基于人工鱼群算法的空调水系统优化控制研究

发布时间：2018-08-25 14:10

【摘要】：变流量空调系统是中央空调系统的主角,而冷冻水系统又是空调系统的主要组成部分,也是影响整个系统能耗的重要因素。随着空调系统的发展和复杂程度的日益提高,对其控制环节的要求也越来越高。对于变流量空调系统,采用较多的是压差控制包括定压差控制和变压差控制,因为实际使用中负荷是不断变化的,系统的工作状态也在不断变化,固定的压差设定值往往使系统的工作状态具有较大起伏,不利于稳定控制和提高设备使用寿命,也造成很多不必要的能耗。而变压差控制可以随负荷变化的情况合理的调整压差设定值,从而提高了系统的适应能力,逐渐成为近些年研究的热点。本文以变流量空调冷冻水系统为控制对象,介绍了最不利热力环路的概念,并在此概念的基础上引入变压差控制策略,详细说明了采用变压差控制的优势和如何实现变压差控制,介绍了变压差设定值的线性调整算法。在控制器方面采用的是基于鱼群算法的PID神经网络控制器,PID控制简单可靠,但是不适用于非线性系统,在复杂系统和时变系统中的应用效果也不尽如人意,抗干扰和自适应能力较弱；神经网络具有很强的非线性能力和容错能力,适应能力较强,但是常规的神经网络结构复杂,计算量大。空调系统是一个复杂的时变的非线性系统,将PID与神经网络优势互补形成的PID神经网络则正好解决空调系统的控制问题。在此基础上,再采用具有强大寻优能力的人工鱼群算法对PID神经网络进行训练,对其进一步优化。将优化好的PID神经网络应用到压差控制回路中,可以使末端压差快速高效地稳定在设定值。另一方面,变压差设定的线性调整算法在负荷大幅度变化时无法很好地跟踪负荷的变化,压差设定值的调整速度不够,精确度也有待加强。为此,再次应用鱼群算法对线性调整算法进行优化,根据负荷变化调整算法的参数,增强算法调整压差设定值的能力。最后,建立相关设备的模型,利用MATLAB仿真,对比定压差控制与变压差控制效果,验证鱼群算法在优化PID神经网络和优化变压差设定值线性调整算法方面的效果。仿真结果表明,鱼群算法的优化效果明显。
[Abstract]:Variable flow air conditioning system is the main role of the central air conditioning system, and the chilled water system is the main component of the air conditioning system, and it is also an important factor affecting the energy consumption of the whole system. With the development of air conditioning system and the increasing complexity, the requirements of its control links are becoming higher and higher. For the variable flow air conditioning system, the pressure difference control includes constant pressure difference control and variable pressure difference control, because the load is constantly changing in actual use, and the working state of the system is also changing. The fixed pressure-difference value often makes the working state of the system fluctuate greatly, which is not conducive to stable control and increase the service life of the equipment, and also results in a lot of unnecessary energy consumption. The variable pressure difference control can adjust the pressure-difference setting value reasonably with the change of load, thus improving the adaptability of the system, and gradually becoming the hot spot of research in recent years. In this paper, the concept of the most unfavorable thermal loop is introduced, and the variable pressure difference control strategy is introduced on the basis of the variable flow air conditioning chilled water system as the control object. The advantages of variable pressure difference control and how to realize variable pressure difference control are described in detail. The linear adjustment algorithm of variable pressure difference setting value is introduced. In the controller, the PID neural network controller based on fish swarm algorithm is simple and reliable, but it is not suitable for nonlinear system, and the application effect in complex system and time-varying system is not satisfactory. The ability of anti-interference and self-adaptation is weak, and the neural network has strong nonlinear and fault-tolerant ability, but the conventional neural network structure is complex and the computation is large. Air conditioning system is a complex and time-varying nonlinear system. PID neural network, which combines PID and neural network, can solve the control problem of air conditioning system. On this basis, the artificial fish swarm algorithm with strong optimization ability is used to train the PID neural network and optimize it further. By applying the optimized PID neural network to the differential pressure control loop, the end pressure difference can be quickly and efficiently stabilized at the set value. On the other hand, the linear adjustment algorithm of variable pressure difference setting can not track the change of load well when the load changes greatly, the adjustment speed of the pressure difference setting value is not enough, and the accuracy needs to be strengthened. Therefore, the linear adjustment algorithm is optimized by using the fish swarm algorithm again, and the ability of adjusting the pressure-difference setting value is enhanced according to the load change adjustment algorithm parameters. Finally, the model of related equipment is established, and the effects of constant pressure difference control and variable pressure difference control are compared by using MATLAB simulation. The effect of fish swarm algorithm in optimizing PID neural network and optimizing linear adjustment algorithm of variable pressure difference setting value is verified. The simulation results show that the optimization effect of fish swarm algorithm is obvious.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB657.2;TP18

【相似文献】