Cogeneration: size does matter

Cogeneration is not a new concept, yet many cogeneration plants perform at a suboptimal level because of sizing issues which determine the utilisation, potential heat recovery and ultimately the return of investment of the cogeneration plant. Many consider sizing a cogeneration plant a ‘black art’ and relegate it to the equipment suppliers. However, it is the customer who has to make the final, informed decision.

Cogeneration is an efficient process where one fuel source is converted into two forms of energy (electrical power and heat) simultaneously on site. In cogeneration, the heat generated is a by-product of the electricity generation process; the exhaust heat and jacket water heat is a result of the combustion process. Therefore, heat recovery is only feasible when the cogeneration plant is in operation.

It can be tragic comedy watching some attempt heat recovery with the gas generator running on idle. The basic idea is to burn the fuel (natural gas) once and use the energy produced twice, ie, using the electrical power generated by the cogeneration plant (gas generator, gas turbine and microturbine) to power the site and recover the waste heat for heating processes; usually in the form of hot water for food manufacturing, heating commercial buildings, hospitals, universities, TAFE colleges, swimming pools, etc.

Since a cogeneration plant is designed to operate continuously between the low-load limits and 100% of its rated capacity, any gaps in the site’s electrical demand is met by the electricity imported from the grid. Since no heat is available for recovery when additional electrical power is imported from the grid, the key to selecting the optimum cogeneration size is to maximise the cogeneration plant utilisation and heat-recovery potential, while minimising the grid import within the constraints of the site’s electrical load profile.

This article will focus on data modelling a cogeneration system using a natural gas generator. Three common approaches to sizing a cogeneration system will then be compared to determine how the cogeneration sizing method affects the overall plant utilisation and heat-recovery potential. The article will examine:

1. How the averages sizing method is close but not close enough;
2. How the minimum sizing method limits the heat-recovery potential of a cogeneration plant; and
3. The flawed conventional wisdom of bigger is better for maximum heat recovery.

When sizing a cogeneration system, three key inputs are required to determine the utilisation:

1. Site load profile;
2. Minimum import level from the grid; and
3. Low-load limits of the cogeneration equipment.

The site’s load profile: Business can obtain its electrical load profile through its utility provider. The site’s load is available at intervals of 15 minutes, 30 minutes or one hour over a one-year period. The data is usually in CSV format which can be exported into a spreadsheet, eg,  Microsoft Excel.

Chart 1: A sample site load profile over 1 year.

The minimum, average and maximum site load profile over a 24-hour period can then be plotted with a pivot table. Chart 2 contains 17,568 rows of data at 30 minute intervals, hence 48 periods over a 24-hour day. Visual representation such as minimum, average and maximum load profile does not necessarily point to the correct cogeneration size.

The minimum import level: Most cogeneration equipment is designed for continuous operation parallel to the grid, and the choice of fuel (natural gas) and fuel system design limits its capability to respond to quick load changes on site. Therefore, any instantaneous electrical load change is taken up by the grid and the cogeneration system follows by ramping up or down. Cogeneration system suppliers usually specify 10~15% of the cogeneration rated capacity for minimum import from the grid for parallel operations, eg, for a 100_ekW generator, 15_ekW minimum import is required from the grid.

The low-load limits: Imagine operating an engine on idle or on low load for prolonged period - unburnt fuel or lubricating oil gets carried over into the exhaust system through incomplete combustion, which eventually leads to poor performance and increase maintenance. The same also applies to operating the cogeneration plant at low load. The equipment supplier can advise the low-load limits of their equipment - it is approximately 50% for a natural gas generator, 30% for gas turbine and 10% for micro-turbines. Most equipment manufacturers will limit the generator’s idle or low-load operation to 30 minutes.

Using Microsoft Excel or other equivalent spreadsheet, the average, minimum and maximum of the site power in KW can then be calculated. From our data, our mean (average) is 140_ekW, our minimum is 35_ekW and our maximum is 350_ekW.  The gas generator which is closest to the figures above should be selected.

Three common sizing methods (minimum, maximum and average) will be used to observe the effect on the ‘Cogen Sizing’ on the data model created. The charts will show the average site power, power imported without the cogeneration (no heat recovery), power imported with cogeneration (no heat recovery) and power generated from cogeneration (heat recovery potential) throughout a 24-hour period.

Mean/Average sizing method: 126_ekW (closest to 140).

Minimum sizing method: From the data, the minimum site load is 35_ekW. However, it seems unambitious to size the cogeneration plant to 35_ekW. Remember that the cogeneration plant (gas generator) low load limit is 50%. At 100% load the cogeneration sizing would be 70_ekW. The closest cogeneration sizing that has been chosen is 86_ekW.

Maximum sizing method: Conventional wisdom assumes that bigger is better since more heat is available for recovery with a bigger cogeneration plant. However, this assumption is flawed because by increasing the ‘cogeneration sizing’ to the maximum of the site load, the ‘Cogen Utilisation’ is considerably reduced because of the corresponding increase in the ‘minimum import’ and the cogeneration generator ‘low-load limits’. The generator will receive the start signal less frequently and parallel the generator to the grid less frequently.

Using data modelling, we are able to test different ‘cogeneration sizing’ such as 115_ekW and 119_ekW.

The data modelling and chart shows that 119_ekW would produce the highest electrical output and maximum potential for heat recovery.

Strategy	Size (electrical kW)	Annual power produced on site (electrical KWh)	Cogeneration utilisation	Heat-recovery potential (thermal kWh)
Minimum	86	729,461	100%	1,314,860 (4th)
Data modelling	115	863,554	100%	1,554,398 (3rd)
Data modelling	119	870,854	99%	1,567,537 (1st)
Average	126	866,672	94%	1,562,183 (2nd)
Maximum	374	28,379	2%	42,442 (5th)

In summary, cogeneration sizing is not a black art, users just need to understand that the waste heat generated is a by-product of the power generation process and heat recovery is only feasible when the generator is on load. Armed with the site’s electrical load profile, understanding of the cogeneration equipment low-load limits, minimum import level and thermal efficiency, users will have the information required to perform a data modelling to optimise the cogeneration sizing selection so that maximum electrical and thermal power can be generated on site. A properly maintained cogeneration plant will be operational for the next 20 years; missed heat-recovery opportunities are magnified several times over. In cogeneration, size does matter.