Font Size: a A A

Research And Application Of Gymnasium Lighting Design Method Based On Neural Network Algorithm

Posted on:2021-03-08Degree:MasterType:Thesis
Country:ChinaCandidate:W Y FanFull Text:PDF
GTID:2392330611989089Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
The three-dimensional distribution space characteristics and dynamic change time characteristics of the stadium lighting system make the lighting systems of different stadiums have strong randomness,dynamics and uncertainty,which causes the lighting problems in the stadiums to show more obvious nonlinear relationships.For lighting designers,it will be a big challenge to study and design a high-level stadium lighting system using traditional lighting design methods.To this end,this article is based on neural network algorithms,respectively fusing fuzzy control,particle swarm optimization algorithm(PSO)and genetic algorithm(genetic algorithm,GA)to calculate and research the stadium illuminance,and seek a kind of The calculation method of optimal illuminance is mainly studied as follows:(1)First,the neural network algorithm(ANN)is used to model the stadium illuminance,and then two methods are used to improve and optimize the neural network algorithm.A method that combines the advantages of fuzzy control to study fuzzy neural network control algorithms can make up for the disadvantages of neural network algorithms that rely heavily on samples and are not good at expressing rule-based knowledge.In addition,a method for fusion of PSO-GA to optimize neural network illuminance prediction is proposed.The initial weights and thresholds of neural network are optimized through successive iterations to minimize the sum of prediction deviations of all samples,and the PSO-GA optimized method is obtained.Neural network illuminance prediction model.(2)Taking the actual project as an example,taking the illuminance,comprehensive cost,energy saving and economic benefits as the evaluation indicators,the illuminance prediction model established by the neural network,the fuzzy neural network(ANFIS)illuminance prediction model and the neural network optimized by PSO-GA The network(FSO-GA-ANN)illuminance prediction model is trained,tested,calculated and compared.The results show that the neural network control algorithm fused with PSO-GA optimization has fast prediction speed,small error and high accuracy in the design of stadium lighting scheme.(3)Use DIALux software to model and simulate a basketball gymnasium of the Fourteenth National Games,and simulate the calculation results of ANN,ANFIS and FSO-GA-ANN.Through analysis and comparison of point illuminance plan and contour map,It is concluded that the average illuminance predicted by the FSO-GA-ANN algorithm is closest to the illuminance value specified by the state,avoiding light waste,and achieving the purpose of energy saving;high uniformity of illuminance and low glare are more conducive to the normal progress of the game.It can meet the scientific,reasonable,safe and comfortable lighting requirements in the stadium,and provide a way of thinking for the stadium lighting designers.
Keywords/Search Tags:stadium lighting, illumination, neural network, fuzzy neural network, DIALux software
PDF Full Text Request
Related items