Date of Award

2012

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Department Not Listed

First Advisor

Kolla, Sri R

Abstract

Fieldbus networks are all-digital, two-way, multi-drop communication systems that are used to connect field devices such as sensors and actuators, and controllers. These fieldbus network systems are also called networked control systems (NCS). Although, there are different varieties of fieldbus networks such as Foundation Field Bus, DeviceNet, and Profibus available in the automation industries, Controller Area Network (CAN) is more widely accepted in automotive applications. The growing popularity of, and demand for, fieldbus networks can be attributed to several advantages they have, such as: reduction in capital costs, interoperability, and greater system functionality. However, as the complexity of modern fieldbus systems continue to increase, the concern on performance, reliability, and security also increases. To better reflect on this concern, the fieldbus based control systems should be extensively studied using simulations before implementing them in hardware. Network induced delays that may result from the bus arbitration schemes of the messages is an issue that needs investigation for these fieldbus networks. The impact of these delays on control system performance measures such as peak overshoot and settling time needs investigation. The purpose of this research was to study the causes of fieldbus network induced delays and their effects on control systems. The existence and causes for network induced delays were studied by other researchers. Previous delay analyses used analytical and stochastic methods to establish relationships for delays. This dissertation, however, uses statistical analysis methods to study the effect of various CAN parameters on network delays. The data for the statistical analysis was obtained from simulations. Though the literature indicates use of general purpose simulation tools such as OPNET, OMNeT++, and Network II, there exist simulation tools that are designed specifically to address a particular type of fieldbus such as CAN. The research in this dissertation uses such a tool called CANoe for simulating an automobile system. The impact of these delays on control system performance was studied by other research on Proportional Integral (PI) controllers. This dissertation extends these studies to Proportional Integral and Derivative (PID) controllers. In this dissertation, the causes of network delays and how these delays are affected by CAN parameters such as baud rate, bus load, and message length were investigated using CANoe simulations of an automobile system. The statistical techniques of descriptive statistics, and analysis of variance (ANOVA) were used to analyze data obtained for this part of the study. The findings of the ANOVA analysis revealed that CAN parameters have effect on CAN message delays. The effect of fieldbus network induced delays on control system performance such as stability and step-response for different PI and PID controllers were studied using a DC motor model. The delays considered were sensor-to-controller delay and controller-to-actuator delay. MATLAB/Simulink tools were used to analyze the effects of these delays. From this study, it was observed that fieldbus network induced delays have an effect on control systems stability and performance as described by the system step response. The results of this performance evaluation will be useful to design PID controller gains, and to verify how sensitive the control loops are under various time delays.

Download

Share

COinS