CIRED

18th International Conference on Electricity Distribution

Turin, 6-9 June 2005

TEMPERATURE CORRECTION FACTOR APPLIED TO SUBSTATION CAPACITY ANALYSIS

Allen WATSON, Rob WELLER EDF Energy - United Kingdom allen.watson@edfenergy.com INTRODUCTION Several substations on EDF Energys networks exhibit summer peaking characteristics. Previous load forecasting methodologies have estimated load growth on a regional basis, which is not suitable for analysing these types of substations. It will be shown that the most appropriate way to illustrate urban load characteristics is to correct demand for average cold spell (ACS) and average hot spell (AHS) temperature conditions at each substation. The IEC transformer loading guide [1] is suited to modelling standard transformers, which are sited outdoors. However, it is not appropriate for power transformers unable to readily dissipate heat such as those enclosed in basements or situated underground. A transformer thermal model has been developed which is based on the loading guide. Modelled transformer winding temperature data will be compared against actual winding temperature data to determine the thermal model correction factor for a particular enclosure. Determination of substation firm capacity can be facilitated by illustrating load and capacity, which are corrected to ACS and AHS conditions, on the same chart. This substation analysis tool is used to plan reinforcement to ensure networks comply with planning standards. It will be shown how the impact of network reinforcement solutions, such as increased transformer capacity and load transfers, can be modelled to achieve the best planning solution. URBAN NETWORKS WITH SUMMER PEAKING LOADS Regional Load Growth Methodologies Grid supply point level. At the UK national transmission level system maximum demand occurs during the winter. The England and Wales transmission operator, NGC, is concerned with the demand at the grid supply points (GSP) for the purposes of planning and operating the transmission network. Load growth forecasts are calculated at each GSP using actual demand data. The peak demand in England and Wales is then corrected to ACS temperature conditions [2]. Bulk supply point level. Historically, EDF Energy applied a regional correction factor approach as a basis for forecasting load growth on its East area network substations. An ACS correction factor was CIRED2005 Session No 5 calculated at each bulk supply point (BSP). Loads at all downstream primary substations were multiplied by this correction factor. Table 1 shows an example of an ACS correction factor applied to primary substations.
Table 1 - BSP ACS correction factor example
Bulk Supply Point Name (132/33kV) ACS Correction Factor Primary substation Name (33/11kV) Broxbourne East Hertford Hoddesdon Much Hadham North Harlow The Cross Turnford Ware Rye House 1.02 Actual Maximum Demand (MVA) 11.9 18.7 22.6 5.9 26.1 9.5 18.9 12.6 ACS Corrected Maximum Demand (MVA) 12.1 19.1 23.1 6.0 26.6 9.7 19.3 12.9

Substations with Summer Peaking Load The NGC and EDF Energy growth forecasting methodologies presented above do not consider AHS correction factors. Therefore they are not suitable for application to summer peaking substations. Forecasts would be more appropriate if both ACS and AHS analysis was carried out at each substation. Figure 1 shows a typical annual daily maximum demand load profile for a substation which serves an urban network. From November to May the weekday maximum demand is close to 54MVA. It increases to 65MVA during August 2003, which coincides with the exceptionally high temperatures experienced in summer 2003.
Figure 1 - Summer peaking substation load profile
70 60 Maximum Daily Demand (MVA) 50 40 30 20 10 0 14/11/2002 28/11/2002 12/12/2002 26/12/2002 09/01/2003 23/01/2003 06/02/2003 20/02/2003 06/03/2003 20/03/2003 03/04/2003 17/04/2003 01/05/2003 15/05/2003 29/05/2003 12/06/2003 26/06/2003 10/07/2003 24/07/2003 07/08/2003 21/08/2003 04/09/2003 18/09/2003 02/10/2003 16/10/2003 30/10/2003 13/11/2003

Date

Figure 1 shows that demand is higher during the summer months than the rest of the year. This implies a positive correlation between demand and ambient air temperature.

CIRED

18th International Conference on Electricity Distribution

Turin, 6-9 June 2005

Figure 2 - Substation demand against ambient air temperature

Weekday Maximum Demand 80 70 Maximum weekday demand (MVA) 60 50 40 30 20 10 0 -5 0 5 10 15 20 25 30 Average ambient air temperature (oC) ACS
54.0

Figure 3 - Block diagram of the transformer thermal model

Substation load profile data

Maximum Demand Trend

AHS
64.5

Ambient air temperature data

Transformer thermal model

Transformer cooling method parameters

Transformer daily cyclic rating capacity

Figure 2 is a scatter chart showing weekday maximum demand plotted against average ambient air temperature [3]. It is produced from demand data used in Figure 1 and air temperature data from EDF Energys weather stations. A correlation coefficient is a numerical value that determines the relationship between two properties. Correlation coefficients lie between -1 up to 1. -1 and 1 indicate that the relationship between two sets of data is a perfect negative and positive linear correlation respectively. A zero correlation coefficient value suggests that there is no relationship between the two sets of data. Figure 2 clearly shows that for ambient temperatures greater than 14oC there is a positive linear correlation between air temperature and demand. The correlation coefficient between maximum demand and ambient air temperatures, greater than 14oC, is 0.87. Correction of winter peaking loads to ACS temperature conditions is a well understood method [2]. The demand ambient air temperature characteristic seen in Figure 2 suggests that summer peaking loads should also be projected or corrected to AHS temperature conditions. Figure 2 shows the demand corrected to ACS and AHS temperature conditions, -1.4oC and 25.3oC respectively, which are defined in reference [3]. TRANSFORMER THERMAL MODEL CORRECTION FACTOR Transformer Thermal Model Figure 3 shows a block diagram of a transformer thermal model which has been implemented in Microsoft Excel using Visual Basic Scripts [3]. It has been based on the algorithms given in reference [1].

Inputs to the transformer thermal model are ambient air temperature data, substation load profile data and transformer cooling method parameters. The model outputs transformer daily cyclic rating capacity based on the winding temperature trip setting. Figure 4 is an example of the transformer thermal model output. As would be expected, the transformer has a higher rating in the winter than during the summer.
Figure 4 - Transformer daily capacity against ambient air temperature
Emergency Cyclic Rating 35 30 25 20 15 10 5 0 -5 0 5 10 15
o

Rating Trend

28.2 24.5 ACS AHS

Maximum transformer capacity (MVA)

20

25

30

Average ambient air temperature ( C)

Figure 4 also illustrates the transformer daily cyclic rating capacity corrected to ACS and AHS temperature conditions. Transformer Enclosure Derating Factor The loading guide gives an optimistic view of the operating capability of transformers that are housed in an enclosure. To obtain a reliable indication of operating capability of transformers in these circumstances it is necessary to measure winding temperature and compare it with modelled results. This will enable the transformer enclosure derating factor to be determined. Comparison of Temperature Modelled and Actual Winding

Naturally cooled transformer. Figure 5 compares actual and modelled winding temperature data for a naturally cooled transformer. The modelled winding temperature includes an enclosure derating correction factor of 0.86. This figure shows that the correction factor has been multiplied by the transformers nameplate capacity so that the modelled data closely matches the actual winding temperature data.

CIRED2005 Session No 5

CIRED

18th International Conference on Electricity Distribution

Turin, 6-9 June 2005

Figure 5 Comparison of modelled and actual winding temperature (a 0.87 correction factor is included in the modelled data)
Modelled Winding Temperature 100 95 90 Winding temperature (oC) 85 80 75 70 65 60 55 50 0 5 10 15 20 25 30 Average ambient air temperature (oC) Measured Winding Temperature

Figure 7 - Combined demand and capacity against temperature

Emergency Cyclic Rating Maximum Demand Trend 70 Weekday maximum demand and capacity (MVA) 60 50 47.3 40 30 Maximum daily demand 20 10 0 -5 0 5 10 15 20 25 30 Average ambient air temperature (oC) ACS 48.3 AHS 63.9 Weekday Maximum Demand Rating Trend Transformer daily capacity 54.3

Forced cooled transformer. Figure 6 compares actual and modelled winding temperature data for a forced cooled transformer. Winding temperature results have been modelled only for winding temperatures less than 90oC where forced cooling is switched off, as shown in Figure 6. Further work needs to be carried out to model the winding temperature when the transformer forced cooling is thermostatically controlled or is switched on permanently.
Figure 6 - Comparison of modelled and actual winding temperature (for winding temperatures less than 90oC)
Modelled Winding Temperature 100 95 90 Winding temperature (oC) 85 80 75 70 65 60 55 50 0 5 10 15 20
o

A substations firm capacity headroom at ACS and AHS temperature conditions can be readily identified from this type of chart. Although the transformer winding is modelled in this paper it may not be the item of plant at a substation with the minimum rating. Therefore, ratings of other plant items are required before a full security assessment can be made. Solutions to Capacity Shortfall Capacity shortfalls are addressed by implementing solutions such as power factor correction, load transfers, and additional transformer capacity. A powerful feature of the substation analysis tool is its capability in modelling the impact of these options to determine the optimal solution. Rather than analysing historical demand it is more useful to confirm the future demand to ensure that the substation will remain firm. Annual load growth, information about new connections and increases in demand for existing customers are used to estimate the future load at substations. The analysis tool can be modified to take into account any future load conditions at a substation. Substation Design Case Study

Measured Winding Temperature

25

30

Average ambient air temperature ( C)

SUBSTATION DESIGN ANALYSIS TOOL This final section combines the two previous sections into a practical context by showing how they can be used as an analysis tool for determining the true emergency cyclic ratings of power transformers and hence the firm capacity of a substation. The analysis tool merges the load - temperature chart, as shown in Figure 2, and the capacity - temperature chart, as given in Figure 4, into a combined demand and capacity - temperature chart. An example of this chart is illustrated in Figure 7. Figure 8 shows an example of the demand and transformer thermal capacity at a substation. The demand exceeds the N-1 capacity for 46 days of the year (i.e. where one circuit is switched out due to a fault or for planned maintenance) at AHS temperature conditions. The substation design tool will be used to model load transfer and increased transformer capacity solutions.

CIRED2005 Session No 5

CIRED

18th International Conference on Electricity Distribution

Figure 10 - Impact of an additional transformer
Emergency Cyclic Rating Maximum Demand Trend 120
65.1

Turin, 6-9 June 2005

Figure 8 - Substation demand and N-1 capacity

Emergency Cyclic Rating Maximum Demand Trend 80 Weekday maximum demand and capacity (MVA) 70 60
59.2

Weekday Maximum Demand Rating Trend

Weekday Maximum Demand Rating Trend

Weekday maximum demand and capacity (MVA)

62.9

100

94.4 80.7

50 40 30 20 10 0 -5

ACS

53.8

80

AHS

60 59.2 40 ACS

65.1 AHS

20

0 0 5 10 15 20 25 30 -5 0 5 10 15 20 25 30 Average ambient air temperature (oC) Average ambient air temperature (oC)

HV load transfer. Figure 9 shows the impact, on the substation depicted in the previous chart, of moving a load of 19.9MVA to another substation. This results in a capacity headroom of 20.6 MVA and 8.8MVA for ACS and AHS temperature conditions respectively. This chart also shows that the reduced substation load profile marginally increases the ACS and AHS transformer cyclic rating capacity by 0.1MVA and 0.2MVA respectively.
Figure 9 - Impact of carrying out the load transfer
Emergency Cyclic Rating Maximum Demand Trend 70 63.0 Weekday maximum demand and capacity (MVA) 60 50 40 30 20 10 0 -5 0 5 10 15 20 25 30 Average ambient air temperature (oC) 45.2 42.4 ACS AHS 54.0 Weekday Maximum Demand Rating Trend

CONCLUSIONS The summer peaking substation example given in this paper is typical of several other substations that serve urban networks. This has led to the need to change the methodology used to forecast load growth to one that is focused on each substation. It has been demonstrated that load at this substation is characterised by a strong positive linear correlation between ambient temperature and demand, for ambient air temperatures greater than 14oC. The best way to illustrate this attribute is to plot weekday maximum demand against average ambient air temperature. Additionally, the demand is corrected to both ACS and AHS temperature conditions. Where power transformers are enclosed or located underground, an enclosure correction factor is required, which derates the capacity of the transformer thermal model. A substation analysis tool has been presented that is capable of illustrating demand and capacity, corrected to ACS and AHS temperature conditions, on the same chart. This tool can be useful for highlighting the need for any future reinforcement at a substation and modelling the impact of network reinforcement solutions. The tool can also beneficially be used to determine more accurately the maximum firm capacity of a substation and hence the optimal timing of reinforcement. Examples have shown how increased transformer capacity and load transfer solutions can be modelled to evaluate the extent to which each would solve an anticipated capacity shortfall at a substation. REFERENCES [1] IEC 60354 Ed. 2.0: 1991, Loading Guide for oil immersed transformers. [2] National Grid Company PLC, 2003, Seven Year Statement. [3] D. Ellis, 2003, The impact of summer peaking demands on distribution network supply security assessments at main substations, ERA Technology.

Increased transformer capacity. The problem could also be solved by increasing the substation capacity by introducing an additional transformer. Figure 10 shows the impact of implementing this solution. This provides a headroom of 35.2MVA and 15.6MVA for ACS and AHS temperature conditions respectively.

CIRED2005 Session No 5