An investigation of brake application delays in Australian train brake systems began with a literature review of pneumatic train braking systems. Data located in the review gave examples of brake application delays of pre 1990 designs from the U.K., India and North America. Information on application delays on later Australian designs was scarce. Reading of literature has shown a difference between the Australian and North American control valves in the way the propagation of the pressure reduction rate in the brake pipe is maintained. Control valves of the North American style allow the brake pipe air to be connected for a short time to a small cavity or quick service volume of each valve. The quick service volume is then released to atmosphere. The action of exhausting a small amount of air from the brake pipe helps to ensure a propagation of an adequate pressure reduction rate as it travels to the next valve. Australian control valves rely on the ratio of the volume of brake pipe between control valves and the size of the quick service volume or ‘bulb’ to ensure the propagation of an adequate pressure reduction as it travels to the next valve. The air in a bulb of an Australian valve is not expelled to atmosphere until a brake release is made. The research explored possible reductions in application delays by utilizing an experimental pipe test rack that included 4 control valves and 120 meters of brake pipe. Experiments with different configurations of exhaust orifices or chokes, valves and branch pipe lengths that supplied the valves gave a record acquired by data acquisition of the timing of each valve and the local pressure drop from a valve or each valve for comparison. Experiments with exhaust chokes that gave a reduction drop rate in the brake pipe that approached the minimum required to operate a control valve resulted in instability of the application operation of the control valve. The quick service volume of different sizes was included in the experiments to give comparisons in the propagation of the pressure reduction toward the end of a long train. Further increases into the size of the bulb of a control valve to enhance the propagation features toward the end of a long train are discussed. The branch pipe with different diameters from 12 mm to 20 mm and lengths from 160 mm to 800 mm when fitted to an adaptor pipe bracket were investigated and results show that larger diameters gave larger gulps in the brake pipe. Other components that were studied included the pipe bracket that is fitted on some control valves. The pipe bracket and isolation cock was found to add 282 mm of additional length to the air path and while not changing the operation of the valve, the results showed a smaller drop in local pressure in the brake pipe to assist the pressure reduction rate than shown in valves without pipe brackets.

This research is focused on developing a control scheme for regulating the pressure in the brake chamber of an air brake system found in most commercial vehicles like trucks, tractor-trailers and buses. Such a control scheme can be used for providing the ground work for future systems such as forward collision avoidance systems, advanced anti-lock brake systems and differential braking systems. The development of this controller involves two tasks. The first task was the development of a control scheme for achieving the desired pressure in the brake chamber. This scheme was based on a mathematical model of the treadle valve of the air brake system. The second task was the implementation of this control scheme on the experimental facility that was set up at Texas A&M University. The results indicate successful control of a desired brake chamber pressure for a demonstrated range of controller gains.

In this dissertation, we focus on development of algorithms for estimating the severity of air leakage and for predicting the out-of-adjustment of pushrod in an air brake system of heavy commercial vehicles. The leakage of air from the brake system causes a reduction in the steady-state pressure in the brake chamber and an increase in the lag of the braking pressure response thereby increasing the stopping distance of the vehicle. Currently a presence of leak in the system is detected for the severities of leak that cause the reservoir pressure to drop below a threshold, such as, the leakage of compressed air due to rupture of the reservoir or of the hoses carrying the compressed air. The leakage of air is also possible due to several other reasons such as, cracks in the hoses, loose couplings between the hoses etc. The severities of leak, corresponding to such situations, do not lead to the reservoir pressure drop below the threshold; therefore, their presence remains undetected. For the detection and estimation of such severities of leak, a diagnostic scheme has been given and is based on a model developed for the mass ow rate of the leakage of air from the air brake system. Out-of-adjustment of the pushrod is the extension of pushrod beyond a prede- ned value and for safety concerns, an extension beyond this value is not desired. Currently no warning system is available for monitoring the out-of-adjustment of pushrod, except, during the safety inspection. Inspection of the pushrod for outof- adjustment is the most labor-intensive and time consuming task during a typical safety inspection procedure. For efficient and continuous monitoring of the pushrod for out-of-adjustment, a diagnostic algorithm for estimating the steady-state pushrod stroke has been developed. The scheme is expected to expedite the inspection process for the out-of-adjustment of pushrod. Experimental data from the air brake test setup at Texas A and M University has been used for corroborating both the models. Also, the problem of parameter estimation of sequential hybrid systems such as the air brake system, has been addressed. The \hybrid" nature of the air brake system stems from the system being in di erent modes corresponding to di erent values of the displacement of the pushrod and is a result of di erent spring compliances associated with the pushrod in di erent ranges of its displacement. The air brake system is \sequential" in the sense that as the pressure increases, the displacement of the pushrod increases and there is a distinct sequence of modes that the system will transition through and upon a reduction in pressure, the sequence of modes is revisited in the reverse order. The mode to mode transition of the air brake system is governed by the parameters, such as, the clearance between the brake pad and the brake drum. The problem of estimation, that has been addressed, is as follows: Suppose the pressure in the air brake system were to be measured and that the motion of the pushrod is not measured. Is it possible to estimate the nal displacement of the pushrod without knowing the parameters, such as the clearance, that govern the system to transition from one mode to another?

This dissertation deals with the development of a model-based diagnostic system for air brake systems that are widely used in commercial vehicles, such as trucks, tractor-trailers, buses, etc. The performance of these brake systems is sensitive to maintenance and hence they require frequent inspections. Current inspection techniques require an inspector to go underneath a vehicle to check the brake system for possible faults, such as leaks, worn brake pads, out-of-adjustment of push rods, etc. Such inspections are time consuming, labor intensive and difficult to perform on vehicles with a low ground clearance. In this context, the development of an onboard/ handheld diagnostic tool for air brakes would be of significant value. Such a tool would automate the brake inspection process, thereby reducing the inspection time and improving the safety of operation of commercial vehicles. In this dissertation, diagnostic schemes are developed to automatically detect two important and prevalent faults that can occur in air brake systems – leaks and out-of-adjustment of push rods. These diagnostic schemes are developed based on a nonlinear model for the pneumatic subsystem of the air brake system that correlates the pressure transients in the brake chamber with the supply pressure to the treadle valve and the displacement of the treadle valve plunger. These diagnostic schemes have been corroborated with data obtained from the experimental facility at Texas A&M University and the results are presented. The response of the pneumatic subsystem of the air brake system is such that it can be classified as what is known as a “Sequential Hybrid System”. In this dissertation, the term “hybrid systems” is used to denote those systems whose mathematical representation involves a finite set of governing ordinary differential equations corresponding to a finite set of modes of operation. The problem of estimating the push rod stroke is posed as a parameter estimation problem and a transition detection problem involving the hybrid model of the pneumatic subsystem of the air brake system. Also, parameter estimation schemes for a class of sequential hybrid systems are developed. The efficacy of these schemes is illustrated with some examples.