ANN control on optimum on/off time of intermittent air conditioning systems in shallow underground projects

Taking four main affecting factors as input and optimum on/off time as output, develops an artificial neutral network (ANN) prediction model. Analyses the variation of indoor thermal environment under the influence of soil heat storage. Based on a CFD simulation, builds up a coupling heat transfer model of indoor air and envelopes when using intermittent air conditioning. The outcomes of simulation are taken as inputs to train the neutral network model. The optimized neutral network model can be used to predict the optimum on/off time under various complicated non-linear conditions. Verifies the ANN control by a small system model.