feeder C 4.00 2.54 5.08 7.62 8.89 10.2 11.43 12.7 Feeder D 2.40 0.59 1.19 1.78 2.10 2.38 2.67 2.97 Feeder D 2.40 0.59 1.19 1.78 2.10 2.38 2.67 2.97

8.2 Assumptions The cost analysis looks at installing the various advanced technologies on a particular feeder section to determine how much it would cost using each technology to achieve the same PV capacity on that particular section. This ensures the results are comparable and there is a baseline metric to compare against. The assumptions made in this analysis are:

• All costs are provided in USD and are Net Present Value (NPV) 2018 dollars; • All costs are assumed to be costs to the State of Minnesota, but this analysis is not meant

to determine who the responsible party is for bearing these costs; • The lifetime of the project is assumed to be 10 years for the storage system, but no

operational costs are included in the analysis; • An installed cost of $541.67/kWh (USD) is assumed for Tesla Powerpacks, including

inverter, commissioning, installation and transportation costs. This is used to scale storage costs for any additional storage capacity;

• The cost of the PV system is not included in these assumptions; • The cost of curtailment is based on an assumption that the wholesale marginal cost of

energy is $40/MWh ($0.04/kWh); • The curtailment figures are provided by the enhanced hosting capacity results for each

feeder, and are inclusive of costs associated with deployment of control infrastructure;


• It is assumed that there is no associated cost with installing smart inverters to simplify the analysis. This study in no way considers remuneration or compensation for this functionality this will be determined by relevant stakeholders in the state of Minnesota. Interested readers are referred the referenced sources for more information on smart inverters;11,12

• Since thermal load shifting alone cannot provide enough additional capacity to meet any of the increased PV hosting capacity scenarios, the benefit of thermal load shifting is presented as the NPV sum of the yearly lost revenue coupled with curtailment versus using curtailment alone;13

• The data provided by CPR was used as an indicator for how much thermal load could be shifted on each feeder. This data only used electric water heaters, so this may not be fully representative of the actual thermal load available on all four feeders;

• The traditional hardware costs are based on costs provided by NREL from their Distribution Grid Integration Unit Cost Database;14

• The traditional hardware costs were determined using Synergi models of each feeder to locate the specific overloads on each feeder experienced at different levels of PV penetration, and then determine what equipment would need to be upgraded or reinforced to allow for that particular level of PV penetration;

• The cost scenarios are based on the data presented earlier in this report indicating the levels of each advanced technology that would be required to achieve that level of PV Capacity on the given feeder;

• The combined control scenario has all of the advanced technologies working together, utilizing thermal load shifting as the lowest-cost resource, then limiting curtailment to 3% of annual generation. Storage is then sized for any additional overloads that cannot be accommodated by the combined thermal load shifting and curtailment scenario.

11 Horowitz, Kelsey A. W., Fei Ding, Barry Mather, and Bryan Palmintier. (April 2018). Distribution-Level Costs Associated with Integration of Increasing Penetrations of Distributed Photovoltaic Systems. Golden, CO: National Renewable Energy Laboratory. NREL/TP-6A20-70710. 12 https://www.edockets.state.mn.us/EFiling/edockets/searchDocuments.do?method=showPoup&documentId={90170661-0000-C61A-99C8-3693818EF1F1}&documentTitle=20181-139051-01 13 Available: https://www.otpco.com/media/1338/mn_1406.pdf 14 Available: https://www.nrel.gov/solar/distribution-grid-integration-unit-cost-database.html

https://www.otpco.com/media/1338/mn_1406.pdf

https://protect-eu.mimecast.com/s/aAwjC1WoqsMGx6tAlZzA?domain=nrel.gov


8.3 Feeder A Table 46: Feeder A NPV cost comparison

PV Capacity (MW)

Traditional Hardware Cost

(CAPEX)

Smart Inverter Cost

Curtailment Cost (Lifetime)

Curtailment Cost with

Thermal Load Shifting

(Lifetime)

Storage Cost (CAPEX)

1.56 2,320,000 0.00 289,000 129,000 1,780,000 2.34 2,320,000 0.00 306,000 299,000 3,770,000 2.73 2,320,000 0.00 658,000 589,000 6,060,000 3.12 2,400,000 0.00 904,000 833,000 8,340,000 3.51 2,400,000 0.00 1,200,000 1,130,000 11,800,000 3.90 2,400,000 0.00 1,530,000 1,460,000 12,800,000

Since the primary constraint on Feeder A feeder is over voltage, and the smart inverter can resolve this constraint as a standalone option up to the 3.90 MW rapid voltage change constraint, smart inverters are by far the cheapest option in every scenario. Additionally, the $129,000 curtailment cost to achieve up to 1.56 MW of PV would require curtailing 12.5% of the total annual generation. In the highest PV capacity scenario (3.90 MW), 37.8% of the total annual generation would have to be curtailed in order to achieve the same benefit as the smart inverters.

Table 47: Feeder A combined technology NPV cost comparison

PV Capacity (MW) Curtailment Cost with Thermal Load Shifting

(Lifetime)

Storage Cost (CAPEX) Total Aggregated Lifetime Cost

0.78 123,000 246,000 368,000 1.56 145,000 1,370,000 1,520,000 2.34 168,000 3,320,000 3,490,000 2.73 181,000 5,590,000 5,770,000 3.12 191,000 7,800,000 7,990,000 3.51 203,000 11,300,000 11,500,000 3.90 214,000 11,900,000 12,100,000

Since the technologies are stacked to limit curtailment to 3%, there is still an overall cost reduction for both curtailment and storage individually; however, the aggregated cost is not a significant reduction from the individual storage costs associated with reducing curtailment to 0% of annual generation.

8.4 Feeder B As presented in Table 26, the only technologies capable of providing meaningful increases in hosting capacity are curtailment and storage. While thermal load shifting is only able to provide a minimal amount of additional hosting capacity, the NPV project lifetime savings are presented below as the cost savings of utilizing thermal load shifting before curtailment.


Table 48: Feeder B NPV cost comparison

PV Capacity (MW)


(CAPEX)

Smart Inverter Cost ($/kW)


Curtailment Cost with Thermal Load Shifting

(Lifetime)


14.4 566,000 N/A 101,900 101,700 1,212,000 16.8 2,320,000 N/A 128,000 126,000 4,630,000 19.2 2,320,000 N/A 306,000 299,000 10,700,000 21.5 2,320,000 N/A 739,000 724,000 17,400,000 24.0 2,400,000 N/A 1,480,000 1,460,000 25,500,000

Storage as a standalone option for Feeder B does not make sense in any scenario, since it is twice as expensive as the traditional hardware cost scenario to allow up to 14.4 MW of PV on the feeder, and this factor increases as PV capacity increases further. The least cost option in every scenario for Feeder B is curtailment which is $920,000 less expensive than the traditional hardware scenario and increases the PV capacity to 24.0 MW. In the 24.0 MW scenario, the lifetime savings of using the thermal load shifting available on the feeder are $20,000. In the 24 MW PV hosting capacity scenario, the $1,480,000 lifetime curtailment cost corresponds with a 5.92% curtailment of annual generation.

Table 49: Feeder B combined technology NPV cost comparison


(Lifetime)


14.4 101,700 39,800 141,000 16.8 121,000 2,230,000 2,350,000 19.2 217,000 7,400,000 7,620,000 21.5 378,000 13,500,000 13,900,000 24.0 793,000 18,500,000 19,300,000

This aggregated scenario limits curtailment to 3% of total annual generation. While combining the advanced technologies reduces the overall costs of both curtailment and storage individually in every scenario, the aggregated cost is still more expensive than curtailment or curtailment with thermal load shifting in the standalone state.

8.5 Feeder C Table 50: Feeder C NPV cost comparison

PV Capacity (MW)


(CAPEX)

Smart Inverter Cost

($/kW)




(Lifetime)


12.7 157,600 0.00 N/A N/A N/A


Beyond the static hosting capacity of 12.2 MW, limited by voltage, Feeder C is limited by the rapid voltage change constraint at 12.7 MW of PV capacity. Since the rapid voltage change constraint defines the limitations of this study, and there are no constraints experienced on the feeder before the 12.7 MW of PV, there is no curtailment, thermal load shifting, or storage that would allow for an increased PV hosting capacity on Feeder C. The use of smart inverters increases the PV capacity from 12.2 MW to 12.7 MW, but the 500 kW of additional PV capacity can also be met by installing voltage regulators. In the instance of Feeder C, the traditional hardware costs are the least expensive option up until the rapid voltage change constraint is met.

8.6 Feeder D Table 51: Feeder D NPV cost comparison

PV Capacity (MW)


(CAPEX)

Smart Inverter Cost ($/kW)




(Lifetime)


0.59 239,000 N/A 100,000 100,000 N/A 1.19 239,000 N/A 100,000 100,000 N/A 1.78 239,000 N/A 100,000 100,000 248,000 2.10 239,000 N/A 101,000 100,000 553,000 2.38 239,000 N/A 102,000 101,000 817,000 2.67 239,000 N/A 104,000 102,000 1,270,000 2.97 239,000 N/A 107,000 104,000 1,700,000

The overloads on Feeder D can primarily be resolved using the available thermal load. While the advanced technologies can increase the PV hosting capacity beyond the static limit, the 2.97 MW rapid voltage change constraint limits the capacity to install any additional PV beyond that limit. Although storage could be used to resolve the constraints experienced on the Feeder D, the lowest-cost storage option is 1.5 times as expensive as the costs associated with thermal load shifting. Additionally, even the most expensive curtailment cost with thermal load shifting is less expensive than the traditional hardware upgrade cost.

Table 52: Feeder D Combined Technology NPV Cost Comparison


(Lifetime)


0.59 100,000 N/A 100,000 1.19 100,000 N/A 100,000 1.78 100,000 N/A 100,000 2.10 100,000 N/A 100,000 2.38 101,000 N/A 101,000 2.67 102,000 N/A 102,000 2.97 104,000 N/A 104,000


Since thermal load shifting and small levels of curtailment can resolve the constraints on Feeder D up to the 2.97 MW rapid voltage change limit, there is no need for storage in the combined technology scenario. Even at 2.97 MW of PV, less than 1% of the annual PV generation would need to be curtailed or shifted, so this still meets the requirement to stay below 3% curtailment of annual PV generation.


9 Conclusion This report has documented the hosting capacity analysis performed for four feeders provided by two of the utility partners in the Minnesota Solar Pathways project. This sub-task of the larger project aimed to evaluate the potential role of technology in enabling higher levels of PV onto the distribution grid, as part of the solution for meeting the long-term energy policy objectives of the state. The project considered stand-alone smart inverters, centrally-coordinated smart inverters, thermal load shifting, energy storage, and a combined technology scenario to compare and contrast the benefits of each across the four different feeders studied. Smart inverters can be used to manage feeder voltages with variable Volt-Watt and Volt-VAr settings, but would require coordination for thermal constraints, and hence were of no value once those constraints were reached. In the circuits where voltage issues were triggered first (Feeder A and Feeder C), stand-alone smart inverters was the preferred technology, although rapid voltage change constraints were triggered in short order, particularly for Feeder C. The smarter inverter curves in the enhanced hosting capacity essentially corroborated the conclusion from the leading power factor case that, generally, this case addresses voltage-based constraints. However, these results should not be applied blindly as dynamics between the inverter operating with volt-var and the grid voltage regulating equipment needs to be carefully analyzed to determine the appropriate settings for both assets. That being said, we support the recommendation from Xcel and OTP that at least leading power factor be used in evaluation of static hosting capacity which provides value across all voltage related constraints. The aggressive smart inverter curves employed in the present study actually lend themselves to higher levels of curtailment, and these curves are outside of IEEE 1547-2018 range of allowable settings, and thus we recommend using the default IEEE 1547-2018 curves unless there is a justification for doing otherwise. At high penetration, interaction with voltage regulation equipment should be careful considered. With these curves in place, voltage concerns will give way to thermal constraints, which could then be managed using curtailment to greatly extend hosting capacity. Implementation of curtailment as a means for extending hosting capacity does require engagement between the utility, the developer, and a technology provider to implement the required functionality. At a minimum, this requires the installation of appropriate monitoring at the identified thermal constraint, appropriate control functionality at the DER site responsible for the thermal constraint, and communication between the two, as well as integration back to SCADA and/or the DMS. There are multiple technology paths to implementation of this concept but all will require field area communications and advanced control systems, possibly but not necessarily based on a network-aware operating model. These systems can be deployed either system wide or—to limit resources and capital investment—more selectively in an incremental fashion to only in those location where DER integration has become limited. These


considerations represent one part of much broader grid modernization initiatives that generally will be utility led but require discussion with multiple stakeholders. Thermal load shifting demonstrated limited benefit relative to the PV capacities considered. For the study, the overall potential was estimated based on data provided by CPR, which considered only hot water load at residential customers, and some feeders had very low percentages of residential customers. In the case of Feeder B there was a very low residential population so little hot water demand to control. While the benefit was small, demand response resources of this type are compelling as part of the overall solution. In contrast, energy storage allowed total flexibility to determine capacity; however, maximizing PV export (limit to zero curtailment) resulted in uneconomic battery sizes. While there could be other value streams to improve the economics of the battery, the current costs of the technology for this application make the justification tenuous at best. In the combined approach, the size of the storage was reduced significantly and by association so were the costs. Nonetheless, the inclusion of more reasonable storage sizes still failed to present a positive economic outlook when compared to curtailment, with or without thermal load shifting. Further work would be required to optimize between the relative costs and benefits, and to investigate the other value streams of energy storage technology, and various scenarios for cost depreciation, to gain better clarity in its overall role in the long-term strategy. The cost analysis aligns with the individual feeder analysis, and shows most importantly that there is not an optimal solution that can meet the needs of every feeder across the board. The least cost option to resolve the thermal overloads on the Feeder B was curtailment coupled with thermal load shifting, while the least expensive solution to resolve the voltage overloads on Feeder A was to use smart inverters. The static hosting capacity on Feeder C by 500 kW can be increased using a traditional hardware upgrade of voltage regulators as the lowest cost solution. Finally, the constraints experienced on Feeder D can be resolved using curtailment and thermal load shifting as standalone, least-cost solutions. The greatest gains in the near-term relate to updating the interconnection process to consider the contribution of inverter reactive power control for addressing voltage related constraints.

Enhanced Hosting Capacity Analysis - MN Solar...

Documents

Transcript of Enhanced Hosting Capacity Analysis - MN Solar...