L. Felipe Blázquez

Fuzzy logic-based decision-making for fault diagnosis in a DC motor

The uncertainty of system models, the presence of noise and the stochastic behaviour of several variables reduce the reliability and robustness of the fault diagnosis methods. To tackle these kinds of problems, this paper presents a decision-making module based on fuzzy logic for model-based fault diagnosis applications. Fuzzy rules use the concept of fault possibility and knowledge of the sensitivities of the residual equations. A fault detection and isolation system, based on the input-output linear model parity equations approach, and including this decision-making module, has been successfully applied in laboratory equipment, resulting in a reduction of the uncertainty due to disturbances and modelling errors. Furthermore, the experimental sensitivity values of the residual equations allow the fault size to be estimated with sufficient accuracy.

Neuro-fuzzy identification applied to fault detection in nonlinear systems

This article describes a fault detection method, based on the parity equations approach, to be applied to nonlinear systems. The input–output nonlinear model of the plant, used in the method, has been obtained by a neural fuzzy inference architecture and its learning algorithm. The proposed method is able to detect small abrupt faults, even in systems with unknown nonlinearities. This method has been applied to a real industrial pilot plant, and good performance has been obtained for the experimental case of fault detection in the level sensor of a level control process in the said industrial pilot plant.

Additive fault detection in nonlinear dynamic systems with saturation

This paper describes the effects of input saturation on the performance of a model-based fault detection method based on the input-output parity equation approach. For this purpose, the level control of a chemical reactor has been chosen as the control process to be analyzed, where the saturation of the dynamic process is due to the inflow control valve, and only additive faults have been considered. This study has been carried out in two ways: first by simulation techniques and second on a real industrial system. In the simulated case, the decrease in the fault detectability due to the saturation effects is shown, and some ways of achieving higher fault detectability are explored. The results obtained in the industrial case complement those obtained in the simulated case, and also reveal the existence of a relation between the control strategy used in the process and additive fault detectability, in the sense that increases in fault detectability are obtained due to the use of faster control strategies.

Traffic sign recognition application based on image processing techniques

Abstract This paper describes a software application for traffic sign recognition (TSR). The application works in four stages. First, an image preprocessing step and the detection of regions of interest (ROIs), which involves a series of steps that include transforming the image to grayscale and applying edge detection by the Laplacian of Gaussian (LOG) filter. Secondly, the potential traffic signs detection, where the ROIs are compared with each shape pattern. Thirdly, a recognition stage using a cross-correlation algorithm, where each potential traffic sign, if validated, is classified according to the data-base of traffic signs. Finally, the previous stages can be managed and controlled by a graphical user interface, which has been specially designed for this purpose. The results obtained show a good performance of the developed application, taking into account acceptable conditions of size and contrast of the input image.

Additive Fault Detection in a Real Nonlinear System with Saturation

Abstract This paper describes an application of the parity equation method to a real nonlinear system with saturation. This nonlinear system is an industrial pilot plant, and the water level control of the pressurized reactor is the control process implemented in this pilot plant. The saturation of the dynamic process is due to the inflow control valve. The parity equation method uses a nonlinear model of the plant to generate residuals, and only additive faults are considered. The goal of this application is to show the effects of saturation over the performance of the model-based fault detection method. In this sense, the results obtained indicate the decrease of additive fault detectability due to presence of saturation in the dynamic process. Furthermore, these results also reveal the existence of a relation between the control strategy used in the process and additive fault detectability, in the sense that increases of fault delectability are obtained the due to use of faster control strategies

Online Monitoring of an Industrial Semi-Batch Vinyl Acetate Polymerization Reaction by Programmable Logic Controllers

Abstract This paper describes a virtual online monitoring of the non-measurable conditions inside a real-life industrial reactor for the production of polyvinyl acetate. These non-measurable conditions are the conversion, the polymerization rate, the viscosity and the solids content. The goal of this paper is to use them for control strategies, which can be implemented and solved in real time on common industrial programmable logic controllers (PLCs). Currently, these non-measurable conditions can be estimated from the measured variables by a detailed dynamic model, which is composed of a set of differential and algebraic equations, which can only be solved by means of iterative numerical methods. In this work, this model has been simplified and discretized to allow its implementation in industrial PLCs, which is the common practice in the industry. As a result, the PLC available at the factory can estimate the internal status of the reactor in real time with a high accuracy.

Application of fuzzy tools to the automatic analysis of system dynamics models: an example of World3

Abstract Even the simple run of a medium size system dynamics model can be a cumbersome process, since the uncertainty of the parameters forces the modeler to consider many runs before being confident of how the model behaves. System dynamics simulation packages include some analysis tools, but there is a lack of a commonly used analysis tool such as fuzzy clustering. In this paper, we explore the possibilities of a programming language, Matlab, and its simulation tool, Simulink, for the above mentioned possibilities. These languages enable the development of customized analysis tools at a very low programming cost. World3 model is a widely known example of a large model whose analysis becomes a difficult task. We have programmed this model in Simulink and, afterwards, some tools, such as fuzzy clustering and pathway participation metric (PPM), have been applied to show their potential capacity to facilitate model simulation analyses.

Neural network based radioisotope discrimination on polyvinyl toluene radiation portal monitors

Radiation portal monitors (RPMs) are an effective mean of detecting radioactive material inside cargo containers. Polyvinyl toluene (PVT) monitors are the most broadly extended mainly due to their cost. The drawback when compared to other detectors is the lower resolution of the measured energy spectra. This low resolution hinders the use of spectrometric analysis to discriminate isotopes and discard nuisance alarms. Every alarm must thus be checked in a second inspection by a handheld detector or a spectroscopy-based radiation portal. The cost of this secondary inspection in terms of throughput can be significant, specially at maritime ports and borders. This paper aim is to assess the ability of neural networks to discriminate radioactive isotopes from the energy spectrum as measured by PVT RPMs. For this purpose, the system proposed preprocesses these energy spectra, dividing them by specific zones and transforming them into information. In a second step, this information is used by the neural network architecture, which allows to classify the radioisotopes in different groups.

Neurofuzzy based temperature prediction of an industrial polymerization reactor in real time

The ultimate purpose of this work is the real-time control of a semibatch emulsion polymerization reaction. The main variable under control is the temperature inside the reactor. The keypoint to get accurate control is the early detection of temperature deviations. A neurofuzzy network has been trained to predict the temperature from some of the previously calculated variables. This approach aims to extend the prediction horizon with which the temperature is predicted to an order of magnitude of minutes, based on variables which are not measured online but rather estimated using a reduced calorimetric model. This methodology has been applied to data from a real life emulsion polymerization reactor, and in order to allow its implementation in this factory, it has been ensured that all of the operations can be performed in real time by a common Programmable Logic Controller (PLC). The resulting set of equations predicts the temperature several minutes in advance with good accuracy.

Fault Detection by Neuro-Fuzzy Identification in a Nonlinear System

: This paper describes a fault detection and identification method, based on the parity equations approach, to be applied to nonlinear systems. The input-output nonlinear model of the plant, used in the method, has been obtained by a neural fuzzy inference system. The method proposed is able to detect both abrupt and incipient faults. This method has been applied with good performance to a real industrial pilot plant for fault detection in a level sensor.