Abstract:Air handling unit (AHU) is a typical special batch control process, exhibiting strong nonlinear property and two-directional dynamic characteristics which are the time-wise and batch-wise dynamic characteristics. Specifically, the time-wise dynamic characteristic corresponds to the evolution of different operating modes caused by the underlying driving forces which vary slowly in each running day (a batch run), while the batch-wise dynamic characteristic relates to the dynamic variations and deviations among different running days (batch runs). In order to further improve the AHU FDD performance through capturing the underlying driving forces of the AHU system and tackling the batch-wise dynamic property between different batch runs, in this paper, an enhanced kernel slow feature analysis (SFA) based FDD scheme is developed to detect and identify the faults of the nonlinear AHU system. Firstly, a three-way data based kernel SFA (TBKSFA) approach is proposed to detect the faults. In the proposed TBKSFA approach, the kernel trick is adopted in the SFA to sufficiently deal with the nonlinearity and the time-wise dynamic characteristic, and the multiway data analysis is employed to cope with the batch-wise dynamics among different batch runs by converting the three-way training dataset into a variable-wise unfolding two-way matrix. In addition, to handle the tough problem of nonlinearly identifying the fault pattern, a novel kernel discriminant SFA (KDSFA) model is further built by combining the kernel SFA with the discriminant analysis method. In the fault pattern diagnosis process, the proposed KDSFA is pairwisely implemented on the normal and fault datasets to calculate the fault direction, and the fault is then identified by computing the similar degrees of its own fault direction and the historical fault directions. At last, experiments and comparisons on the FDD performance of the developed approach are made using the experimental data provided by ASHRAE Research Project RP-1312. To be specific, the proposed TBKSFA based fault detection method is compared with the popular kernel principal component analysis (KPCA) method, the closely related kernel SFA method and the emerging manifold learning based kernel locality preserving projections (KLPP) method. While the developed KDSFA based fault pattern diagnosis scheme is compared with the conventional jointed angle analysis technique, the strongly linked DSFA based method and the rising artificial neural networks based long short-term memory (LSTM) classifier. Experimental results demonstrate that significant improvements can be achieved by the proposed approach compared with some other popular methods.

Fast and Accurate Classification of Time Series Data Using Extended ELM: Application in Fault Diagnosis of Air Handling Units

Fault diagnosis of rotating machinery components using a deep kernel extreme learning machine under different working conditions

Fault diagnosis in industrial processes using principal component analysis and hidden Markov model

Feature Learning for Fault Detection in High-Dimensional Condition-Monitoring Signals

Fault Diagnosis in Wind Energy Management System using Extreme Learning Machine: A Systematic Review

Fault detection and diagnosis of the air handling unit via an enhanced kernel slow feature analysis approach considering the time-wise and batch-wise dynamics

Enhanced Fault Diagnosis in Time-series Data by Deep Learning

An Intelligent Time-Adaptive Data-Driven Method for Sensor Fault Diagnosis in Induction Motor Drive System

A Data-Driven Fault Detection Framework Using Mahalanobis Distance Based Dynamic Time Warping

Weighted Time Series Fault Diagnosis Based on a Stacked Sparse Autoencoder

ELM-KL-LSTM: a robust and general incremental learning method for efficient classification of time series data

Reliable composite fault diagnosis of hydraulic systems based on linear discriminant analysis and multi-output hybrid kernel extreme learning machine

Fault Diagnosis Method for Hydropower Units Based on Dynamic Mode Decomposition and the Hiking Optimization Algorithm–Extreme Learning Machine

Extended attention signal transformer with adaptive class imbalance loss for Long-tailed intelligent fault diagnosis of rotating machinery

A Novel Time-domain Fault Diagnosis Method with ELM for Aviation Intermediate Frequency Inverter

Gas Path Fault Diagnosis of Turboshaft Engine Based On Novel Transfer Learning Methods

Time-series Processing of Large Scale Remote Sensing Data with Extreme Learning Machine

Fault Diagnosis for Electro-Mechanical Actuators Based on STL-HSTA-GRU and SM

Sensitive time series prediction using extreme learning machine

Medical rolling bearing fault prognostics based on improved extreme learning machine

A Hybrid Data-Driven Simultaneous Fault Diagnosis Model for Air Handling Units