Plant Model Uncertainties Research Articles

Performance Analysis of Learning-Based Disturbance Observer for Pulsed Superconducting Cavity Field Control

The use of Disturbance Observer-based (DOB) control is widespread in stabilizing the electromagnetic field in superconducting radio frequency cavities to facilitate beam acceleration in particle accelerators. Repetitive disturbances such as beam loading and Lorentz force cavity detuning are compensated by DOB control, and their suppression is enhanced through the incorporation of a learning scheme into the conventional disturbance observer. This paper evaluated the performance of a learning-based disturbance observer for compensating beam loading and cavity detuning in pulsed superconducting radio frequency cavities and proposes modifications for better field stability. A superconducting cavity baseband model for π-mode was simulated in Matlab/Simulink with a trapezoidal beam pulse as the input disturbance and different cavity detuning values to analyze the controllers’ performance. The simulations were conducted for multiple observer filter bandwidths to evaluate the performance of the learning-based disturbance observer under plant model uncertainties and different detuning values. The results demonstrate that the learning-based disturbance observer yields faster convergence to the reference input and lower tracking errors during the flat top of pulse voltage in comparison to conventional disturbance observer control.

Robust Mixed H2-H∞ Control Synthesis for Dual-Stage Hard Disk Drives Using Convex Optimization

This paper presents a convex optimization-based control synthesis method for a dual-stage hard disk drive system that is robust to model uncertainty. A mixed H2-H∞ approach is used that minimizes the steady-state variance of the piezoelectric (PZT) micro-actuator stroke through the H2 norm, while constraining the closed-loop sensitivity through the H∞ norm. Robustness of the closed-loop sensitivity to model uncertainty is ensured through the use of a time-domain integral quadratic constraint (IQC) quantification of the voice coil motor (VCM) plant model uncertainty and a circle criterion-based matrix inequality condition. The synthesis method is derived in discrete time and involves iteratively solving semidefinite programs with linear matrix inequality (LMI) constraints. Numerical simulations demonstrate the benefits of the proposed synthesis method over other approaches in the literature that do not explicitly consider robustness to model uncertainty or the minimization of PZT micro-actuator stroke.

Hybrid Fault Prognostics for Nuclear Applications: Addressing Rotating Plant Model Uncertainty

Nuclear plant operators are required to understand the uncertainties associated with the deployment of prognostics toolsin order to justify their inclusion in operational decision making processes and satisfy regulatory requirements. Operationaluncertainty can cause underlying prognostics models to underperform on assets that are subject to evolving impactsof age, manufacturing tolerances, operating conditions, and operating environment effects, of which may be capturedthrough a condition monitoring (CM) system that itself may be degraded. Sources of uncertainty in the data acquisitionpipeline can impact the health of CM data used to estimate the remaining useful life (RUL) of assets. These uncertaintiescan disguise or misrepresent developing faults, where (for example) the fault identification is not achieved until it hasprogressed to an unmanageable state. This leaves little flexibility for the operator’s maintenance decisions and generallyundermines model confidence.
 One method to quantify and account for operational uncertainty is calibrated hybrid models, employing physics, knowledgeor data driven methods to improve model accuracy and robustness. Hybrid models allow known physical relations tooffset full reliance on potentially untrustworthy data, whilst reducing the need for an abundance of representative historicaldata to reliably identify the monitored asset’s underlying behavioural trends. Calibration of the model then ensuresthe model is updated and representative of the real monitored asset by accounting for differences between the physics orknowledge model and CM data. 
 In this paper, an open-source bearing knowledge informed machine learning (ML) model and CM datasets are utilizedin an illustrative bearing prognostic application. The uncertainty incurred by the decisions made at key stages in thedevelopment of the model’s data acquisition and processing pipeline are assessed and demonstrated by the resultant impacton RUL prediction performance. It was shown that design decisions could result in multiple valid pipeline designswhich generated different predicted RUL trajectories, increasing the uncertainty in the model output.

A Building Automation and Control micro-service architecture using Physics Inspired Neural Networks

In this work, we present a micro-service architecture which defines a Digital Twin (DT) framework for adaptive building automation and control. The DT framework primarily involves the orchestration of several containerized micro-services, promoting the scalability and deployability of the proposed framework within the industrial context. In the proposed framework, containerized microservices facilitate: (i) model-based control strategies; (ii) data-driven learning; (iii) data management; (iv) the inclusion of an internal High-Fidelity Simulator (HFS) to enable bootstrapped learning; and (v) a User Interface/User Experience (UI/UE) micro-service orchestrator. To validate the usefulness of the proposed framework, we implement a Physics Inspired Neural Network (PINN) to adapt the model-based control strategies for plant-model uncertainty and utilize bootstrap sampling against an internal HFS.

Handling large model uncertainty in adaptive feedback noise attenuation by overparametrization

Adaptive feedback noise attenuation is a very efficient way of strongly attenuating multiple tonal and narrow band disturbances with unknown and time varying characteristics. These adaptive schemes implement the internal model principle (IMP) for canceling disturbances combined with the Youla Kucera (YK) parametrization which allows to directly tune the disturbance compensation filter without explicit identification of the disturbance model. Efficient use of these schemes requires a good knowledge of the compensatory path model, which can be obtained by experimental system identification. However, there are potential applications where the characteristics of the compensatory path may change significantly during operation and this may lead to the instability of the system. The paper addresses the problem of handling large plant model uncertainties by overparametrization of the adaptive disturbance compensation filter. A methodology for designing adaptive feedback noise cancelers in the presence of large model uncertainties is proposed. In addition to the overparametrization, a specific design of the linear feedback controller has to be done in order to satisfy a frequency condition in the range of variations of the frequencies characteristics of the compensatory path model. Experimental validation of the design is done on a relevant active noise control bench test.

Adaptive rejection of narrow-band disturbances in the presence of plant uncertainties—A dual Youla–Kucera approach

The stability of adaptive disturbance rejection schemes using Youla–Kucera (YK) parametrization and the internal model principle (IMP) in the presence of plant model uncertainties is investigated. The problem is approached by using the dual Youla–Kucera parametrization for the description of the plant model uncertainties. The known disturbance case is discussed first, emphasizing the need of over parametrization of the Youla–Kucera filter used for control in order to both solving the IMP and the stability problems. Then this solution is extended for the case of unknown disturbances leading to the use of a parameter adaptation algorithm with projection. A stability analysis of the adaptive scheme is provided. Simulation results on relevant examples and experimental evaluation on an active noise attenuation system illustrate the possibilities of this approach for handling significant plant-model mismatch.

Design and industrial implementation of fuzzy logic control of level in soda production

The linear control techniques often fail to satisfy the high standards in the control of the contemporary complex plants accounting for the emerging importance of the plant nonlinearity, model uncertainty, variables coupling and disturbances from the industrial environment. The fuzzy logic controller (FLC) successfully ensures control system stability, robustness and desired performance in the whole range of plant operation. Despite the rapid theoretical development the reported FLC industrial applications are still only few. The aim of the present investigation is to design and implement a model-free FLC for the liquid level in a carbonisation column in soda ash production using a low-cost general purpose industrial programmable logic controller (PLC). The novelty is the design of a single-input-single-output FLC, the FLC transformation to enable its programming in a PLC and implementation of the PLC-FLC for the real time control of the level in a carbonisation column in “Solvay Sodi” SA – Devnya, Bulgaria. The conducted real time experiments on the industrial plant show that the FLC system outperforms the existing model-based fuzzy parallel distributed control system in an increased dynamic accuracy and a reduced settling time.

Síntesis secuencial de controladores PID basada en el método LQR

The linear quadratic regulator (LQR) method is generalized to allow synthesis of PID controllers in MIMO processes. The proposed method is sequentially applied to produce proportional, integral and derivative actions. Three major usages are conceived for the proposed methodology: (i) <<de novo>> design of PID controllers, (ii) addition of derivative action to existing PI controllers and (iii) diagonalization of PID gain matrices. The two last procedures can be applied to controllers designed with different methodologies. The developed method was applied to the <<de novo>> design of a centralized PID controller for a three input-three output distillation column as well as the addition of derivative action to existing both centralized and multiloop PI controllers for a nonlinear CSTR. Proposed LQR method allowed the synthesis of centralized and multiloop PID controllers with better characteristics for set-point tracking, disturbance rejection, limited use of control signal and insensitivity to plant model uncertainty than those reported by other authors.

Signal-to-noise ratio constrained feedback control: Robust stability analysis

The explicit consideration of a communication channel model in a feedback control loop is known to be constrained by a fundamental limitation for stabilizability on the channel signal-to-noise ratio (SNR) when the linear time invariant (LTI) plant model is unstable. The LTI modelling approach for real, usually nonlinear, processes compromises accuracy versus complexity of the resulting model. This in turn introduces a gap between the proposed model and the real process, which is known as the model uncertainty. In this paper we then study SNR limitations by considering the continuous-time scenario and the case of an additive coloured Gaussian noise (ACGN) channel with bandwidth limitation, for which we then quantify the infimal SNR subject to the simultaneous presence of plant, channel and noise model uncertainties. We observe, for the special case of memoryless additive white Gaussian noise (AWGN) channels, that the obtained SNR limitation subject to plant model uncertainty can be redefined as a channel capacity limitation for stabilization.

Performance degradation due to measurement noise in control systems with disturbance observers and saturating actuators

Augmenting feedback control systems with disturbance observer (DOB) is a widely used technique in system design to compensate for the effect of exogenous disturbances as well as plant model uncertainties. In practice, actuator saturation should be taken into account in the design of control systems with DOB. In such cases, we have observed performance degradation due to zero mean measurement noise in the form of tracking loss. This phenomenon has never been reported in DOB literature. This paper reports the phenomenon, analyzes the conditions under which the tracking loss occurs, and also presents design guidelines to avoid the tracking loss. Experimental verification is also provided using a BLDC motor drive testbed.

Parallel distributed compensation for improvement of level control in carbonization column for soda production

The liquid level control is essential in many production installations but the classic approaches often fail to ensure the desired performance. The reasons are the plant nonlinearity, the level oscillations and the plant model uncertainties. The aim of the present investigation is to improve the existing linear control of the level in the carbonization columns for soda ash production by employing fuzzy logic using parallel distributed compensation (PDC). The design of the PDC is based on a nonlinear Takagi–Sugeno–Kang (TSK) plant model which is derived via genetic algorithms optimization and validated using the data from the real time linear level control. The PDC control performs soft blending of the outputs of several parallel local linear controllers each developed for the local linear plant of the TSK model. The fuzzy rules are represented by ordinary logics conditions to enable the PDC programming and use by an industrial programmable logic controller. The PDC increases the dynamic accuracy in the level control and reduces the frequency of the control oscillations compared to the previous linear control thus prolonging the lifetime of the expensive pneumatic actuators used.

Modified Phase-Scheduled-Command FxLMS Algorithm for Active Sound Profiling

Active sound profiling, or active noise equalization strategies have been proposed to achieve spectral shaping of a primary disturbance signal. The control algorithms proposed to achieve such spectral shaping have either suffered from poor robustness to plant modeling uncertainties or required high levels of control effort. To improve the robustness of active sound profiling to uncertainties in the plant model, while avoiding increased control effort, a modified phase-scheduled-command filtered -x least-mean-square algorithm is proposed in this paper. The new algorithm provides improved stability, while requiring the minimum control effort. This improvement is achieved by replacing the plant model with an intelligent adaptive-hysteresis switching mechanism to allow the necessary estimation of the disturbance signal phase. The improved performance and robustness of the proposed algorithm is demonstrated through a series of simulations using measured acoustic responses.

Algebraic robust control of a closed circuit heating-cooling system with a heat exchanger and internal loop delays

This study demonstrates the use of a simple algebraic controller design for a cooling-heating plant with a through-flow air-water heat exchanger that evinces long internal delays with respect to the robustness to plant model uncertainties and variable ambient temperature conditions during the season. The advantage of the proposed design method consists in that the delays are not approximated but fully considered. Moreover, the reduction of sensitivity to model parameters’ variations yields the better applicability regardless modeling errors or environmental fluctuations. The infinite-dimensional mathematical model of the plant has been obtained by using anisochronic modeling principles. The key tool for the design is the ring special of quasipolynomial meromorphic functions (RQM). The Two-Feedback-Controllers (TFC) rather than the simple negative control feedback loop is utilized, which enables to solve the reference tracking and disturbance rejection independently and more efficiently. The eventual controller is then tuned such that robust stability and robust performance requirements are fulfilled. The tuning procedure is supported by a performance optimization idea. Since the originally obtained controller is of the infinite-dimensional nature, a possible way how to substitute it by a simplified finite-dimensional one is proposed for engineering practice. The functionality of both the controllers is compared and verified by simulations as well as by real measurements which prove a very good performance.

Composite Adaptive Internal Model Control and Its Application to Boost Pressure Control of a Turbocharged Gasoline Engine

Internal model control (IMC) explicitly incorporates the plant model and its approximate inverse and offers an intuitive controller structure and calibration procedure. In the presence of plant-model uncertainty, combining the IMC structure with parameter estimation through the certainty equivalence principle leads to adaptive IMC (AIMC), where either the plant model or its inverse is identified. This paper proposes a composite AIMC (CAIMC) that explores the IMC structure and simultaneous plant dynamics and inverse dynamics identification to achieve improved performance of AIMC. A toy plant is used to illustrate the feasibility and potential of CAIMC. The advantages of CAIMC are later demonstrated on the boost-pressure control problem of a turbocharged gasoline engine. The design of the CAIMC assumes that the plant model and its inverse are represented by the first-order linear dynamics. The unmodeled dynamics and uncertainties due to linearization and variations in operating conditions are compensated through adaptation. The resulting CAIMC is first applied to a physics-based high-order and nonlinear proprietary turbocharged gasoline engine model, and then validated on a turbocharged 2-L four-cylinder gasoline engine on a vehicle with vacuum-actuated wastegate. Both the simulation and experimental results show that the CAIMC cannot only effectively compensate for uncertainties but also auto-tune the IMC controller for the best performance.

Decentralized discrete sliding mode controller for TITO processes with time delay with experimental application

Multiple input multiple output systems posses interaction among the variables and hence are difficult to control. In addition, if these systems have slow and non-linear dynamics with time delays; it becomes more difficult to control them under the effect of parametric uncertainty. This paper presents the decentralized discrete time sliding mode controller (DSMC) for the robust tracking of two input two output (TITO) chemical processes with time delay. The ideal decoupler is used to reduce interaction among the system variables by forming decentralized structure consisting decoupled subsystems. Then for each subsystem, DSMC is independently designed and the control signal is applied to the TITO system after passing it through the decoupler. In the design procedure, an optimal sliding surface is chosen as a linear function of the system’s state error and the coefficients of sliding surface are computed by minimizing the quadratic performance index. A delay ahead predictor–corrector is used to handle system’s time delay and plant-model uncertainties. The control law is derived from the discrete time state model and sliding surface. A variable gain switching control law is proposed which results into power rate reaching law with power equal to 1. The methodology integrates ideal decoupler, optimal sliding surface and delay ahead prediction–correction; and therefore results optimal performance of the systems with less interaction. The stability condition is derived using Lyapunov approach. The stability condition results into chatter free convergent quasi sliding mode. Two benchmark simulation examples are used to demonstrate effectiveness of the proposed controller and an experimentation is performed on interacting coupled tank level control system to show its applicability to real life non-linear TITO systems.

Discrete sliding mode control for robust tracking of time-delay systems

This paper presents the discrete time sliding mode controller for the robust tracking of time-delay systems. In this, an optimal sliding surface is chosen as a linear function of the system-state error and the coefficients of sliding surface are computed by minimizing the quadratic performance index. A delay ahead predictor and corrector is used to handle system's time-delay and plant–model uncertainties. The control law is derived from the discrete time-state model and sliding surface with predicted states for general class of delay-time systems. The methodology integrates optimal sliding surface and delay ahead prediction; and therefore results in optimal performance of the systems. The stability condition is derived using the Lyapunov approach. Simulation examples are included to show the usefulness of the proposed controller.

State estimation and synchronization of pendula systems over digital communication channels

The recent results on nonlinear systems synchronization and control under communication constraints are applied to the remote state estimation and synchronization for a class of exogenously excited nonlinear Lurie systems. State estimation of the chain of diffusively coupled pendulums over the digital communication channel with limited capacity is experimentally studied. Advantage of the adaptive coding procedure under the conditions of the plant model uncertainty and irregular disturbances is shown. Quality of the estimation is evaluated by means of the experiments with the multi-pendulum set-up. Experimental study of master-slave synchronization over network (local network, wireless network) for the system with two cart-pendulums is presented.

Lessons learnt from the application of a multi-variable controller for nitrogen removal in the Mekolalde wastewater treatment plant: good simulation practices in control

Although often perceived as tools for use by scientists, mathematical modelling and simulation become indispensable when control engineers have to design controllers for real-life wastewater treatment plants (WWTPs). Nonetheless, the design of effective controllers in the wastewater domain using simulations requires effects, such as the nonlinearity of actuators, the time response of sensors, plant model uncertainties, etc. to have been reproduced beforehand. Otherwise, control solutions verified by simulation can completely underperform under real conditions. This study demonstrates that, when all the above effects are included at the outset, a systematic use of simulations guarantees high quality controllers in a relatively short period of time. The above is exemplified through the Mekolalde WWTP, where a comprehensive simulation study was conducted in order to develop a control product for nitrogen removal. Since its activation in May 2011, the designed controller has been permanently working in the plant which, from this time onwards, has experienced significant improvements in the quality of water discharges combined with a lower utilization of electricity for wastewater treatment.

Robust Stabilizer Design for High-Precision Tracking of a Linear Gantry

In this paper, we develop a robust ℋ∞ stabilizer design method for a parallel structure of internal model-based tracking control for a Voice Coil Motor (VCM) driven servo gantry. The physical model of the VCM gantry is discussed and the system augmentation with the internal-model control structure and the exosystem immersion is generated, where the plant model uncertainties are taken into consideration as well. A mixed sensitivity optimization method is introduced for the synthesis of robust stabilizers. Simulations and real time experiments demonstrate good performance of the proposed method for high precision tracking of the servo gantry system.

Robust common rail pressure control algorithm for light-duty diesel engines

This paper proposed a robust common rail pressure control algorithm for light-duty diesel engines. In this study, rail pressure and the driving current of pressure control valve are selected as a performance variable and a control input, respectively. The proposed algorithm is composed of feedforward and feedback controllers. In order to achieve fast tracking performance, the feedforward control algorithm is derived from the empirical model which represents physical relation between the driving current of the pressure control valve and the common rail pressure. The feedback algorithm is designed using integral control method to compensate the modeling error of the feedforward algorithm. According to the quantitative feedback theory, the feedback algorithm is robustly designed with consideration of the plant model uncertainty which is originated from the metering unit's operating conditions. The proposed common rail pressure control algorithm was validated through various conditions of engine experiments. From the experimental results, it is observed that the algorithm shows a good degree of tracking performance and robustness.