Numerical simulation of a new biomass/solar microcogeneration

Transcription

1 Numerical simulation of a new biomass/solar microcogeneration ORC system Ana I. Palmero-Marrero 1, João Soares 2 and Armando Oliveira 3 1 University of Porto, Dept Mech.Eng., Rua Dr. Roberto Frias, , Porto, Portugal, apalmero@fe.up.pt 2 University of Porto, Dept Mech.Eng., Rua Dr. Roberto Frias, , Porto, Portugal, joaosoares@fe.up.pt 3 University of Porto, Dept Mech.Eng., Rua Dr. Roberto Frias, , Porto, Portugal, acoliv@fe.up.pt Abstract: This paper presents numerical simulation results for a new prototype of a 6 kwe solar/biomass system. The system uses a micro-cogeneration Organic Rankine Cycle (ORC), driven by a combination of solar thermal and biomass sources. Both sources may be used separately or combined. The solar thermal energy is obtained through medium-temperature concentrating compound parabolic solar collectors (CPC) with evacuated tubes. This system is able to generate electrical energy and useful heat at two different temperature levels. The computer model of the overall system was developed by combining EES (Engineering Equation Solver) with TRNSYS software. EES was used for the power cycle calculation, and TRNSYS for simulating the solar circuit and overall system. The ORC/power cycle model consists of a set of algebraic equations representing the thermodynamic behaviour of each ORC component. The ORC simulated by EES includes the following components: evaporator, expander, generator, regenerator, condenser and pump. For an electrical output equal to 6 kw, a fluid inlet temperature in the ORC evaporator of 180ºC was assumed, achieved with the solar collectors and biomass boiler. The ORC is expected to operate at nearly steady-state conditions, as the cycle driving temperature is kept approximately constant. However, TRNSYS is needed to take into account the transient behaviour of solar energy. The TRNSYS model includes the following components: solar collector, boiler, thermal storage, economizer, pumps and control systems. The system is under development and a prototype will be installed and tested in Benguerir (Morocco), in the framework of the REELCOOP project, funded by the EU. For solar resource assessment purposes, Meteonorm software was used to generate climatic data of Benguerir. Results of system simulation under different operating conditions are presented. Annual solar fraction and annual global electrical efficiency for different collector areas and storage capacities were calculated. Keywords: organic Rankine cycle, micro-cogeneration, solar energy, biomass

2 1. INTRODUCTION Micro-generation is the decentralised production of electricity, through different means (micro-turbines, fuel cells, Stirling engines, small internal combustion engines, PV cells) with an electrical power output up to 50 kw. Microcogeneration, or micro-chp, is the combination of micro-generation with useful heat. The use of solar energy with conventional energy sources, for combined heat and power (CHP) in buildings, reduces pollutant emissions and offers energy savings. Distributed power generation in small, decentralised units, is expected to help reducing emissions and saving grid capacity, while also providing opportunities for renewable energy. It could thus form a constituent part of a more sustainable future (Pehnt et al., 2006). Among the several existing technological solutions for micro-cogeneration, Organic Rankine Cycle (ORC) systems are an interesting solution in cases where the heat demand is significantly larger than electricity needs, which is the case of residential and also other buildings (Oliveira et al., 2014). Micro-ORC systems driven by renewable energy sources have attracted the attention of many researchers. Either theoretical or experimental works have been carried out, including developments and improvements in system components (Facão et al., 2008), (Palmero-Marrero and Oliveira, 2009), (Quoilin et al., 2011), (Qiu et al., 2011), (Twomey et al., 2013), (Jradi et al., 2014). In the present paper, a micro-cogeneration ORC system driven by a combination of solar thermal and biomass sources is simulated. Both sources may be used separately or combined. High efficiency vacuum-cpc solar collectors are used, and the power cycle uses a highly efficient rotary lobe expander. The system is under development and a prototype is being built, under the framework of the REELCOOP project, funded by the EU (REELCOOP, 2015). Preliminary simulation results were presented in the SET 2014 (Oliveira et al., 2014). In the next sections, detailed numerical simulations for specific climatic and operational conditions are presented. 2. DESCRIPTION OF THE SYSTEM AND MODEL In this work, the computer model of the overall system was developed by combining EES (Engineering Equation Solver) (EES, 2015) with TRNSYS (TRNSYS, 2013). EES was used for the power cycle calculation, and TRNSYS for simulating the solar circuit and overall system. In this system, the power cycle is based on the Organic Rankine Cycle (ORC) and is driven by solar thermal energy and biomass combustion. The operating temperature of the solar thermal collector varies between 100ºC and180ºc, and the utilization of the biomass boiler guarantees an evaporator inlet temperature of 180ºC.The first step was the development of the ORC / power cycle model in EES. The model consists of a set of algebraic equations representing the thermodynamic behaviour of each ORC component, under steady state. The developed program may be used in connection with TRNSYS, as discussed later, or as a standalone program, able to predict ORC efficiency and output electrical power, besides relevant cycle properties, using as inputs the useful heat input and the efficiencies of several components such as expander, generator or heat exchangers. Concerning the development of the TRNSYS model, for solar resource assessment purposes METEONORM (Meteotest, 2015) was used to generate data of Global Horizontal Irradiation, GHI [kwh/m2], solar azimuth and height [º] and ambient temperature [ºC], on an hourly basis for a typical year. The chosen location was Benguerir (Morrocco), where the prototype will be installed and tested. For this city, the data from Meteonorm and local data were compared - see Fig. 1. In this case, the Meteonorm data were obtained with interpolation from nearby meteorological stations. Horizontal Irradiation (kw h/m 2 ) G T a G d Month METEO Figure 1: Global (G) and diffuse (Gd) solar radiation on horizontal surface and ambient temperature (Ta) with Meteonorm data and local data for Benguerir. Local Ta (ºC) 2

3 The major differences between Meteonorm and local data concern horizontal global irradiation (G) and ambient temperature (Ta) for July, August and September (difference lower than 10% for these months) Definition of system components The system was presented and schematically represented in Oliveira et al., The solar circuit was simulated by TRNSYS and included these components: solar collector, thermal storage, boiler, economizer, pumps and control systems. The solar collector is a new CPC collector with evacuated tubes developed in the framework of the REELCOOP project. The Organic Rankine Cycle (ORC) was simulated by EES and included these components: evaporator, expander, generator, regenerator, condenser and pump. The expander was connected to a generator with a power output of 6 kw. It is assumed that the ORC operates under steady-state conditions, as the cycle driving temperature will be kept approximately constant. Even if the condenser temperature changes, due to changes in ambient temperature, this is a reasonable simplification due to the low system inertia. Different working fluids are used in the system: thermal oil for the solar sub-system, water for the condenser, and Solkatherm (SES36) fluid for the ORC. The Solkatherm fluid is a non-flammable refrigerant with low toxicity. Another important reason for using SES36 in the ORC system is the possibility of using energy at a relatively low temperature. The characteristics of each component simulated in TRNSYS are taken as follows: - Solar collector. This is a medium-temperature non-tracking concentrating compound parabolic solar collector (CPC) with evacuated tubes. The CPC collectors have an external reflective concentrator located behind an evacuated tube which houses the absorber and heat transfer fluid. The performance is obtained by interpolating bi-axial IAM data. - Thermal storage. The thermal storage was modelled by assuming that the tank consists of N (N = 6) fully-mixed equal volume segments. The degree of thermal stratification is determined by the value of N. The performance of the fluid-filled sensible energy storage tank was obtained considering negligible heat losses (perfect insulation). The fluid is oil (TERMOL 5HT). - Boiler. The boiler is a heater that uses biomass as combustible to elevate the temperature of a fluid using external proportional control. External proportional control (an input signal between 0 and 1) is in effect as long as a fluid set point temperature is not exceeded. In this case, the fluid is oil, the set point temperature is 180ºC and the boiler efficiency is Economizer. The economizer was modelled as a zero capacitance counter-flow sensible heat exchanger. The cold side input was assumed to be water. The effectiveness was calculated by using the overall heat transfer coefficient (UA). - Pumps. The pump model computes a mass flow rate using a variable control function and a fixed maximum flow capacity. For the solar circuit simulated in TRNSYS, two pumps are considered: one between the thermal storage tank and the solar collectors and another between the Organic Rankine Cycle (ORC) circuit and solar circuit. - Control system. The on/off differential controller generates a control function which can have a value of 1 or 0. The value of the control signal is chosen as a function of the difference between upper and lower temperatures, compared with two dead band temperature differences. In the solar circuit two control functions are considered. One control allows the flow between the solar collector and storage tank, when the collector outlet temperature is higher than the temperature of the fluid at the bottom of the storage. Another control acts between the boiler and storage tank. In this case, the boiler is running when the tank outlet temperature is lower than 180ºC. The inputs of TRNSYS components and EES components are shown in Table 1. As can be seen in Table 1, the solar collector areas selected for the numerical simulation were 100, and 200 m 2, being that m 2 will be the collector area installed in the Benguerir prototype (in Morocco). 3

4 Table 1: Definition of the inputs for each component. Programs Components Inputs Climatic data (Meteonorm) Local: Benguerir Concentration factor (C) = 2.5 Efficiency curve parameters: ƞ 0=0.68; a 1=0.4 W/m 2 K; a 2=0.004 W/m 2 K 2 Flow rate = kg/s/m 2 CPC collector with evacuated tubes Collector area: 100 m 2, m 2, 200 m 2 Collector tilt: adjusted seasonally (twice per year) in the range of solar height values (at solar noon in the solstices).tilt = 9ºN in Summer and 56ºN in Winter Collector azimuth = 0º Fluid specific heat = 2610 J/kgºC (oil TERMOL 5HT) TRNSYS IAM data: file created through the numerical simulation of the novel CPC collector Tank volume (V tank): 2 m 3, 5 m 3, 8 m 3 Stratified: 6 fully-mixed equal volume segments Thermal storage Tank losses: negligible Cold-side temperature (from Evaporator) : varies with evaporator heat Cold-side flowrate (from Evaporator)= 0.20 kg/s Rated capacity = 60 kw Boiler Set point temperature = 180ºC Boiler efficiency = 0.9 Economizer Pump Overall heat transfer coefficient of exchanger (UA) = 2.8 W/K Maximum power = 0.6 kw Useful heat: varies with expander efficiency (45.5 kw, 42.5 kw and 40 kw) Evaporator Outlet temperature = 145ºC Superheating = 5ºC Expander Efficiency (ƞ exp): 0.65, 0.70, 0.75 EES Generator Regenerator Efficiency = 0.95 Net electrical power = 6 kw T min = 10ºC Efficiency = 0.82 Condenser Condenser temperature = 45ºC 3. SIMULATION RESULTS Using the EES program, different ORC cycle outputs depending on expander efficiency were obtained, as shown in Table 2. The ORC cycle efficiency can be defined as net electricity divided by evaporator heat. 4

5 Table 2: Outputs of the ORC cycle Components Outputs η exp = 0.65 η exp = 0.70 η exp = 0.75 Condenser ( cond) kw kw kw Regenerator ( reg) kw kw kw Evaporator ( evap) 45.5 kw 42.5 kw 40 kw Expander ( Exp) 5.89 kw 6.27 kw 6.65 kw Generator ( G) 5.60 kw 5.96 kw 6.31 kw ORC cycle effic. (Ƞ elec,net,orc) For TRNSYS, the input values shown in Table 1 were considered. The monthly and annual performances for different collector areas are shown in Table 3, with: a) A col=100 m 2, b) m 2, and c) 200 m 2, in all cases with an expander efficiency (ƞ exp) equal to 0.75 and a tank volume of 5 m 3. The tables contain values of incident solar radiation on the collector surface (G coll), useful heat gain on the solar collector fluid (Q usef_coll), energy obtained by the boiler (Q aux) and solar fraction (f, defined as the ratio between the useful solar heat and total heat provided to the power block). For the calculations, a collector tilt equal to 9º was used for the summer period (from April to September) and 56º for the winter period (from January to March and from October to December). The operating time considered for the system was equal to 24 hours/day. Note that if the system operates during a shorter daily period, as in the case where electrical needs are not continuous, the solar fractions will be much higher, assuming the system is run during sunshine hours. For instance, with A col=100 m 2 and operation during 16 hours/day, the annual solar fraction would increase from 0.14 to If the operating period was 12 hours/day, the annual solar fraction would increase to 0.28 for this collector area. Table 3.a): Monthly and annual performance results for A col = 100 m 2 (η exp = 0.75, V tank = 5 m 3 ) Month G coll (MWh) Q usef_coll (MWh) Q aux (MWh) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual

6 Table 3.b): Monthly and annual performance results for A col = m 2 (ηexp = 0.75, V tank = 5 m 3 ) Month G coll (MWh) Q usef_coll (MWh) Q aux (MWh) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual Table 3.c): Monthly and annual performance results for A col = 200 m 2 (η exp = 0.75, V tank = 5 m 3 ) Month G coll (MWh) Q usef_coll (MWh) Q aux (MWh) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual For each expander efficiency, the solar fractions for the different collector areas were calculated. The results show that the annual solar fractions for each collector area are practically constant for the different expander efficiencies (f annual=0.14 for A col=100 m 2, f annual=0.22 for A col=168.3 m 2 and f annual=0.25 for A col=200 m 2 ). This is due to a decrease in the tank inlet temperature (from the evaporator) with the decrease of expander efficiency (ƞ exp), leading to an increase in collector useful energy (Q usef_coll). Q usef_coll and the auxiliary energy from the boiler (Q aux) increase as ƞ exp decreases, resulting that the annual solar fraction (f annual) remains practically constant with expander efficiency, for each collector area. The annual thermal performance can be related to the annual net electricity generation of the power block. Assuming that the solar driven CHP plant is running non-stop throughout the year (helped by the boiler), the annual net generated electricity value is 52.6 MWh, with an operating period of 24 h/day. One could define an average solar electrical efficiency, defined as the net electricity divided by auxiliary energy, with the values shown in Table 5. 6

7 Table 5: Annual auxiliary energy, annual solar fraction and annual average solar electrical efficiency (η sol,electr) for different collector areas and V tank= 5 m 3. A col (m 2 ) Q aux_annual (MWh) f annual η sol,electr 100 m m m For each solar collector area, different storage capacities (2 m 3 and 8 m 3 ) were also analysed. For a collector area of 100 m 2, the variation of storage capacity has a small influence on the solar fraction and on the solar electrical efficiency. For a higher collector area, the solar fraction and solar electrical efficiency decreased considerably when the tank volume changed from 5 to 2 m 3. For V tank=2 m 3, when A col=168.3 m 2, f annual=0.178 and η sol,electr=0.183; when A col=200 m 2, f annual=0.180 and η sol,electr= Note that in this case (A col=200 m 2 ), the tank volume of 2 m 3 is insufficient for the utilization of incident solar radiation. The annual global electrical efficiency (η electr,annual) is defined as the total electrical energy generated during one year, divided by the sum of all annual energy inputs boiler input energy plus solar incident radiation. Figure 2 shows the calculated values for different collector areas for V tank=8 m 3. Figure 2: Annual solar fraction (f annual) and annual average global electrical efficiency (η electr,annual) for different collector areas with V tank=8 m 3. We may note that the average (annual) electrical efficiency may reach 9%, with the best expander efficiency value considered (ƞ exp = 0.75) and the lower collector area. The results show that it is not interesting to use higher collector areas, as the efficiency drops. In fact, the increase in incident solar energy, which is proportional to collector area, is not translated into useful energy, as a significant part is dumped (energy excess). This excess could be reduced by increasing storage capacity. Nevertheless, areas above 168 m 2 are not recommendable, additionally when they represent a significant increase in the system initial cost. Figure 3 shows the daily variation of the incident solar radiation on the collector surface (G coll), useful heat gain on the solar collector fluid (Q usef_coll), energy supplied by the boiler (Q boiler) and electrical efficiency (ƞ electr), for 21 st December (a) and 21 st June (b), when A col = m 2 and V tank = 8 m 3. Note that the tilt of the collector was optimized for the solstices (tilt=56º in December and 9º in June). So, the maximum incident solar radiation on the collector were similar for both days (approximately 670 MJ/h) 7

8 Q (MJ/h) Dec. (tilt=56º) Hours Gcoll Qusef coll Qboiler η electr η electr Q (MJ/h) June (tilt=9º) a) b) Hours Gcoll Qusef coll Qboiler η electr η electr Figure 3: Incident solar radiation (G coll), useful heat gain (Q usef_coll), energy supplied by the boiler (Q boiler) and electrical efficiency (ƞ electr), for: a) 21 st December and b) 21 st June, when A col = m 2 and V tank=8 m 3. Note that with an operating period of 24 h/day, the higher electrical efficiency was achieved when the boiler was running without solar radiation (between 8 pm and 6 am). The utilization of biomass as combustible in the boiler for this biomass/solar micro-cogeneration ORC system, is very important for the full day operation. Figure 4 shows the results (solar fraction and average global electrical efficiency) for different storage capacities with A col=168.3 m 2. Figure 4: Annual solar fraction and annual average global electrical efficiency (η electr,annual) for different storage capacity with A col=168.3 m 3. Increasing storage capacity implies a decrease in collector inlet temperature, leading to an increased collector efficiency and annual solar fraction. However, excessive storage capacities don t improve system performance. Also, when the storage capacity varied between 2 to 8 m 3, no significant effect on η electr,annual was obtained, as can be seen in Figure CONCLUSION In the present paper, the numerical simulation of a novel hybrid biomass/solar micro-cogeneration ORC system was presented. It uses a rotary lobe expander coupled to a generator for electricity generation. Its nominal electrical power is equal to 6 kw, and the ORC, using SES36 as working fluid, is driven by thermal oil at a maximum temperature of 180ºC. This temperature is achieved by a combination of two renewable energy sources: solar energy, through the use of highly efficient novel vacuum-cpc thermal collectors, and a biomass boiler. The monthly and annual performances for different collector areas (100 m 2, m 2 and 200 m 2 ) and different expander efficiencies (0.65, 0.70 and 0.75) were calculated. The results showed that the annual solar fractions (f annual) for each collector area were practically constant for the different expander efficiencies. When the operating 8

9 period considered for the system was 24 hours/day and with a storage capacity of 5 m 3, f annual=0.14 for A col=100 m 2, f annual=0.22 for A col=168.3 m 2 and f annual=0.25 for A col=200 m 2. For each solar collector area, different storage capacities (2 m 3 to 8 m 3 ) were studied. Increasing storage capacity implies a decrease in collector inlet temperature, leading to an increased collector efficiency and annual solar fraction. However, excessive storage capacities don t improve system performance. Also, for a given collector area, when the storage capacity varied between 2 to 8 m 3, no significant effect on the average (annual) electrical efficiency (η electr,annual) was obtained. For the higher storage capacity and the best expander efficiency value considered (η exp = 0.75), η electr,annual may reach 9%, with the lower collector area. The results show that it is not interesting to use higher collector areas, as the efficiency drops. Nevertheless, areas above 168 m 2 are not recommendable, additionally when they represent a significant increase in the system initial cost. A prototype of the system is under construction and will be tested in Morocco in 2016/17. More complete results, namely comparison between numerical and experimental results for the operating conditions, will be presented in the near future. ACKNOWLEDGMENTS The REELCOOP project receives funding from the European Union Seventh Framework Programme (FP7/ ), under grant agreement nº All consortium partners are acknowledged, especially those involved with the development of this prototype: University of Evora (Portugal), IRESEN (Morocco), MCG Solar (Portugal) and Termocycle (Poland). 5. REFERENCES EES - Engineering Equation Solver, Program Manual, F-Chart Software (available at Facão J., Palmero-Marrero A., Oliveira A.C., 2008, Analysis of a Solar Assisted Micro-Cogeneration ORC System, International Journal of Low Carbon Technologies, vol.3 n.4, pp Jradi M., Li J., Liu H., Riffat S., 2014, Micro-Scale ORC-Based Combined Heat and Power System Using a Novel Scroll Expander, International Journal of Low Carbon Technologies, Advance Access published February 20, 2014, available online (doi: /ijlct/ctu012). Meteotest, Meteonorm Handbook, Parts I, II and III. Bern, Switzerland (available at Oliveira A.C. et al., 2014, Presentation and Preliminary Simulation of a Biomass/Solar Micro-Cogeneration ORC System, Proceedings SET2014. Palmero-Marrero A., Oliveira A.C, 2009, Economic Analysis of a Solar Assisted Micro-cogeneration Organic Rankine Cycle System, Proceedings SET2009. Pehnt M., Cames M., Fischer C., Praetorius B., Schneider L., Schumacher K., Voβ Jean-P., 2006, Micro Cogeneration: Towards decentralized energy systems, Ed. Springer, Germany. Qiu G., Liu H., Riffat S., 2011, Expanders for Micro-CHP Systems with Organic Rankine Cycle, Applied Thermal Eng, vol.31, pp Quoilin S., Orosz M., Hemond H., Lemort V., 2011, Performance and Design Optimization of a Low-Cost Solar Organic Rankine Cycle for Remote Power Generation, Solar Energy, vol.85, pp REELCOOP project: TRNSYS 17 - A Transient System Simulation Program, Solar Energy Lab, University of Wisconsin-Madison, (available at Twomey B., Jacobs P.A., Gurgenci H., 2013, Dynamic Performance Estimation of Small-Scale Solar Cogeneration with an Organic Rankine Cycle Using a Scroll Expander, Applied Thermal Eng, vol.51, pp