This dissertation presents a new type of fuel control algorithm for gaseous fuelled vehicles. Gaseous fuels such as hydrogen and natural gas have been shown to be less polluting than liquid fuels such as gasoline, both at the tailpipe and on a total cycle basis. Unfortunately, it can be expensive to convert vehicles to gaseous fuels, partially due to small production runs for these vehicles. One of major development costs for a new vehicle is the development and calibration of the fuel controller. The research presented here includes a fuel controller which does not require an expensive calibration phase.<p>The controller is based upon a two-part model, separating steady state and dynamic effects. This model is then used to estimate the optimum fuelling for the measured operating condition. The steady state model is calculated using an artificial neural network with an online learning scheme, allowing the model to continually update to improve the controller's performance. This is important during both the initial learning of the characteristics of a new engine, as well as tracking changes due to wear or damage.<p>The dynamic model of the system is concerned with the significant transport delay between the time the fuel is injected and when the exhaust gas oxygen sensor makes the reading. One significant result of this research is the realization that a previous commonly used model for this delay has become significantly less accurate due to the shift from carburettors or central point injection to port injection.<p>In addition to a description of the control scheme used, this dissertation includes a new method of algebraically inverting a neural network, avoiding computationally expensive iterative methods of optimizing the model. This can greatly speed up the control loop (or allow for less expensive, slower hardware).<p>An important feature of a fuel control scheme is that it produces a small, stable limit cycle between rich and lean fuel-air mixtures. This dissertation expands the currently available models for the limit cycle characteristics of a system with a linear controller as well as developing a similar model for the neural network controller by linearizing the learning scheme.<p>One of the most important aspects of this research is an experimental test, in which the controller was installed on a truck fuelled by natural gas. The tailpipe emissions of the truck with the new controller showed better results than the OEM controller on both carbon monoxide and nitrogen oxides, and the controller required no calibration and very little information about the properties of the engine.<p>The significant original contributions resulting from this research include:<br> -collection and summarization of previous work,<br> -development of a method of automatically determining the pure time delay between the fuel injection event and the feedback measurement,<br> -development of a more accurate model for the variability of the transport delay in modern port injection engines,<br> -developing a fuel-air controller requiring minimal knowledge of the engine's parameters,<br> -development of a method of algebraically inverting a neural network which is much faster than previous iterative methods,<br> -demonstrating how to initialize the neural model by taking advantage of some important characteristics of the system,<br> -expansion of the models available for the limit cycle produced by a system with a binary sensor and delay to include integral controllers with asymmetrical gains,<br> -development of a limit cycle model for the new neural controller, and<br> -experimental verification of the controller's tailpipe emissions performance, which compares favourably to the OEM controller.

This dissertation presents a new type of fuel control algorithm for gaseous fuelled vehicles. Gaseous fuels such as hydrogen and natural gas have been shown to be less polluting than liquid fuels such as gasoline, both at the tailpipe and on a total cycle basis. Unfortunately, it can be expensive to convert vehicles to gaseous fuels, partially due to small production runs for these vehicles. One of major development costs for a new vehicle is the development and calibration of the fuel controller. The research presented here includes a fuel controller which does not require an expensive calibration phase.<p>The controller is based upon a two-part model, separating steady state and dynamic effects. This model is then used to estimate the optimum fuelling for the measured operating condition. The steady state model is calculated using an artificial neural network with an online learning scheme, allowing the model to continually update to improve the controller's performance. This is important during both the initial learning of the characteristics of a new engine, as well as tracking changes due to wear or damage.<p>The dynamic model of the system is concerned with the significant transport delay between the time the fuel is injected and when the exhaust gas oxygen sensor makes the reading. One significant result of this research is the realization that a previous commonly used model for this delay has become significantly less accurate due to the shift from carburettors or central point injection to port injection.<p>In addition to a description of the control scheme used, this dissertation includes a new method of algebraically inverting a neural network, avoiding computationally expensive iterative methods of optimizing the model. This can greatly speed up the control loop (or allow for less expensive, slower hardware).<p>An important feature of a fuel control scheme is that it produces a small, stable limit cycle between rich and lean fuel-air mixtures. This dissertation expands the currently available models for the limit cycle characteristics of a system with a linear controller as well as developing a similar model for the neural network controller by linearizing the learning scheme.<p>One of the most important aspects of this research is an experimental test, in which the controller was installed on a truck fuelled by natural gas. The tailpipe emissions of the truck with the new controller showed better results than the OEM controller on both carbon monoxide and nitrogen oxides, and the controller required no calibration and very little information about the properties of the engine.<p>The significant original contributions resulting from this research include:<br> -collection and summarization of previous work,<br> -development of a method of automatically determining the pure time delay between the fuel injection event and the feedback measurement,<br> -development of a more accurate model for the variability of the transport delay in modern port injection engines,<br> -developing a fuel-air controller requiring minimal knowledge of the engine's parameters,<br> -development of a method of algebraically inverting a neural network which is much faster than previous iterative methods,<br> -demonstrating how to initialize the neural model by taking advantage of some important characteristics of the system,<br> -expansion of the models available for the limit cycle produced by a system with a binary sensor and delay to include integral controllers with asymmetrical gains,<br> -development of a limit cycle model for the new neural controller, and<br> -experimental verification of the controller's tailpipe emissions performance, which compares favourably to the OEM controller.

Thesis (Ph. D.)--University of Hong Kong, 2002. / Includes bibliographical references (leaves 379-401).

Thesis (Ph. D.)--University of Hong Kong, 2002. / Includes bibliographical references (leaves 354-392).

published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

The objective of this thesis is to study a cycle-to-cycle control approach called Terminal Iterative Learning Control (TILC) and apply it to the process of plastic sheet heating in a thermoforming oven. Until now, adjustments to the oven heater temperature setpoints have been made manually by a human operator following a trial and error approach. This approach causes financial losses, because plastic sheets are wasted during the period of time when the adjustments are made at the beginning of a production run. Worse, the heater setpoints are subject to modification because of variation in the ambient temperature, which has an important impact on the sheet reheat process. / The TILC approach is analyzed by studying the closed-loop system in the discrete cycle domain through the use of the z-transform. The system, which has dynamic behaviour in the time domain, becomes a static linear mapping in the cycle domain. One can then apply on this equivalent system a traditional control approach, while considering that the system output is sampled once at the end of the cycle. On the other hand, from the standpoint of the real system, this control approach can be viewed as cycle-to-cycle control. / The stability and rate of convergence of the TILC algorithm can be analyzed through the location of the closed-loop system poles in the cycle domain. This analysis is relatively easy for a first-order TILC but becomes more complex for a higher-order TILC algorithm. The singular value decomposition (SVD) is used to simplify the convergence analysis while decoupling the system in the cycle domain. The SVD technique can be used to facilitate the design of higher-order TILC algorithms. / Internal Model Control (IMC) is another approach that can make the ILC design easier, because there is only one parameter per filter to adjust. The IMC technique has an interesting feature. In the case where the system is nominal, the closed-loop transfer function of the system is the same as the IMC filter's transfer function. Therefore, the adjustment of the filter parameter allows the designer to select the desired system response. / For industrial processes such as thermoforming ovens, it is important that the systems controlled by TILC algorithms are stable and have good performance. For thermoforming ovens, the terminal sheet temperature response must not be too oscillatory from cycle to cycle, since this may lead to high heater temperature setpoints. In the most serious case, high heater temperatures can cause the sheet to melt and spill on the heating elements at the bottom of the oven. / The performance aspect must not be neglected, since it is important to minimize the number of wasted plastic sheets, particularly at process startup. To avoid such waste of time and material, it is necessary that the TILC algorithm converge as quickly as possible. However, the robustness and performance objectives are conflicting and an acceptable compromise must be achieved. The control engineer must define specifications to describe these two constraints. Tools such as the Hinfinity Mixed-Sensitivity Analysis and mu-Analysis can be used to check the compliance of a given TILC algorithm with the robustness and performance specifications defined before the analysis. One can therefore compare various TILC algorithms quantitatively, through a computed measure obtained with one of the two approaches. These same tools can be used for the design of TILC algorithms, using weighting functions representing the specifications. / Simulation and experimental results obtained on industrial thermoforming machines show the effectiveness of the various approaches in this thesis. Many examples are also presented throughout the chapters.

The objectives of this project are (1) to reduce the excursions between real heater temperatures and the desired values, and (2) to realize cyclic production of plastic sheets on the AAA thermoforming machine. / At first, present relevant knowledge and modeling of the AAA machine are covered. A programmable pre-processing module is inserted before the heaters to prevent the input commands from delaying in heater response. The results prove that the excursions can be theoretically reduced to zero. / Based on the Terminal Iterative Learning Control (TILC) algorithm, a hybrid dual-mode cascade-loop system is designed: the inner loop monitors the real-time temperatures for PID control in Mode 0, while the outer loop reads temperatures and commands once per cycle to decide the necessity of switching mode. A more realistic version with flexible cycle length and instant response to operator's commands is also designed to simulate the real operating circumstance.

No description available.

No description available.

Thesis (PhD) -- University of South Australia, 1999