基于运行数据的离心式机组故障诊断研究

发布时间：2018-04-28 18:57

本文选题：离心式机组 + 故障检测与诊断　；参考：《湖南大学》2011年硕士论文

【摘要】：随着科学技术的快速发展和工业自动化水平的不断提高,空调系统在国民经济各部门和人民生活中应用日益广泛,空调系统的能耗越来越大。节能和环保成为当今世界关注的两个重大主题,提高能源利用效率和减少温室气体的排放变得越来越重要。暖通空调系统在实际的运行中,常常由于各种各样的故障导致系统无法满足设计的预期要求,导致大量的能源浪费,甚至经过正确调试的空调系统在运行一段时间之后也时常发生这样的情况。故障检测与诊断可以降低系统能耗并帮助系统恢复到正常的运行状态。因此,故障诊断对空调系统十分重要。本文提出了基于模型方法的离心式机组故障检测与诊断策略,该策略主要包含数据预处理器,回归模型的建立,实时阈值的产生和故障诊断分类器几个部分。机组故障检测与诊断策略是选取六个有典型物理意义的特性参数来指示冷水机组的运行情况,利用机组正常运行数据建立特性参数预测模型作为基准值,比较每个特性参数的计算值和基准值得到残差。故障阈值结合数据分析,采用统计分析方法计算得到,随机组运行工况变化而变化。当一个或多个特性参数的残差超出了相应阈值并有持续偏离的趋势时,就检测出机组存在故障,再由故障诊断分类器诊断出具体的故障。在机组系统的基本制冷循环和各个部件的热力学原理基础上,结合大量的实验数据分析了机组各个故障与每一个特性参数的相互影响,得到机组故障诊断分类器规则表。故障检测与诊断策略运用ASHRAE研究项目的实验数据进行了验证,并将该策略应用于香港某建筑真实空调系统,利用采集的机组历史运行数据分析了机组冷凝器结垢故障。最后,运用冷水机组故障检测与诊断策略原理在Visual Basic6.0环境下开发了冷水机组性能诊断软件
[Abstract]:With the rapid development of science and technology and the continuous improvement of the level of industrial automation, the air conditioning system is widely used in the national economy and people's daily life, and the energy consumption of the air conditioning system is increasing. Energy conservation and environmental protection have become two major themes in the world today. It is becoming more and more important to improve energy efficiency and reduce greenhouse gas emissions. In the actual operation of HVAC system, due to various failures, the system can not meet the expected requirements of the design, resulting in a large amount of energy waste. Even properly adjusted air-conditioning systems often occur after running for some time. Fault detection and diagnosis can reduce system energy consumption and help the system to return to normal operation state. Therefore, fault diagnosis is very important to air conditioning system. In this paper, a fault detection and diagnosis strategy for centrifugal units based on model method is proposed. The strategy includes data preprocessor, establishment of regression model, generation of real time threshold and fault diagnosis classifier. The strategy of unit fault detection and diagnosis is to select six characteristic parameters with typical physical significance to indicate the operation of the chiller, and to establish the prediction model of characteristic parameters by using the normal operation data of the chiller as the reference value. The residual error is obtained by comparing the calculated value of each characteristic parameter with the reference value. Combined with the data analysis, the fault threshold is calculated by statistical analysis method, and it varies with the operating condition of the unit. When the residuals of one or more characteristic parameters exceed the corresponding threshold and have a tendency to deviate continuously, the fault of the unit is detected and the specific fault is diagnosed by the fault diagnosis classifier. On the basis of the basic refrigeration cycle of the unit system and the thermodynamic principle of each component, combined with a large number of experimental data, the interaction between each unit fault and each characteristic parameter is analyzed, and the rule table of unit fault diagnosis classifier is obtained. The strategy of fault detection and diagnosis is verified by the experimental data of the ASHRAE research project. The strategy is applied to the real air conditioning system of a building in Hong Kong. The scaling fault of condenser is analyzed by using the historical operation data of the unit. Finally, using the principle of fault detection and diagnosis strategy of chillers, the performance diagnosis software of chillers is developed under the environment of Visual Basic6.0.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【参考文献】