论文部分内容阅读
为了提高温室环境测控系统中传感器数据的准确性,针对温室环境参数变化的时间相关性和空间相似性特点,本文提出了一种基于PCA(Principal component analysis)的故障检测与基于时空信息比较的温室环境监测系统传感器故障识别方法。首先利用基于PCA的传感器故障检测方法,通过监控统计量T~2和SPE的变化实现传感器系统故障检测;再针对检测出故障的传感器节点,对该时刻传感器节点采用基于时空特性的节点信息比较实现不同传感器的故障识别。分别对比基于时间尺度、空间尺度、时空尺度的节点信息比较方法对传感器故障识别的影响进行了分析与试验验证,验证结果表明:基于PCA的传感器故障检测方法能够有效地实现对传感器系统的初步故障检测,提出的基于时空信息比较的传感器故障识别方法,融合考虑时间尺度和空间尺度的节点信息,能够有效地实现具体故障传感器定位;本文所建立的传感器故障识别方法准确检测率CDR为98.37%、平均虚警率FAR为1.72%,较传统的传感器故障识别方法准确检测率CDR提高了22.07%,而平均虚警率FAR则降低了15.76%,能够有效地保证故障诊断效率、提高故障诊断精度、降低虚警率,具有可靠性和准确性。
In order to improve the accuracy of sensor data in greenhouse environment monitoring and control system, this paper presents a PCA (Principal Component Analysis) -based fault detection and greenhouse based on spatio-temporal information comparison for the time-dependent and spatial similarity of greenhouse environmental parameters. Environmental monitoring system sensor fault identification method. Firstly, based on the PCA-based sensor fault detection method, the sensor system fault detection is implemented by monitoring the changes of statistics T ~ 2 and SPE. Then, the sensor node with fault is detected, and the node information based on spatio-temporal characteristics is compared Different sensors fault identification. The comparison of node information based on time scale, space scale and spatio-temporal scale is carried out to analyze and verify the sensor fault identification respectively. The results show that PCA-based sensor fault detection method can effectively realize the initial failure of the sensor system Detection and identification of sensor fault based on space-time information comparison, the fusion of the node information considering the time scale and the space scale can effectively locate the specific fault sensor. The accurate detection rate CDR of the sensor fault recognition method established in this paper is 98.37% The average false alarm rate (FAR) is 1.72%, which is 22.07% higher than that of the traditional sensor fault detection method, while the average false alarm rate (FAR) is reduced by 15.76%. It can effectively ensure the fault diagnosis efficiency, improve the fault diagnosis accuracy, Reduce false alarm rate, with reliability and accuracy.