Evaluation and parametric modeling of 50 kW organic rankine cycle for waste heat recovery from rural Alaska diesel generator power plants

Abstract

In rural Alaska, there are about 180 villages that run independent electrical power systems using diesel generator sets. A diesel engine generator loses fuel energy in the form of waste heat through the charge air cooler (after cooler), the jacket water cooler, friction, and exhaust. Diesel engine jacket water and exhaust account for about 20% and 30% of the total fuel energy, respectively. In previous studies it has been demonstrated that about 80% of the heat present in jacket water and 50% of the heat from exhaust gases can be recovered for useful purposes such as heating, power generation, refrigeration, and desalination. In this study, the diesel engine waste heat application selected was power generation using an organic Rankine cycle (ORC) heat engine. The basic principle of an ORC system is similar to that of the traditional steam Rankine cycle; the only difference is the working fluid. The working fluids generally used in an ORC are refrigerants, such as R11, R113, R123, R134a, R245fa, and HFE-7000. The working fluid in the ORC system under study is R245fa. A typical ORC consists of a pump, preheater, evaporator, expansion machine (expander), and condenser. The working fluid is pressurized through the pump and supplied to the preheater and evaporator, where it is heated by the heat source. The working fluid exits the evaporator as vapor or liquid/vapor. It expands in the expander, generating power. The low-pressure working fluid exiting the expansion machine is liquefied in the condenser by a cooling source, returned to the pump, and the cycle repeats. At the University of Alaska Fairbanks (UAF) power plant, a lab experimental setup was designed: a hot water loop (heat source) and cold water loop (heat sink) for testing the 50 kW ORC power unit. Different diesel engine waste heat recovery conditions were simulated to study the unit's reliability and performance. After lab testing, the ORC system was installed permanently on a 2 MW Caterpillar diesel engine for jacket water heat recovery in Tok, Alaska, and tested further. These two tests provide for the goals of the present dissertation which are: (i) testing of a 50 kW ORC system for different heat source and heat sink supply conditions, (ii) develop guidelines on applying the present 50 kW ORC system for individual rural Alaska diesel gen-sets, (iii) develop empirical models for the screw expander, (iv) develop heat transfer correlations for single-phase and two-phase evaporation, and two-phase condensation for refrigerant R245fa in the preheater, evaporator and condenser, respectively, and (v) parametric modeling and validation of the present ORC system using the empirical correlations developed for a screw expander and R245fa in heat exchangers to predict the performance of the ORC system for individual diesel generator sets. The lab experimental data were used to plot performance maps for the power unit. These maps were plotted with respect to hot water supply temperature for different ORC parameters, such as heat input to power unit in evaporator and preheater, heat rejection by power unit in condenser, operating power output, payback period, and emissions. An example of how performance maps can be used is included in this dissertation. As detailed in this dissertation, the resulting lab experimental data were used to develop guidelines for independent diesel power plant personnel installing this ORC power unit. The factors influencing selection of a waste heat recovery application (heating or power) are also discussed. A procedure to find a match between the ORC system and any rural diesel generator set is presented. Based on annual electrical load information published in Power Cost Equalization data for individual villages, a list of villages where this ORC system could potentially be beneficial is included. During lab work at the UAF power plant, experimental data were also collected on the refrigerant side (R245fa) of the ORC system. Inlet and outlet pressures and temperatures of each component (evaporator, pump, and expander) of the ORC were measured. Two empirical models to predict screw expander power output were developed. The first model was based on polytropic work output, and the second was based on isentropic work output. Both models predicted screw expander power output within ±10% error limits. Experimental data pertaining to the preheater, evaporator, and condenser were used to develop R245fa heat transfer correlations for single-phase and two-phase evaporation and two-phase condensation in respective heat exchangers. For this study the preheater, evaporator, and condenser were brazed plate heat exchangers (BPHEs). For single-phase heat transfer in the preheater, a Dittus-Boelter type of correlation was developed for R245fa and hot water. For R245fa evaporation in the evaporator, two heat transfer correlations were proposed based on two-phase equation formats given in the literature. For condensation of R245fa in the condenser, one heat transfer correlation was proposed based on a format given in the literature. All the proposed heat transfer correlations were observed to have good agreement with experimental data. Finally, an ORC parametric model for predicting power unit performance (such as power output, heat input, and heat rejection) was developed using the screw expander model and proposed heat transfer correlations for R245fa in heat exchangers. The inputs for the parametric model are heating fluid supply conditions (flow rate and temperature) and cooling fluid supply conditions, generally the only information available in rural Alaska power plant locations. The developed ORC parametric model was validated using both lab experimental data and field installation data. Validation has shown that the ORC computation model is acceptable for predicting ORC performance for different individual diesel gen-sets.