Design optimization and optimal control of energy systems in nearly/net-zero energy buildings

Abstract

Nearly/net zero energy buildings (nZEBs) have been attracted increasing attention particularly when high performance is required in terms of energy-saving,indoor thermal comfortable, environment-friendly and grid-friendly. Increasing attention has been paid on how to design nZEBs in a cost/energy efficient and environment-friendly way.However, there is no exact approach at present for the design and control of the buildings to achieve nearly/net zero energy targets. This is mainly due to the complex interplay of electricity generation/consumption system and energy storage system, automatically and manually controlled systems/elements in the highly integrated buildings. Effective optimization methods are essentially needed for the optimal design and control of energy systems in nZEBs. The aim of this PhD project is to study and develop design optimization methods and optimal scheduling strategies for the energy systems in nZEBs. A comprehensive literature review is presented first. Then, a nZEB simulation platform is developed for the test and analysis of system design and control optimization. Validation of system models is made on the basis of Hong Kong Zero Carbon Building.The performance of nearly/net zero energy buildings is largely affected by the renewable energy system design. The sizes of renewable energy systems for nZEBs are optimized by two optimization methods, including a single objective optimization using Genetic Algorithm and a multi-objectives optimization using Non-dominated Sorting Genetic Algorithm (NSGA-II). Building energy system models and renewable energy system models are developed and adopted, allowing the consideration of the interaction between building energy systems and renewable energy systems in optimization. The performance of the buildings with the optimized renewable energy systems is much better than that of the benchmark building in most scenarios. The single objective optimization can provide the "best" solution directly for a given objective while the multi-objective optimization provides rich information for designers to make better compromised decisions. Due to the intermittent and unstable nature of renewable energy resources, the performance of net zero energy buildings may suffer a great degree of uncertainty compared to traditional buildings without renewable energy systems. Sensitivity analysis is conducted on an optimized renewable energy system (photovoltaic/wind turbine/bio-diesel generator) to investigate the impacts of the variations of input variables on the building performance. Four important design inputs regarding working conditions are concerned in the study, including wind velocity, other load, cooling load and solar radiation. Results show that, with 20% variations in the four variables, the maximum change of the combined objective is about 26.2%. In addition, wind velocity is the most influential factor on the building performance regarding the total cost and/or CO2 emissions, while the building loads (other load and cooling load) should be considered with top priority at the design stage concerning the overall building performance. The performance of the energy system, which integrates photovoltaic and bio-diesel generator, has been found not to be the most efficient. But, compared with the other three design options, its performance is the most robust when the working condition changes. The results also indicate that the use of active electricity generation systems in net zero energy buildings could increase the performance robustness of the building energy systems significantly. The increasing complexity of building energy systems integrated with renewable energy systems requires essentially more intelligent scheduling strategy. An optimal scheduling using nonlinear programming is proposed for the control of energy systems in buildings integrated with electricity generation and thermal energy storage. A case study is conducted to validate the proposed strategy based on the Hong Kong Zero Carbon Building. Two types of grid-connections (i.e., selling electricity to grid is allowed and forbidden respectively) are considered. Results show that significant reductions in carbon dioxide emissions, primary energy consumption and operation cost are achieved by the proposed optimal scheduling strategy.Considering the discrete working ranges of some energy systems, the mixed-integer nonlinear programming (MINLP) approach is further used to solve the optimal scheduling problems. The enhanced scheduling strategy based on MINLP minimizes the overall operation cost day-ahead, including operation energy cost and cost concerning the plant on/off penalty. Four scenarios are investigated and compared to evaluate the performance of the enhanced scheduling strategy. Results show that the strategy can reduce operation energy cost greatly (about 25%) compared with a rule-based strategy and the reduction is even increased to about 47% when a thermal energy storage system is used. The strategy can also reduce the on/off frequency of chillers significantly.