On-line adaptive control of a direct expansion air conditioning system using artificial neural network

A common issue to all controllers, including the previously developed artificial neural network (ANN)-based controller for a direct expansion (DX) air conditioning (A/C) system, developed based on system identification is limited controllable range. To address the issue, an ANN-based on-line adaptive controller has been developed and is reported. The ANN-based on-line adaptive controller was able to control indoor air temperature and humidity simultaneously within the entire expected controllable range by varying compressor and supply fan speeds. The controllability tests for the controller were carried out using an experimental DX A/C system. The test results showed the high control accuracy of the ANN-based on-line adaptive controller developed, within the entire range of operating conditions. It was able to control indoor air dry-bulb and wet-bulb temperatures both near and away from the operating condition at which an ANN-based dynamic model in the ANN-based on-line adaptive controller was initially trained.

► We develop an ANN-based on-line adaptive controller for a DX A/C system.
► It addressed limited controllable range for an earlier ANN-based controller.
► The controller was validated experimentally by controllability tests.
► The controller could work well within the entire range of operating conditions.