PID autotuning using neural networks and model reference adaptive control

This paper describes the application of artificial neural networks for automatic tuning of PID controllers using the Model Reference Adaptive Control approach. The effectiveness of the proposed method is shown through a simulated application.

Linear-wavelet models for system identification

A model structure comprising a wavelet network and a linear term is proposed for nonlinear system identification. It is shown that under certain conditions wavelets are orthogonal to linear functions and, as a result, the two parts of the model can be identified separately. The linear-wavelet model is compared to a standard wavelet network using data from a simulated fermentation process. The results show that the linear-wavelet model yields a smaller modelling error when compared to a wavelet network using the same number of regressors.

Hierarchical control and optimisation of the Tennessee Eastman process

Based on integrated system optimisation and parameter estimation a method is described for on-line steady state optimisation which compensates for model-plant mismatch and solves a non-linear optimisation problem by iterating on a linear - quadratic representation. The method requires real process derivatives which are estimated using a dynamic identification technique. The utility of the method is demonstrated using a simulation of the Tennessee Eastman benchmark chemical process.

Imaging the structure of activated carbon using aberration corrected TEM

The precise atomic structure of activated carbon is unknown, despite its commercial importance. Here we show that the structure of a commercial activated carbon can be imaged directly using aberration corrected transmission electron microscopy. Images are presented both of the as-produced carbon and of the carbon following heat- treatment at 2000°C. In the 2000°C carbon clear evidence is found for the presence of pentagonal rings, suggesting that the carbons have a fullerene-related structure.

Imaging the atomic structure of activated carbon

The precise atomic structure of activated carbon is unknown, despite its huge commercial importance in the purification of air and water. Diffraction methods have been extensively applied to the study of microporous carbons, but cannot provide an unequivocal identification of their structure. Here we show that the structure of a commercial activated carbon can be imaged directly using aberration-corrected transmission electron microscopy. Images are presented both of the as-produced carbon and of the carbon following heat treatment at 2000 degrees C. In the 2000 degrees C carbon clear evidence is found for the presence of pentagonal rings, suggesting that the carbons have a fullerene-related structure. Such a structure would help to explain the properties of activated carbon, and would also have important implications for the modelling of adsorption on microporous carbons.

Acrylamide is formed in the Maillard reaction

Reports of the presence of acrylamide in a range of fried and oven-cooked foods have caused worldwide concern because this compound has been classified as probably carcinogenic in humans. Here we show how acrylamide can be generated from food components during heat treatment as a result of the Maillard reaction between amino acids and reducing sugars. We find that asparagine, a major amino acid in potatoes and cereals, is a crucial participant in the production of acrylamide by this pathway.

P-Prism: a computationally efficient approach to scaling up classification rule induction

Top Down Induction of Decision Trees (TDIDT) is the most commonly used method of constructing a model from a dataset in the form of classification rules to classify previously unseen data. Alternative algorithms have been developed such as the Prism algorithm. Prism constructs modular rules which produce qualitatively better rules than rules induced by TDIDT. However, along with the increasing size of databases, many existing rule learning algorithms have proved to be computational expensive on large datasets. To tackle the problem of scalability, parallel classification rule induction algorithms have been introduced. As TDIDT is the most popular classifier, even though there are strongly competitive alternative algorithms, most parallel approaches to inducing classification rules are based on TDIDT. In this paper we describe work on a distributed classifier that induces classification rules in a parallel manner based on Prism.

Improved MATLAB GUIs for learning frequency response methods

The author developed two GUIs for asymptotic Bode plots and identification from such plots aimed at improving the learning of frequency response methods: these were presented at UKACC Control 2012. Student feedback and reflection by the author suggested various improvements to these GUIs, which have now been implemented. This paper reviews the earlier work, describes the improvements, and includes positive feedback from the students on the GUIs and how they have helped their understanding of the methods.

The cyber challenge: a robotics project to enthuse

Robotics is a key theme in many of the degrees offered in Systems Engineering. The topic has proved useful in attracting students to the University, and it also provides the basis of much practical and project work throughout the degrees. This paper focuses on one aspect, a Part 2 project in which students doing various degrees work together to develop a mobile robot which is controlled remotely to navigate an environment and perform specific tasks. In addition to providing practical experience of relevant academic topics, this project helps to contribute to key teaching and learning priorities including problem based learning, motivation and important employability skills.