Optimizing and Defending Capital Allocations for Distribution - - PowerPoint PPT Presentation

Optimizing and Defending Capital Allocations for Distribution Assets

Steve Bubb Metering Americas April 25, 2006

2

Load Data is critical to a number of processes & systems

Distribution Planning Engineering Operations Asset Management Transmission Planning DSM Projects Account Management

GIS OMS Circuit Analysis Tool

Load Data

Bank Feeder Protection Device Transformer Customer

Protection Analysis Transmission Load flow

3

Loading impacts transformers, wire and cable loss-of-life

Understanding asset status is key to optimization.

0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 0% 20% 40% 60% 80% 100% 120% 140% 160% 180% 200% Value % of Transformers at each loading level Traditional Approach With Customer Data

Ideal Loading Over Utilized Under Utilized Reduced Reliability Wasted Capital

4

Substation T1 T2 NO

C

C C C C SW SW SW SW SW

SCADA/Meters/Read Sheets

Circuit Amps

SCADA/Meters/Read Sheets

Bank Loads

Monthly/ daily customer usage data typically not used Interval data for large customers typically not used

Typical Existing Distribution Load Data Sources

5

Common Utility Comments

Efficiency

“I should be spending most of my time doing analysis, not gathering data.” “I would like to see quicker turn around on engineering analysis.” “We don’t have time to review all of our circuits every year.”

Accuracy and Consistency

“Each engineer has their own spreadsheets and methods.” “Almost all of our experienced planners are leaving in the next 3 years. I need some consistency in the process to makeup for lack of experience.” “We don’t use our TLM because we don’t believe the data.”

Justification

“We stopped our TLM program a couple of years ago and went to run to failure. We don’t know if there is a large scale problem looming out there or not” “We can’t afford to build for a 1 in 100 year heat wave, nor would the commission want us to. We would like to be able to show the commission what failure rate we would expect for certain defined weather conditions and decide together what we should build to withstand, rather than reacting after the

• fact. ”

6

Distribution data challenges

Issues with substation data SCADA data is either inaccessible, hard to get to or not reliable for analysis (not validated) Metering is not read often enough and/or not readily accessible Issues with data downstream of the breaker There is still relatively little measured data between the customer and the breaker Fixed network or mesh network AMR is providing more granular data at the customer level, but it is hard to manage and package for T&D Issues with TLM Data is not validated bill data is not necessarily good usage data Connectivity data quality varies significantly within the utility GIS and CIS input processes impact connectivity data quality

7

Meter data can help fill the information gap

8

Increasing the granularity of data is one key ingredient to better decision making, but weather is the driver of surprises.

9

If it was a cool year, how do you combat the “we aren’t having any loading issues” question?

Xcel Energy Feeder MEL089 Loading By Weather Scenario

2 3 4 5 6 7 8 9 10 1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161

Hours of Week Beginning July 17, 2004 Demand (MW) Extreme Typical Actual

10

Summer Cumulative Hourly Temperature Distribution Minneapolis Airport

500 1000 1500 2000 2500 3000 3500 4000 20 30 40 50 60 70 80 90 100

Temperature (°F) Number of hours exceeding temperature

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Weather scenarios can be defined and simulated

Summer Cumulative Hourly Temperature Distribution Minneapolis Airport

500 1000 1500 2000 2500 3000 3500 4000 20 30 40 50 60 70 80 90 100

Temperature (°F) Number of hours exceeding temperature Extreme 1 in 10 Scenario Typical Scenario

Example Applications

12

Typical Annual Capacity Planning Process

Gather Peak Load Data Filter Out Switching Peaks Check Load Transfers and Known Loads Load Projections Analyze Overloads Recommend Solution

• SCADA
• Manual Reads
• Substation Reads

from Interval Meters

Data Sources

• Excel
• HG

Applications

• Estimating/

Engineering

• New service

requests

• Excel
• TLM
• SCADA

Tools

• PI
• Individual meter

interfaces

• Spreadsheets/

databases

• Paper sheets/charts
• Customer
• Transformer
• Feeder
• Bank
• Mainframe
• Database
• Database
• Individual

Spreadsheets

• LDC
• Spreadsheets
• HG Application
• 3D Chart
• Circuit analysis tools

Issues

• Multiple systems
• Gaps in data
• No validation
• Labor

intensive process

• Reconciliation
• f multiple

data sources

• Consistency
• Loads not

weather normalized

• Often relies on

communications between departments

• Labor intensive
• Consistency
• Estimated load

data downstream

• f breaker
• No weather risk

analysis

13

Typical Annual Area Planning Process

Annual Growth

Substation Name Ban k # CCT # Limit Device Voltage Capacity Limit 1998 1999 2000 2001 2002 5 yr AGR of actual loads (%) Kentucky 1 813221 OH 12470 10000 8043 8164 7862 7681 9865 4.17 Kentucky 1 813223 OH 12470 10000 10281 10402 10402 10766 12784 4.45 Kentucky 1 813225 OH 12470 10000 2873 2721 2721 2721 2955 0.56 Kentucky 2 813222 OH 12470 10000 9797 9555 9313 9556 10874 2.11 Kentucky 2 813224 OH 12470 10000 7076 6713 6410 6894 8048 2.61 Pennsylvania 1 815121 UG 12470 10000 13834 10054 8845 12020 10841

• 4.76

Pennsylvania 1 815125 UG 12470 10000 10765 10765 10281 10403 10781 0.03 Pennsylvania 2 815132 UG 12470 10000 10522 10286 10523 10765 11166 1.20 Pennsylvania 2 815136 UG 12470 10000 9857 9737 9374 10887 9801

• 0.11

Pennsylvania 3 815123 UG 12470 10000 10659 11037 10281 10054 13551 4.92 Pennsylvania 3 815127 UG 12470 10000 10583 10428 12548 6290 10651 0.13 Pennsylvania 4 815134 UG 12470 10000 9827 10054 8240 8241 6190

• 8.83

Pennsylvania 4 815138 UG 12470 10000 7681 7801 7439 8345 5998

• 4.83

SW 64th Street 1 812921 OH 12470 10000 6985 7741 9676 9193 10525 8.55 SW 64th Street 1 812923 OH 12470 10000 10341 8799 8074 8345 10142

• 0.39

SW 64th Street 2 812922 OH 12470 10000 10070 10160 9616 9979 10875 1.55 SW 64th Street 2 812924 OH 12470 10000 10039 9918 9737 10039 13816 6.60 TOTALS 159233 154335 151342 152179 168863 1.18 Circuit Load History** Circuits

14

Kentucky 224 Load Growth Projection

5000 5500 6000 6500 7000 7500 8000 8500 9000 1998 1999 2000 2001 2002 2003 2004 2005 Year kVA

Typical Annual Area Planning Process

A growth rate is determined by taking the peak kVA (or Amp) reading for each year and calculating a trend

15

Assess Weather Impacts

Substation Name Ban k # CCT # Limit Device Voltage Capacity Limit 2002 Peak Time Stamp 2002 Base Load 2002 Cooling Load 2002 Heating Load Typical Extreme 1 in 4 yr Very Extreme 1 in 10 yr Kentucky 1 813221 OH 12470 10000 9865 8/21/02 18:00 2615 7250 11646 12430 14240 Kentucky 1 813223 OH 12470 10000 12784 8/22/02 17:00 4605 8179 14664 15225 17116 Kentucky 1 813225 OH 12470 10000 2955 8/26/02 17:00 2454 502 2986 2982 3013 Kentucky 2 813222 OH 12470 10000 10874 8/21/02 16:00 3295 7580 12575 13354 14712 Kentucky 2 813224 OH 12470 10000 8048 8/21/02 16:00 2857 5191 9407 9608 10908 Pennsylvania 1 815121 UG 12470 10000 10841 8/21/02 18:00 3446 7395 12409 13490 15242 Pennsylvania 1 815125 UG 12470 10000 10781 8/1/02 16:00 4107 6674 12596 13242 14898 Pennsylvania 2 815132 UG 12470 10000 11166 8/21/02 18:00 4652 6514 12427 13599 14529 Pennsylvania 2 815136 UG 12470 10000 9801 8/21/02 18:00 2726 7075 11071 12014 13536 Pennsylvania 3 815123 UG 12470 10000 13551 8/21/02 16:00 6123 7428 15683 15476 17637 Pennsylvania 3 815127 UG 12470 10000 10651 8/21/02 17:00 3806 6845 12319 13161 14560 Pennsylvania 4 815134 UG 12470 10000 6190 8/21/02 17:00:00 2298 3892 6816 7533 8055 Pennsylvania 4 815138 UG 12470 10000 5998 8/21/02 17:00 2213 3785 6878 7337 7964 SW 64th Street 1 812921 OH 12470 10000 10525 8/21/02 18:00 2719 7685 12329 13300 14928 SW 64th Street 1 812923 OH 12470 10000 10142 8/21/02 17:00 3188 7229 11616 12788 14223 SW 64th Street 2 812922 OH 12470 10000 10875 8/21/02 17:00 5282 6999 12162 13335 14332 SW 64th Street 2 812924 OH 12470 10000 13816 8/21/02 18:00 5239 10895 15322 17327 19273 TOTALS 168,863 61625 111118 192906 206201 229166 160,420 8/1/02 18:00 Weather Pattern Forecasts Timestamp of Coincident Peak Circuits 2002 Coincident Loads at Area Peak

16

Assess Weather Impacts Now weather scenarios can be applied to your base load trends to assess weather impacts. Develop a planning criteria based on which weather scenario warrants capital investment.

Kentucky 224 2003 Forecasts

7076 6713 6410 6894 8048 8048 9407 9608 10908

4000 5000 6000 7000 8000 9000 10000 11000 12000

1998 1999 2000 2001 2002 2003*

Year kVA

Actual Typical 1 in 4 yr 1 in 10 yr

17

Knowing how much and for how long for alternate weather scenarios provides insight into what the solutions should be

18

Data provides the ability to estimate loss-of-life and prioritize replacements

With the availability of hourly loading and ambient temperature information, transformer hot spot temperatures can be estimated using the equations defined in the IEEE Guide for Loading Mineral-Oil-Immersed Transformers (IEEE Std C57.91). IEEE equations reasonably predict actual hot spot temperatures for a specific transformer. Calculating these hot spot temperatures enables estimates

• f the transformer loss-of-life for any combination of loading levels and ambient

temperatures.

Transformer Hot Spot Temperature Assuming a 145% Overload

0.00 °C 20.00 °C 40.00 °C 60.00 °C 80.00 °C 100.00 °C 120.00 °C 140.00 °C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hours Temperature °C Ambient Actual IEEE Std Calculation

19

# Over H Number of Overload Hours # Over E Number of Overload Events # Over D Number of Overload Days Avg O/Day Average Overload Hours per Overload Day

Knowing when and how many times and how long your asset is

• verloaded adds valuable information to your prioritization process.

Applying weather and growth what-if scenarios will predict magnitude and conditions of “looming” issues.

20

Targeting failed transformers for replacement

Targeted replacements reduce failures at lower cost than complete replacement Key decision is tradeoff between reduced failures and number of proactive replacements Predictive ability is improving rapidly

21

Customer level models make a difference

Custom er Data Hourly Profile Perform Regression Analysis Load Equation Daily Usage Customer 1 Customer 2 Customer N

What was the actual utilization for 2003? Transformer Load Profiles How much of my load is weather related? What was my real baseline growth? Transformer Load Equations 1000+ customer profiles matched to individual customer characteristics

100+ validation tests Assign Library Profile

22

Every customer is different

Utility bills of 1 customer Slope indicates heating sensitivity Slope indicates cooling sensitivity 63 ºF balance temp 69 ºF balance temp

23

Customer level models allow the establishment of growth trends for each customer

Calculate % deviations of bills vs. Model Mean deviation as 12-month moving geometric mean Growth trend of 4.5%/yr

Growth Trend Extracted from Three-year Model

5000 10000 15000 20000 25000 30000 35000

9/1/01 9/1/02 9/1/03 8/31/04

kWh per Bill

• 5.0%

0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% Bills Model Mean Deviation Average Growth = 4.6%/yr % Deviation

24

Transformer Growth Rates

Transformer growth rates vary widely!

Transformer Load Growth Rate Distribution, Growth Rates (N=5,788)

0% 5% 10% 15% 20%

• 20%
• 15%
• 10%
• 5%

0% 5% 10% 15% 20% 25%

Annual Load Growth Rate (% ) % Transformers in 1% bin

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Cumulative % Transformers

Transformers per 1% bin Cumulative # of Transformers

25

Commercial Growth Rate (N=2,901)

20 40 60 80 100 120 140

• 100%
• 50%

0% 50% 100% 150% 200% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Num CumNum

Growth Rates by Sector

Median residential growth is 1.8%/yr (3.3% if weighted by energy Median commercial growth is -0.5%/yr (- 0.1%)

Residential Growth Rate (N=40,240)

200 400 600 800 1000 1200 1400 1600

• 100%
• 50%

0% 50% 100% 150% 200% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Num CumNum

26

Feeder Energy and Growth Rates

High growth for some feeders – let’s drill down to FEDE2726 (growth=9.84%/yr)

Feeder Energy and Growth Rates 20,000,000 40,000,000 60,000,000 80,000,000 100,000,000

G L E N 1 9 1 3 G L E N 1 9 1 6 G L E N 1 9 1 5 F E D E 2 7 2 4 F E D E 2 7 2 8 G L E N 1 9 1 1 G L E N 1 9 1 8 F E D E 2 7 2 3 G L E N 1 9 1 2 F E D E 2 7 2 5 F E D E 2 7 2 2 F E D E 2 7 2 1 F E D E 2 7 2 7 G L E N 1 9 1 4 G L E N 1 9 1 7 F E D E 2 7 2 6

Annual Energy (MWh)

• 6%
• 4%
• 2%

0% 2% 4% 6% 8% 10% Annual Growth Rate

AnnualMWh AnnualGrowth

What’s with FEDE2726?

27

3-meter, 500 kVA Transformer

Actual Bills (M Wh) vs. Split M

• del for Xform #13513861
• 50

100 150 200 250 300 350 9/1/01 9/1/02 9/1/03 8/31/04 MWh

Actual Bills 2003 Model 2002 Model What causes this jump in March 2003?

28

Largest of 3 Customers

Model was enhanced to detect and quantify sudden changes in usage Significant increase or decrease can be automatically flagged within weeks

Actual Bills vs. Three-year Model

20000 40000 60000 80000 100000 120000 140000 160000 180000

9/1/01 9/1/02 9/1/03 8/31/04

kWh per Bill

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Bills 2002 Model 2003 Model Loadfactor

What causes this jump in March 2003?

29

Feeder Growth Rates, Revised

Treat sudden, large growth (positive or negative) individually After filtering out sudden, large growth, underlying growth rates are in the single %

Feeder Energy and Robust Growth Rates

20,000,000 40,000,000 60,000,000 80,000,000 100,000,000

G L E N 1 9 1 3 F E D E 2 7 2 2 G L E N 1 9 1 2 F E D E 2 7 2 4 G L E N 1 9 1 5 F E D E 2 7 2 1 F E D E 2 7 2 5 F E D E 2 7 2 3 G L E N 1 9 1 1 F E D E 2 7 2 6 F E D E 2 7 2 8 G L E N 1 9 1 8 G L E N 1 9 1 6 F E D E 2 7 2 7 G L E N 1 9 1 7 G L E N 1 9 1 4

Annual Energy (MWh)

• 6%
• 4%
• 2%

0% 2% 4% 6% 8% 10%

Annual Growth Rate

AnnualMWh AnnualGrowth

30

Data provides the opportunity to transform average to precise

Traditional

Annual or monthly peak load estimates for feeders Device and segment peaks estimated No weather or time correlations No weather or growth risk analysis Multiple un-reconciled data sources

Possible

Hourly loads aggregated to system devices Single repository and reporting for enterprise- wide, true coincident distribution data Weather and growth risk analysis applying Individual customer load characteristics Quantify risk by knowing hours of overload

Circuit 1101 Month Peak KVA January 38 February 41 March 43 April 48 May 52 June 55 July 57 August 59 September 56 October 49 November 44 December 41

Annual Peak

31

Indications are that more precise information provides insights that save money and increase reliability

• Deferred 46% of area capital budget in the

first year (pilot area)

• Deferred 25% of area capital budget for 5

years (pilot area)

• Replaced only 1/3 of the overloaded

transformers based on the additional qualitative data

• Changed the allocation of feeder loads a

minimum of 33% and a maximum of 79%

• Reduced transformer outages 75% with

proactive replacements

• Generated over 100 work orders in one

year to fix overloaded protection devices avoiding untold number of outages

• Assessed and changed engineering

standards based on new information readily available

Increase Reliability Decrease Costs

32

Precise, accurate and consistent data combined with intelligent analytics can achieve key

• bjectives such as:

Increase regulatory and internal confidence in capital allocations with defensible consistent processes for evaluating growth and weather risk for large and small capital projects Reduce subjectivity in distribution planning and asset management processes by automating data reconciliation and complex calculations Increase engineering consistency and productivity therefore shortening analysis turn around time Predict and locate asset loading issues heading off future customer, regulatory and budget surprises

Thank You!