| Application of ice storage technology in the building is a key and most efficient method to alleviate the contradiction between the supply and need of power and improve the shape of power load. According to relevant research results, among the mature technologies in solving peak and valley difference of power, the conversion efficiency of ice storage is the highest. But surveys of EPRI and ASHRAE RP-766 on the ice storage systems show that because of improperly control, many of them run poorly. Properly use of optimization and control technology is crucial and of great value to the ice storage system.1. Based on the meteorological parameters of test reference year (TRY) in Xi'an, the dynamic simulation program calculates the hourly cooling loads of an office building between April and September. These results become the basic data in the prediction of meteorological parameters and cooling loads. The statistical analysis and data process of prediction parameters show mat the nonlinear relationship between cooling load and its influential factors is complicated. Among them, temperature has a great impact on the cooling load.2. Accurate prediction of meteorological parameters and cooling loads of building is the foundation and premises of the optimization and control in ice storage system. In China and abroad, almost all of the dynamic load simulation software cannot calculate the hourly cooling load of the building, under random weather and building conditions. So, the research of hourly meteorological parameters and cooling load prediction is a must.The hourly meteorological parameters are the basic inputs in the determination of the hourly cooling load of the building. Several hourly temperature prediction methods are compared and analyzed. It shows mat among short-term prediction methods, ASHRAE coefficients method is the best. The moving-average of temperature takes advantage of the online measured data to predict the temperature of the next stage. This makes the prediction become more accurate. The prediction of solar radiation and humidity ratio is also discussed.3. In China, for the first time, the training data is generated through dynamic load program DeST,which is authoritative and reliable. The hourly meteorological parameters of test reference year (TRY) are scientific, full and accurate. Develop the artificial neural network (ANN) program to predict the hourly cooling load of the building under arbitrary meteorological conditions. The result of cooling load prediction is true, believable and satisfying. The prediction precision of ANN network is almost the same with that of the others introduced in foreign documents.4. The ice storage system is a large-scale system. The optimization of large-scale system must use the method of decomposition-coordination. The hierarchical structure, status and function of optimization and control are different. Optimization is the coordination of varied controls, and control is the concretion and implementation of optimization rules.5. The optimization problem of ice storage system is nonlinear in both objective function and constraint conditions. Under reasonable assumptions, the mathematical model of optimization of ice storage system is founded. Nonlinear optimization is implemented and analyzed in an office building to minimize the operating cost. Nonlinear optimization is closer to the real situation than the simplified linear ones.6. Since the error in both the meteorological parameters and cooling load prediction is unavoidable, the online correction of prediction and offline optimization results is needed. At present, the practical control of ice storage based on the correction of load prediction and off-line optimization is not seen in literature. Since the recent and short-range prediction is always better than the mid-range and long-range prediction, the solution of online correction is brought forward. It integrates the short-range prediction, online measurement and recent-range correction of prediction. The operati...
|
PDF Full Text Request