Next: Campus Analysis  Previous: Theoretical Analysis

Subsections

GEORGIA TECH STUDY

Introduction

By using the concepts and guidelines presented in previous chapters, the economic feasibility of a cogeneration system on the Georgia Tech campus will be evaluated. Differing operational strategies were studied as well as differing sizes of prime movers.

Campus Energy Usage Characteristics

Figure 6 shows the relative sizes of the electrical and gas usage on campus for the 1995 year. It is clear that the campus energy usage is seasonal in nature. The thermal load is highest in the winter months while the electrical load is highest in the summer months. This is clearly due to the space conditioning requirements of the campus. The electrical usage is higher during a greater portion of the year, which is consistent with the seasonal weather patterns in Atlanta, Georgia.
Figure 6 - Total Monthly Electrical and Gas Usage for Georgia Tech for 1995 year.

Electrical Usage

Georgia Tech's campus has experienced a burst of growth caused by the Olympic Games held in Atlanta during the Summer of 1996. The campus served as the host for all the Olympic Athletes. New dorms and associated facilities have been constructed, adding to both the campus electrical load and thermal load. Roughly 1.3 million sq. feet of building space has been added since the Olympic planning began.14 Assuming the campus reverts to a more normal growth rate over the next fifteen years, another 5500 tons (~5500 kW ) of cooling alone will be added to the campus, as well as lighting and other loads generated by new construction.

As of 1995, the campus power consumption varied anywhere from 7500 kW on winter holidays to 24000 kW on peak summer days. While this is a significant spread, the campus did not spend more than 500 hours below 10000 kW for the entire year of 1995. Figure 7 shows a plot characterizing the nature of power usage on campus. Note that the campus maintains an hourly usage above 10 MW for over 8000 hours per year. To further characterize campus usage, the load factor for each month was calculated. Figure 8 shows the campus load factors for 1995. The load factor was determined by dividing the actual kWh usage for the campus in a given month by the kWh usage if the campus operated at the month's peak electrical demand for each hour the entire month.

Figure 7 - Frequency Plot of Hourly Power Consumption for Georgia Tech.
Figure 8 - Georgia Tech Monthly Load Factors for 1995.
 

As you can see, the load factors are quite high, with December 1995 turning in the lowest value of 68.4% and August 1995 the highest of 80.4%. Such high load factors mean that the difference between the highest and lowest demand for a given month is quite small. So from season to season consumption may differ, but is generally flat throughout a month.

Electrical Rate Structure

The campus is served under what is known as the Real Time Pricing-Day Ahead (RTP-DA) rate structure provided by Georgia Power Company. The DA simply represents that prices are available one day prior to power usage. The RTP rate structure consists of two parts, a standard bill, and an incremental bill. The standard bill is calculated under the campus' previous rate structure, Time of Use - 4 (TOU-4) by using a load shape generated by the campus in 1991, the year before they began the RTP rate. The RTP incremental portion of the bill is calculated by comparing the current year power consumption to the 1991 power consumption for each hour, then either charging or crediting the campus account, depending upon whether usage was more or less. A detailed discussion of the RTP rate structure follows.

TOU-4 Electric Rate

In order to fully discuss the new RTP rate structure, it is first necessary to understand the TOU-4 rate structure, since the RTP rate acts as a rider on top of the TOU-4 rate. The TOU-4 rate structure consists of both a demand charge and an energy charge. The demand charge varies based upon the amount of peak half-hour demand incurred in a given month and the energy charge uses a rate dependent upon the time of day the energy is consumed. Included also is a monthly base charge of $475.

The demand charge varies depending upon the magnitude of customer demand: the first 2000 kW of demand is charged at one rate, the next 3000 kW is charged at another rate, and finally, all kW above 5000 is charged at a final rate. The rate decreases as the demand increases, and the rate is lower for the off-peak months (October - May) than the on-peak months (June-September). Also, there is a provision which allows for an economy demand charge if the peak demand for a given month occurs in a off-peak time during the day (8:00 AM -12:00 PM).

The energy charge varies depending upon the time of day it is used. The late night and early morning hours are least expensive, while the afternoon and early evening hours are the most expensive. Also, weekend power usage comes at a discount. Like the demand charges, energy charges are lower in the off-peak months, and there is an adjustment to the time of day structure. A copy of the TOU-4 rate is available in the Appendix B. It discusses the variations in time periods as well as the specific charges for each time period for both demand and energy.

RTP-DA Rate Structure

The RTP rate rides on top of the TOU-4 rate, creating two parts, a standard TOU-4 charge and an incremental RTP charge. The standard charge is calculated under the old TOU-4 rate structure, based upon the power consumption of a contractually agreed to year. Specifically, the Tech campus' base year is 1991, the year before the RTP rider began. The power consumption for the base year is known as the customer baseline load (CBL). The CBL can be thought of as the dividing line between the high cost TOU-4 electricity and the low cost RTP electricity. The incremental RTP charge is the charge calculated by comparing a given half hour's power consumption from the current month to the same half hour period of the same month in the base year. The incremental charge may be either positive or negative, depending upon the relative power between years.

To compile each month's bill, both the TOU-4 demand and the energy charges are calculated based upon the 1991 year. The demand charge is calculated by using the highest power reading for a given meter in a given month in 1991. The resulting charges for demand and energy are then added together to form the standard charge, or TOU-4 portion, of the bill.

The RTP portion of the bill, the incremental charge, is calculated by first determining the difference between the power consumption for the current year and the 1991 base year (can be either positive or negative) for each half hour period during the month. The difference is then multiplied by the cost of the RTP power for that half hour, then added (or subtracted) from the running total for the month. At the end of the month, the cumulative total incremental charge is added (or subtracted) from the standard bill to determine the final monthly bill. The cost of RTP power, which changes hourly, is determined by Georgia Power one day ahead and is available for downloading at 4:00 PM of the preceding day.

The only adjustment that is made to the 1991 data which makes up the CBL is to shift the days so that the weekdays for the 1991 year and the current year are aligned. The CBL, once agreed to, is permanent and usually only changed for a permanent removal of equipment, installation of energy efficient equipment, or a demonstrated decrease in total power consumption that causes the new power consumption for the year to fall below that of the CBL. In order to establish a new CBL, the consumer must revert to the old tariff for an entire year. The new year's power data can then be used as the new CBL.

Gas Usage

Georgia Tech purchases its gas on an interruptible contract. During times of interruption, No. 2 fuel is used as a substitute. The total campus boiler steam capacity is 160, 000 pounds per hour of steam, generated at 150 psig, but delivered at 40 and 15 psig.

The campus base thermal load is roughly 15,000 pounds per hour, with a peak load of greater than 75,000 pounds per hour. Figure 9 and 10 show how the campus gas requirements vary with temperature.

Both plots use degree days to characterize the daily temperature. Degree days are a measure of how far the average daily temperature deviated form a 65 degree Fahrenheit norm. Therefore, 0 degree days represent a day for which the average daily temperature was 65 degrees, 10 degree days represent a day for which the average daily temperature was 75 degrees, and -10 degree days represent a 55 degree average daily temperature.

It is obvious from Figure 9 that gas usage is high during the cold winter months and decreases steadily to the base load throughout the spring and summer, then begins increasing again as fall progresses. Figure 10 shows a regression plot which directly relates gas usage to degree days. As the degree days decrease (negative degree days), gas usage increases, and as the degree days increase (positive degree days), gas usage decreases, but only to a limiting value. This limit can be attributed to campus hot water needs or other non-space heating loads. The figure shows all gas usage days from the past year, including the curtailment days. However, the curtailment days were not used when calculating the regression plot in Figure 10.

Figure 9 - Degree Days and Total Daily Gas Usage for 1995.
Figure 10 - Regression Plot for Gas Usage as a Function of Degree Days.
 

Prime Mover Selection

Once all of the criteria have been considered, a choice of prime mover for the cogeneration system must be made. A simple gas turbine without regeneration appears to be the best choice for the Georgia Tech Campus. For the Georgia Tech campus, the regenerator would limit the quality of the thermal energy leaving the turbine thereby making the recovery of heat more difficult, so it was excluded from the study. First, the electrical load of the campus is large enough to exceed the 3 MW maximum size generally limiting reciprocating engine generators. Although it would be possible to operate more than one reciprocating engine generator to satisfy the large campus demand, installation considerations, maintenance requirements, and limited life span all tend to make the reciprocating generator a poor choice.

The steam turbine generator also seems to be a poor choice. Steam turbines have the lowest fuel-to-electrical ratio of any of the types of cogeneration systems. In order to produce a sizable amount of electricity with steam turbines, the boilers would have to operate at full load and a great deal of steam would have to be discarded. It would be possible to install small steam turbines intended to directly replace electrical drives, but the number of steam turbines required would require a great deal of maintenance, and the increased cost of installing smaller turbines throughout campus makes steam turbines an economically unattractive option.

The gas turbine naturally lends itself to the campus requirements. The large variety of sizes in the 3 MW to 15 MW range allows for precise sizing of the turbine unit. Since the campus boilers are already fired with natural gas, with No. 2 oil as a backup, the fuels used by the gas turbine are already available. Gas turbine units are compact and can be installed either inside or outside. The high reliability of a gas turbine unit means that it can operate continuously with only a minimum of downtime or unscheduled maintenance. The exhaust from the turbine is rich enough in oxygen to sustain supplementary firing, which allows greater flexibility in meeting the changing campus steam loads. Finally, the low capital cost of a gas turbine generator system makes it an extremely attractive option form an economic point of view.

Turbine Sizes

Four sizes of turbines have been chosen for the alternative case studies. The largest is a 10000 kW unit. A 7500 kW unit, a 5000 kW unit, and a 4000 kW unit were also used. The 10000 kW unit is the largest unit that can be operated in a base load strategy. Recall Figure 7 which shows that the campus electrical load is at least 10000 kW for more than 8000 hours per year (91% of the time), and since the campus demand is increasing with the addition of the new facilities constructed for the Olympic Games, the minimum demand is expected to exceed 10000 kW by the end of 1996.15 The turbines sizes are then chosen in evenly decreasing increments down to 5000 kW. The 4000 kW unit was used as the smallest size because of the increasing cost of turbines on a dollar/kW basis and the decreasing efficiency as they approach 3 MW.
 

Operational Strategies

Once the prime mover has been chosen, a number of strategies were evaluated to determine the best economic return. The economic attractiveness of each strategy is inescapably linked with the utility rate structure under which the location is served. Each strategy can return good economic results if a given rate structure lends itself to the strategy.

The following three strategies were chosen for the study; base loading, peaking, and thermal following. For the base loading strategy, the turbines are operated at full load at all times. As much steam is recovered as is required to meet the campus steam load, with the remainder discarded in a condenser. During times when the steam load exceeds the turbine's ability to generate steam, additional steam is generated in the existing boilers. Obviously the heat recovery for a base loading strategy is higher during the winter when the thermal load is high, and lower in the summer, when the thermal load is low.

The peaking strategy studied requires that a certain minimum electrical price be met before the generator is operated. This minimum price was determined by using an assumed cost of fuel to calculate the cost of generating 1 kWh of electricity using the gas turbines. When the minimum price of electricity being charged to campus by Georgia Power is exceeded, the generator is switched on and operated at maximum load, with enough heat recovery to meet the campus steam load. A more detailed discussion of the economic calculations used in the peaking strategy is carried out in Chapter VII.

Finally, the thermal following strategy varies the gas turbine load to meet the campus thermal load. During the times when the generator cannot meet the steam load, the generators operate at full load, with the remainder of the steam being generated by the existing boilers. All electricity generated during the thermal following strategy is used within the campus.