Roger Fales

Robust control design for a wheel loader using H∞ and feedback linearization based methods

The heavy equipment industry is building more and more equipment with electro-hydraulic control systems. The existing industry practices for the design of control systems in construction machines primarily rely on classical designs coupled with ad-hoc synthesis procedures. Such practices produce desirable results, but lack a systematic procedure to account for invariably present plant uncertainties in the design process as well as coupled dynamics of the multi-input multi-output (MIMO) configuration. In this paper, two H(infinity) based robust control designs are presented for an automatic bucket leveling mechanism of a wheel loader. In one case, the controller is designed for the base plant model. In another case, the controller is designed for the plant with a feedback linearization control law applied yielding improved stability robustness. A MIMO nonlinear model for an electro-hydraulically actuated wheel loader linkage is considered. The robustness of the controller designs are validated by using analysis and by simulation using a complete nonlinear model of the wheel loader linkage and hydraulic system.

Modeling and Control of a Wheel Loader With a Human-in-the-Loop Assessment Using Virtual Reality

This paper presents dynamic modeling, controller design, and virtual reality (VR)-based human-in-the-loop real-time simulation for a wheel loader control system. In particular, a loader with electrohydraulic actuation is considered. A detailed nonlinear dynamic model is developed for the hydraulic system and the loader linkage. The hydraulic model includes a load sensing pump, valves, and cylinders. The linkage model represents a two degree of freedom loader with lift and tilt functions. A linear quadratic Gaussian based robust controller is designed for automatic bucket leveling to assist the operator by keeping the angle of the bucket leveled while the boom is in motion. The closed-loop control system design is tested with a nonlinear model in a real-time VR simulation. In the VR simulation, the operator interacts with the model using a joystick input. The loader linkage geometry is displayed to the operator in real time using a VR display. The controller performance was assessed in the VR environment. As expected, the controller was found to provide a significant improvement in the accuracy of the bucket leveling, particularly in the case of a novice operator controlling the linkage motion. While prototypes cannot be eliminated, the VR simulation combined with realistic physics and control dynamics provided a useful tool for evaluating hydraulic systems and controls with less reliance on prototype machines.

Robust control design for a wheel loader using mixed sensitivity h-infinity and feedback linearization based methods

The existing industry practices for the design of control systems in construction machines primarily rely on classical designs coupled with ad-hoc synthesis procedures. Such practices lack a systematic procedure to account for invariably present plant uncertainties in the design process as well as coupled dynamics of the multi-input multi-output (MIMO) configuration. In this paper, an H/sub /spl infin// based robust control design combined with feedback linearization is presented for an automatic bucket leveling mechanism of a wheel loader. With the feedback linearization control law applied, stability robustness is improved. A MIMO nonlinear model for an electro-hydraulically actuated wheel loader is considered. The robustness of the controller designs are validated by using analysis and by simulation using a complete nonlinear model of the wheel loader system.

Design and Analysis of a Two-Stage Poppet Valve for Flow Control

Abstract This paper explores dynamic modelling and design of a typical two stage metering poppet valve system. In particular, nonlinear and linear models of a spring force feedback configuration are developed and parameters tuned through the use of root locus techniques. Typical steady state conditions as well as extreme high and low pressure drops are simulated in attempts to uncover instabilities and other possible undesirable performance characteristics of the valve. Finally the nonlinear model is used to produce Bode magnitude plots at various pressure drops in order to estimate the system bandwidth. Results indicate that increasing the size of the orifice at the inlet of the pilot stage of the valve increases performance in terms of rise time at the cost of a more oscillatory response. High pressure differences between the inlet and outlet of the valve were found to cause performance to increase significantly as well as move poles into a region indicated less damping. A scheme for controlling the inlet orifice area to the pilot stage is presented and shown to improve performance capability (bandwidth) of the valve while maintain a damped response.

Design and analysis of a two-stage poppet valve for flow control

This paper explores dynamic modeling and design of a typical two stage metering poppet valve system. In particular, nonlinear and linear models of a forced feedback configuration are developed and parameters tuned through the use of root locus techniques. Typical steady state conditions as well as extreme high and low pressure drops are simulated in attempts to uncover instabilities and other possible undesirable performance characteristics of the valve. Finally the nonlinear model is used to produce Bode plots at various pressure drops in order to estimate the system bandwidth

Uncertainty modeling and predicting the probability of stability and performance in the manufacture of dynamic systems.

In this work, a method for determining the reliability of dynamic systems is discussed. Using statistical information on system parameters, the goal is to determine the probability of a dynamic system achieving or not achieving frequency domain performance specifications such as low frequency tracking error, and bandwidth. An example system is considered with closed loop control. A performance specification is given and converted into a performance weight transfer function. The example system is found to have a 20% chance of not achieving the given performance specification. An example of a realistic higher order system model of an electro hydraulic valve with spring feedback and position measurement feedback is also considered. The spring rate and viscous friction are considered as random variables with normal distributions. It was found that nearly 6% of valve systems would not achieve the given frequency domain performance requirement. Uncertainty modeling is also considered. An uncertainty model for the hydraulic valve systems is presented with the same uncertain parameters as in the previous example. However, the uncertainty model was designed such that only 95% of plants would be covered by the uncertainty model. This uncertainty model was applied to the valve control system example in a robust performance test.

Modern Control Design for a Variable Displacement Hydraulic Pump

In this work robust control methods are used to design and analyze control systems for a variable-displacement hydraulic pump. More accurate uncertainty descriptions are derived by using an uncertainty model with some structure as opposed to an unstructured uncertainty model. The system studied includes one variable-displacement swash-plate hydraulic pump with a constant drive speed model. The input to the system is the current actuating the control valve position, while the system output is the discharge pressure of the pump. A PD controller and an H-infinity two degrees-of-freedom controller were designed. Frequency domain analysis compares the robustness of the two designs. Time domain results show that performance of the PD controlled system is better than the two degrees-of-freedom controlled system. Time domain simulations also show improved robustness to parametric variation from the modern control method.

Experimental Modelling and Control of a Servo-Hydraulic Force Control System

Abstract The objective of this work is to model a hydraulic force control servo system and then improve upon the performance of the system through feedback control design. The hydraulic system is first constructed and tested. Experimental data based linear models of the system are found through input-output measurements. The models contain a right-half- plane zero; therefore, a bandwidth limitation is placed on the control design (i.e., the bandwidth frequency of the control system is limited). Three types of controllers (P, PID and Hm) are designed specifically for the linear models. The closed-loop time domain and frequency domain performance of each control system is found and compared for the models and system. Uncertainties and performance weights are finally used in finding the nominal/robust stability and performance.

Modeling and Feedback Control of Inspired Oxygen for Premature Infants

Premature infants are commonly treated for respiratory problems due to their underdeveloped lungs. Due to Respiratory Distress Syndrome, the infant requires mechanical ventilation or increased inspired oxygen. If the blood oxygen saturation is kept a too high of a level, the infant is at risk for retinopathy of prematurity. A safe level for the infant’s blood oxygen saturation is between 85–92%. An automatic control system would aid nurses in care of premature infants. Since each infant is different, the control system must be robust enough to achieve adequate control of the percentage of oxygen in inspired air administered to the patient. Clinical data is acquired from patient bedside monitors. A parameter estimating extended Kalman filter assuming a first order model is applied to the data to calculate a range of system gains and time constants. An error model is then created using the resulting ranges of parameters. Performance specifications are defined and a μ-synthesis controller is developed to automatically control the oxygen percentage of inspired air. The control system is analyzed using H∞ methods to determine whether robust stability and robust performance are achieved in the presence of system uncertainty described by the error model.Copyright

Stability and Performance Analysis of a Metering Poppet Valve

Poppet type metering vales have many benefits including low leakage and an economical design. These benefits make the poppet valve an appealing alternative to spool valves in a valve stack. The fact that the metering element is not hydrostatically balanced as in a spool valve leads to control design challenges. In this work, a model of an electro hydraulic metering poppet valve is considered. Due to design compromises, the response of production metering poppet valves tends to be too slow to maintain a desired flow rate when there are fast upstream pressure variations. Re-designing to speed up the response of the valve may lead to stability issues which can be traced to plant uncertainty. Frequency response analysis of the valve model shows that the model varies greatly depending on the operating point chosen for the linearization. The analysis presented will help define the problem of designing hardware and control systems for higher performance but still reliable metering poppet valves.Copyright