Examining the effect of heat storage in a cogeneration system

Small power plants of cogeneration of power, heat and cooling are good solutions of increasing the efficiency of energy consumption for fossil fuels in order to protect natural resources and the environment. However, at moments when heat demand is lower than the heat production of the CHP module, the excess heat has to be rejected to the environment and this fact results in waste of energy. Also, since CHP modules are basically heat driven, when heat demand is lower than a certain value, the module will be switched off just to be switched on later when heat demand increases. This cycle of switching on and off is harmful for the CHP module if it happens repeatedly. A solution is to use heat storage and an alternative control method. In this paper, a CHP system is chosen for an educational building and the design is carried out in two forms, with and without heat storage and the results are compared and judgment is made about the optimal system.


Introduction
When power is produced traditionally, a large portion of original energy of the fuel is wasted as heat and hardly more than 40 pe rcent of this energy is transformed into electricity.Moreover, usually consumers are located far away from the power plant and this distance causes more waste of energy in distribution of electricity.One way to tackle these problems is using local cogeneration.In this modern method of power generation, power is produced at the location of consumption and the majority of lost heat is recovered to supply heat demands of the user.This results in a considerable improvement in efficiency.Furthermore, since power is generated at the same location where it is consumed, distribution losses will be avoided.The total efficiency of cogeneration power plants amounts up to 90%, while the electrical efficiency of a traditional power plant hardly reaches 40%.Among different options of power generation in the form of cogeneration, reciprocating engines seem to be the most suitable for buildings which essentially have small demands.They have high power to heat ratios compared to gas turbines and due to advances made in automotive industry, enjoy a higher degree of modernization [9].Although stationary reciprocating engines have traditionally been diesel engines but some issues like environmental issues and good access, have been promoting the users in recent years to use natural gas as the fuel instead.In Iran, a Persian gulf country with the second largest resource of natural gas in the world, even automobiles are increasingly using gas burning and dual fuel engines.
X Q Kong et al (2004) optimized a trigeneration system (cogeneration of heat, power and cooling) based on gas turbine.In their research a trigeneration system was modeled and then, after specifying constraints and an objective function, the solution was optimized using a linear modeling program [2].In another work, they examined a co generation system and presented the results as graphs and tables [3].In 2005 P .Arcuri et al designed optimally a trigeneration system using a mixed integer model.They optimized a trigeneration system for a hospital employing a reciprocal engine as its prime mover [4].In 2006 E .Cardona and A. Piacentino designed and optimized a trigeneration system for a hospital application from the thermoeconomic point of view [5].The same researchers carried out another analysis for an apartment building using the thermoeconomic method [6].In 2008, Behbahani Nia et al. [7] optimized a cogeneration system based on gas turbine with the aim of minimizing the capital cost in which they considered electricity, heat and cooling demands for each month.
In this paper, a cogeneration system is designed and optimized for the building of mechanical engineering faculty of K.N.Toosi University of technology in Tehran, Iran, using two different strategies, with heat storage and without heat storage.First, energy simulation is carried out using the software Carrier HAP 4.2 resulting in values of electricity and heating demands in all 8760 hours of the year.Later, based on these demands, the main components of the CHP system are designed based on products of the Austrian manufacturer, Jenbacher®.Products of this company are cogeneration modules including the reciprocating engine, heat recovery system and electrical generator all in one, covering a range of capacities from 400kWth to 3MWth.

A description of the building
The building of mechanical engineering faculty of K.N.Toosi University of technology is a ten-floor building, including 3 unde rground floors and covering about 20 thousand square meters of area.The second and third floors contain classes, fourth and fifth floors contain administrative rooms, almost all of which benefit from natural light during daytime.The sixth floor is dedicated to professors' rooms about half of which have access to natural light.The library and some laboratories are placed on the first floor.Ground floor primarily contains public places like the big lobby, the pray place, computer services hall and so forth.The floor -1 contains laboratories, cafeteria, the big restaurant and the amphitheatre.The floors -2 and -3 are for workshops and labs and also sport activity salons.Table 1 shows a list of areas of these floors.

Calculation of loads
Thermal and electrical loads have been calculated using the energy simulation function of the software Carrier HAP 4.2.All parts of the building were modeled and wattages of lights, electrical equipments, geometrical and heat transfer features of rooms were entered in the software.A total of 270 s paces were defined in the process.Another important issue in determining loads is the presence of people in different spaces.Schedules were defined for presence of people in different types of spaces including classes, amphitheatre, computer services salon, corridors, administrative rooms, pray place, restaurant and security compartments, and also for lighting for each of these types of places, based on percentages of full presence or full lighting in different hours of the day and different days of the year.National holidays and weekends were considered based on t he year 2009 which covers portions of Persian years 1387 a nd 1388.T he difference of intensity of natural light in summer and winter days and different levels of presence of students and employees in different months of the year and different hours of the day were all considered based on personal observation of the second author who has been a studying in the same building for two years.The monthly distribution of heating and cooling loads resulting from this energy simulation is as shown in figure 1.The weather conditions were defined based on simulation information of www.Carrier.com of Tehran including hot and cold bulb temperatures and sunlight situation throughout the year.

Selection of cogeneration modules
Selection was carried out based on pr oducts of Jenbacher, including 13 m odels of CHP modules.The manufacturer did not reply requests of price quotation and purchase equipment costs and O&M charges were estimated using the information in [9] and by curve fitting.The cost of natural gas and electricity were taken 690 Rials per m3 and 773 Rials per kWh, equal to Iranian unsubsidized rates.
Another issue which was considered in this optimization was the environmental issue.
According to [9], emission of pollutants imposes costs which are in fact costs of reduced performance of human beings caused by these pollutants.This fact is considered as costs assigned to pollutants CO, CO2 and NOx.According to catalogs of the manufacturer, using the lean combustion system and SCR catalysts, emissions of CO and NOx caused by their products are limited to 100 mg/Nm3 for Nox and 300 mg/Nm3 for CO.CO2 emission from natural gas combustion is equal to 1.15m3/1m3 Natural Gas according to [11] which by considering the density of carbon dioxide in normal conditions equals to 20420mg/Nm3.Values of emissions of CO and NOx for small boilers are 641mg/Nm3 and 1506 m g/Nm3 respectively, according to [12].As calculated in [9], the social cost associated with these emissions is 81750 Rials/kg for carbon monoxide, 240 Rials/kg for carbon dioxide and 64240 Rials/kg for Nitrogen oxides.Therefore, the social costs for burning of each cubic meter of natural gas for Jenbacher® reciprocating engines and the boiler are as shown in tables 2 and 3.

The case without heat storage
In this section, sizing is carried out in two different strategies, one is the absence of heat storage and the other is its presence.In both strategies, modules of cogeneration and their annual working durations are determined so that the total annual cost is minimized.
For the case where there is no heat storage system, the CHP system is designed based on loadduration curves.These curves are constructed using the hourly load data taken from energy simulation, i.e. first values of heating and electrical loads for all 8760 hours of the year are taken from outputs of Carrier HAP and then, those numbers are put in descending order and plotted against duration, from 1 hour to 8760 hours.According to [13], the largest rectangle which can be circumscribed in that curve represents the optimal choice of the CHP system, in terms of capacity (on the vertical axis) and number of total working hours throughout the year, on the horizontal axis.Here, the basic idea is quiet similar.However, this curve is used here to determine the capacity of the supplementary boiler which is the difference of maximum load with the heat production of the CHP module and its total heat production throughout the year being equal to all heat demand not satisfied by the module.
Electricity is considered as a bi-product of the system that can be used locally or sold to the network.The rates of buying and selling power to the network are very close to each other in Iran [20] and both are assumed to be 773 Rials.If a CHP system is independent from the grid, it can employ batteries to store excess electricity to be used later but when selling power to the grid is possible, using storage of electrical energy is not economical [9].
The control strategy used for the case where there is no he at storage system is as follows: When the number of working hours of the CHP module determined from optimization is plot with load-duration curve, the point where it intersects that curve shows the value of minimum load for operation of the module, i.e. when the thermal load is lower than that value, the module will be switched off and when the load exceeds that value, the module will be switched back on.
In manufacturer's catalogs, two heuristics are suggested: • The thermal power of the cogeneration power should be between 30 to 50 percent of the peak value of thermal power demand.• The module of cogeneration should work at least 4000 hours during a year.Figure 2 shows an example of load-duration curve.

Figure 2. Load duration curve for heating load
Naturally there will be times when the heat demand is higher than the production of modules and at these times this heat shortage is covered by the auxiliary boiler.
Load-duration curves for heating, and electrical loads of our building are shown in figures 3 and 4. Now, the objective function for the optimization is defined as the total annual cost of the system.To calculate the total annual cost, first we should annualize capital investments using the capital recovery factor (CRF) .
Where i r , the interest rate, according to [14] is taken 12 percent, and n is the number of years of life time of the system, here taken 20 years.Thus, the objective function is: (1) Where C Ann is the total annual cost, C TCM is the total capital investment for the CHP module, C TCAB is the total capital cost for the auxiliary boiler, C TCST Is the total capital cost for the storage tank (if included), C O&M Module is yearly O&M plus fuel costs for the cogeneration modules, C O&MAB is the yearly O&M plus fuel costs for the auxiliary boiler, C Emi is the yearly emission cost and C el is yearly cost of electricity production which is the profit of the system and therefore appears with a negative sign in the total annual cost.The optimization is carried out using the direct search method.For this optimization, decision variables are taken to be capacities of CHP modules and their durations of operation throughout the year.Constraints are defined based on heuristics provided by the manufacturer, namely each module should not operate less than 4000 hours in the year, and the values of capacities of modules and the boiler, naturally may not be negative and the values of working hours of each of modules cannot be more than 8760 hours.Results are as presented in the next section.

The case with heat storage
If we decide to employ heat storage in our system for more smooth operation and less waste of energy, a different design and operation strategy has to be used.Heat is stored as hot water (90 o C) in a well insulated storage tank.Its cost data is taken from [14] and ( 1) is also used for cost estimation, using two different values of the exponent α (0.3 and 0.65) based on t he calculated volume.The cost data is available in terms of volume of the storage tank while in the optimization, the capacity in terms of energy storage is considered.As mentioned in [15], the CHP module receives cooling water at 40 o C and sends it out at 90 o C. Thus, in order to determine the volume of the storage tank conservatively, we take the unit volume energy of the water stored in this tank as the difference of enthalpy of water in those input and output states.
Thus, by storing each cubic meter of water in the storage tank, we have stored 58.167kWh thermal energy.
As a r esult of the above mentioned strategy, there will be fewer start-stop cycles and probably, less heat rejection to the surroundings.When optimizing the system in this case, working duration of the module will no longer be a decision variable but instead, the volume of the storage tank will be searched for its optimal value and working duration of the module will be determined from the volume of the storage tank and load-time curve.The other decision variable will be the size of the CHP module, as before.The results of optimization of this case are presented in the following section.As it is evident from tables 5 and 6, heat dissipation to surroundings and total annual cost are both higher for the case with heat storage than the simple case.Moreover, as illustrated in results, curves of electrical and thermal loads have more consistency with curves of energy production of the module in the simple case.However, in the case with the possibility of heat storage, more electricity is produced and the module works for a longer total duration, representing a smaller number of switching off and on cycles which is better for durability of the reciprocal engine and the whole module.

6 B Conclusion
Heating and electrical loads were calculated for a 10-floor educational building using energy simulation of Carrier HAP®, and based on those loads, cogeneration systems were designed to provide electricity and heating needs of the building.The CHP module was selected 13 models of a globally renowned manufacturer.
Firstly, a simple CHP system was designed containing a CHP module and an auxiliary boiler.Secondly, the possibility of heat storage was taken into account using a storage tank as heat accumulator.Two different control strategies were considered for these two cases and consequently, design and optimization were also carried out differently.
Comparison of results showed that the simple system excluding heat storage had a lower total annual cost and heat dissipation to surroundings.On the other hand, it had a lower work duration for the CHP module and consequently, a larger number of switching on a nd off cycles representing its disadvantage to the system with heat storage.

Fig. 1 .
Fig. 1.Heating and cooling loads throughout the year.

Figure 3 .
Figure 3. Load-duration curve for heating load

Table 1 .
Area of each floor of the building

Table 2
Emissions and their costs for natural gas-burning boiler

Table 3
Emissions and their costs for natural gas-burning engine

Table 5 .
Optimization results for a CHP module without heat storage

Table 6 .
Optimization results for a CHP module with heat storage