Optimal Control Study Of Air Conditioning Water System Based On Artificial Fish Algorithm

Variable flow air conditioning system is the leading role of the air conditioning system,however, Chilled water system is the main part of it, is also the important factors influencing the energy consumption of the whole system. With the development of the air conditioning system and the complexity is increasing day by day, the requirement of the control links also became more and more higher. For variable flow air conditioning system, mainly using the differential pressure control including the constant pressure difference control and variable pressure difference control, because of the load changing constantly in practical use, the working state of the system are also changing, what the differential pressure be setted to be fixed is tend to make the system working state volatility, that is not conducive for stabling control and improveing the service life of the equipment, also caused a lot of unnecessary energy consumption. Variable pressure difference control can adjust reasonable pressure difference value with load change, so it can improve the adaptability of the system and become a research hotspot gradually in recent years.This paper put the variable flow chilled water air conditioning system as the object of control, introduced the concept of the most unfavorable thermodynamic loop, introduced into the variable pressure difference control strategy on the basis of this concept, elaborated the advantage of variable pressure difference control and how to realize it, this paper introduced the linear adjustment algorithm of variable pressure difference value.Controller is using the PTD neural network based on fish algorithm, PID control is simple and reliable, but is not suitable for nonlinear system, the control effect is not very desirable in the application of complex systems and time-varying system, the anti-interference and adaptive ability is weak; Neural network has strong nonlinear ability and fault tolerance, self ability to adapt is stronger, but the conventional neural network's structure is complex, calculation is lager. Air conditioning system is a complex nonlinear time-varying systems,PID neural network that with PID and neural network complementary advantages just solve the problem of air conditioning system control. Based on this, and then use the artificial fish algorithm with strong optimization ability training PID neural network for the further optimization. Apply PID neural network be well optimized to the differential pressure control loop, it can make the pressure difference at the end stable at the set value. On the other hand, the linear adjustment algorithm of variable differential pressure setting can't tracking the change of the load well when the load change dramatically, differential pressure set point adjustment speed is not enough, the precision is also need to be strengthened. Therefore, optimizing linear adjustment algorithm by fish algorithm, adjust the parameters of the algorithm according to load variation to enhance the ability of algorithm.Finally, established the model of the related equipment, use of MATLAB simulation, compared to constant pressure difference control and variable pressure difference control, verify the effect of fish algorithm in optimizing PID neural network and linear adjustment algorithm on variable pressure difference value, the results showed that the fish algorithm optimization effect is obvious.