Novel Process Monitoring Approach Enhanced by a Complex Independent Component Analysis Algorithm with Applications for Wastewater Treatment

Abstract

The process monitoring of industries by means of multivariate statistical methods has gained popularity in academic and industrial communities. However, nonlinearity, autocorrelation, and high dimensionality can render traditional approaches inadequate. This necessitates a more powerful method for implementing better process monitoring in reality. Therefore, a novel independent component analysis (ICA) algorithm, termed complex dynamic independent component analysis (CD-ICA), is proposed for information refinement and feature extraction in this paper. This proposed algorithm extracts dominant features from a complex-valued matrix containing raw data information and their changing rates through complex ICA and properly shifting phase operations. The novel fault detection index, based on the three traditional monitoring statistics of the ICA algorithm, is enhanced and applied to monitor real chemical and biological processes efficiently by combining the aforementioned algorithm. The performance of the presented method is assessed based on data from a simple multivariate mathematical simulation and data from a real wastewater treatment plant (WWTP). The results show that this approach provides higher efficiency and performance than traditional approaches.

A Supervised Bidirectional Long Short-Term Memory Network for Data-Driven Dynamic Soft Sensor Modeling

“…In this case study, we use the urban WWTP dataset in the UCI Machine Learning Repository, which is a set of field data collected from daily sensor measurements in an urban WWTP [31]. The WWTP dataset has been widely used by many researchers to evaluate data-driven models on a practical effluent prediction problem [32]- [35]. Instead of installing sensors at different stages of a process, the WWTP dataset collects data from repeated measurement of a fixed set of process variables throughout the process.…”

Modeling an Adaptive Hybrid Soft Sensor with Co-training Learning toward Applications in Wastewater Treatment

“…Soft sensor technology has been proven to be an effective way of solving the abovementioned problems and then supporting the optimization and diagnosis of WWTPs. 2 Various data-driven soft sensors are widely used for process monitoring, including statistical learning 3 (partial least square, PLS, 4 principal component analysis, PCA 5 ), kernel learning (support vector machine, SVM, 6 relevance vector machine, RVM 7 ), single hidden layer neural network (radial basis function neural network, RBFNN or RBF, 8 back propagation neural network, BPNN or BP 9 ), mixture modeling (Gaussian mixture model, GMM 10 ), etc.…”

Data-Driven Drift Detection in Real Process Tanks: Bridging the Gap between Academia and Practice

“…Mali and Laskar [19] proposed an optimized Monte Carlo deep neural network and were able to detect faults of low magnitude in simulated data. Xu et al [20] proposed a version of ICA called complex-valued ICA. The method was both evaluated for simulated data and for data from a real plant.…”