Kernel methods are a class of learning machines for the fast recognition of nonlinear patterns in any data set. In this paper, the applications of kernel methods for feature extraction in industrial ...process monitoring are systematically reviewed. First, we describe the reasons for using kernel methods and contextualize them among other machine learning tools. Second, by reviewing a total of 230 papers, this work has identified 12 major issues surrounding the use of kernel methods for nonlinear feature extraction. Each issue was discussed as to why they are important and how they were addressed through the years by many researchers. We also present a breakdown of the commonly used kernel functions, parameter selection routes, and case studies. Lastly, this review provides an outlook into the future of kernel-based process monitoring, which can hopefully instigate more advanced yet practical solutions in the process industries.
Rapid localization of hazardous gas leak sources is a vital task regarding the security of human society. Over the past few decades, autonomous search using mobile sensing devices equipped with gas ...sensors has become a promising field. Cognitive search strategies, such as Infotaxis and Entrotaxis, have been successfully used to search an unknown source autonomously in turbulence. However, they are time-consuming and may not be efficient because all source parameters (locations and strength) are estimated via particle filtering, especially when the prior information of source strength is rarely known. To address this inefficiency, this paper proposes a hybrid autonomous source searching approach, named as regression-enhanced Entrotaxis, which incorporates the Entrotaxis algorithm with the regression methods. Only the position of leak source is inferred by particle filtering, while the strength is estimated by the least squares based on the historical measurements. Through this way, the searching space in particle filtering is converted from 3-dimensions (source position in x–y plane and source strength) to 2-dimensions (source position in x–y plane), speeding up the computation significantly. Then, the reward function according to maximum entropy sampling principles is utilized to guide the robot’s next move. The performance of the proposed method is compared to that of the Entrotaxis algorithm. The results of Monte Carlo simulations demonstrate that fewer particles will be used to reach a specific success rate, saving time in the search process. Besides, the proposed strategy relies on less prior information regarding source strength, which is more in line with the actual scenario. In addition, the proposed approach is adaptable to a broader search region.
•Regression-enhanced Entrotaxis is proposed as an autonomous source searching method.•The posterior distribution of source location is updated using particle filter (PF).•The source emission rate is estimated using the regression method.•The number of estimated terms in PF is reduced from 3 to 2, speeding up computation.•Less prior information about source emission rate is required.
The steady-state thermal environment creates obvious challenges to health and energy, which arouse people's concern on the dynamic thermal environment. As an important way to improve thermal comfort ...and reduce energy consumption, airflow has been widely concerned. Being a highly acceptable airflow, simulating natural wind can be adopted to set up a dynamic thermal environment. Influencing factors of the dynamic thermal environment are much more spread out than the steady-state thermal environment and the human thermal comfort problem gets start from the relatively simple steady-state thermal comfort, those two determiners work and led to the consequence that research on the human comfort model of the dynamic thermal environment is relatively limited. Based on the fuzzy comprehensive evaluation method, this paper establishes the human thermal comfort model of simulating natural wind environment. Combined with the human thermal comfort experiment, the evaluation indexes of Mean Thermal Sensation Vote (MTSV) and Dissatisfaction Rate (DR) are obtained. By comparing the PMV-PPD index and MTSV-DR index with the experimental results of human thermal comfort, it was found that there was a large deviation between the PMV-PPD index and the experimental results, but the MTSV-DR index was in good agreement. According to the range of dissatisfaction rate, the appropriate combination of environmental parameters is recommended. It provides a new way of thinking for human thermal comfort evaluation in dynamic thermal environment.
•The simulating natural wind is introduced into the steady-state air-conditioning environment to form a dynamic simulating natural wind environment.•The human thermal comfort model is established by fuzzy comprehensive evaluation, and the MTSV-DR indexes are put forward to evaluate the comfort of the environment.•The MTSV-DR is compared with PMV-PPD and the experimental result to verify the reliability of the MTSV-DR.•Adding simulating natural wind in steady-state air-condition environment can increase the indoor comfortable temperature and achieve the purpose of energy saving.•When the dissatisfaction rate of the population is required to be less than 10% (DR is less than 10%), the MTSV is between -0.05 and +0.05.6)It is recommended that the indoor temperature be 28℃, the turbulence intensity be 0.68, and the blowing time be 5 minutes.
Crystalline polymer powder inevitably incorporates certain impurities, including decomposed polymers or foreign particles. An essential criterion for assessing the quality of polymers is the quantity ...of contaminants in the powder. However, it is challenging to discern powder contaminants through machine vision due to the poor quality of images taken at production sites. Inspired by the spectral properties of crystalline polymers, this paper proposes an efficient image-based impurity detection method, which seeks to precisely and robustly detect contaminant content. Based on the changes in absorbance during polymer decomposition, a highly selective channel-weighted image enhancement approach is designed to emphasize the difference between impurities and normal particles. Then, using the prior information on the powder’s attributes, an adaptive thresholding method is employed to categorize pixels belonging to impurities. Finally, a dataset of 119 12-megapixel photos from a chemical facility, where the average size of contaminants in images is 43 pixels, is used to evaluate the performance of the proposed algorithm. The results of the detection demonstrate that the proposed strategy for image enhancement has better selectivity to impurities than typical image enhancement methods.
•Detecting powder impurities as small targets in low contrast images is a challenge.•The channel-weighted image enhancement approach is highly selective to impurities.•The proposed method is evaluated on a real dataset and achieves acceptable results.
Self-optimizing control is a strategy for selecting controlled variables, where the economic objective guides the selection and design of controlled variables, with the expectation that maintaining ...the controlled variables at constant values can achieve optimization effects, translating the process optimization problem into a process control problem. Currently, self-optimizing control is widely applied to steady-state optimization problems. However, the development of process systems exhibits a trend towards refinement, highlighting the importance of optimizing dynamic processes such as batch processes and grade transitions. This paper formally introduces the self-optimizing control problem for dynamic optimization, termed the dynamic self-optimizing control problem, extending the original definition of self-optimizing control. A novel concept, ”dynamic controlled variables” (DCVs), is proposed, and an implicit control policy is presented based on this concept. The paper theoretically analyzes the advantages and generality of DCVs compared to explicit control strategies and elucidates the relationship between DCVs and traditional controllers. Moreover, this paper puts forth a data-driven approach to designing self-optimizing DCVs, which considers DCV design as a mapping identification problem and employs deep neural networks to parameterize the variables. Three case studies validate the efficacy and superiority of DCVs in approximating multi-valued and discontinuous functions, as well as their application to dynamic optimization problems with non-fixed horizons, which traditional self-optimizing control methods are unable to address.
•The concept of dynamic controlled variables (DCVs) is proposed.•An implicit control policy based on DCVs is proposed.•It is theoretically proven that the implicit policy can represent the set-valued mapping.•Dynamic self-optimizing control is systematically and formally proposed for the first time.•A data-driven method for designing self-optimizing dynamic controlled variables is proposed.
This work considers to achieve near-optimal operation for a class of batch processes by employing self-optimizing control (SOC). Comparing with a continuous one, a batch process exhibits stronger ...nonlinearity with dynamics because of the non-steady operation condition. This necessitates a global version of SOC to achieve satisfactory performance. Meanwhile, it also makes the existing global SOC (gSOC) not directly applicable to batch processes due to the causality amongst variables. Therefore, it is necessary to extend the original gSOC to batch processes. In addition to the nonconvexity challenge of the original gSOC problem, the new extension for batch processes has to face even more challenges. Particularly, the causality due to dynamics of batch processes brings in structural constraints on controlled variables (CVs), making a CV selection problem even more difficult. To address these challenges, the gSOC problem is recast in a vectorized formulation and it is proved that the structural constraints considered are linear in the vectorized formulation. Moreover, a novel shortcut method is proposed to efficiently find sub-optimal but more transparent solutions for this problem. The effectiveness of the new approach is validated through a case study of a fed-batch reactor, where CVs are constructed through a combination matrix with a repetitive structure, resulting in a simple SOC scheme. This simplicity facilitates the implementation of the SOC approach and enhances its practical applicability and robustness.
•Extends global SOC (gSOC) methodology to batch processes for the first time.•Proves certain structural constraints on CV combination matrix are linear.•Develops new convex approximation algorithm to solve gSOC with constraints.
Combustion state monitoring in industrial municipal solid waste incineration (MSWI) power generation process is very difficult. There are two main reasons for this dilemma: (1) there is a large time ...lag between the combustion chamber and the steam generation stage; (2) reactions continue to occur and raw waste is fed into the incinerator continuously, so there are strong correlations and remarkable dynamic characteristics among process variables. The complex correlations in time and in space cause difficulties in combustion state monitoring for incinerators. Those correlations are common in industrial processes because the large scale equipment is usually needed and a whole process is realized by equipment in sequence. To overcome those difficulties, a multi-block dynamic enhanced canonical correlation analysis (MBDCCA) method is proposed to help monitor the MSWI process. Firstly, canonical correlation analysis (CCA) is extended to MBDCCA to model the MSWI process. Secondly, a combustion state online monitoring approach is designed based on the proposed MBDCCA algorithm. Finally, the proposed method is applied in an industrial MSWI plant to demonstrate the effectiveness.
•CCA enhanced by dynamics is used for fault detection in an MSWI process.•This work offers a promising method for systems with large time lags.•The proposed method is verified by using industrial online data.•This work reduces the false alarm rate by dividing the system into multi-blocks.
Cyber-physical system (CPS) has been widely adopted in modern industrial productions. Safety and security (S&S) play an important role in CPS, which assists the reliability of the system. ...Traditionally, safety and security risks were managed independently. As the development of cyber technology, S&S issues become complex and could affect each other in multiple ways. There is a strong need to develop a systematic method to manage safety and security risks simultaneously. In this work, a systematic method to integrately analyze S&S risks is proposed. Firstly, attack route models (ARM) as the root cause of typical cyber-threats are summarized from the literature together with their corresponding consequences in CPS. Secondly, in addition to commonly adopted physical safety prevention route (PSPR), cyber security prevention route (CSPR) based on ARM is developed to investigate the safety hazards and security threats. Then, safety critical variable analysis (SCVA) is proposed to quantify the S&S risk. Finally, SCVA, CSPR and PSPR are integrated via the bowtie method. The key contribution of the work is the method which simultaneously consider safety and security risk for CPS. In parallel. SCVA represents the working status of CPS devices, which would be useful to quantitatively determine the severity of consequence and further level of risk.
•A new occupational health and safety training method is proposed, especially for long-term health risks.•The training method is developed based on the integration of diminished quality of life (DQL) ...and virtual reality (VR).•The DQL-VR training improves the training quality in both hazard identification and risk perception.
Safety training using Virtual Reality (VR) can be widely found in many industries. They are majorly focused on educating sequential operations, standard working procedures, and short-term safety risk assessment. Occupational health and safety (OHS) issues at the workplace contain both short-term safety incidents and long-term health issues. Comparing to short-term safety training, long-term health training is little discussed in the existing literature. One of the reasons is that the health issues are more complex than safety, as the biological outcomes may take time to develop, affected by multiple resources, and the injuries may be cumulative. The key contribution of the current paper is the development of a solid training method based on the integration of diminished quality of life (DQL) risk assessment and VR, namely DQL-VR. The use of DQL-VR would assist in delivering OHS knowledge to trainees with a special focus on chronic health components. The use of DQL increases the ability of trainees to determine long-term health hazards and perceive long-term health risks. The use of VR improves the learning efficacy and users’ acceptance of the safety training. A pilot test is investigated to examine the efficacy of DQL-VR comparing with traditional safety training. The result shows that DQL-VR has great positive efforts in teaching hazard identification and risk perception, especially for long-term health aspects. The findings also indicate that the use of DQL-VR would benefit trainees’ recall of knowledge. The development of DQL-VR also offers opportunities for trainees to foresee the hazards and experience the harmful consequences in a safe environment.
Air pollution episodes (APEs) caused by excessive emissions from chemical industry parks (CIPs) have resulted in severe environmental damage in recent years. Therefore, it is of great importance to ...detect APEs timely and effectively using contaminant measurements from the air quality monitoring network (AQMN) in the CIP. Traditionally, APE can be detected by determining whether the contaminant concentration at any ambient monitoring station exceeds the national environmental standard. However, the environmental standards used are unified in various ambient monitoring stations, which ignores the source–receptor relationship in the CIP and challenges the effective detection of excessive emissions in some scenarios. In this paper, an approach based on a multivariate statistical analysis (MSA) method is proposed to detect the APEs caused by excessive emissions from CIPs. Using principal component analysis (PCA), the spatial relationships hidden among the historical environmental monitoring data are extracted, and the high-dimensional data are projected into only two subspaces. Then, two monitoring indices, T2 and Q, which represent the variability in these subspaces, are utilized to monitor the pollution status and detect the potential APEs in the CIP. In addition, the concept of APE detectability is also defined, and the condition for APE detectability is derived, which explains when the APEs can be detectable. A simulated case for a CIP in Zhejiang province of China is studied to evaluate the performance of this approach. The study indicates that the method can have an almost 100% APE detection rate. The real-world measurements of Total Volatile Organic Compounds (TVOC) at a 10-min time interval from 3 December 2020∼12 December 2020 are also analyzed, and 64 APEs caused by excessive TVOC emissions are detected in a total of 1440 time points.