Estimation of hybrid systems using discrete sensors

International audience

This paper deals with the formal verification of reconfigurable discrete event systems. We use the formalism called reconfigurable timed net condition/event systems (R-TNCESs) which is a Petri net pattern that deals with reconfiguration properties. Systems can experience malfunctioning due to hardware failures or software errors. Model-based diagnosis algorithms are widely used in academia and industry to detect faulty components and ensure systems safety. The application of these methods on reconfigurable systems is impossible due to their special behavior. In this paper, we propose accomplishing techniques of backward reachability to make reconfigurable systems model-based diagnosis possible using R-TNCESs. The flexibility among reconfigurable systems allows them to challenge recent requirements of markets. However, such properties and complicated behavior make their verification task being complex and sometimes impossible. We deal with the previous problem by proposing a new methodology based on backward reachability of RDECSs using (R-TNCESs) formalism including improvement methods. The proposed methodology serves to reduce as much as possible redundant computations and gives a package to be used in model-based diagnosis algorithms. A real case study on smart electrical grids is adopted in order to demonstrate the paper’s contributions. Finally, a performance evaluation is achieved using different factors and sizes to study benefits and limits of the proposed methodology among large-scale systems.

This paper deals with reconfigurable discrete event control systems (RDECSs). We model RDECSs using reconfigurable timed net condition/event systems (R-TNCESs) formalism which is an extension from Petri nets to deal with reconfiguration properties. Model-based diagnosis algorithms are widely used in academia and industry to detect faulty components and ensure systems safety. The application of these methods on reconfigurable systems is impossible due to their special behavior. In this paper, we propose accomplishing techniques of backward reachability to make reconfigurable systems model-based diagnosis possible using R-TNCESs. The flexibility among reconfigurable systems like RDECSs allows them to challenge recent requirements of markets. However, such properties and complicated behavior make their verification task being complex and sometimes impossible. We deal with the previous problem by proposing a new methodology based on backward reachability of RDECSs using (R-TNCESs) formalism including improvement methods. The proposed methodology serves to reduce as much as possible redundant computations and gives a package to be used in model-based diagnosis algorithms. The paper’s contribution is applied to a benchmark modular production system. Finally, a performance evaluation is achieved for different sizes of the problem to study benefits and limits of the proposed methodology among large-scale systems.

This paper develops an integrated state estimation and control methodology of networked battery systems that require only terminal voltage and current measurements for reduced cost and complexity. Without direct internal voltage/current measurements and individual control on each battery pack in the system, a networked battery system is inherently uncontrollable and unobservable. By employing bypass power electronics, we show that the networked system can be made controllable and observable by switching control. To coordinate switching control for integrated state of charge (SOC) estimation and control, this paper introduces switching control algorithms that use time division, duty cycle control, and periodic switching. It is shown that under noisy terminal voltage and current measurements, state estimation algorithms are convergent, and the control algorithms can achieve balanced SOC control during operation. The proposed hardware and methodology framework is useful for networked battery systems that can accommodate battery packs of different types, ages, and power/capacity ratings. Case studies using production batteries demonstrate convergence behavior for both state estimation and SOC consensus control under time-varying charging current profiles. The approach will be especially useful for managing battery storage systems to support power grids with renewable energy sources where the battery systems are required to operate continuously.

Sampling density design for particle filters

Particle filtering has been proved effective for the state estimation of nonlinear and non-Gaussian systems. To solve the problems of sample degradation and depletion in standard particle filtering tracking algorithm, a novel immune particle filtering target tracking method is proposed in Passive Millimeter Wave (PMMW) imaging. Particle filtering provides a framework in which the posterior density of PMMW target state is represented by a weighted sample set. By using the artificial immune algorithm combined with the Mean Shift algorithm, the samples are optimized during the evolution process. To achieve robust description of the PMMW targets, both gray and gradient orientation distributions are taken into account. Besides, the observation density is established by computing the Bhattacharyya distance between the distribution of the target model and that of the candidate. Experimental results demonstrate that the proposed algorithm is superior to traditional ones when tracking scale changing PMMW targets.

This paper presents a Hybrid Particle Filter based RNN method for state estimation of non-linear dynamical system with knowledge of its input and output measurements. Particle filters are sequential Monte Carlo methods based on point mass (or particle) representations of probability densities, which is used to train Recurrent Neural Networks for estimation problems. The performance this method is compared with EKF based estimation and RNN based estimation. An Induction motor is considered as typical non-linear system and is implemented in MATLAB environment.

State and parameter estimation of nonlinear stochastic systems by evolution strategies based particle filter

Improved Particle Filter for Target Tracing Application based on ChinaGrid

Control of a mixed-flow drier part 1: Design of the control algorithm

State estimation of river system includes the flow and stage estimation and used in the estimation of the final sensor's velocity which is proportional to the river velocity. The estimation process includes the system model and measurements. The measurements can be carried out using Eulerian sensor or Lagrangian sensor. The Lagrangian sensor is considered in this research since it able to observe the overall system compared to the Eulerian sensor that perform measurement in fixed position. Besides that, the estimation can be carried out using DA method such as Extended Kalman Filter (EKF) and Particle Filter (PF). The Smooth Variable Structure Filter (SVSF) is introduced to give better response in terms of error reduction by having convergence rate and the smoothing boundary layer width. The convergence analysis of SVSF shows the effectiveness of the proposed method in terms of mathematical representation. The simulation results proved the SVSF have better response compared to the forward simulation, EKF and PF. This shown by 25.6% for RMSE, 25.92% for SD, 15.34% for MAE and the computational time is reduced by 95.51 % if compared with PF.

In this paper, a new algorithm for active control design of structures is proposed and investigated. The algorithm preserves the decoupling property of the modal vibration equation and eliminates the spillover problem, which is the main shortcoming in the independent modal space control (IMSC) algorithm. With linear quadratic regulator(LQR) control law, the analytical solution of algebraic Riccati equation and the optimal actuator control force are obtained, and the control design procedure is significantly simplified . A numerical example for the control design of a tall building subjected to wind loads demonstrates the effectiveness of the proposed algorithm in reducing the acceleration and displacement responses of tall buildings under wind actions.

Real-time estimation of the annular pressure profile is crucial for the planning and execution of well control operations, particularly while drilling wells in formations having narrow mud windows. Currently, the industry employs either simplified heuristic techniques, or complicated and often cost prohibitive commercial simulators, which are essentially "black boxes" with no options for further change by the user. Neither of the two approaches is particularly suitable for precise real-time well control analysis or design of robust automatic well control algorithms. Therefore, a simple transient multi-phase simulator amenable to real-time applications, which can represent essential dynamics with minimal cost and computational requirements, is highly desirable. In this paper, we present a novel reduced Drift-Flux approach to model multi-phase well control situations. The model still preserves the transient multi-phase behavior of liquids and gas in the well. The proposed formulation consists of a lumped parameter model of the pressure dynamics, a transport equation for gas bubble migration, and the associated closure relations. Together, they form a system of coupled ordinary and partial differential equations with limited complexity. We then employ an explicit numerical solution algorithm that is fast and thereby reduces computational time and cost, making it suitable for control purposes. The validity of the model is tested against experimental data from a test well with a simulated gas kick event. The model is also compared against a well control scenario generated by a commercial multi-phase flow simulator extensively used in the industry. There is good agreement between the results of the proposed model and the experimental data, on the one hand, and the commercial simulator results, on the other, thereby justifying the applicability of this model to real-world scenarios. The model provides a powerful tool to estimate and manage the surface pressures during a well control event for a given kick size and intensity. It has evident applications in well control analysis, automated Managed Pressure Drilling and well control management. As such, it has the potential to improve well control risk mitigation, enhance rig safety, and reduce non-productive time associated with well control-related trouble events.

During the integration phase of a system development, we are often concerned as to whether the designed control algorithm could be performed on the selected controller in real-time. One of the tools to test and validate the control scheme is the Hardware-in-the-Loop (HiL) simulation technique, which is a part of a model-based design methodology. This approach requires a simulation model of a controlled system running in a real-time loop with an intended controller and a control algorithm, which are objects of interest in this method. To perform the test, the control algorithm must be deployed to the controller such as a PLC. This paper presents a use case of the HiL technique in the design of the Stewart platform control, where the controller is PLCnext from Phoenix Contact. The control algorithm was first verified in the Model-in-the-Loop simulation (MiL) and then generated as a code from the Matlab/Simulink environment and deployed to the PLCnext, which resulted in a smoother transition from the design to the integration and testing phase. The presented method is also applicable to other controllers that support code generation.

System state estimation and prognostics are the key issues in dynamic system monitoring and management. Although the particle filter (PF) has been applied to model the nonlinear degradation feature of the system aging mechanism in several studies, it has two potential problems: the sample degeneracy and the impoverishment. To tackle these problems, a new adaptive particle filter (APF) technique is proposed in this article to enhance the performance of PFs. In the APF, a self-evaluation method is suggested to track the posterior distribution and detect the low-weight particles (sample degeneracy). A new adaptive weight adjustment approach is proposed to adaptively explore the posterior space, process those low-weight particles, and tackle the sample degeneracy. The effectiveness of the proposed APF technique is validated by simulation tests using several model conditions. It is also implemented for battery-health monitoring and prognosis. Test results show that the proposed APF technology can effectively capture and track the system’s dynamic characteristics.