| Nowadays,with the improvement of urbanization in our country,the urban population is increasing rapidly and the urban construction land is also in short supply.Super high-rise building has many advantages,such as improving the utilization ratio of urban construction land,comprehensive and iconic advantages,and so on.The super high-rise building more and more favored by city managers can greatly increase the capacity of urban settlement and effectively solve the problem of land.It is more and more necessary to study the super high-rise building because of its lots of advantages.In recent years,a large number of domestic and foreign research projects on super high-rise building have achieved fruitful results.In this paper,the maximum load of super high-rise building is studied.the super high-rise building refers to building which is more than 100 meters in height.The maximum load of the study is the average of the maximum load of 30 minutes in a complete year.Through the analysis of the electricity consumption data of many super high-rise buildings throughout the year,and the analysis of the maximum load data,the BP neural network algorithm and the support vector machine algorithm are used to establish the forecasting model respectively.By comparing the forecasting results of the two models,the forecasting method for the maximum load of super high-rise buildings is found.This paper first introduces the classification and research status of power load forecasting methods and the characteristics and research background of super high-rise building.Then it introduces the BP neural network algorithm and support vector machine algorithm and their establishment of mathematical model and forecasting process in power load forecasting.In the modeling of BP neural network algorithm,the optimal number of neurons in the hidden layer is found by empirical method and trial method.In the modeling of support vector machine algorithm,the RBF kernel function is selected and the parameter g and the penalty coefficient C are optimized by using the grid-search method to obtain the optimal parameters.The two forecasting methods are modeled using MATLAB toolbox respectively and the simulation of the maximum load of super high-rise building is carried out.Through the quantitative analysis of the accuracy and stability of the forecasting results of the two algorithms,it is concluded that the advantages of the support vector machine algorithm are more obvious on the maximum load forecasting of the super high-rise building.Finally,the analysis and summary are made. |
