M. Čech

Running discrete Fourier transform and its applications in control loop performance assessment

Control loop performance assessment (CLPA) techniques are crucial for optimizing any plant or machine. They can bring huge energy and material savings and increase product quality. In this paper, the employment of running discrete Fourier transform (RDFT) in CLPA field is discussed. The first part of the paper documents the development of new RDFT function block which is suitable for CLPA. The paper focuses on implementation aspects whose aim is to minimize the number of arithmetic operations and to avoid numerical errors which are cumulated in many algorithms when running over longer time period. Then three RDFT applications are introduced. They are mostly dedicated to CLPA area: The changes in RDFT output help to detect increasing valve stiction or reveal a cause of oscillations in the loop. RDFT can be also used for continuous monitoring of process changes at particular frequencies. The most advanced problem presented is the estimation of special performance indices. More specifically, key samples of sensitivity function are gained and compared to the reference ones. Inspired by the model free design techniques, only a minimum a priori information about the process is assumed. The authors believe that the presented ideas will be suitable for both academic and industrial sphere.

Interval PID Tuning Rules for a Fractional-Order Model Set

Abstract The paper describes new PID tuning rules suitable for both researcher and industrial practice. Compared to other ones, the new method provides an admissible interval of all controller parameters satisfying the required closed loop performance. The novel PID tuning technique consists of an exact identification and design part. The identification is based on the model set combining a priori information about the process with experimental data. More specifically, a class of fractional-order-pole processes is a priori assumed and moments of the impulse response are taken as characteristic numbers. In the design part, a generalized robustness regions method is employed to compute the boundary of the PID parameters region ensuring common frequency domain requirements. The described procedure was partly implemented and packed into the interactive Java applet freely accessible at http://www.pidlab.com .

New tools for teaching vibration damping concepts: ContLab.eu

Abstract In last two decades, virtual laboratories help to teach students and train technicians in a broad range of engineering areas including automation and control. However, there is still a lack of laboratories that present more complex control schemes directly related to industrial problems. In this paper, new virtual laboratories presenting feedback and feedforward vibration damping techniques are described. Thanks to the automatic code generation, the presented algorithms are directly deployable to various real-time platforms suitable for teaching/training or professional industrial applications ranging from industrial PCs towards deep embedded boards like Raspberry Pi. The interactive tools are freely accessible at www.ContLab.eu . The authors believe that those virtual labs may be useful for both academic and industrial sphere.

Generalized robust stability regions for fractional PID controllers

Fractional-order PID controller (FOPID) design and implementation is one of emerging research areas. This paper presents a method for computing generalized robust stability regions in controller parameter plane. The method can cope with arbitrary linear process model of integer or fractional order. It allows to fulfill essential frequency domain design specifications, namely gain and phase margins, closed loop bandwidth, etc. Further, it can operate simultaneously with number of processes hence can work with uncertainty given e.g. by model set or by parameter intervals. Moreover, the regions can be computed even for selected filter in derivative part of the FOPID controller. The method described is partly available in the interactive Java applet freely accessible at www.pidlab.com. The illustrative example demonstrates that FOPID controller can fulfill stricter design specifications compared to traditional PID.

REX — Rapid development tool for automation and robotics

The paper described possibilities of REX control system in rapid development of plants, machines and their control systems. It is shown how to pass almost automatically through all types of model based simulations (MIL, SIL, PIL, HIL) and how to develop in parallel both HMI and virtual model. The whole approach is explained on two case studies - simple educational model and professional industrial manipulator.

Novel tools for model-based control system design based on FMI/FMU standard with application in energetics

The paper presents novel tools for model-based control system design based on FMI/FMU standard (Functional Mock-up Interface / Unit). It is focused on application of FMI standard for easy integration of control system development cycle starting with Model-in-the-Loop (MIL) simulation and finishing with Hardware-in-the-Loop (HIL) simulation. It is shown, how the Functional Mock-up Units (FMU) containing dynamic differential-algebraic equations of various parts of the device (mechanical, electrical, hydraulic, thermal, etc.) can be easily deployed to unified simulation environment where the control system is designed, consequently. The procedure allows to combine inputs from various Modelica-based tools at the process model side, utilizing power of Matlab/Simulink for design, analysis and optimization of control system and perform final test via HIL scenario where both the model and control system are simulated in real-time on separated HW units. The pros and cons of both FMI concepts, i.e. Co-Simulation and Model Exchange are discussed in detail. The whole procedure is demonstrated on a steam turbine example combining component-based and equation based modeling. Both the turbine model and the full control loop are validated in all phases of control system development. It is shown, that monolithic simulation block with proprietary solver reduces computational burden compared to automatic FMU concept.

Innovated PID pulse autotuner for fractional-order model set: The method of moments

Recently, a successful PID pulse autotuner based on model set approach was developed. Despite high reliability, several requirements for significant improvement have appeared after a decade of its industrial usage. Two of them are handled in this paper. Firstly, the set of a priori admissible processes was extended to fractional-order all-pole models to cover wider range of real plants. Secondly, a new parameter affecting the model set span was introduced. It helps markedly the practitioners to reach the proper robustness/performance ratio especially in the case when the initial controller tuning is too conservative. The authors believe that the new theory will be soon transferred to industrial practice.

Advanced input shaping filter 3D virtual laboratory

In this paper, a 3D virtual laboratory presenting all features of advanced input shaping filter is described. By proper choice of filter coefficients, various design requirements can be achieved (damping at given frequencies, bandwidth, robustness to uncertainties, etc.). Hence, one can handle the trade-off between the filter performance and signal delay which is added to the loop. The filter principle can be evaluated on gantry crane 3D model by using virtual laboratory. The control aim is to avoid load oscillations during load transport. The shaping filter works as a band stop filter which attenuates the natural frequency by modifying setpoint changes. The 3D crane model development is based on Java3D package (rendering package) and a VRML (Virtual Reality Modeling Language) Java loader which creates a bridge between the CAD system and the rendering engine. The interactive tool is freely accessible at www.contlab.eu. The authors believe that the virtual lab may be useful for both academic and industrial sphere.

Optimal loop shaping compensators for fractional-order model set

The paper deals with parametrization of shaping compensators (filters) ensuring the Bodes ideal control loop shape for an exactly defined class of process models. The class of essentially monotone fractional-order processes is considered. The fractional integrator is used as the optimal open loop reference model. Several aspects of shaping filter implementation on given frequency band are discussed. In contrast to the other known approaches, the exactly computed filter frequency response is approximated by the integer order zero/pole transfer function. The comparison with other methods shows that the direct approximation leads to the significant reduction of the filter order. Arisen filters are applicable especially for processes with large gain or load variations.

Minimally invasive control loop performance evaluation

Nowadays, companies are facing a strong pressure for industrial plant and machine optimization in order to achieve energy and material savings and increase product quality. Control loop performance assessment techniques create one cornerstone of this challenge. In process control applications, the performance is frequently compared just to the minimum variance controller. It is known that when optimizing process controllers having fixed structure (e.g. PIDs) different concepts must be applied. In authors recent work, the systematic approach for a class of fractional-order processes was presented. The method uses only a limited a priori information about the process. The performance index is defined as a difference between reference and actual sensitivity function at selected frequencies. In this paper, a minimally invasive method for performance index estimation is proposed. It employs running discrete Fourier transform. Next, the paper discusses various practical aspects and verifies the method on real temperature process.