变风量空调系统故障诊断及容错控制方法研究

In VAV air-conditioning systems, the problem related to faults including the increase of energy consumption, the deterioration of heat comfort and the increase of maintain costs is more and more serious. Aiming at faults of VAV air-conditioning system, methods of fault diagnosis and fault-tolerant control will be intensively studied in this research project for VAV air-conditioning systems. In development of fault detection methods,the effects of serial correlation in monitoring data, noises,dynamic characteristics and normal transient processes will be considered. Focus on improving the fault detection accuracy, robust fault detection method will be studied for real VAV air-conditioning systems.Aiming at the key problem and difficulty of fault diagnosis: a fault symptom may exist many potential fault causes, rule-based proactive fault diagnosis method will be studied for real VAV air-conditioning systems by comprehensively adopting the cause and effect diagram method,system analysis method, the proactive diangosis technique,and expert rule method. In development of fault-tolerant control methods,the nonlinear characteristics of the VAV air-conditioning system will be considered. Support vector machine regression model will be used to online recover fauty measurement data of sensor.Moreover, online self-correction methods will be developed for soft faults of VAV air-conditioning systems. This research will provide a theoretical basis and technical guidance for development the fault diagnosis and fault-tolerant control systems of VAV air-conditioning systems.

变风量空调系统故障引起的能源浪费、热舒适性降低和室内空气品质下降等问题日益严重。针对变风量空调系统故障，本项目研究变风量空调系统故障诊断及容错控制方法。在研究故障检测方法时，考虑数据自相关性、噪声、系统动态性和瞬变过程等因素的不利影响，研究适用于实际系统的变风量空调系统鲁棒故障检测方法，着重于提高故障检测结果的准确性和可靠性。针对故障诊断的关键问题及难点：一个故障征兆可能存在多种潜在的故障原因，综合运用因果分析图、系统分析、主动式故障诊断技术和专家规则等方法，研究适用于实际系统的基于规则的变风量空调系统主动式故障诊断方法。在研究在线容错控制方法时，考虑变风量空调系统的非线性特点，研究传感器故障测量数据在线恢复方法，并开发适用于实际系统的变风量空调系统软故障自纠正方法。本项目将为开发变风量空调系统故障诊断及容错控制系统提供理论依据和技术指导。

针对变风量空调系统故障，本项目深入研究了变风量空调系统故障诊断方法及容错控制方法，取得的主要研究成果包括：1）给出基于残差EWMA控制图的变风量空调系统鲁棒性故障检测方法，设计了残差EWMA控制图上限值和下限值，可以消除时序分析模型建模误差对残差EWMA控制图性能的不利影响，基于残差EWMA控制图的变风量空调系统故障检测方法可以消除系统动态性、数据自相关性、噪声、瞬变过程和建模误差等因素的不利影响，有效提高了变风量空调系统故障检测结果的准确性和可靠性。2) 将基于规则的故障诊断方法拓展为变风量空气处理机组多故障诊断方法，设计了用于诊断22种变风量空气处理机组故障的26条专家规则，将26条专家规则分为故障征兆相关的规则和能耗相关的规则，可以分别分离出故障征兆相关的故障集合和能耗相关的故障集合，融合这两个故障集合里的故障就可以给出变风量空气处理机组多故障诊断结果。3) 给出了基于规则的变风量空气处理机组主动式故障检测与诊断方法，初步研究了变风量空气处理机组主动式故障检测与诊断的机理，综合运用因果分析图、系统分析、主动式故障诊断技术和专家规则等技术，给出了一种基于专家规则的变风量空气处理机组主动式故障诊断方法，使变风量空气处理机组系统进入预先设定的主动式故障诊断模式，逐项排除潜在的故障原因，分离和诊断变风量空气处理机组故障。4) 给出了变风量空调末端装置在线故障检测与诊断方法，开发了变风量空调末端装置参数自整定模型，给出了利用遗传算法优化模型参数的方法，计算模型预测值与系统实测值之间的残差，通过比较残差与故障检测阈值在线检测变风量空调末端装置故障，利用设计的18条专家规则分离与诊断12种变风量空调末端装置故障。