Font Size: a A A

Research On Energy Consumption Data Analysis Model Of Public Institutions Of Tianjin-based On Genetic Neural Network

Posted on:2015-04-07Degree:MasterType:Thesis
Country:ChinaCandidate:Z H ZhangFull Text:PDF
GTID:2298330467955352Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of China’s economy and society, energy issues have becomeincreasingly prominent society, public institutions as an important group of social energyconsumption analysis and forecast of its energy consumption is especially important.Energyconsumption of the in public institutions was influenced by many factors such as statisticalnumbers, the construction area, the number of vehicles and other factors include energymanagement systems. Other non-identified factors bring some energy analysis difficult. Inthis paper, based on the platform of Tianjin public institutions of historical data on energyconsumption statistics Tianjin public institutions were analyzed and predicted. The main workincludes the following areas.1) Overview of Tianjin public agencies to analyze the trend in recent years, drawn fromquantitative public institutions, composition, etc.Analysis of the types of public institutions and energy expenditure type, draw publicinstitutions consumption characteristics: non-profit, lack of motivation and relative stabilitycontrol. Factors that affect energy consumption from public bodies, including: theconstruction area, the number of public vehicles, with the number of energy, publicinstitutions and other aspects of energy consumption data types in Tianjin from2005to2010public bodies described in detail. The total energy consumption of public institutions tostabilize little change in per capita energy consumption reduced year by year.2) Determine the impact of factors affecting the energy consumption of public institutions.Many factors that affect energy consumption of public institutions, including not onlybuilding type, structure HVAC, lighting systems and other hardware facilities include nonuncertainties. Based on energy consumption data obtained using gray relational theory of theexisting energy consumption indicators gray correlation analysis, key indicators of the impactof public institutions draw power consumption are: building area, with a number of energy,the number of preparation and organization types.3) Energy analysis model of public institutions was established based on genetic neuralnetwork. The energy consumption of public institutions is highly nonlinear characteristics,and artificial neural network has good linearity, self-learning and adaptive ability, and issuitable for processing multivariable systems and good fault tolerance, BP neural network canselect consumption forecast. BP neural network’s own shortcomings, the initial weights canlead to blindness selected network into a local minimum, while the genetic algorithm basedon genetics and natural selection have global optimization capability, select the genetic algorithm to optimize BP neural network. By generating the initial population of the geneticalgorithm, selection, crossover and mutation determine the initial weights and thresholds ofthe neural network and the training of the network structure to overcome the shortcomings ofBP neural network.4) Using MATLAB language to simulate the energy consumption prediction model, selectgroup of100public institutions in Tianjin consumption statistics for genetic neural networktraining, validation of the model, and the model with the standard BP neural network tocompare the results BP neural network model is superior to the standard, and the use ofmodels for energy consumption five public institutions were predicted.
Keywords/Search Tags:Genetic Algorithms, BP neural network, Public institutions, Energy analysis
PDF Full Text Request
Related items