AHU sensor fault diagnosis using principal component analysis method

The paper presents a strategy based on the principal component analysis (PCA) method, which is developed to detect and diagnose the sensor faults in typical air-handling units. Sensor faults are detected using the Q-statistic or squared prediction error (SPE). They are isolated using the SPE and Q-contribution plot supplemented by a few simple expert rules. Two PCA models are built based on the heat balance and pressure–flow balance of the air-handling process, aiming at reducing the effects of the system non-linearity and enhancing the robustness of the strategy in different control modes. The fault isolation ability of the method is improved using the multiple models. Simulation tests and site data from the building management system (BMS) of a building are used to verify the PCA-based strategy for automatic validation of AHU monitoring instrumentations and detecting/isolating AHU sensor faults under typical operating conditions. The robustness of the PCA-based strategy in detecting/diagnosing AHU sensor faults is verified. Effects of sensor faults and the strategy energy efficiency of an automated AHU are evaluated using simulation tests.     

基于主成分分析法的空调系统传感器自动故障诊断

由于各种各样的故障,空调系统在整个建筑生命周期内其性能很少能够达到设计目标。为了改善空调系统的性能,在整个建筑生命周期内对其进行定期或连续检测是很必要的。随着空调系统的日益复杂化和对快速可靠检测的需求,人工检测已远远不能满足要求,这使自动检测成为了必须。目前国际上研究建筑空调系统自动检测的技术很热门,但较少考虑空调系统传感器的准确性。传感器读数准确性是空调系统可靠控制和检测的先决条件。本文提出了一种基于主成分分析法的鲁棒故障诊断策略,该策略可以有效地对空气处理单元中的传感器进行自动检测。结合建筑自动化系统,该诊断策略可以实现空调系统传感器故障的在线检测。     

A sensor fault detection strategy for air handling units using cluster analysis

Sensors are an essential component in the control systems of air handling units (AHUs). A biased sensor reading could result in inappropriate control and thereby increased energy consumption or unsatisfied indoor thermal comfort. This paper presents an unsupervised learning based strategy using cluster analysis for AHU sensor fault detection. The historical data recorded from sensors is first pre-processed to reduce the dimensions using principal component analysis (PCA). The clustering algorithm Ordering Points to Identify the Clustering Structure (OPTICS) is then employed to identify the spatial separated data groups (i.e. clusters), which possibly indicate the occurrence of sensor faults. The data points in different clusters are then checked for temporal separation in order to confirm the occurrence of sensor faults. The proposed sensor fault detection strategy is tested and evaluated with the data collected from a simulation system. The results showed that this strategy can detect single and non-simultaneously occurred multiple sensor faults in AHUs. The fault detection results were not strongly affected by the selection of the user defined input parameters required in OPTICS.     

A rule-based fault detection method for air handling units

Air handling unit performance assessment rules (APAR) is a fault detection tool that uses a set of expert rules derived from mass and energy balances to detect faults in air handling units (AHUs). Control signals are used to determine the mode of operation of the AHU. A subset of the expert rules which correspond to that mode of operation are then evaluated to determine whether a fault exists. APAR is computationally simple enough that it can be embedded in commercial building automation and control systems and relies only upon the sensor data and control signals that are commonly available in these systems. APAR was tested using data sets collected from a “hardware-in-the-loop” emulator and from several field sites. APAR was also embedded in commercial AHU controllers and tested in the emulator.     

一次回风空调箱冷却除湿过程计算

分析了舒适性空调系统中由空调箱选型不当引起的空气处理偏离设计状态点的现象。验证了风机盘管表冷器修正公式应用于空调箱的可行性。提出了空调箱特性指标(S0B00.162)的概念,建立了一次回风系统空调箱空气热湿处理过程数学模型,并使用MATLAB软件编程求解模型,在一定设计条件下计算出空调箱的出风参数、供回水温差,并以供回水温差作为空调箱选型的依据。运用模型对两个气候差异较大地区的候车厅空调箱进行选型计算。建立的数学模型及求解方法同时可作为空调箱运行调节的依据。     

A robot fault diagnostic tool for flow rate sensors in air dampers and VAV terminals

A robot fault diagnostic tool for flow rate sensors in air dampers and VAV terminals is presented to ensure well capacity of energy conservation in building air conditioning systems. Principal component analysis (PCA) is used to detect the sensor faults including fixed bias, drifting bias and complete failure. To improve the detection efficiency, several PCA models are built through employing the conservation equations and control relations of the system. With the historical data, PCA models are trained to capture most useful information of normal operation. As a result, the training models can identify whether the present condition is abnormal through comparing the residues with the thresholds. Since the principal component subspace and residue subspace of the operation data space are obtained using PCA decomposition, these two subspaces are used to develop the fault isolation scheme. The new fault detected and the known ones in the library are all projected into the principal component subspace and residue subspace decomposed by PCA. The joint angle plot, illustrating the direction relations of the projections in both subspaces between the new fault and the known ones, is used to diagnose the fault source.     

住宅夏季空调行为对空调能耗的影响研究

本文针对北京地区的城镇住宅,采用问卷调查的方式,从居住者使用的角度分析了住宅空调行为在空调设定温度、空调运行时间、空调同时运行台数及空调运行时门窗开关状况等方面的分布,研究发现了住宅空调行为的变化特点以及不同住户空调行为模式的差异。在此基础上,采用模拟分析的方法,计算确定了空调行为模式对住宅空调能耗的影响程度。最后,对空调实际耗电量与模拟计算结果进行了比较,验证了空调行为调查统计结果的可靠性。     

A hybrid FDD strategy for local system of AHU based on artificial neural network and wavelet analysis

This paper presents a self-adaptive sensor fault detection and diagnosis (FDD) strategy for local system of air handing unit (AHU). This hybrid strategy consists of two stages. In the first stage, a fault detection model for the AHU control loop including two back-propagation neural network (BPNN) models is developed. BPNN models are trained by the normal operating data of system. Based on sensitive analysis for the first BPNN model, the second BPNN model is constructed in the same control loop. In the second stage, a fault diagnosis model is developed which combines wavelet analysis method with Elman neural network. The wavelet analysis is employed to process the measurement data by extracting the approximation coefficients of sensor measurement data. The Elman neural network is used to identify sensor faults. A new approach for increasing adaptability of sensor fault diagnosis is presented. This approach gains clustering information of the approximations coefficients by fuzzy c-means (FCM) algorithm. Based on cluster information of the approximation coefficients, the unknown sensor fault can be identified in the control loop. Simulation results in this paper show that this strategy can successfully detect and diagnose fixed biases and drifting fault of sensors for the local system of AHU.     

Predicting construction crew productivity by using Self Organizing Maps

A Self Organizing Map (SOM), is a machine learning method that represents high-dimensional data in low-dimensional form without losing topological relations of the data. After an unsupervised learning process, it organizes the data on the basis on similarity. In the current study, a SOM based algorithm has been developed which not only produces 2-D maps to analyze the relationship between various factors and crew productivity, but also predicts productivity under given conditions. Validation of the model has been achieved both by using artificial data set and data from 144 concrete pouring, 101 formwork and 101 reinforcement crews. The results show that maps which are produced by the model are satisfactory in clustering the data and prediction performance of the model is superior to similar artificial neural network models.     

软土边壁(坡)流鼓破坏模式研究

依据相似模型原理及所建立的相似法则,对软土边壁（坡）破坏模式进行了实验研究。试验研究证实,刚体转动假设不适用于软土边壁（坡）稳定性分析;软土地壁（坡）不取圆弧破坏模式,而是取流鼓破坏模式。模型实验结果与现场试验及理论分析计算结果规律基本和,数据较为吻合。     

关于构建中国环境模式语言的思考

在全球化发展的今天,如何研究传统、继承传统,成为摆在每一位中国设计师面前的一项极为迫切与重要的任务。本文试图借鉴与应用C·亚历山大的“模式语言”理论,通过提炼与构建中国环境模式语言来解读传统原型环境,实现对传统的继承与超越;同时探索一种研究中国传统建筑与城市景观的方法,为当代有中国特色的建筑与城市景观的创作提供启示。     

Model-based operational constraints for fluidised bed crystallisation

Fluidised bed reactors are used for water softening in water treatment plants. Recent research shows that under current operation of reactors the crystallisation of calcium carbonate can be hampered. Until now the operational constraints on the fluidised bed have not been exactly known. Experiments were carried out to investigate the fluidisation behaviour of calcium carbonate pellets in water. The results of the fluidisation experiments are compared to two commonly used modelling approaches of Ergun and Richardon-Zaki. Using the experimental data the models are calibrated. The calibrated Richardson-Zaki model is used to determine operational constraints on pellet size at the bottom of the reactor and water flow through the reactor. The model-based constraints are compared to operational data of the Weesperkarspel full-scale treatment plant of Waternet (The Netherlands). It can be concluded that the current operation of the treatment plant violates the calculated constraints with consequences for effluent quality and corrective maintenance. By using models for determining the operation of the fluidised bed, the softening process can thus be improved.     

建筑能耗模拟典型年中气象参数权重的确定

在建筑能耗模拟用典型年的生成中,传统的Finkelstein-Schafer统计方法对气象参数赋予了固定的权重因子,但有关研究表明,由于地域间气象资源不同,气象参数权重因子的固化有待商榷。针对挑选典型年时气象参数权重因子统一与否对典型年挑选结果和建筑能耗模拟准确性的影响问题,选同一建筑热工分区中寒冷地区的代表城市北京和拉萨,分别使用FS方法和主成分法进行了典型年的挑选,并对典型公共建筑建模进行了能耗模拟分析。结果表明:FS统计方法适于表征单独气象参数的长期相似性,但存在对太阳辐射参数权重赋予过大的问题,适用于太阳能丰富地区;而主成分法适合寻求当地气象资源的本质特征,使用主成分法时对气象参数的选择尤为重要。     

Nonlinear multivariable control and performance analysis of an air-handling unit

To maintain satisfactory comfort conditions in buildings with low energy consumption and operation cost, control of air-conditioner units is required. In this paper, nonlinear control of an air-handling unit (AHU) is investigated and compared for two control approaches: gain scheduling and feedback linearization. A nonlinear multi input-multi output model (MIMO) of an air-handling unit (AHU) is considered. Both indoor temperature and relative humidity are controlled via manipulation of valve positions of air and cold water flow rates. Using an observer to estimate state variables, a hybrid control system including regulation system for disturbance rejection and nonlinear control system for tracking objectives is designed. Achievement of tracking objectives is investigated for various desired commands of indoor temperature and relative humidity; including a sequence of steps and ramps-steps. According to results, more quick time responses with a bit more overshoot in tracking set-points/paths are achieved by using feedback linearization method (especially for temperature). However, valves position as input control signals are associated with less oscillation (and consequently less energy consumption) when the controller designed based on gain scheduling approach is used. Finally, it is shown through phase portrait of the system that the controller designed based on feedback linearization shows a robust performance in the presence of random uncertainty in model parameters.     

Multivariate and multiscale monitoring of wastewater treatment operation

In this work extensions to principal component analysis (PCA) for wastewater treatment (WWT) process monitoring are discussed. Conventional PCA has some limitations when used for WWT monitoring. Firstly, PCA assumes that data are stationary, which is normally not the case in WWT monitoring. Secondly, PCA is most suitable for monitoring data that display events in one time-scale. However, in WWT operation. disturbances and events occur in different time-scales. These two limitations make conventional PCA unsuitable for WWT monitoring. The first limitation can be overcome by use of adaptive PCA. In adaptive PCA. the PCA model is continuously updated using an exponential memory function. Variable mean, variance and co-variance are thus adapted to the changing conditions. The second problem can be solved by time-scale decomposition of data prior to analysis. The time-scale decomposition methodology involves wavelets and multiresolution analysis (MRA) in combination with PCA. MRA provides a tool for investigation and monitoring of process measurement at different time-scales by decomposing measurement data into separate frequency bands. Time-scale decomposition increases the sensitivity of the monitoring, which makes it possible to detect small but significant events in data displaying large variations. Moreover, time-scale information is sometimes important in the interpretation of a disturbance to determine its physical cause. Also, by decomposing data, the problem of changing process conditions is partly solved. All the presented methods are illustrated with examples using real WWT process data.     

EPS在1∶500地形数据整理中的应用

随着信息化测绘的不断发展,各行各业对测绘技术的要求越来越高,旧的作业方式已经不能满足当前的工作需要。对于数据的生产,要求生产入库一体化。本文结合1∶500数据整理生产,介绍了EPS在1∶500地形数据整理中的作业流程,总结了1∶500地形数据的作业方法,为数据的直接入库做好准备工作。     

离心式制冷机系统传感器故障诊断的试验研究

传感器的可靠性及准确性对制冷机系统的可靠控制和系统的最优运行起着至关重要的作用。同时,传感器的读数也是进行部件故障诊断的基础。本文提出了基于主元分析法的制冷机传感器故障诊断方法,该方法的主元分析模型由离心式制冷机系统中的相关测量变量在正常条件下的观测样本构成。这一方法利用这些变量在正常条件下的相关性来对传感器的测量观测值进行故障检测与诊断及测量重构,并分别用Q-统计及Q-分布图来对传感器故障进行检测及诊断。本文利用实验室离心式制冷机的试验数据对这一基于主元分析法的传感器故障诊断方法进行了验证。     

Comparative study of protozoan communities in full-scale MWTPs in Beijing related to treatment processes

The potential influence of process principles and system conditions on shaping protozoan community structures in eight full-scale municipal wastewater treatment systems in Beijing, including four process types, i.e., anaerobic/anoxic/aerobic or anoxic/anaerobic/aerobic process (A2O), anoxic/aerobic or anaerobic/aerobic process (AO), oxidation ditch (OD) and sequencing batch reactor (SBR), was evaluated with the aid of cluster analysis and principal components analysis (PCA). The species richness and abundance distribution of protozoa varied significantly with the process types: the A2O ecosystems harbored more diverse protozoan communities with higher relative abundance of crawling and sessile ciliates than the other systems. Cluster analysis revealed that the protozoan community structures were in high coordination with the process types, i.e., different systems with the same process principles exhibited similar community structures. The A2O processes displayed a distinctively higher similarity of protozoan community structures than the AO processes, suggesting that the A2O ecosystems were more stable than those of AO. The PCA analyses demonstrated that swimming and carnivorous ciliates were correlated with poor settleability of sludge, and that amoebae were sensitive to DO level. We therefore concluded that protozoan community structures were primarily shaped by treatment process principles, whilst they were also modified by system conditions in terms of operational properties and water quality.     

Optimizing the Alignment of Inspection Data from Track Geometry Cars

Track geometry car inspections play a vital role in assessing railroad track quality and scheduling track maintenance. Effective use of inspection data depends on accurate location measurement. Field surveys reveal that measured milepoint positions can be off by up to 200 m. Previous efforts to correct milepoint errors resulted in the development of the Key Equipment Identification model, which reduces errors to below 5 m and in most cases to below 1 m for new inspections. However, analysis of track segment deterioration requires the alignment of historical inspection data. This article presents an improved method for aligning these historical inspections. The core of the approach is an optimization model termed Dynamic Sampling Position Matching (DSPM). DSPM overcomes limitations of existing methods and their assumptions that milepoint shifts between inspections are constant and that no track maintenance is carried out between inspections. A case study is presented using inspection data from the Jinan Bureau of China Railways demonstrating improved performance over two widely used inspection alignment models. Results indicate that DSPM better tolerates noisy measurements and that inspections processed by DSPM align precisely. And it takes the developed model 2.82 seconds on average to align inspection data for a track segment of 1 km.