Validating the performance simulation program SOLDES using data from an operating solar desalination plant

A computer program (named SOLDES) was developed to simulate the operation of solar desalination plants which utilize evacuated tube collectors, heat accumulators and multiple-effect distillation (MED) systems. The heat accumulator used is of the thermally stratified type using pure water as the storage fluid. The procedure was written in Fortran language and consists of a main program, 22 sub-programs, two system data files and four meteorological data files. The absorber area of the solar collector field can be varied between 500 m(2) and 20,000 m2; the storage capacity per unit collector area of the heat accumulator can vary between 0.05 and 1.00m(3)/m(2); the capacity of the evaporator can be varied between 100m(3)/d to 2000 m(3)/d. The heat collecting system uses a bypass circuit to allow the heat collecting fluid (pure water) to recirculate back to the solar collector field when the outlet temperature from the collector field is below a setpoint. When the collector outlet temperature rises above the set-point, operation is switched over to the accumulator side. A solar-cell-type controller is used to start and stop the water circulating pump of the collector field. The operation of the MED evaporator is controlled by the state of charge of the heat accumulator by the use of set-point switches which allow the evaporator to start up when the accumulator water temperature is above a set-point and to shut down if the water temperature drops below the set point. In order to validate the SOLDES program, a comparison was made between the predicted results of the program and the actual measured data from a solar plant of similar design features to the simulation program. The selected plant was the one in actual operation in Abu Dhabi, UAE, which has almost identical design features as the simulation program and has been in operation since 1984. The data from the plant collected during 1985 were used to compare the simulation results for the months of January and June. These two months were found to be typical of a winter month (January) and of summer months (June). Except for days when a plant interruption took place, such as a power failure, the agreement between the measured and simulation data appears to be quite good.