Eurosun conference 2006, Glasgow, UK
Design method for solar heating systems in combination with pellet boilers/stoves
Frank Fiedler*1, Chris Bales1, Tomas Persson1, Alexander Thür2
1Solar Energy Research Center SERC, Dalarna University College, Sweden
2Department of Civil Engineering,Technical University of Denmark
* Corresponding Author, email : ffi@du.se
Abstract
In this study an optimization method for the design of combined solar and pellet heating systems is presented and evaluated. The paper describes the steps of the method by applying it for an example of system. The objective of the optimization was to find the design parameters that give the lowest auxiliary energy (pellet fuel + auxiliary electricity) and carbon monoxide (CO) emissions for a system with a typical load, a single family house in Sweden. Weighting factors have been used for the auxiliary energy use and CO emissions to give a combined target function.
Different weighting factors were tested. The results show that extreme weighting factors lead to their own minima. However, it was possible to find factors that ensure low values for both auxiliary energy and CO emissions.
Keywords: Design method, system optimization, pellet and solar heating systems 1. Introduction
The use of wood pellets for domestic heating is increasingly popular in Europe especially in Austria, Germany and Sweden [1]. With about 60000 installed domestic pellet heating systems Sweden is so far the largest market for these systems in Europe [2]. This is due to the long tradition of wood heating in Sweden but also due to the low prices of wood pellets compared to other energy sources [3]. Several studies have shown that the combination of conventional heating systems with solar heating is beneficial in terms of fuel energy savings and lower emissions since the boiler usually in the summer can be turned off when it’s efficiency is low [4],[11]. Simulation studies of existing combined solar and pellet heating systems have also shown that there is still a large optimization potential for these systems [4],[5].
The aim of this study was to demonstrate and evaluate an optimization method for the design of combined solar and pellet heating systems.
2. Method
The method presented in this study consists of the following steps:
• Definition of the boundaries for the system optimization.
• Parameter identification for the model of the system from measurements, former studies and other sources.
• Definition of the objective function – discussion.
• Choice of the optimization tool.
• Sensitivity analysis: which parameters have the largest impact on the objective?
• Definition of optimization method.
• Choice of system on which to apply the optimization method.
• Application of the optimization tool.
3. Boundary conditions and system description
The boundary conditions for the whole system such as weather data, space heating (SH) and domestic hot water (DHW) load need to be specified as well as which part of the system is going to be optimized.
3.1 Boundary conditions
The combined solar and pellet heating system has been simulated with hourly weather data for Stockholm, Sweden. The space heating and DHW demand has been modelled with load files from IEA task 26 with about 12000 kWh/year for space heating and 3300 kWh/year for DHW with 47.5°C hot water temperature [6].
3.2 System description
The system that is studied and optimized in this article is modelled in the simulation environment TRNSYS [7]. The solar part of the modelled system represents the typical size of such a system in Sweden, with a collector area of 10 m2 and a store volume of 730 ltr [8]. The main auxiliary heat source is a pellet boiler that has been developed and tested within the REBUS project [9].
The characteristics of this pellet boiler are shown in figure 1. More specific information about the pellet boiler can be found in [10]. The boiler is not (yet) a product that can be found on the market but as can be seen in [10] that its characteristic is similar to other pellet boiler with relatively small combustion power. The CO emissions per MJ fuel increase significantly with combustion powers lower than the maximal combustion power. The boiler is modelled in TRNSYS with Type 210 [12] and the store with type 140 [13]. The size and heat loss coefficients for the store have been taken from an existing store where the parameters have been identified in [14].
0 2 4 6 8 10 12 14
0 2 4 6 8 10 12 14
Combustion power [kW]
Heat [kW], Air factor
0.0 0.2 0.4 0.6 0.8 1.0
Efficiency, CO [g/MJ]
heat to room
heat to water efficiency
air factor CO-emissions
start emissions of CO = 4.4 g/start stop emissions of CO = 1.4 g/stop
Fig.1. Characteristics of the pellet boiler [10]
The aim of this study is to minimize the auxiliary energy demand and the CO emissions by optimizing the boiler operation and the interaction between the boiler and the store.
Consequently, a theoretically optimal charging from solar loop and discharging from the load side is assumed by using stratified inlet pipe for the solar loop and stratified return pipe from the SH/DHW circuit. For simplification of the system, the outlet height for both SH and DHW preparation has been fixed to the top of the store. The hydraulic scheme of the studied system can be seen in figure 2. The grey area of the system represents the part to be optimized. Chapter 5 gives details of the studied parameters.
Fig. 2. Schematic hydraulic plan of the system, in grey boundary for the optimization study 4. Objective function and optimization tool
The objective function is the output function of the system simulation that is going to be
optimized. In this study it is a combination of auxiliary energy and CO emissions. In this case the auxiliary energy is the bought energy that includes the pellets fuel and the electricity for the electrical backup heater. A mathematical function containing both outputs is used for the optimization and explained in chapter 6.
The program Genopt has been used as optimization tool [15]. The program was developed to optimize cost functions in building simulations. It has already been applied in other simulation studies for parameter optimization with TRNSYS [8]. Genopt has several optimization algorithms inbuilt for different kinds of optimization problems. For this study the Hooke-Jeeves algorithm has been applied [16], an algorithm that has been used in the previous mentioned studies - with good results.
5. Sensitivity analysis
In this chapter the impact of several system parameters on the performance of the system are shown and discussed. These were chosen based on previous experience and on work in the literature. The impact of these parameters on the objective function, the annual auxiliary energy (defined in chapter 4) and the CO emissions, are illustrated in figure 3: auxiliary heated standby volume (3a); the hysteresis for the boiler control of this standby volume, dT (3b); the minimum boiler power, Pbmin (3c), with the maximum always fixed to 12 kW; and the set temperature for the standby volume, Tboil (3d). The lower limit for the size of the standby volume and the boiler temperature is determined by the DHW load that has to be covered. Thus the analysis has been performed only for the parameter range where the complete load can be covered. The default values were: standby volume 140 litres; dT of 15 K; Pbmin of 3.4 kW;and Tboil of 70°C. All studied parameters influence the number of starts and stops of the boiler which in turn influence the boiler efficiency and the CO emissions. The CO emissions are not a linear function of the number start/stop of the boiler. This can be seen in figure 3c. In fact, the total CO emissions are a function of the start/stop emission and the emissions during operation that vary with the
combustion power. Operating the boiler with low modulation power reduces the number of
DHW
Space heating Solar
collectors
10 m2 Standby
volume
Pellet boiler
Buffer store 730 ltr
start/stop emissions but not necessarily the total emissions if the CO emissions for low combustion power are significantly higher than for the nominal combustion power [5].
From figure 3 it can be seen that for the fuel consumption the set temperature of the standby volume and for the CO emission modulation band width have the highest impact. The modulation band width influences the number of starts and stops of the boiler and the average combustion power which in turn both influence the CO emissions.
Fig. 3a
536 485 992
673
10 12 14 16 18 20 22
50 100 150 200 250 300
Standby volume [ltr]
CO Emissions [kg]
17.0 17.5 18.0 18.5 19.0 19.5 20.0
Auxiliary energy [MWh]
Auxilliary energy CO emissions
Number of start/stop Tboil=70°C
Pbmin=3.4 kW dT=15 K
Fig. 3b
400 317 582468
699 886
10 12 14 16 18 20 22
5 10 15 20 25 30 35 40
dT [K]
CO Emissions [kg]
17.0 17.5 18.0 18.5 19.0 19.5 20.0
Auxiliary energy [MWh]
Number of start/stop
CO emissions Auxilliary energy Vsb=140 ltr
Pbmin=3.4 kW Tboilmin=60°C
Fig. 3c
2758 2646 2450 1831
971 699
10 12 14 16 18 20 22
3 5 7 9 11 13
Pbmin [kW]
CO Emissions [kg]
17.0 17.5 18.0 18.5 19.0 19.5 20.0
Auxiliary energy [MWh]
Number of start/stop
CO emissions
Auxilliary energy
Tboil=70°C Vsb=140 ltr dT=15 K
Fig. 3d
644 620 702 676
805
10 12 14 16 18 20 22
60 70 80 90
Tboil [°C]
CO Emissions [kg]
17.0 17.5 18.0 18.5 19.0 19.5 20.0
Auxiliary energy [MWh]
Number of start/stop
CO emissions
Auxilliary energy Vsb=140 ltr
Pbmin=3.4 kW dT=15 K
Fig. 3. Parametric study of the standby volume size (a), dT of the boiler start/stop (b), the lower limit of the boiler modulation band width (c) and the set temperature of the standby store (d).
6. Optimization method
The sensitivity analysis has shown that each of the studied parameters has a relativly high impact on the auxiliary energy consumption and/or the CO emissions. Consequently all parameters have been included in the optimization process. The variation range of the parameters and the step sizes were defined as input data for GenOpt. The chosen range and step size during the optimization were similar to those used in the sensitivity analysis; however, the algorithm can reduce the step size by 50% near a minimum.
The objective function has been modelled as a function of the auxiliary energy and the CO emissions (mco). Both values shall be as low as possible. The auxiliary energy is the amount of yearly pellet fuel (Qbpel) and the yearly electrical energy (Qauxel) of the electrical back up heater that is heating the standby volume when the boiler is turned off during the summer months.
Fig. 3 shows that the auxiliary energy demand varies between 17.3 and 19.5 MWh and the CO emissions between 12.7 and 20.6 kg. For the optimization procedure 1 MWh auxiliary energy was given the same nominal value as 1 kg CO. The part of the auxiliary energy that is provided from
the electrical heater has been assigned with a primary energy factor of 3. The objective function used in the study uses Xaux and Xco as weighting factors, and where Xaux = 1-Xco.
CO co auxel
bpel
aux Q Q X m
X
ObjFunc= ⋅( +3⋅ )+ ⋅ eq.1
7. Optimization results for one system
The optimization results with GenOpt have been obtained for the combined solar and heating system described in section 3.2. The results are based on the boiler characteristic and the control strategy of the boiler and the interaction between boiler and store. It is obvious that for boilers with different characteristics or control strategies, different optimization results might be obtained. Thus, the achieved parameter combination is only valid for this specific configuration.
Nevertheless, the way the boiler is connected to the store is very common and the characteristics of the boiler are similar to other pellet boiler of this size [4]. Consequently, it can be expected that with the same parameter configuration also good result would be achieved for boilers with similar characteristics. The optimization method can be applied easily for other boilers and system configurations if the characteristics of the pellet boiler are available.
1865
487 487 487
792
10 12 14 16 18 20 22
0.00 0.25 0.50 0.75 1.00
Xaux, 1-XCO
CO Emissions [kg]
17.0 17.5 18.0 18.5 19.0 19.5 20.0
Auxiliary energy [MWh]
Number of start/stop
CO emissions
Auxilliary energy
Fig. 4. Results of the optimization, depending on weighting factor Xaux and XCO
Table 1. Results of the parameter optimization
Xaux Vsb [ltr]
Tboil
[°C] Pbmin [kW]
dTboiler [K]
0 240 75 12.04 20
0.25 182 70 3.4 15 0.5 182 70 3.4 15 0.75 182 70 3.4 15 1 110 65 3.4 10
The results for the studied system can be seen in figure 4. The optimization has been performed for four weightings of the auxiliary energy and the CO emissions. It shows that the auxiliary energy demand can be reduced to 17.7 MWh and the CO emission to 13.7 kg if the system is optimized for both targets. If the focus is only on the CO emissions, the boiler should be on/off operated with the maximal combustion power and with a standby volume of 240 ltr, 75 °C forward temperature and a dT of 20 K.
8. Discussion and conclusion
A method for the optimization and design of combined solar and pellet heating systems has been proposed. The method has been applied for the optimization of one system. The objective function consists of two parameters that have been weighted: auxiliary energy consumption and CO emissions. If the optimisation is performed for non-zero weighting factors, the results are higher for the parameter than that obtained when the relevant factor is zero. However, the differences are not large: 17.3 MWh contra 17.7 MWh for auxiliary energy; and 12.8 kg CO contra 13.7 kg for CO emissions. The results could be further refined by decreasing the step size in the setup of the optimization tool or by changing the optimisation algorithm to one more suited
for finding a global minimum. The optimization runs are rather time demanding since the system is modelled in detail with a time step of 1.5 min. This is certainly a disadvantage of the method but can be compensated with increased calculation power. The method can in principal be used for all kinds of combined systems provided that the input data for the TRNSYS components are available or can be obtained from measurements.
The optimization method delivers data for the design of the system. The optimal standby store size for the studied system is 182 ltr, boiler temperature is 70 °C, the hysteresis for the boiler control is 15 K and the complete modulation width of the boiler (3.4-12 kW) should be used. For the studied system a relatively large standby volume on a minimum temperature level is more advantageous than a small volume on high temperature, even though this would give additional solar gains.
Acknowledgement
We are grateful to the Nordic Energy Research for their financial support for this work.
References
[1] Fiedler, F. (2004). "The state of the art of small-scale pellet-based heating systems and relevant regulations in Sweden, Austria and Germany." Renewable and Sustainable Energy Reviews 8(3).
[2] SBBA - Swedish Heating Boilers and Burners Association, Stockholm, Sweden, http://www.sbba.se/
[3] Äfab, Bioenergy consulting, “Pelletseldning” 2003-2004, http://www.afabinfo.com/
[4] Persson, T., F. Fiedler, et al. (2006). Increasing efficiency and decreasing CO-emissions for a combined solar and wood pellet heating system for single-family houses. Pellets 2006 Conference, Sweden.
[5] Fiedler, F., S. Nordlander, et al. (2006). "Thermal performance of combined solar and pellet heating systems." Renewable Energy 31(1): 73-88.
[6] Weiss, W., Ed. Solar Heating Systems for Houses, A Design Handbook for Solar Combisystems.
International Energy Agency, IEA, Solar Heating & Cooling Programme, James & James, Ltd, London, United Kingdom, 2003. ISBN 1 902916 46 8.
[7] Klein, S., et al. (2004). TRNSYS 16.0 Transient Simulation Program, SEL, University of Winsconsin, Madison, WI, USA.
[8] Bales, C. (2003). Reports On Solar Combisystems Modelled in Task 26 (System Description, Modelling, Sensitivity, Optimisation), Appendix 6: Generic System #11: Space Heating Store With DHW Load Side Heat Exchanger(S) And External Auxiliary Boiler. Paris, France, IEA-SHC Task 26.
[9] Furbo, S., Nordic energy research cooperation on solar combisystems, EuroSun 2006, Glasgow [10] Persson, T., F. Fiedler, and S. Nordlander, (2006). Methodology for identifying parameters for the
TRNSYS model Type 210 – wood pellet stoves and boilers,SERC Report, ISRN DU-SERC--92--SE.
2006, Borlänge, Sweden, (Available at: http://dalea.du.se/research/)
[11] Thür, A., S. Furbo, L.J. Shah, Energy savings for solar heat-ing systems. Procedings 1 of Eurosun 2004, pp. 715-724, CD-ROM, 2004, Freiburg, Germany
[12] Nordlander, S. (2003). TRNSYS model for Type 210, Pellet stove with liquid heat exchanger. SERC Report, ISRN DU-SERC--78--SE. Borlänge, Sweden, (Available at: http://dalea.du.se/research/) [13] Drück, H., Multiport Store - Model for TRNSYS, Type 140, Version 1.99B. Institut für
Thermodynamik und Wärmetechnik, Universität Stuttgart, Stuttgart, Germany, 2000.
[14] Bales, C., COMBITEST. A New Test Method for Thermal Stores Used in Solar Combisystems.
Doctorial thesis, 2004, Department of Building Technology, Chalmers University of Technology, Göteborg, Sweden. ISBN 91-7291-465-3.
[15] Wetter, M. (2002). GenOpt, Generic Optimization Program, Lawrence Berkeley Nat. Lab. CA, USA.
[16] Hooke, R. and T. A. Jeeves (1961). “’Direct search’ solution of numerical and statistical problems.” J.
Assoc. Comp. Mach 8((2)): 212–229.
