Study Of Building Cooling And Heating Load Forecast Model Based On Electrical Equivalent Model

With the rise in the proportion of building energy consumption accounts for the proportion of global energy consumption, building energy conservation has come to a critical state under the International Energy circumstances. Based on the building energy consumption simulation which include building cooling and heating load, particularly with reference to improve the control strategy of construction equipment is a fundamental measure to implement building energy saving. Therefore, the establishment of a building load forecasting model is a prerequisite for precise prediction of building energy consumption. White box model forward modeling according to the physical characteristics of the building heat, the precision of the model can't be such high if not obtain very detailed construction parameters and has a wealth of architectural modeling knowledge. Black box model which named data driven model establish statistical models based on a large number of historical data and time-consuming, even though the application is limited. Gray box modeling method combine with the two above mentioned advantages, so has a well trend in the field of building load model.Currently, with the increase on building structure and description of environment, parameter identification of black box model become the key to influencing the accuracy of the model. Traditional optimization algorithm of identification results generally present larger dispersion, the identification results can't correspond with the actual system, which influence the accuracy of the models. In this article, the subject background is the National Natural Science Foundation "Energy Net with Variable Mode and Variable Structure in the Framework of Cloud-basic Message". The method of modeling building load forecasting model is studied. In the research, the heat transfer model of capacitors and resistors which according to the physical characteristics of the building is established, in order to make sure the heat transfer theory is consistent with the real model and improve the accuracy of the model, model parameter identification method based on genetic algorithm is put forward, this method can accurately describe the architecture characteristics so as to improve the prediction precision.The specific work of this paper can be summarized briefly as follows: Firstly, according to the electrical equivalent of the envelope method, the building RC model is established based on Kirchhoff's law. In the process of determining the envelope of R, C parameters, through the application of genetic algorithm can adapt to the complex boundary conditions of a global search to identify the characteristics of the model parameters, the theoretical model of envelope and the model after identification have a similar frequency domain characteristics. Modeling of the indoor thermal mass is unpredictable factors, so the 2R2 C empirical model is applied in the modeling of indoor thermal mass, then identification of its parameters comes from training application load data and by the way of inverse modeling. Building load data include two parts: simulated by the TRNSYS simulation and measured by cigarettes factory air conditioning system, data fusion method used to deal with load data in such process. Finally, based on the law of conservation of energy, the total differential equation load model of the building is determined. Known the temperature, radiation values and human occupancy and other parameters the building load can be calculated. In this paper, the application of the electrical equivalent model established in this paper in Qianjiang cigarette factory library achieved a good prediction result. By comparing with the prediction data and the simulated data, we analysis the main factors which leading to the prediction error. At the same time according to the sensitivity analysis of RC Load Forecasting Model, the method to improve load forecasting precision is discussed.