Qualitative process trend representation is an useful approach to model the temporal evolution of sensor data and has been applied in areas such as process monitoring, data compression and fault diagnosis. However, th...
详细信息
Qualitative process trend representation is an useful approach to model the temporal evolution of sensor data and has been applied in areas such as process monitoring, data compression and fault diagnosis. However, the sheer volume of realtime sensor data that needs to be processed necessitates an automated approach. The first step of recovering the important temporal features is made difficult because of the absence of a priori knowledge about trend characteristics such as noise and varying scales of evolution. In this paper, we propose a novel approach to automatically identify the qualitative shapes of trends using an interval-halving based polynomial-fit technique. To estimate the significance of fit-error, wavelet-based denoising is used. The procedure identifies piecewise unimodals represented as quadratic segments. Finally, a unique assignment of qualitative shape is made to each of the identified segments. The application of the technique is illustrated on both simulated and industrial data.
A multi-agent system (MAS) can be viewed as a group of entities interacting to achieve individual or collective goals. Communication is a central issue in this interaction between agents. This paper uses the COOL lang...
详细信息
A multi-agent system (MAS) can be viewed as a group of entities interacting to achieve individual or collective goals. Communication is a central issue in this interaction between agents. This paper uses the COOL language to define an extended version of the contract net protocol to address issues in one particular MAS-sensible agents. The proposed implementation of our protocol and its use in the sensible agent testbed is also discussed.
Principal component analysis (PCA) has emerged as a useful tool for process monitoring and fault diagnosis. The general approach requires the user to identify the root cause by interpreting the measured variable contr...
详细信息
Principal component analysis (PCA) has emerged as a useful tool for process monitoring and fault diagnosis. The general approach requires the user to identify the root cause by interpreting the measured variable contributions to the residual and/or principal components. This could be tedious and often impossible for a large process. It also hinders the automation of high level supervisory tasks like choice of corrective actions and their associated costs. In this paper, the interpretation of the PCA-based contributions is automated using signed directed graphs (SDGs). The implementation of the PCA-SDG based fault diagnosis algorithm is done using Gensym's expert system shell G2. Its application is illustrated on the Amoco Model IV Fluidized Catalytic Cracking Unit (FCCU).
The Batch Design-Kit (BDK) environment is a software system for the development of batch manufacturing processes. Given a set of reaction pathways, BDK enables chemists and chemical engineers to investigate possible p...
The Batch Design-Kit (BDK) environment is a software system for the development of batch manufacturing processes. Given a set of reaction pathways, BDK enables chemists and chemical engineers to investigate possible production routes, refine promising candidates and develop a detailed chemical recipe in an interactive manner. Automatic facility allocation of equipment for these conceptual recipes help determine the feasibility and efficiency of the processing idea at existing manufacturing sites. Material assessment reveals the environmental impact caused by the utilized chemical compounds and process mixtures. Automatic waste treatment synthesis generates efficient treatment strategies for all process wastes and calculates the residuals emitted to the environment. With process analysis based on the facility allocation, material assessment and treatment selection, alternative processing schemes can be ranked and potential problems, such as examination of equipment costs, regulatory compliance and expenditures for waste treatment, can be identified at an early stage of the process development. Through its material-centered design concept, BDK maintains a cause-effect relationship between process mixtures, operations and their economic and environmental performance. In that way, process improvement can be focused specifically on the replacement of offending chemicals and/or the improvement or modification of problematic operations or operation sequences.
Multi-scale models of processing systems offer an attractive alternative to the conventional models in the time or frequency domain for process simulation, estimation and control. Defined on trees, these models captur...
Multi-scale models of processing systems offer an attractive alternative to the conventional models in the time or frequency domain for process simulation, estimation and control. Defined on trees, these models capture the essential features of the systems' dynamic behavior, localized in time and scale. In this paper we introduce a formal framework for the formulation of multi-scale models, which leads naturally to a multi-scale systems theory with ensuing definitions for transfer functions, stability, controllability and observability notions. The resulting formulations of (a) system identification, and (b) model predictive control, offer certain very attractive properties, e.g. low computational complexity, natural integration of measurements and control actions at different scales, modeling errors and disturbance descriptions at various scales, handling of multi-rate monitoring and control, and other. In this paper we will provide a brief introduction on how to construct multi-scale models and how they can be used in process estimation and control, summarize some of the key results, and sketch the directions for further work.
We discuss the development of a wavelet theory-based adaptive system for trend analysis (W-ASTRA). W-ASTRA performs process-monitoring and diagnosis. The main contributions of this paper are two fold. A wavelet theory...
详细信息
We discuss the development of a wavelet theory-based adaptive system for trend analysis (W-ASTRA). W-ASTRA performs process-monitoring and diagnosis. The main contributions of this paper are two fold. A wavelet theory based nonlinear adaptive algorithm has been developed for identification of trends from sensor data. In order to perform diagnosis using the identified trends, a knowledge base is required. Our second contribution is the development of an automated framework for knowledge base development. W-ASTRA uses the adaptive algorithm for identification of sensor trends and the knowledge base generated by the automated framework for diagnosing fault origins from the identified trends. The application of W-ASTRA is demonstrated on the Amoco Model IV FCCU.
Response data from a novel micro-hotplate gas sensor were used to test a robust classification technique based on an artificial neural network. Four different compounds were identified even when the sensor signal was ...
详细信息
Response data from a novel micro-hotplate gas sensor were used to test a robust classification technique based on an artificial neural network. Four different compounds were identified even when the sensor signal was corrupted by high levels of noise and drift. Additional verification rules in doubtful cases were provided by examining unique binary patterns of outputs of the hidden layer neurons. Similarities to proposed models of the human nose are noted.
The purpose of this review is three-fold. First, sketch the directions that research and industrial applications of ''intelligentsystems'' have taken in several areas of processengineering. Second, i...
详细信息
The purpose of this review is three-fold. First, sketch the directions that research and industrial applications of ''intelligentsystems'' have taken in several areas of processengineering. Second, identify the emerging trends in each area, as well as the common threads that cut across several domains of inquiry. Third, stipulate research and development themes of significant importance for the future evolution of ''intelligentsystems'' in processengineering. The paper covers the following seven areas: diagnosis of process operations;monitoring and analysis of process trends;intelligent control;heuristics and logic in planning and scheduling of process operations;modeling languages, simulation, and reasoning;intelligence in scientific computing;knowledge-based engineering design. Certain trends seem to be common and will (in all likelihood) characterize the nature of the future deployment of ''intelligentsystems''. These trends are: (1) Specialization to narrowly defined classes of problems. (2) Integration of multiple knowledge representations, so that all of relevant knowledge is captured and utilized. (3) Integration of processing methodologies, which tends to blur the past sharp distinctions between AI-based techniques and those from operations research, systems and control theory, probability and statistics. (4) Rapidly expanding range of industrial applications with significant increase in the scope of engineering tasks and size of problems.
Hazard and Operability (HAZOP) analysis is the most widely used and recognized as the pre ferred safety analysis approach in the chemical process industry. A model-based framework and an expert system called HAZOPExpe...
详细信息
Hazard and Operability (HAZOP) analysis is the most widely used and recognized as the pre ferred safety analysis approach in the chemical process industry. A model-based framework and an expert system called HAZOPExpert has been developed recently for automating this analysis. The performance of HAZOPExpert was evaluated on a sour water stripper plant and a hydrotreator plant case studies. The expert system was found to successfully emulate the human experts reasoning and identify the hazards similar to the HAZOP team. But, it generated a large number of unrealizable hazardous consequences compared to the team, due to the strict qualitative seasoning approach implemented. In order to filter and rank the results from HAZOPExpert using additional quantitative knowledge similar to the team, a semi-quantitative reasoning methodology is developed using the quantitative design and operating specifications of the process units, and process material property values. Significant reduction in the number of consequences was obtained using this approach on an ethylene plant case study.
暂无评论