Recent developments in neural network PID autotuning

PID controllers are widely used in industrial applications. Because the plant can be time variant, methods of autotuning of this type of controllers, are of great economical importance, see (Astrom, 1996). Since 1942, with the work of Ziegler and Nichols (Ziegler and Nichols, 1942), several methods have been proposed in the literature. Recently, a new technique using neural networks was proposed (Ruano et al., 1992). This technique has been shown to produce good tunings as long as certain limitations are met.

Neural network assisted autotuning of industrial PID controllers

A recent servey (1) has reported that the majority of industrial loops are controlled by PID-type controllers and many of the PID controllers in operation are poorly tuned. poor PID tuning is due to the lack of a simple and practical tuning method for avarage users, and due to the tedious procedurs involved in the tuning and retuning of PID controllers.

Comparison of alternative approaches to neural network PID autotuning

In this paper, a scheme for the automatic tuning of PID controllers on-line, with the assistance of trained neural networks, is proposed. The alternative approaches are presented and compared.

A neural network PID autotuner

Proportional, Integral and Derivative (PID) regulators are standard building blocks for industrial automation. Their popularity comes from their rebust performance and also from their functional simplicity. Whether because the plant is time-varying, or because of components ageing, these controllers need to be regularly retuned.

A neural network controller

Proportional, Integral and Derivative (PID) regulators are standard building blocks for industrial automation. The popularity of these regulators comes from their rebust performance in a wide range of operating conditions, and also from their functional simplicity, which makes them suitable for manual tuning.

A new formulation of the learning problem for a neural network controller

In this paper we consider the learning problem for a class of multilayer perceptrons which is practically relevant in control systems applications. By reformulating this problem, a new criterion is developed, which reduces the number of iterations required for the learning phase.

A Fuzzy-Wavelet Neural Network Model for the Detection of Meat Spoilage using an Electronic Nose

Food product safety is one of the most promising areas for the application of electronic noses. The performance of a portable electronic nose has been evaluated in monitoring the spoilage of beef fillet stored aerobically at different storage temperatures (0, 4, 8, 12, 16 and 20°C). This paper proposes a fuzzy-wavelet neural network model which incorporates a clustering pre-processing stage for the definition of fuzzy rules. The dual purpose of the proposed modeling approach is not only to classify beef samples in the respective quality class (i.e. fresh, semi-fresh and spoiled), but also to predict their associated microbiological population directly from volatile compounds fingerprints. Comparison results indicated that the proposed modeling scheme could be considered as a valuable detection methodology in food microbiology

LMP based bid formation for virtual power players operating in smart grids

Power system organization has gone through huge changes in the recent years. Significant increase in distributed generation (DG) and operation in the scope of liberalized markets are two relevant driving forces for these changes. More recently, the smart grid (SG) concept gained increased importance, and is being seen as a paradigm able to support power system requirements for the future. This paper proposes a computational architecture to support day-ahead Virtual Power Player (VPP) bid formation in the smart grid context. This architecture includes a forecasting module, a resource optimization and Locational Marginal Price (LMP) computation module, and a bid formation module. Due to the involved problems characteristics, the implementation of this architecture requires the use of Artificial Intelligence (AI) techniques. Artificial Neural Networks (ANN) are used for resource and load forecasting and Evolutionary Particle Swarm Optimization (EPSO) is used for energy resource scheduling. The paper presents a case study that considers a 33 bus distribution network that includes 67 distributed generators, 32 loads and 9 storage units.

ANN based day-ahead spinning reserve forecast for electricity market simulation

Electricity market players operating in a liberalized environment require adequate decision support tools, allowing them to consider all the business opportunities and take strategic decisions. Ancillary services represent a good negotiation opportunity that must be considered by market players. This paper deals with short-term predication of day-ahead spinning reserve (SR) requirement that helps the ISO to make effective and timely decisions. Based on these forecasted information, market participants can use strategic bidding for day-ahead SR market. The proposed concepts and methodologies are implemented in MASCEM, a multi-agent based electricity market simulator. A case study based on California ISO (CAISO) data is included; the forecasted results are presented and compared with CAISO published forecast.

Aplicação de redes neuronais artificiais à deteção e isolamento de falhas em processos industriais

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia Mecânica

Classificação e previsão de cardiotocogramas usando algoritmos baseados em redes neuronais

Mestrado em Computação e Instrumentação Médica

Previsão de avarias em transformadores de potência através da análise de ensaios físico-químicos

Os Transformadores de potência são máquinas de elevada importância ao nível dos Sistemas Elétricos de Energia (SEE) uma vez que são estas máquinas que possibilitam a interligação dos diferentes níveis de tensão da rede e a transmissão de energia elétrica em Corrente Alternada (CA). Geralmente, estas máquinas são de grandes dimensões e de elevado nível de complexidade construtiva. Caracterizam-se por possuírem períodos de vida útil bastante elevados (vinte a trinta anos) e preços elevados, o que conduz a um nível de exigência de fiabilidade muito elevada, uma vez que não e viável a existência de muitos equipamentos de reserva nos SEE. Com o objetivo de tentar maximizar o período de vida útil dos transformadores de potência e a sua fiabilidade, tenta-se, cada vez mais, implementar conceitos de manutenção preventiva a este tipo de máquinas. No entanto, a gestão da sua vida útil e extremamente complexa na medida em que, estas máquinas têm vários componentes cruciais e suscetiveis de originar falhas e, quase todos eles, encontram-se no interior de uma cuba. Desta forma, não e possível obter uma imagem do seu estado, em tempo real, sem colocar o transformador fora de serviço, algo que acarreta custos elevados. Por este motivo, desenvolveu-se uma técnica que permite obter uma indicação do estado do transformador, em tempo real, sem o retirar de serviço, colhendo amostras do óleo isolante e procedendo a sua análise físico-química e Analise Gases Dissolvidos (DGA). As análises aos óleos isolantes tem vindo a adquirir uma preponderância muito elevada no diagnóstico de falhas e na analise do estado de conservação destes equipamentos tendo-se desenvolvido regras para interpretação dos parâmetros dos óleos com carácter normativo. Considerando o conhecimento relativo a interpretação dos ensaios físico-químicos e DGA ao oleol, e possível desenvolver ferramentas capazes de otimizar essas mesmas interpretações e aplicar esse conhecimento no sentido de prever a sua evolução, assim como o surgimento de possíveis falhas em transformadores, para assim otimizar os processos de manutenção. Neste campo as Redes Neuronais Artificiais (RNAs) têm um papel fundamental

Neural control of an autonomous robot

This article aims to apply the concepts associated with artificial neural networks (ANN) in the control of an autonomous robot system that is intended to be used in competitions of robots. The robot was tested in several arbitrary paths in order to verify its effectiveness. The results show that the robot performed the tasks with success. Moreover, in the case of arbitrary paths the ANN control outperforms other methodologies, such as fuzzy logic control (FLC).

Training deep convolutional architectures for vision

Les tâches de vision artificielle telles que la reconnaissance d’objets demeurent irrésolues à ce jour. Les algorithmes d’apprentissage tels que les Réseaux de Neurones Artificiels (RNA), représentent une approche prometteuse permettant d’apprendre des caractéristiques utiles pour ces tâches. Ce processus d’optimisation est néanmoins difficile. Les réseaux profonds à base de Machine de Boltzmann Restreintes (RBM) ont récemment été proposés afin de guider l’extraction de représentations intermédiaires, grâce à un algorithme d’apprentissage non-supervisé. Ce mémoire présente, par l’entremise de trois articles, des contributions à ce domaine de recherche. Le premier article traite de la RBM convolutionelle. L’usage de champs réceptifs locaux ainsi que le regroupement d’unités cachées en couches partageant les même paramètres, réduit considérablement le nombre de paramètres à apprendre et engendre des détecteurs de caractéristiques locaux et équivariant aux translations. Ceci mène à des modèles ayant une meilleure vraisemblance, comparativement aux RBMs entraînées sur des segments d’images. Le deuxième article est motivé par des découvertes récentes en neurosciences. Il analyse l’impact d’unités quadratiques sur des tâches de classification visuelles, ainsi que celui d’une nouvelle fonction d’activation. Nous observons que les RNAs à base d’unités quadratiques utilisant la fonction softsign, donnent de meilleures performances de généralisation. Le dernière article quand à lui, offre une vision critique des algorithmes populaires d’entraînement de RBMs. Nous montrons que l’algorithme de Divergence Contrastive (CD) et la CD Persistente ne sont pas robustes : tous deux nécessitent une surface d’énergie relativement plate afin que leur chaîne négative puisse mixer. La PCD à "poids rapides" contourne ce problème en perturbant légèrement le modèle, cependant, ceci génère des échantillons bruités. L’usage de chaînes tempérées dans la phase négative est une façon robuste d’adresser ces problèmes et mène à de meilleurs modèles génératifs.