Real-time fault diagnosis using knowledge-based expert system

Nan, Cen; Khan, Faisal; Iqbal, M. Tariq

doi:10.1016/j.psep.2007.10.014

“…Retaining only the wavelet coefficients from each scale for both the training and testing data sets. 4. Retaining only the wavelet coefficients for the training data set, but retaining the entire scale of wavelet coefficients for the testing data set.…”

Section: Multiscale Shewhart Chart Fault Detection Algorithmmentioning

confidence: 99%

“…Fault detection methods can be classified according to their dependency on models used. These can be quantitative model-based methods, such as observers or parity space [2], qualitative model-based methods, such as fault trees, digraphs, or even process engineering experts [3][4][5][6], or process-history (data-based) methods, such as Principal Component Analysis (PCA) and neural networks [7,8]. Accurate qualitative and quantitative model-based methods may not always be available, especially for complex processes with multiple process variables, and therefore data-based methods are often employed.…”

Section: Introductionmentioning

confidence: 99%

Improved Fault Detection and Process Safety Using Multiscale Shewhart Charts.

Sheriff¹,

Nounou²

2017

J Chem Eng Process Technol

View full text Add to dashboard Cite

Process safety is a critical component in various process industries. Statistical process monitoring techniques were initially developed to maximize efficiency and productivity, but over the past few decades with catastrophic industrial disasters, process safety has become a top priority. Sensors play a crucial role in recording process measurements, and according to the number of monitored variables, process monitoring techniques can be classified into univariate or multivariate techniques. Most univariate process monitoring techniques rely on three fundamental assumptions: that process residuals contain a moderate level of noise, are independent, and are normally distributed. Practically, however, due to a variety of reasons such as modeling errors and malfunctioning sensors, these assumptions are violated, which can lead to catastrophic incidents. Fortunately, multiscale wavelet-based representation of data inherently possesses characteristics that are able to deal with these violations of assumptions. Therefore, in this work, multiscale representation is utilized to enhance the performance of the Shewhart chart (which is a well-known univariate fault detection method) to help improve its performance. The performance of the developed multiscale Shewhart chart was assessed and compared to the conventional chart through two examples, one using synthetic data, and the other using simulated distillation column data. The results of both examples clearly show that the developed multiscale Shewhart chart provides lower missed detection and false alarm rates, as well as lower ARL 1 values (i.e., quicker detection) for most cases where the fundamental assumptions of the Shewhart chart are violated. Additionally, the relative simplicity of the proposed algorithm encourages its implementation in practice to help improve process safety.

show abstract

“…In quality management there are papers on fuzzy CBR on quality management in GPRS networks (Pulkkinen et al 2008) , a fuzzy DEA-neural approach to measuring design services performance in PCM projects (C.-H. ) and a neural-genetic algorithm for reservoir water quality management (Kuo et al 2006 Nan et al (2008) present a knowledge based expert system for process industry using FL to make inferences based on the acquired information and knowledge. Rafiee et al(2009) presents an optimised gear fault identification system using GAs and NNs.…”

Section: Hybrid Ai In Quality Maintenance and Fault Diagnosismentioning

confidence: 99%

A survey of AI in operations management from 2005 to 2009

Kobbacy

¹

,

Vadera

²

2011

Journal of Manufacturing Technology Management

View full text Add to dashboard Cite

Purpose: The use of AI for operations management, with its ability to evolve solutions, handle uncertainty and perform optimisation continues to be a major field of research. The growing body of publications over the last two decades means that it can be difficult to keep track of what has been done previously, what has worked, and what really needs to be addressed. Hence this paper presents a survey of the use of AI in operations management aimed at presenting the key research themes, trends and directions of research.Design/methodology/approach: The paper builds upon our previous survey of this field which was carried out for the 10 year period 1995 to 2004(Kobbacy et al. 2007. Like the previous survey, it uses the Elsevier's ScienceDirect database as a source. The framework and methodology adopted for the survey is kept as similar as possible to enable continuity and comparison of trends. Thus the application categories adopted are: (a) design, (b) scheduling, (c) process planning and control and (d) quality, maintenance and fault diagnosis. Research on utilising neural networks, case based reasoning, fuzzy logic, knowledge based systems, data mining, and hybrid AI in the four application areas are identified. Findings:The survey categorises over 1400 papers, identifying the uses of AI in the four categories of operations management and concludes with an analysis of the trends, gaps and directions for future research. The findings include: (a) The trends for Design and Scheduling show a dramatic increase in the use of GAs since 2003-04 that reflect recognition of their success in these areas, (b) A significant decline in research on use of KBS, reflecting their transition into practice, (c) an increasing trend in the use of fuzzy logic in Quality, Maintenance and Fault Diagnosis, (d) surprising gaps in the use of CBR and hybrid methods in operations management that offer opportunities for future research.Originality/value: This is the largest and most comprehensive study to classify research on the use of AI in operations management to date. The survey and trends identified provide a useful reference point and directions for future research.

show abstract

“…These charts may also require one or more tuning parameters to be modified, thus making the design of these charts increasingly computationally expensive and tedious. Process noise is often a concern in most process industries and can adversely affect the accuracy of fault detection techniques [4,19,20]. Charts with process memory such as CUSUM and EWMA may be able to address concerns related to process noise as they utilize filters.…”

Section: Introductionmentioning

confidence: 99%

Systematic Preparation of Artificial Cells (DNA Crown Cells)

Inooka¹

2017

J Chem Eng Process Technol

View full text Add to dashboard Cite

Process safety is a critical component in various process industries. Statistical process monitoring techniques were initially developed to maximize efficiency and productivity, but over the past few decades with catastrophic industrial disasters, process safety has become a top priority. Sensors play a crucial role in recording process measurements, and according to the number of monitored variables, process monitoring techniques can be classified into univariate or multivariate techniques. Most univariate process monitoring techniques rely on three fundamental assumptions: that process residuals contain a moderate level of noise, are independent, and are normally distributed. Practically, however, due to a variety of reasons such as modeling errors and malfunctioning sensors, these assumptions are violated, which can lead to catastrophic incidents. Fortunately, multiscale wavelet-based representation of data inherently possesses characteristics that are able to deal with these violations of assumptions. Therefore, in this work, multiscale representation is utilized to enhance the performance of the Shewhart chart (which is a well-known univariate fault detection method) to help improve its performance. The performance of the developed multiscale Shewhart chart was assessed and compared to the conventional chart through two examples, one using synthetic data, and the other using simulated distillation column data. The results of both examples clearly show that the developed multiscale Shewhart chart provides lower missed detection and false alarm rates, as well as lower ARL 1 values (i.e., quicker detection) for most cases where the fundamental assumptions of the Shewhart chart are violated. Additionally, the relative simplicity of the proposed algorithm encourages its implementation in practice to help improve process safety.

show abstract

Real-time fault diagnosis using knowledge-based expert system

Cited by 111 publications

References 13 publications

Improved Fault Detection and Process Safety Using Multiscale Shewhart Charts.

Improved Fault Detection and Process Safety Using Multiscale Shewhart Charts.

A survey of AI in operations management from 2005 to 2009

Systematic Preparation of Artificial Cells (DNA Crown Cells)

Contact Info

Product

Resources

About