DRA/tcr Page 1 of 7

APPENDIX B

Summary of DRA Analysis of Late-Filed ALJ Exhibits

A. Estimation of hydrotest costs used to calculate disallowances for pipe replacement projects

1. The PD uses an overly simplistic estimate of hydrotest costs to adjust pipe replacement project costs

Late filed exhibits show that the PD used a constant value of $95.8 per foot for hydrotesting to determine the disallowance for pipe replacement projects,1 regardless of whether the disallowed test project is 3 feet or 3,000 feet long.2 The PD does not describe how the “estimated cost of pressure testing” should be calculated, but DRA determined that the $95.8 value is PG&E’s estimated total cost for all hydrotests, less a correction for removed AFUDC, $396.1 million, divided by the total length of hydrotests proposed by PG&E, 783.3 miles.3 This method provides a very different value from the “average cost per foot of PG&E’s hydrotest projects”, which is $289.78, and the median cost per foot for the 165 hydrotest projects, $149.23.4 As discussed below, these differences illustrate the significant inaccuracies generated if any single estimate for the hydrotest cost per foot is applied indiscriminately to all replacement projects.

This section describes why the PD is inconsistent with the explicit and implied goals to have PG&E shareholders pay for costs caused by PG&E mismanagement, and ratepayers pay where they receive an incremental benefit for new pipe.5

As a preliminary point, it should be clear that the adjustment of replacement costs should be consistent with the costs adopted for hydrotesting elsewhere in the PD. If the Commission adopts PG&E’s hydrotest costs as reasonable, despite evidence provided by DRA to the contrary, it must use the same cost estimates when adjusting pipe replacement costs.

2. Hydrotest costs per foot are inversely proportional to project or test section length, and are much higher than $95.8 for “short” pipes

The first point supporting an alternative estimation method is that PG&E’s testimony showed that hydrotest costs per foot vary widely, from $47 to $2,646 per foot, 6 and DRA explained that this is a direct result of PG&E’s inflated and unsupported fixed hydrotest costs,

1 Italics are used throughout this Attachment to focus on important differences, such as “program” vs. “project” data, and “hydrotest” length vs. “replacement” project length. 2 Late filed exhibit ALJ-1, Table 2, heading for the ninth column from the left. 3 Late filed exhibit ALJ-1, Table 3. Total value of “Gross project cost…. with AFUDC removed” divided by “Total Project Footage”. These summations, which should be shown in Line 168 of the table, were calculated by DRA. 4 The average developed by PG&E and used in the PD considers PG&E’s hydrotest program as a whole, and obtains an average price at the program level ($396.1 million/783.3 miles/5,280 feet per mile). There is no median value of the program level because there is only one hydrotest program proposed. Statistics can also be calculated at the project level, as was done in the cited data above. In this case, the cost per foot for each project is calculated, and average and median statistics can be calculated on the 165 project level costs per foot. 5 PD, p.62. 6 Exhibit 9, p.3-42.

F I L E D11-16-1204:59 PM

DRA/tcr

which adhydrotestThis resuthe mediacost of a simplificfoot are p

Inof a projehydrotestof each p

T

a cost per

3

Gcompare projects i

7 Exhibit $715,000 test heads8 For examapproxim9 Data frodivided by10 Data froproposed

dd a minimumt program inults in the sigan values is hydrotest, pation which primarily depn addition, thect, as fixed t costs per fo

project:9

This chart shor foot higher

. The a17 timestab

Given this shthe length o

illustrated ab

144, TCR tesminimum, in

s, Exhibit 2, p

mple, project ately $45.6 p

m late filed ey “Total Proje

om late filed ehydrotest pro

m of $715,0ncludes a fewgnificant diffshown abov

particularly ufails to inco

pendent on phe record shocosts are spr

oot for PG&

ows that shor than $95.8.

average lenmes shorterblish the $9

owing that hof PG&E’s pbove. For hy

timony, pp. 4

ncluding $500p.3E-17 for 12

L-187TEST ier foot.

exhibit ALJ-1ect Footage”

exhibit ALJ-1oject.

P

00 to every tw projects thfferences betwve. Thus, theusing PG&Eorporate evidproject and sows that hydread over a lE’s 165 prop

ort test projec.

ngth of pipr than in th5.8 value u

hydrotest cosproposed repydrotesting, t

46-48 and 64-0,000 for Mob2” and under p

is a 39.2 mile

, Table 3. Vafor each prop

1, Table 3. M

Page 2 of 7

test regardleat are very lween the tw use of any s’s inflated fi

dence from thsegment lengdrotest costs larger lengthposed hydro

cts, includin

es in pipelihe hydroteusing in the

sts depend olacement prothe median p

-65. PG&E’s b/Demob, $20pipes.

e project with

alue of “Grosposed project.

Median value o

ess of lengthong and thu

wo methods osingle value ixed hydrotehis proceedigth. per foot dec

h of pipe. Thotests project

ng all those s

ine replaceest projectse PD

on test sectioojects with tproject lengt

estimated fix00,000 for Mo

h a PG&E esti

s project cost.

of “Total Proj

.7 PG&E’s ps have a low

of calculatingto represent

est costs, is aing that hydr

crease with ihe followingts, as a funct

shorter than

ement projs which wer

on length, it ithe length ofth is 9,195 fe

xed costs per ove Around,

imated cost of

t…. with AFU

oject Footage”

proposed w cost per foog the averagt the estimata gross over rotest costs p

increased leng chart showtion of the le

10,000 feet,

ects is nearre used to

is necessary f hydrotest eet.10 A

hydrotest areand $15,000

f $9,681,000,

UDC removed

” for each

ot.8 e and ed

per

ngth ws ength

have

rly

to

e for

, or

d”

DRA/tcr

histogram(x axis):1

Tand 58 prshown ab

Inover 17 t

11 Each prThe final long. 12 9,195 fe“Total Pro

m of the 16511

This chart shorojects are lobove, drive tn comparisontimes less tha

roject length bbin is an exce

eet above divoject Footage

hydrotests s

ows that apponger than 2the average cn, the mediaan the media

bin is 500 feeeption, which

ided by 509 fe” for each pro

P

shows the nu

proximately 20,000 ft. Thcost per footan length of Pan length of

et wide, such th includes all

feet. Data frooposed replac

Page 3 of 7

umber of tes

2/3 of the hyhese long prt used in the PG&E’s prohydrotest pr

that the first b25 hydrotest

om late filed ecement projec

sts (y axis) w

ydrotest projrojects, whicPD.

oposed replarojects:12

bin includes at projects that

exhibit ALJ-1ct.

within each p

jects are longch have a low

acement proj

are project 0 –t are longer th

1, Table 2, me

project length

ger than a mw cost per fo

ects is 509 f

– 500 feet lonhan 20,000 fee

edian value of

h bin

mile, oot as

feet,

ng. et

f

DRA/tcr

T

1,000 feeN

PG&E oron the 14in Sectiofeet, and

This d

projects, use for re

4

To caldeterminA standaPG&E’s

13 Costs arMAOP va14 By coin15 This eqoption.

This chart shoet long.

Not all replacr DRA’s cos42 replacemen II.A.2(a) othe frequenc

data shows thand that the

eplacement p

. The cproje

lculate a morne if a simplerd MS Exceestimated hy

re disallowedalidation, or a

ncidence, the

quation was ge

ows that the

cement projest allocation ent projects wof DRA’s Cocy distributio

hat PG&E pi single estimprojects, the

cost per fooect length w

re accurate ee equation col function yiydrotest cost

d for segmentsa test date afte

50 percentile

enerated using

P

majority of

ects have segcriteria are uwith disallowomments on on are simila

ipe replacemmated hydrot

majority of

ot can be awith minim

estimate of pould be usedielded a powts:15

s with missiner 1955.

occurs in the

g the “trendli

Page 4 of 7

proposed rep

gments that aused. DRA wances usinthe PD.13 T

ar to the full

ment projectstest cost of $f which have

ccurately cmal addition

pressure test d to estimate wer series equ

g installation

e same 500 to

ine” function

eplacement t

are disallowetherefore peg the criteria

The median vl set of repla

s are much s$95.8 per foo

short pipe l

calculated nal effort

costs, a curvhydrotest co

uation that p

n date or MAO

o 1,000 foot b

of MS excel,

tests (84 of 1

ed, regardleserformed a sia of Scenariovalue of thes

acement proj

shorter than hot is inappropengths.

as a functi

ve fit was peost based onprovides a re

OP validation

in.

and selecting

168) are less

ss of whetheimilar analyso 3 as descrise projects, 5ects:14

hydrotest priately low

ion of test

erformed to n project lengeasonable fit

n data; partial

g the “power”

than

er sis ibed 512

w for

gth. to

”

DRA/tcr

T

green genline equafunction more accparticularreplacemestimatio

5

Destimate function pipeline rindependlatter bei

Tfootage fmore accdisallowahigh for s

16 For exawhile the 17 The proproposed $715,000 pipe, fixed18 See DR

This chart shonerated usination” shownof the hydro

curate estimarly true for s

ment projectson method un

. Adjuthe es

DRA used thethe costs forgiven abovereplacement

dent variableng for two p

These projectfor each replcurate methoance increasshort project

ample, PG&Eequation show

ojects are L-3by PG&E waminimum ared costs of $1,

RA Comments

ows PG&E hg the equatio

n in the chartotest project ate of hydrotshort pipe se as describednderestimate

ustment to tstimated co

e more accurr replacemene (y=33693xt project, usie “x.” These projects that t-specific hydlacement prood of estimates from $16ts, the disallo

estimates thewn in the cha

01GREPL anas 325 feet lone applied. If P,040,000 wou

s on the PD at

P

hydrotest coon shown, ant, Y=33693*length, “x”.

test costs thaections, whicd above. It ses the cost pe

the disallowost of hydr

rate, and slignt projects w

x-.576) was useing the total festimated hyare each 3 fedrotest costs

oject to get thting hydrotes.5 million toowance is st

e cost of 325 art yields a co

nd DFM-2412ng, which proPG&E’s hydr

uld be divided

t Section II.A

Page 5 of 7

st per foot dnd the PD’s *X-.576, givesThe chart sh

an the singlech constituteshould also ber foot of the

wance for protesting

ghtly more cwith disallowed to calculafootage of eydrotest cost

feet long.17 s per foot wehe disallowast costs and to $32.3 millitill less than

foot long testost of $1,204.1

2-01REPL. Novided 100 timrotest cost mo

d by 3 ft for a

A.2(b).

data in blue, fixed cost o

s the project hows that th

e figure prop the majoritybe noted thae shortest hy

pipe replac

complicated,wed segmentsate a hydroteeach replacets range from

ere then mulance per replthe PD’s dision.18 Note tPG&E’s pro

t DFM-7204-19/foot.

Note that the smes the lengthodel we used cost per foot

calculated cof $95.8 in re

cost per foohis equation yosed in the Py of PG&E’

at even this mydrotest proj

cement pro

, method diss. Specificaest cost per fment project

m $62/ft to $

ltiplied by thlacement prosallowed foothat while thoposed repla

01TEST to b

shortest hydroh on which fifor L-301G, of $346,666/

ost per foot ed. The “Treot, “y”, as a yields a mucPD. This is s proposed

more accuratects.16

ojects based

cussed abovally, the curvfoot for eacht as the

$17,894/ft, th

he disallowedoject. Using

otages, the tohe cost per foacement cost

e $2,550.77/f

otest project ixed costs of which is a 30/ft.

in end

ch

te

d on

ve to ve fit h

he

d this

otal oot is ts.

foot,

0”

DRA/tcr Page 6 of 7

For example, PG&E’s proposed cost for replacement project L-301G, the 3 foot long project mentioned previously, was $105,000.19 The PD proposes a de minimus disallowance of $ $287.40 (3 ft x $95.8/ft),20 while the more accurate method proposed herein would disallow $53,683 (3 ft x $17,894/ft). The more accurate method proposed by DRA still leaves ratepayers funding $51,317 for the new pipe.

B. Illustrations of Errors in PD cost allocation

Below is an example of pipe segment data for line (Route) L-220, duplicated from DRA testimony:21

This table shows the data used by PG&E and DRA to allocate costs between PG&E

shareholders and ratepayers.22 Column S shows that the three bottom segments were originally included in PG&E’s hydrotest project L-220 TEST, and DRA confirmed that these three segments are still planned for hydrotest at ratepayer expense per the PD.

Data for segment 140, row 15, illustrates one of the types of cost allocation errors raised by DRA: segment tested after 1955 without complete test records. Test records were not found for all 1,199 feet of this segment (Partial Mileage in Column P) which was tested after 1955 (Column L). This data indicates that segment 140 was likely hydrotested when segment 139.5 was replaced in 1962 (Row 14), which was after GO112 was adopted. The record in this proceeding supports that this test was originally funded by ratepayers, and that it should not be charged again because PG&E cannot account find complete test records.

Examples of the remaining cost allocation errors are noted below, and can be seen in DRA’s workpapers supporting these comments:

1. Segments missing installation date charged to ratepayers: Project DFM-0141-01TEST,

segment 180.23

19 Exhibit 8, PG&E workpapers supporting Chapter 3, p.WP3-232. 20 Late filed exhibit ALJ-1, Table 2, line 74, column titled “cost of disallowed footage equivalent hydro (ave test $98.5/ft).” 21 Exhibit 144, Figure 5, p.33. Column S data from Figure 6, p. 38. 22 This table doesn’t show the footage per segment which is actually used to allocate costs for the segment between PG&E shareholders and ratepayers. 23 DRA Hydrotest Workpaper, Scenario 1. “Calculations” tab, Row 1008 shows that the PD charges this segment to ratepayers, even though line 1008 of the “Input Data” tab shows the installation date is missing.

DRA/tcr Page 7 of 7

2. Segment installed after 1955 with missing MAOP Validation status charged to ratepayers: Project L-138, Segment 114. 24

3. Segment installed after 1955 with “partial” MAOP Validation status charged to ratepayers: Project L-300B, segment 427. This 3.73 mile long, 34” diameter segment was installed in 1957, and tested in 1965, after GO 112 was in place.25

The above examples are all for hydrotest projects, but similar examples exist for

replacement projects. DRA opening testimony provided a summary of the overall scope of these issues.26

24 DRA Hydrotest Workpaper, Scenario 1. “Calculations” tab, Row 417 shows that the PD charges this segment to ratepayers, even though line 417 of the “Input Data” tab shows the MAOP validation field (Column L) is blank. DRA’s understanding of the PD’s cost allocation logic is that this 2,580 foot segment was incorrectly allocated to ratepayers based on a “Test_Pressure” of 1,468 [psi] (Column O). 25 DRA Hydrotest Workpaper, Scenario 1. “Calculations” tab, Row 3155 shows that the PD charges this segment to ratepayers, even though line 3155 of the “Input Data” tab shows “Partial Mileage” for MAOP status (Column L) and a 1965 Test Date. 26 Ex. 144, p.83.