Albany Expeller project (2)

26
Reliability & Integrity Group Continuous Improvement Project : Albany, Repeller pump, leakage Paul Richardson, Stewart Stephen, HMC SERT Team.

Transcript of Albany Expeller project (2)

Page 1: Albany Expeller project (2)

Reliability & Integrity Group Continuous Improvement Project : Albany, Repeller pump, leakage

Paul Richardson, Stewart Stephen,HMC SERT Team.

Page 2: Albany Expeller project (2)

2

Project StructureProject Structure

● DMAIC

DMAIC (Define, Measure, Analyse, Improve, Control) is a Lean/Six Sigma tool, which focuses on process improvement. With strategic goals in mind, a steering team identifies areas/processes to focus on. An improvement project is then initiated.

Through a rigorous methodology, a project team determines customer needs (DEFINE), and gathers information on the current (as-is) process performance (MEASURE), which leads to the identification of issues and problems.

In ANALYSE, root causes are identified and confirmed with data. The development and implementation of solutions (IMPROVE) are followed by actions to ensure that the improved process is standardised and gains are maintained(CONTROL).

DMAIC is supported by quality tools and advanced statistical tools, which facilitate fact-based decision making.

Page 3: Albany Expeller project (2)

3

Contents1. Define – What’s the problem?

2. Measure – Is it really a problem?

3. Analyse – What can we do about it?

4. Improve – What we did about it

5. Control – Did it work?

6. What’s the Forward plan?

7. Frequently Asked Questions

8. Appendices

Page 4: Albany Expeller project (2)

4

Define● What?

- Lost production or lost opportunity due to Albany Impkem mk5 centrifugal pump failures

● Where?

- Huddersfield Manufacturing Centre

● When?

- Data available from 2002 - Present

● Significance?

- Lost production (“breakdown” and “start now” SAP data) : 213 occurrences is an estimated 54 days downtime Lost sales cost from $250K - $600K per day dependent on which plant is affected. It is not known how affected sales have been but

at worst it equates to $4.5M/year.

- Lost opportunities (“breakdown” and “all others” SAP data) : 313 occurrences is an estimated 52 days downtime $2.2M/year potential sales

- HS&E : The Albany repellor/lip seal sealing arrangement has a lip seal “wear out” rate of months, therefore any stopped pump will leak after this period. The risks of a loss of containment are high for Intermittently operated equipment and where unexpected shutdown is needed.

Losses of containment threaten our license to operate and damages our reputation.

Page 5: Albany Expeller project (2)

5

● Select 3 Trial pump berths;

o MI - 66/213 (M062623)

o R5 - 52/117 (M062696)

o R5 - 41/212 (M062680)

● Trial duration (6–8 months)

● Special instruction for the SERT Team to stop and monitor on a weekly frequency.

Define

Page 6: Albany Expeller project (2)

6

Primary Measure

The above is a control chart of three HMC Albany pumps. They have been selected as a potential trial because they operate continuously but can be stopped for short periods to assess the operation of any seal under trial. A successful trial would be one that either ran in excess of 7000hrs or had 3 failures in succession above the MTBF value of 1700hrs

Page 7: Albany Expeller project (2)

7

Secondary Measure

The above control chart is compiled from SAP data. An Albany pump has a breakdown intervention, on average, every 3 days. The HMC equivalent for centrifugal pumps is 10 days.

Special cause failures have not been investigated.

Page 8: Albany Expeller project (2)

8

Albany Impkem Mk V Pumps

1. Simple.

2. Over Engineered and Robust

3. Mechanically reliable

4. Better than any other HMC pumps in terms of leakage whilst running

Analyse

Page 9: Albany Expeller project (2)

9

AnalyseRepellor

• Acts like an impellor – “Throws” product into the outer casing preventing leakage down the shaft.

Page 10: Albany Expeller project (2)

10

AnalyseGarlock Lip seals

• Two Lip seals onto the pump shaft prevent leakage whilst the pump is stopped

Page 11: Albany Expeller project (2)

11

Analyse10TC Mechanical Seal

• Used to prevent leakage whilst the pump is stopped

• Metal “wings” use centrifugal force to pull the seal away from the stationary face whilst the pump runs

Page 12: Albany Expeller project (2)

12

Analyse Failure due to lip seal wear

Root causes:

The current seal is dry running, has a small sealing area, and is constantly pressed against the shaft

Page 13: Albany Expeller project (2)

13

AnalyseSolution

Use a “lip seal”;

● With self lubricating properties● A large surface area to seal

against● That pulls away from the

sealing area when running● Uses the product to increase

the sealing force when stationary

- The Flowserve FXP seal had this potential

Seal Chamber

Gland

Rotating Assembly

GF PTFE Disk

Page 14: Albany Expeller project (2)

14

Analyse● We needed to prove the potential before we introduced to site.

Expeller seal trialHeld at Flowserve Test facility, Newark

Page 15: Albany Expeller project (2)

15

Analyse

24th March to 30th March 2009 - Flowserve Team

● Ran 9 hours, had 20 stop/starts varying the delivery pressure from +6m to – 9m head. A Small weep after initial set-up then ok. The rotating assembly was pulled away from the stationary face twice and re-set.

● Pump ran 15 hrs overnight at 2 m head, then subjected to the same stop/start tests as the 24th.

● Pump ran 24 hours a day at 2m head – Stopped over the weekend.

31th March / 1st April 2009 – Flowserve/Syngenta Team

● In addition to stop starts of 24th we removed the stationary seal for inspection, witnessed a number of stop starts - varying the delivery pressure from +6m to – 10m head. No leakage.

Page 16: Albany Expeller project (2)

16

AnalyseObservations:

● Only one test showed leakage from the seal, this was stopped by pushing up the rotating assembly against the PTFE

● Further testing showed that if the rotating assembly is pushed too far, it cuts its way into the PTFE face, bedding in very quickly, still maintaining its seal

● These tests have showed that the seal is capable of doing its job under test conditions (varying from + 6 to -10m head pressure) while pumping water – this warrants an on-plant test

We were unable to make the pump cavitate. At a negative head of 10m, it appeared from the readings that air was being drawn in at the seal face “dampening” the cavitating effect.

Page 17: Albany Expeller project (2)

17

Implement

Under the Syngenta “Management of Change” system;

● 9th July 2009 : 66/213 - A fully overhauled unit was installed

● 30th July 2009 : 41/212 - Seal modification done on site after vibration

analysis verified the pump in berth was ok

● 6th August 2009 : 52/117- Seal modification done on site after vibration

analysis verified the pump in berth was ok

Page 18: Albany Expeller project (2)

18

Control

FXP fitted

Page 19: Albany Expeller project (2)

19

Control

FXP fitted

Page 20: Albany Expeller project (2)

20

Control – External Validation (ABB)

Currently with Flowserve design but with the contingency of doing in-house tests

Procedure already in use – to document and make official.

Page 21: Albany Expeller project (2)

21

ControlSeal certified for zone 2 by Flowserve – waiting documentation (expected July 2010)

Waiting Flowserve temperature documentation (expected July 2010) – no problems anticipated; There is no evidence to support that the disk melted

Flowserve To document (expected July 2010)

Ongoing at Flowserve

Page 22: Albany Expeller project (2)

22

Control

Covered in 3.3

Covered in 3.3

Seal beds in straight away by cutting action (observed by trial) and temperature covered by 3.4

Page 23: Albany Expeller project (2)

23

Forward Plan

● Modify the spares – there are currently 19 variations, this will be reduced to 5 (but 8 spares) Completed Jan 2010

● Remove the 11 site bad actors – overhaul, Modify to FXP and replace immediately. Completed Mar 2010

● The other pumps on site will be change out, as they fail. To facilitate this;- Prepared “pre-assessed” Management of Change forms for;

1. Non hazardous applications

2. Flammable applications

3. Corrosive and biological applications

This has the potential to delay repairs and will require managing

MoC’s in place and being applied successfully - June 2011

30% of population modified – Jan 2012

Page 24: Albany Expeller project (2)

24

Forward PlanFlowserve Costs  

FXP Seal £450OEM shaft £380

Guard £35Pump Modification £626

Year 1   

Purchased components - made stock-able and free issued to Flowserve as each existing seal fails - any future FXP failure replaced by Flowserve as part of the healthcare contract

40 £34,600

"On shelf spares" needing modification 8 £4,768Pumps on the current bad actor list to be modified straight away 11 £6,556

TOTAL £47,124Year 2    Purchased components - made stock-able and free issued to Flowserve as each existing seal fails - any future FXP failure replaced by Flowserve as part of the healthcare contract

20 £17,300

TOTAL   £17,300Year 3    Purchased components - made stock-able and free issued to Flowserve as each existing seal fails - any future FXP failure replaced by Flowserve as part of the healthcare contract

10 £13,150*

TOTAL £13,150

Notes The HMD project should "recycle" some FXP seals therefore this projection takes into account the need to buy only 70 seals instead of the potential 104. We may need to buy more in year three, but consider the risk acceptable

* Single FXPs  seals  are approx £900 - the current price is based on purchasing lots of 20

Re-engineered shafts are £250 ea, but using OEM parts are a guiding principle of the HMC pump strategy

The price in year 2 and 3 are not fixed and may rise with inflation, etc 

Page 25: Albany Expeller project (2)

25

FAQ● How does this modification impact ATEX? The current design has already been risk assessed, The function

is no different (i.e. it’s still essentially a lip seal). The FXP seal is now certified Zone 2 by Flowserve.

● We are changing a lot of pumps on the site how confident are you of the changes? External verification by ABB undertaken

● Aren't we changing all the Albany Mk5 pumps to HMD anyway? No. At least 60 pumps, plus spares will be converted to FXP....... But, single seals, of any type, are not good practice on flammable and corrosive duties. A further 36 pumps, in use, on site fall into this category. There is currently a project to trial three HMD canned pumps, in with the EDO, the trial will prove the higher integrity pump option. When successful, the roll out of these replacements will be over a number of years, therefore the interim solution will be to install the improved FXP option to these also.

● The strategy is to change the pumps as the existing ones fail. How do you manage the “urge” to prematurely change a pump for it’s increased reliability? If the pump isn't broken you'd hope this wouldn't happen but there is a risk. We are therefore going to pull and modify the 11 site bad actors straight away. The practice of replacing good pumps will have the effect of dropping the MTBF which will ultimately increase the monthly cost. It also increases lead times on pumps in the workshop potentially delaying your important pumps returning to site.

● Where there any problems encountered in the trial? Yes. There were a number of leaks caused by installation issues. A setting collar is now fitted to address this. Update ...There have been a number of infancy failures during the Roll out. an Apollo RCA was carried out to understand why. The root causes drove the implementation of more robust installation and commissioning procedures and a review and modification of the setting distance.

● Will my monthly bill go down if the pump is more reliable? Yes the proposed charge is £189/year/pump. The current average is £470.

Page 26: Albany Expeller project (2)

26

Appendices

There were special causes associated with the trial pumps – Working with ABB, these were removed from the Primary measure