Bilinear Controller Research Articles

Bilinear controller with PID structure

The motivation for, and the implementation of, a bilinear control strategy having a PID structure is described. Improved consistency is automatically achieved by exploiting the fact that local bilinearisation holds for a wider range about a given operating point than a conventional approach in which local linearity is assumed. In terms of reduced activity of the control variable and tighter tracking of the process variable, the bilinear controller provides improved consistency of performance over a wider operating range. The benefits of employing the bilinear approach, which offers continuous scheduling of gains through an assumed bilinear controller model structure, are highlighted when the controller is applied to a continuously operated high-temperature furnace. Performance is compared to that of an optimally tuned industrial PID controller.

Design of an adaptive control of bilinear system with guaranteed stability

The design of an adaptive bilinear control system and its application to a class of combustion process is described. Its closed loop stability is guaranteed and proved with respect to bounded disturbances and modelling errors. The adaptive controller is applied to control a class of combustion process with one input (flow of air) and one output (oxygen content), which is difficult to control due to its nonlinearity. Numerical simulation results show that the bilinear adaptive controller has satisfactory performance.

A METHOD FOR PREVIEW VIBRATION CONTROL OF SYSTEMS HAVING FORCING INPUTS AND RAPIDLY-SWITCHED DAMPERS

In a variety of applications, especially in large scale dynamic systems, the mechanization of different vibration control elements in different locations would be decided by limitations placed on the modal vibration of the system and the inherent dynamic coupling between its modes. Also, the quality of vibration control to the economy of producing the whole system would be another trade-off leading to a mix of passive, active and semi-active vibration control elements in one system. This termactiveis limited to externally powered vibration control inputs and the termsemi-activeis limited to rapidly switched dampers. In this article, an optimal preview control method is developed for application to dynamic systems having active and semi-active vibration control elements mechanized at different locations in one system. The system is then a piecewise (bilinear) controller in which two independent sets of control inputs appear additively and multiplicatively. Calculus of variations along with the Hamiltonian approach are employed for the derivation of this method. In essence, it requires the active elements to be ideal force generators and the switched dampers to have the property of on-line variation of the damping characteristics to pre-determined limits. As the dampers switch during operation the whole system's structure differs, and then values of the active forcing inputs are adapted to match these rapid changes. Strictly speaking, each rapidly switched damper has pre-known upper and lower damping levels and it can take on any in-between value. This in-between value is to be determined by the method as long as the damper tracks a pre-known fully active control demand. In every damping state of each semi-active damper the method provides the optimal matching values of the active forcing inputs. The method is shown to have the feature of solving simple standard matrix equations to obtain closed form solutions. A comprehensive 9-DOF tractor semi-trailer model is used to demonstrate the effectiveness of the method. Time domain predictions are made to compare performance of ride and tyre-to-road contact in the model for the presented method with those of some other active and semi-active suspension designs.

Adaptive control of bilinear systems with bounded disturbances, and its application

In this paper, an adaptive controller is put forward for bilinear systems with bounded disturbances. Its closed-loop stability is proved and the upper bound of the control error is given. The bilinear adaptive controller has been used in the control of a pH process, which is difficult to control due to its highly nonlinear nature. Experiments on a pilot glutamic acid crystalization plant show that the bilinear adaptive pH controller has satisfactory static and dynamic performances, and can adapt to changes in process conditions.

On Bilinear Control Methodology

Aspects of bilinear system (BLS) control methodology are discussed. New results on global controllability and asymptotic stabilization of time-invariant BLS are discussed. A bilinear self-tuning controller is applied to the parametric control of electric power networks in the form of Flexible AC Transmission Systems (FACTS). Bilinear control structure is demonstrated in a hierarchical multimodel controller.

A bilinear self-tuning controller for multimachine transient stability

Bilinear time-series model-based self-tuning control is proposed for a multimachine power system, controlled by a single variable series capacitor. When the faults of concern are large and DC-stabilizing, it is proposed that nonlinear model-based controllers can enhance the region of stability of the power system, and return the states to their stable equilibrium. A simple predictive nonlinear self-tuning controller using local measurements of relative rotor angles is examined. It is shown to perform well in stabilizing different faults on a 45-bus, 17-generator low order model with the dynamic characteristics of the Western System Coordinating Council (WSCC) system. >

Self-tuning control of non-linear plant - a bilinear approach

This paper describes a collaborative project involving British Gas plc, Midlands Research Station, Solihull, the Control Theory & Applications Centre of Coventry University, the Energy Technology Support Unit, Harwell and Eurotherm Controls Ltd, Worthing. The project is concerned with the design, development and real-time evaluation of a new bilinear self-tuning controller, which is primarily being aimed at improving temperature control of high temperature gas fired industrial heating plant. The largest energy savings are anticipated on plant experiencing variable operating conditions which may occur under continuous operation or during batch process changes. To provide for a realistic comparison, the performance of the bilinear self-tuner is compared to that achieved using a standard industrial auto-tuned PID controller when applied to an industrial test furnance.

Efficient Usage of Energy in High Temperature Heating Plant

This paper describes a collaborative project involving the Control Theory & Applications Centre of Coventry University. British Gas pic. Midlands Research Station, Solihull, the Energy Technology Support Unit, Harwell and Eurotherm International Ltd, Worthing. The project is concerned with the design, development and real-time evaluation of a new bilinear self-tuning controller, which is primarily being aimed at improving temperature control of high temperature gas fired industrial heating plant. The largest energy savings are anticipated on plant experiencing variable operating conditions which may occur under continuous operation or during batch process changes. To provide for a realistic comparison, the performance of the bilinear self-tuner is compared to that achieved using a standard industrial auto-tuned PID controller when applied to an industrial test furnace

An Adaptive Bilinear Controller for Closed Circuit Anesthesia

Administration of gaseous anesthesia using a closed circuit is cheaper and ecologically safer than the traditional seni-closed circuit. However, gas concentrations are more difficult to control. This paper proposes an adaptive bilinear controller for the automatic control of a closed circuit anesthesia system. Alveolar anesthetic gas concentration, alveolar oxygen concentration and circuit volume are controlled using a liquid anesthetic (Halothane), oxygen and nitrous oxide. Oxygen consumption, nitrous oxide uptake and Halothane uptake are estimated by a leastsquares parameter estimator. Derivation of the model used, results from simulations and results from preliminary tests on dogs are presented.