The Application Of Fuzzy Discrete Event System In Air Conditioning Energy Saving Control

At present, 50 percent of the total energy consumption of a building goes to air conditioning system. In the increasingly energy issues concerned world, the energy saving of the air conditioning system becomes very urgent. Usually the air conditioning system is designed according to the largest cooling load, but in the actual run time, load is always changing. There is a statistic shows that the air conditioning system which is designed according to the largest cooling load runs about 90% of the time below 80% load operation. It means that most of the time the load of air conditioning system do not match with the actual load requirements, resulting in energy waste. The existing control methods are all closed loop control that relatively independent to the air supply system、chilled water system and cooling water system. As the air conditioning is a system with large time delay, if we determine the cooling load of the airside according to the difference of the chilled water supply and return water temperature, it will increase the time delay of the system, but if we can predict the load and adjusting the system in advance. It not only can supply the load demand on the requirements, reducing the energy loss, but also can solve the time delay of the system.This paper studies the dynamics methods of forecasting the air conditioning load, analyzing the advantages and disadvantages of common forecasting methods and proposing the method about how to determine the delay time of air conditioning systems. Taking the actual measurement value of the wind sub-system as its historical data and using the neural network algorithm with momentum and changing learning rate. Then predicting the load and evaluating it with root mean square error and average relative error.Modeling of fuzzy discrete event system and FDES automata synthesis have been researched. The synthetic which based on average tensor has been optimized, making the new membership can reflect system requirements better and applying the online control of the FDES to the air conditioning system control.The relation between the load and every sub-system of air conditioning has been analyzed in this paper. Taking the air conditioning system as a whole and determining the optimal chiller load distribution. Then establish the FDES model of the variable (difference of chilled water supply and return water temperature, difference of cooling water supply and return water temperature, difference of supply and return air temperature) which has a reflection to the load changes. Through their synthesis and defuzzification we can get the specific control frequency. At last the method to amend the deviation and how to implement are given.The paper proposes methods to save the energy of air conditioning. There is practical significance to the air conditioning system supply energy on demand to decrease the energy consumption and reduce the time delay.