Methodology for Validation

To best demonstrate the reliability of the campus power system during a disaster, the worst case scenario has to be identified and selected. This worst case scenario corresponds to the maximum loading condition (highest power consumption) that has occurred in the past. This way, by simulating for the worst case scenario we could see how system would react in extreme conditions during a natural disaster. A historical study of the overall trend on each of the building and substation was done including the seasonal load data that Western Facilities

Management provided. It is seen that peak power demands occur mainly during the months of January and July.

January has a high power demand due to fact that it is one of the coldest months and students are returning from a holiday break. For this reason, electricity consumption trends are much higher as compared to other months. July has been historically the hottest month and the increase in electricity use is primarily caused by the increased usage of the campus chillers. Both cases present equal likelihood of a worst-case scenario in terms of total power consumptions.

In this research work, for validating the model, three scenarios have been selected for steady state analysis in EMTDC/PSCAD. Each of these three case studies has the same system diagram, but different data is used for the load flow studies. These three case studies represent different loading conditions of the Western campus electrical systems, as follows:

 Peak Load Conditions

 Light Load Conditions

 Medium Load Conditions

For all the three loading conditions, following steps were followed for validation.

 Once the model was developed in EMTDC/PSCAD, the model is executed for getting the values of different electrical quantities.

 For all the three different loading condition, active power (P), reactive power (Q) and voltage (V) values are calculated from the model developed in EMTDC/PSCAD.

 For each of the three loading conditions, the values for different electrical quantities were calculated at three different locations, start, middle and end of the feeder.

 The actual measured values for same electrical quantities (P, Q and V) were obtained from the Western facilities.

 Comparison is done between the actual and measured values obtained from EMTDC/PSCAD.

 For the validation, the values for three different electrical quantities (P, Q and V) are compared and percentage error is calculated for nine different cases.

4.2.1 Peak Load Conditions

For modelling the peak load conditions, the data provided by the campus physical plant was studied and analyzed. The data provided for all three substations was studied on a weekly basis, analyzing the trends of different days of the week. Physical Plant personnel also provided data for all of the different buildings on the Western campus, from the year 2002 to present. After analysis, it was observed that different weeks of a particular month had a common trend from Monday to Sunday. For the majority of the weeks, the load of campus reaches its peak value for Mondays and then the load gradually decreases day by day as the weekend approached. The peak value on a particular Monday was observed as 22.9 MW. Figure 4.1 presents a typical total power curve for the Western campus for a particular week in 2011. It illustrates that the campus uses more power during the day as compared to night time.

The seven peaks in the Figure 4.1 represents seven days of a week. The first peak represents Monday and the last two lowest peaks represent Saturday and Sunday. Another factor, which was included, is the concept of critical buildings and critical electrical devices that require electrical power 24 hours per day. The critical buildings of the campus include the power plant, the hospital, medical science building, animal care, etc. As such, it was made sure that during the modelling of campus, critical buildings should receive power above the threshold value. The data used for the modelling of the Western campus during peak load conditions is provided in Appendix A. Table 4.1 represents the buildings which have a peak load, more than 5 MW.

4.2.2 Light Load Conditions

The second scenario used for validating the campus electrical model corresponds to the light load conditions. The light load scenario was determined by analysing the total substation data, which is the combination of north, south, and east substations. The basic concept of analyzing the data remains the same as in peak load conditions. As explained in Section 4.2.1, different weeks of a particular month had a common trend from Monday to Sunday.

0 1 2 3 4 5 6 7 13 14 15 16 17 18 19 20 21 22 23 Time (Days) P o w e r (M W )

Figure 4.1: Typical total power curve for Western campus for a particular week in 2011 (On x-axis 0-7 represents Monday to Sunday).

Table 4.1: Campus buildings having peak load more than 5MW

Serial Number Building Name Average Peak Load (KW)

1 Social Science 1833.8194

2 West Valley 942.2667

3 Support Services 881.3778

4 Recreation Centre 722.6389

5 Medical Science 670.3364

6 Spencer Engineering Building 654.0648

7 Dental 614.1991

8 Weldon 599.1821

9 Biotron 572.3241

For all of the weeks, the load of campus reaches its lowest value for nighttime mainly on Wednesday and Thursday, during the weeknights. For the light load case study, the simulation diagram remains the same, but the data is different, which is provided in Appendix B. Also, one of the major worst load conditions was observed in November 2011 for approximately two weeks. This is the incident when the south substation was shut down for roughly two weeks and the north substation was compensating for the closing of the south substation.

4.2.3 Medium Load Conditions

The third scenario used for validating the campus electrical model relates to medium load conditions. Both the building and substation data was analysed to determine the appropriate medium loading conditions. By analyzing different maximum and minimum loading conditions, an average value for medium loading conditions was obtained. In addition, the corresponding loading conditions were obtained from the building load data provided by the physical plant. In actual practice, medium load conditions represent the late afternoons and early evenings during the weekdays. Also, at times the medium loading conditions can be represented by the peak load values during Saturdays and Sundays. In this scenario, the same system diagram is utilized during the simulation except with changed load values, as provided in Appendix C.