The invention relates to a multi-split
air conditioning system control method based on prediction. The method comprises the steps that firstly, the temperatures and
humidity outside a building are predicted, and the space enclosing structure load and the solar
radiation load of an
air conditioning system are calculated; secondly, the periodic law of the number of people inside the building is predicted; thirdly, the
operating time of the
air conditioning system is predicted; fourthly, operating parameters including the outdoor unit
wind speed, the compressor rotating speed, the
refrigerant flow rate, the indoor unit
wind speed and wind amount, and the indoor unit flow rate are monitored in real time; fifthly, the temperatures and
humidity outside the building, the number of people in the building, the operating parameters and hourly COPs in multiple sampling periods are counted up as a training sample, a statistical or
machine learning method is adopted for calculating all characteristic factors, comfort pre-control time and temperatures, and operating parameters corresponding to optimal COPs; and sixthly, the operating parameters corresponding to the optimal COPs in the to-be-controlled time points are predicted, and thus, the system operating parameters and pre-control time are adjusted. According to the multi-split air conditioning system control method based on prediction,the control precision is high, and the control comfort and energy-saving effect are improved.