| Cogeneration of electric and thermal energy through use of combined solar photovoltaic/thermal (PV/T) collectors is a method for improving the overall efficiency of solar electric energy systems. Active cooling maintains the photovoltaic cells closer to their optimum achievable electric output while use of the heat generated by cooling of the cells for satisfying onsite thermal loads improves net energy usage and overall system performance. However, with this system which produces both thermal energy and electric power, maximizing one output can be detrimental to the other. Thus, a proper balance between subsystems must be maintained when satisfying thermal and electric loads in order to maintain comfort conditions in the structure, keep control energy requirements within bounds, and minimize auxiliary energy usage and pump parasitic losses. In this investigation the situation is particularly complex since both system outputs and loads are dependent upon two exogenous weather parameters, which cannot be known a priori, namely, incident solar radiation and ambient dry-bulb temperature. The control of a PV/T system in general and optimization of performance in particular through use of modern (state space) control methods, stochastic weather inputs, and second law of thermodynamics analysis is addressed in this study.; Significant improvement in system performance is noted using optimal control when compared to conventional on/off, multilevel, or proportional controllers for deterministic weather forcing functions. Optimal system control, analyzed first through use of Pontryagin's Minimum Principle and then implemented by specification of a quadratic performance index and solution of matrix Riccati equations, is shown to be a viable and useful strategy for these hybrid systems. Stochastic weather techniques which incorporate temperature/insolation joint probability density matrices and least square constants are found to be a valid method for reducing simulation requirements as long as weather persistence effects are taken into account through use of information derived from Markov transition matrices. System performance optimization and control is successfully implemented by tracking the separate component irreversibilities and by minimizing a performance index based on total system irreversibility. Use of the second law in the analysis of complex energy systems such as this one is shown to be an effective, and increasingly necessary system performance optimization technique. |
