Simulation-based optimization of solar thermal systems

In this PhD thesis application of numerical methods to simulation-based optimization of solar thermal systems both in the planning process and in operation is investigated. The optimization process starting from definition of a target function and specification of optimization parameters, moving forward with application of an optimization algorithm and ending up by sensitivity analysis at the found optimum, is described in detail. A hybrid genetic CHC – binary search algorithm is proposed and applied. As a combination of reliable global genetic algorithm with fast local binary search, the hybrid algorithm is computationally efficient, especially due to good parallelization, and reliable in finding the global optimum. Application of the algorithm to design optimization of the solar heating combisystem shows optimization potential of around 13% in terms of solar energy costs or 19 percent points in terms of extended fractional energy savings when compared to the system configuration planned by the experts. Pareto front is built showing the optimal solar energy costs for desired energy savings of the system, or vice versa. Influence of variation of domestic hot water and space heating demand as well as geographical location on Pareto front and optimal combisystem configuration is investigated. To determine the most important parameters and quantify their influence on the solar energy costs function, three methods of the global sensitivity analysis: multiple linear regression, Morris method and extended Fourier amplitude sensitivity test are applied in two parameter spaces around the optimum. To overcome dimensionality problem when optimizing solar heating system in operation, splitting long year optimization into many short ones is proposed and applied to optimization of flow rates on hourly basis. Only negligible potential of 0.3 per cent points in terms of extended fractional energy savings is determined. A significant potential is shown in another way for dynamic optimization of the auxiliary heater control settings. In the last part of the thesis a control-based anti-stagnation approach consisting of induced inefficient daily collector operation and nightly cooling is proposed and theoretically investigated. Minimal specific store volume required for stagnation-free operation of the solar heating system is calculated for ten sunniest days in a row without heat consumption. Influence of location, collector thermal loss coefficient, solar radiation, day and night ambient temperature is analysed. Practical implementation of the approach reduces the excess thermal energy by 33% during the induced inefficient collector operation compared to the usual control strategy.