| Time-delay systems can arise due to inherent time-delays in the system or a deliberate introduction of time-delays into the system for control purposes. Such systems frequently occur in engineering and science. Time-delays can cause significant problems (e.g., instability and inaccuracy) and, thus, limit and degrade achievable performance. Time-delay terms lead to an infinite number of roots of the characteristic equation, and make analysis difficult using classical methods, especially, in determining stability and designing stabilizing controllers. Thus, such problems have been addressed mainly by using approximate, numerical, and graphical methods. However, such approaches constitute limitations, for example, on accuracy and robustness.;The objective of this research is to develop an effective approach to analyze and control time-delay systems. Using the Lambert W function, free and forced analytical solutions to delay differential equations are derived. The main advantage of this solution approach lies in the fact that the solution has an analytical form expressed in terms of system parameters and, thus, one can explicitly determine how each parameter affects each eigenvalue and the solution. Also, each eigenvalue in the infinite eigenspectrum is associated individually with a branch of the Lambert W function. Solutions are obtained, for the first time, for systems of delay differential equations using the matrix Lambert W function. The obtained solutions are used to analyze essential system properties, such as stability, controllability and observability, and to design controllers for stabilizing systems, improving robustness and/or meeting time-domain specifications. Then, these methods are applied to biological systems to analyze the immune system via eigenvalue sensitivity analysis, to automotive powertrain systems to design feedback control with observers for improvements in fuel economy and emissions, and to manufacturing processes to improve productivity via stability analysis. The newly developed approach based on the matrix Lambert W function provides a tool for analysis and control, which is accurate (i.e., no approximation of time-delay terms), robust (i.e., no prediction of responses from models), and easy to implement (i.e., no need for complex nonlinear controllers). |
