No description available.

No description available.

Model predictive control (MPC) is an advanced control that has found widespread use in industries, particularly in process industries like oil refining and petrochemicals. Although the basic premise behind MPC is easy to comprehend, its inner workings are still generally viewed or regarded as too advanced for actual plant operator understanding. This lack of understanding is exposed when MPC performance deteriorates sometime after commissioning, as is often the case in some commercially operated process plants. Currently operators might have difficulty over reasoning about MPC performance degradation and formulating steps to investigate the cause. A tool is described that aims to make MPC more transparent to the operators. Commonly reported causes of MPC performance degradation are discussed and ways in which the operator can recognise them when they occur are outlined. Issues that are addressed include: making the set of controlled variables to be used for a given set of manipulated variables simpler and clearer; ways to recognise when a MPC controller is performing poorly and to identify the source of performance deterioration. An aim is to determine under what instances the operator can return the MPC performance to previous levels or request for specialist support or simply switch the MPC off. A goal is to avoid the kind of often reported situation where the operator gets worried that the controller is deteriorating and ends up taking knee jerk actions that cause further problems in MPC. At the top of the maintenance tool hierarchy is the trends comparison group, where MPC reference graphical performance trends are compared with actual graphical performance trends counterpart. If any abnormality is observed in these trends, the operator is encouraged to choose an option from a list of preliminary diagnostic questions contained in a group below trends comparison group, which best describes the observed abnormality. Each abnormality is associated with a list of suspected causes. When a suspected cause is chosen from the associated list, the operator is led to the symptoms investigation window, which contains scripts detailing steps for systematic examination of each symptom, with a view to either rejecting or confirming the suspicion. Assisted in the investigation are four background information windows: the virtual plant without MPC window, the virtual plant with MPC window, the transfer function matrix window and steady state gain, relative gain array (RGA) and relative weight array (RWA) window. The windows contain information and guidance that the operator might refer to from time to time during symptom investigation. Development of the maintenance tool is still at the design stage. The key components described have been research implementing MPC on three nonlinear process models, a continuous stirred tank reactor (CSTR), an evaporator process and a fluid catalytic cracking unit (FCCU). Case studies representing different MPC degradation scenarios are simulated, followed by a systematic procedure for diagnosing, isolating and recovering from such degradation, based on assumed operator’s perspective and expert’s technical reasoning. The knowledge gained from the case studies is used to develop an outline of a vision for a data-driven model predictive maintenance tool to help the operator make sensible judgements about performance degradation, the form and direction of diagnosis and fault isolation, and possibly, the recovery procedure.

No description available.

No description available.

The research discussed in this thesis explores a new method for the detection of grinding burn temperature using a laser irradiation acoustic emission (AE) sensing technique. This method is applicable for the grinding process monitoring system, providing an early warning for burn detection on metal alloy based materials (specifically nickel alloy based materials: Inconel718 and MarM002). The novelty in this research is the laser irradiation induced thermal AE signal that represents the grinding thermal behaviour and can be used for grinding burn detection. A set of laser irradiation experiments were conducted to identify key process characteristics. By controlling the laser power, the required grinding temperatures were simulated on alloy test materials. The thermal features of the extracted AE signal were used to identify the high, medium and low temperature signatures in relation to the off-focal laser distances. Grinding experiments were also conducted to investigate burn conditions. The extracted AE data was used to identify grinding burn and no burn signatures in relation to the depth of cuts. A new approach using an artificial neural network (ANN) was chosen as the pattern recognition tool for classifying grinding burn detection and was used to classify grinding temperatures by extracting the mechanical-thermal grinding AE signal. The results demonstrated that the classification accuracy achieved was 66 % for Inconel718 and 63 % for MarM002 materials. The research established that the wheel wear has a large influence on the creation of burn within the workpiece surface. The results demonstrated that the AE signals in each grinding trial presents different levels of high, medium and low temperature scales. This type of information provides a foundation for a new method for monitoring of grinding burn and wheel wear.

Data mining refers to the automation of data analysis to extract patterns from large amounts of data. A major breakthrough in modelling natural patterns is the recognition that nature is fractal, not Euclidean. Fractals are capable of modelling self-similarity, infinite details, infinite length and the absence of smoothness. This research was aimed at simplifying the discovery and detection of groups in data using fractal dimension. These data mining tasks were addressed efficiently. The first task defines groups of instances (clustering), the second selects useful features from non-defined (unsupervised) groups of instances and the third selects useful features from pre-defined (supervised) groups of instances. Improvements are shown on two data mining classification models: hierarchical clustering and Artificial Neural Networks (ANN). For clustering tasks, a new two-phase clustering algorithm based on the Fractal Dimension (FD), compactness and closeness of clusters is presented. The proposed method, uses self-similarity properties of the data, first divides the data into sufficiently large sub-clusters with high compactness. In the second stage, the algorithm merges the sub-clusters that are close to each other and have similar complexity. The final clusters are obtained through a very natural and fully deterministic way. The selection of different feature subspaces leads to different cluster interpretations. An unsupervised embedded feature selection algorithm, able to detect relevant and redundant features, is presented. This algorithm is based on the concept of fractal dimension. The level of relevance in the features is quantified using a new proposed entropy measure, which is less complex than the current state-of-the-art technology. The proposed algorithm is able to maintain and in some cases improve the quality of the clusters in reduced feature spaces. For supervised feature selection, for classification purposes, a new algorithm is proposed that maximises the relevance and minimises the redundancy of the features simultaneously. This algorithm makes use of the FD and the Mutual Information (MI) techniques, and combines them to create a new measure of feature usefulness and to produce a simpler and non-heuristic algorithm. The similar nature of the two techniques, FD and MI, makes the proposed algorithm more suitable for a straightforward global analysis of the data.

Fabrication processes for the high volume production of parts with micro and nano scale features are very important in the global research and industry efforts to meet the increasing needs for device miniaturisation in numerous application areas. Processes for the replication of surface geometries are promising technologies that are capable to meet the demand of manufacturing products at a low cost and in high volume. Among these technologies, hot embossing is a process which relies on raising the temperature of a sheet of polymer up to its melting range and on pressing a heated master plate into the polymer for triggering a local flow of the material to fill the cavities to be replicated. This technique has attracted increased attention in recent years in particular due to the relatively simple set-up and low cost associated with its implementation in comparison to other replication techniques. The present work is concerned with investigating the process factors that influence hot embossing outcomes. In particularly, a detailed study of the process parameters’ effect on the cavity pressure, demoulding force and uniformity of the residual layer for different materials is conducted to analyse the further potential of this process. A review of the current state of the art on these topics reported in Chapter 2, is also used to assess the capability of this replication technology. Chapter 3 presents an experimental study on the effects of process parameters on pressure conditions in cavities when replicating parts in PMMA and ABS. To measure the pressure state of a polymer inside mould cavities, a condition monitoring system was implemented. Then, by employing a design of experiment approach, the iii pressure behaviour was studied as a function of different process conditions. In particular, the effects of three process parameters, embossing temperature and force and holding time, on the mould cavity pressure and the pressure distribution were investigated. In addition, using a simple analytical model, the minimum required embossing force to fill the cavities across the mould surface was calculated. The theoretical value obtained was then used to inform the design of the experiments. It was shown that cavity pressure and pressure distribution were dependent on both materials and processing conditions. The obtained results indicate that an increase in temperature and holding time reduced the pressure in the central and edge cavities of the mould and the pressure distribution while the opposite effect takes place when considering the embossing force. Also, it was observed that an increase of the embossing force has a positive effect on cavity filling but a negative influence for homogenous filling. In Chapter 4, a theoretical model was proposed to predict demoulding forces in hot embossing by providing a unified treatment of adhesion, friction and deformation phenomena that take place during demoulding of polymer microstructures. The close agreement between the predicted results and those measured experimentally suggests that the model successfully captures the relationship between mould design, feature sidewall, applied pressure, material properties, demoulding temperature and the resulting demoulding force. The theoretical results have been confirmed through comparisons with the demoulding experiments. The temperature at which the demoulding force is minimised depends on the geometry of the mould features along with the material properties of the mould and replica. The applied pressure has an important influence on the demoulding force iv as the increase in pressure augments the adhesion force due to changes in material dimensions and reduces the friction force due to resulting decrease in the thermal stress. Furthermore, the relationship between the residual layer uniformity and three process parameters was investigated in Chapter 5, using simulation and experimental studies when processing PMMA sheets. In particular, the characteristics of the residual layer thickness of embossed parts were analysed as a function of the moulding temperature, the embossing force and the holding time. Increasing the moulding temperature resulted in a reduction on the average residual layer thickness and on its non-uniformity. An increase in the embossing force led to a decrease in the homogeneity of the residual layer. Also, an improvement of the residual layer thickness uniformity was also observed when embossing with a longer holding time. The results of the conducted experimental and simulation studies were analysed to identify potential ways for improving the hot embossing process. Finally, in Chapter 6 the results and main findings from each of the investigations are summarised and further research directions are proposed.

No description available.

The primary aim of this thesis is to develop a novel procedure for an intelligent automatic diagnostic condition monitoring system for rolling element bearings. The applicability of this procedure is demonstrated by its implementation in a particular electric motor drive system. The novel bearing condition diagnostic procedure developed involves three stages combining the merits of advanced signal processing techniques, feature extraction methods and artificial neural networks. This procedure is the effective combination of these techniques and methods in a holistic approach to the rolling element bearing problem which provides the novelty in this thesis. Maintenance costs account for an extremely large proportion of the operating costs of machinery. In addition, machine breakdowns and consequent downtime can severely affect the productivity of factories and the safety of products. It is therefore becoming increasingly important for industries to monitor their equipment systematically in order to reduce the number of breakdowns and to avoid unnecessary costs and delays caused by repair. The rolling element bearing is an extremely widespread component in industrial rotating machinery and a large number of problems arise from faulty bearings. Therefore, proper monitoring of bearing condition is highly cost-effective in reducing operating cost. The advanced signal processing techniques used here are bispectral-based and wavelet-based analyses. The bispectral-based procedures examined are the bis-pectrum, the bicoherence, the bispectrum diagonal slice, the bicoherence diagonal slice, the summed bispectrum and the summed bicoherence. The wavelet-based procedure uses the Morlet wavelet. These methods greatly enhance the ability of an automated diagnostic process by linking the increased capability for signal analysis to the predictive capability of artificial neural networks. The bearing monitoring scheme based on bispectral analysis is shown to provide greater insight into the structure of bearing vibration signals and to offer more diagnostic information than conventional power spectral analysis. The wavelet analysis provides a multi-resolution, time-frequency approach to extract information from the bearing vibration signatures. In order to effectively interpret the wavelet map, the time-frequency domain is used instead of the time-scale domain by plotting the associated time trace and power spectrum.