Multi-model Predictive Control Research Articles

Fused multi-model predictive control with adaptive compensation for proton exchange membrane fuel cell air supply system

Regulating the air supply is crucial for high efficiency and reliable operation of proton exchange membrane fuel cell systems (PEMFCs). In this study, a fused multi-model predictive control (FM-MPC) with an adaptive compensation is proposed for the oxygen excess ratio (OER) regulation of the air supply system. The FM-MPC is designed based on the linearized PEMFC model at low and high power phases, with two linear MPCs implemented and combined using adaptive featured weights. An adaptive compensation strategy is created to address the imbalance of the two MPCs and external load disturbances. The stability of the proposed control is analyzed using Lyapunov's second law. Simulation results demonstrate that the proposed method exhibits less overshoot and faster response than conventional MPCs, with the OER total sum-of-squares error (TSSE) reduced by 59.4% and 87.7% for New European Driving Cycle (NEDC) and Urban Dynamometer Driving Schedule (UDDS) conditions, respectively. Finally, a Hardware-In-the-Loop (HIL) experiment verifies the real-time application potential of the proposed controller, with a mean relative error (MRE) of 1.12% between experiment and simulation.

Tracking Control of Wheeled Mobile Robot Based on MMPC

In order to solve the problems of low tracking accuracy and large steady-state error in track tracking of wheeled mobile robots, a MPC control scheme combining feedforward and feedback control is designed on the basis of Traditional Model Predictive Control (TMPC). First, when establishing the WMR kinematic error model, the linearization error is retained, and the control input and incremental constraints are considered. Secondly, a feedforward controller is designed for a given desired trajectory of wheeled mobile robot; At the same time, the multi constraint model predictive control (MMPC) strategy with speed optimization is added to design the feedback controller. Then, the stability of the designed MMPC controller is analyzed based on Lyapunov theory. Finally, MMPC control scheme is compared with TMPC control scheme and PID control scheme. The results show that compared with TMPC control scheme, MMPC control scheme reduces the lateral error by 17.87%, the longitudinal error by 17.12%, and the heading angle error by 9.81%; Compared with the PID control scheme, the lateral error is reduced by 20.82%, the longitudinal error is reduced by 23.26%, and the heading angle error is reduced by 5.87%.

Data-Driven Multimodel Predictive Control for Multirate Sampled-Data Nonlinear Systems

This paper is concerned with the asynchronous control problem of multi-rate sampled-data nonlinear systems. To solve this problem, the data-driven multi-model predictive control (DMMPC) strategy is proposed. First, a multi-model predictive control structure is designed such that each state variable of the multi-rate sampled-data nonlinear system can be controlled synchronously at all sampling instants. In this structure, the fuzzy neural network (FNN) is introduced to build the multi-model. Then, the prediction outputs of each state variable at all sampling instants are obtained to provide control information for the controller. Specially, the objective function with adaptive weight matrix (AWM) is designed to reduce the influence of the prediction error caused by nonlinear fitting on control performance. Then, the optimal control laws are calculated to improve the control precision. Finally, the convergence and stability of DMMPC are proved in detail. The numerical example and industrial application reveal that the proposed DMMPC can obtain considerable control performance for the multi-rate sampled-data nonlinear systems. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —The asynchronous control problem of the multi-rate sampled-data nonlinear system (MRSNS) may degrade the operation performance of closed-loop system. In this paper, a data-driven multi-model predictive control (DMMPC) strategy is designed for MRSNS without mechanism model. The strategy mainly consists of three parts: First, a multi-model prediction structure based on fuzzy neural network (FNN) is established to obtain the prediction output of all variables at each sampling point. The parameters of FNNs are corrected at each sampling point to ensure prediction accuracy. Second, an objective function with an adaptive weight matrix (AWM) is designed to compute the control law, in which AWM is used to reduce the influence of prediction error on control performance. Third, the effectiveness of DMMPC is verified by a numerical example and an industrial application of the wastewater treatment process. The experimental results show that DMMPC can achieve satisfied operation performance in control accuracy.

105-LB: Safety and Performance of an Advanced Hybrid Closed-Loop (AHCL) Algorithm with Ultra-rapid Lispro (URLi) and Lispro

Tuning a multi-model predictive control AHCL algorithm to the PK/PD properties of URLi may result in improved performance vs Lispro with no increased safety concerns. This was a randomized, double-blind, inpatient, 2-period crossover study in 22 adults with T1D on insulin pumps. Each URLi and Lispro treatment period had a 4-day inpatient visit with insulin doses calculated using the tuned AHCL algorithm. To challenge the algorithm, tests included unannounced exercise, 1 hr delayed bolus, underbolus, and overbolus. Participants (59% male) had mean ±SD age 43.4±13.77 years, T1D duration 25.9±9.61 yrs, and HbA1c 7.6±0.47%. Mean time in range (TIR; 70-180 mg/dL) during the overall inpatient period: URLi 79.4%; Lispro 78.7 % (p=0.80) with a trend towards decreased time &amp;lt;70 mg/dl (1.5 vs 2.4, p=0.08) for URLi. URLi had a significant difference in Level 1 hypoglycemia (30 vs 55 events, p = 0.01) but not Level 2 (4 vs 9 events, p=0.3). TIRs were similar including during meal/challenge periods (Table) with a trend towards greater TIR during the postprandial period and reduced time &amp;lt;70 mg/dl during exercise for URLi. There was greater TIR during overbolus with Lispro. No severe hypoglycemia events occurred. Results suggest reduced hypoglycemia with URLi vs Lispro with comparable glycemic control. The AHCL algorithm showed promising glycemic performance in a supervised setting when used with URLi or Lispro. Disclosure R. Pollom: None. E. Dassau: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. M. Katz: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. L. Nosek: None. H. Coester: None. E. Zijlstra: Other Relationship; Eli Lilly and Company, Speaker's Bureau; Gan &amp; Lee Pharmaceuticals, Novo Nordisk A/S. A. Ghosh: None. A. Haidar: Consultant; Eli Lilly and Company, Other Relationship; Bigfoot Biomedical, Inc., Research Support; Adocia, Dexcom, Inc., Tandem Diabetes Care, Inc. R. Jones: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. J. Leohr: None. S. E. Gleissner: None.

Flexible operation of the pressurized water reactor nuclear power system using multi-model predictive control over a wide nonlinear operating range

The nonlinearity problem caused by the large load variation is a significant challenge for the pressurized water reactor nuclear power system (PWRNPS) to realize its flexible operation over a wide operating range when it participates in peak regulation in the power grid. Thus, this work proposes a novel multi-model predictive control (MMPC) strategy for PWRNPS to address its nonlinearity problem in multivariable control within a wide range of conditions, in which the coupling effect between the average coolant temperature and the steam turbine power is considered. Given the insufficient understanding of the nonlinearity of PWRNPS in previous work, an in-depth investigation in this regard is performed to support the control design, in which the nonlinearity of PWRNPS at the system level is analyzed both on quantitative and qualitative aspects for the first time. Considering the captured uneven nonlinearity distribution of PWRNPS, a new approach integrating hierarchical clustering with gap metric is proposed to guide the decomposition of the operating space of PWRNPS systematically, based on which MMPC is designed to realize the nonlinearity control of PWRNPS under the large load variation. Three types of simulations demonstrate the effectiveness of the proposed methods.

Formula omitted]-gap metric based multi-model predictive control of an ocean current turbine system with blade pitch failures

In this article, a method of combining multiple linear models is developed and used by an advanced model predictive control algorithm for ocean current turbine flight control. The developed model bank consists of faulty and healthy linear models created for the ocean current turbine system. Each linear model is used to compute individual control actions using the model predictive control framework, which are combined to generate the aggregate control action that is applied to the system. A weighted average defined by the ν-gap distance between the current linear approximation of system dynamics and each individual linear system is used to compute the aggregate control action. The multi-model predictive control strategy is applied to control the ocean current turbine to mitigate performance degradation that may occur due to failures in the system. The proposed control framework’s effectiveness is shown with the renewable power generation improvement during a faulty case.

Study on an adaptive multi-model predictive controller for the thermal management of a SOFC-GT hybrid system

A SOFC temperature control system based on adaptive multimodel predictive control (MMPC) method is designed for a solid oxide fuel cell-gas turbine (SOFC-GT) hybrid system with anode and cathode ejectors. Two multi-input and multi-output MPCs (under 100% and 90% load) are designed to control the anode and cathode inlet temperatures. The accuracy of the identified linear models are both more than 95%. The control performance of the designed MMPC is compared with a single MPC and traditional PI. The comparison results demonstrate that the proposed MMPC is most effective and competitive in SOFC thermal management. During the load following, the controller overshoot is less than 1.19K. The settling time is about 2000s, and the integral of time-weighted absolute error is less than 472.

Model Predictive Control With Disturbance Observer for Marine Diesel Engine Speed Control

The application of model predictive control (MPC) algorithm in the fixed phase control of marine diesel engine speed is studied under the premise of considering model mismatch and external disturbance. Firstly, the steady-state error problem of conventional MPC controller is solved by changing nonlinear model to incremental form. Furthermore, discrete disturbance observer (DO) is introduced in the feedback correction, which can filter out the high-frequency disturbance and reduce the requirement of algorithm on the accuracy of model. Then, considering that nonlinear MPC based on DO (DONMPC) requires a large amount of online computation, the algorithm is simplified by preliminarily converting the nonlinear model to linear model. Through analysis, the controller performance of the two models is similar. Furthermore, considering that the speed of marine diesel engine is usually set to a few fixed reference values, a linear multi-model predictive controller based on DO (DOLMMPC) with less online calculation is proposed. Finally, the designed controllers are verified by experiments. The software simulations of the designed controllers and the PID controller are carried out on the cylinder-by-cylinder mean value engine model (MVEM). It is proved that the algorithm simplification method retains the control performance of the DONMPC algorithm, and the control performance of the designed two controllers is better than the PID controller. Moreover, the DOLMMPC controller and PID controller are tested on the semi-physical simulation platform. The results demonstrate that the DOLMMPC controller can meet the computational power limit of the microprocessor in practical engineering.

Handling model uncertainty in control of a pressure swing adsorption unit for syngas purification: A multi-model zone control scheme-based robust model predictive control strategy

The application of a robustly stabilizing model predictive control (MPC) strategy to a pressure swing adsorption (PSA) unit, hitherto unexplored in the literature, is addressed in this work. Here, a set of linear models to handle model uncertainties describes the PSA nonlinear dynamic behaviour. Each model represents a different operating condition of the PSA, generating a multi-model uncertainty description. In addition to incorporating the multi-model uncertainty, the proposed MPC control law deals with a controlled output zone tracking scheme to systematically accommodate typical cyclic steady states of a PSA process. From the industrial standpoint, economic targets on process variables (controlled outputs, manipulated inputs, or both) are also incorporated into the robust control method. A Real-Time Optimizer usually defines these targets, making up a two-layer control scheme. The effectiveness of the robust MPC is evaluated in a syngas purification process via a single bed, six-step PSA unit composed of a porous amino-functionalized titanium terephthalate MIL-125-type MOF adsorbent.

ANN based Multi Model Predictive Control for pH-Control

Artificial neural network (ANN) has many uses when non-linear behaviour is modelled. Here we are training a feedforward ANN that will mimic the behaviour of a Robust Model Predictive Controller (RMPC) for use in pH control. The training dataset were generated from running multiple tests on RMPC for different requirements and cases of pH-control. The training data focused on the control-inputs relating to the other process inputs. The training algorithm used in this neural network is Levenberg-Marquardt algorithm which is the most widely use algorithm in current machine learning industry. This neural network was trained by using the deep learning toolbox in Matlab®. Eight different cases is presented: four is for deploying neural network purpose, while the other four is for verification purpose. The result shows good control as long as the ANN-controller has been given a similar input and there are no multiplicity in the process input data.

1005-P: A Multimodel Model Predictive Controller Can Improve Glucose Control in Children with Type 1 Diabetes (T1D)

Background: Many youth with type 1 diabetes do not achieve glycemic targets motivating the development of new therapeutic options for this population. We tested an automated insulin delivery system (AID) that uses a multi-model model predictive controller in children with T1D. Methods: We conducted a 40-hour free-living clinical trial to assess our automated insulin delivery algorithm in children aged 7-13 years old with T1D (mean age 10.5±2.0 years, HbA1c 7.9±0.7%, duration of T1D 5.2±2.8 years) . Participants underwent a 7-day run-in period on sensor-augmented pump therapy followed by two 40-hour intervention periods composed of four hours of open-loop and 36 hours of AID use. Controller settings were identical in both intervention periods. Mean time-in-ranges were compared by paired T-test. Results: Time in target (3.9-10mmol/L) and time in hyperglycemia (&amp;gt;10mmol/L) improved during intervention periods vs. run-in (52.2% vs. 65.6%; p&amp;lt;0.and 45.1% vs. 30.8%; p=0.respectively) while time in hypoglycemia (&amp;lt;3.9mmol/L) during intervention periods was unchanged from the run-in (1.8 vs. 2.0%, p=0.96) . Conclusion: Early results suggest our advanced model predictive controller improves glucose control in a 7-13-year-old population with T1D. Larger and longer trials are warranted. Disclosure E.Palisaitis: Other Relationship; Eli Lilly and Company. R.Jones: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. M.Katz: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. H.Wolpert: Employee; Lilly Diabetes. A.El fathi: Other Relationship; Eli Lilly and Company. A.Haidar: Consultant; Eli Lilly and Company, Research Support; ADOCIA, Dexcom, Inc., Eli Lilly and Company, Tandem Diabetes Care, Inc. J.E.Von oettingen: None. L.Legault: Advisory Panel; Abbott Diabetes, Insulet Corporation, Novo Nordisk A/S, Other Relationship; Eli Lilly and Company, Research Support; AstraZeneca, Merck &amp; Co., Inc. Funding Eli Lilly

A multi-model predictive control method for the Pichia pastoris fermentation process based on relative error weighting algorithm

The Pichia pastoris fermentation process is with highly nonlinear and time-varying and strong coupling characteristics, traditional control methods such as classic PID cannot satisfy the actual needs of the fermentation process control. In order to effectively solve the control problem of the Pichia pastoris fermentation process, a multi-model predictive control method is presented based on the relative error weighting algorithm. First, the prior sample data are divided into multiple training sample sets (sample cluster) by fuzzy C-means clustering algorithm (FCM). Then, the corresponding sub prediction model is obtained by using least square support vector machine (LSSVM) and Improved particle swarm optimization (IPSO) algorithm for each sample cluster. Finally, the control strategy of predictive model is constructed based on multi model relative error weighting algorithm. The simulation in the Pichia pastoris fermentation process shows that the algorithm proposed in this paper could improve the transient response and perform good output tracking in a wide range. It improves the adaptive ability of the model and makes it more accurate to describe the actual state of the nonlinear system.

Multi-model adaptive predictive control of superheated steam temperature of DSG solar parabolic-trough collector based on heat-steam ratio and its reference trajectory

The direct steam generation (DSG) solar thermal power generation system has obvious inertia and nonlinearity in the response of superheated steam temperature. Due to the intermittent and random fluctuations of solar irradiation, the DSG system frequently undergoes large changes in operating conditions, which results in a very complicated problem of controlling the superheated steam temperature at the collector outlet. This paper proposed a multi-model adaptive control scheme based on the heat-steam ratio and its reference trajectory for the control problem of the outlet steam temperature of the DSG solar parabolic-trough collector. Firstly, through analyzing the dynamic characteristics of the superheated steam temperature, the main factors and the main nonlinear characteristics that lead to the nonlinearity of the steam temperature response are revealed. Secondly, taking the ratio of solar irradiation to the outlet steam flow (heat-steam ratio) as the characteristic parameter based on the nonlinear characteristics of the system, the DSG collector system is divided into several linearized subspaces, and then the corresponding steam temperature predictive sub-model set and steam temperature adaptive predictive control scheme are established. Finally, a smooth switching strategy based on the reference trajectory of the nonlinear characteristic parameter (heat-steam ratio) is established in response to the frequent or large switching of the controller caused by the large changes in the operating conditions, which effectively improved the stability of the temperature control process of the superheated steam at the DSG collector outlet. In this paper, the effectiveness of the above-mentioned control scheme has been preliminarily verified through simulation experiments.

Optimal Frequency Control Management of Grid Integration PV/Wind/FC/Storage Battery Based Smart Grid Using Multi Objective Particle Swarm Optimization MOPSO and Model Predictive Control (MPC)

This article forecasts the performance of smart-grid electrical transmission systems and integrated battery/FC/Wind/PV storage system renewable power sources in the context of unpredictable solar and wind power supplies. The research provided a hybrid renewable energy sources smart grid power system electrical frequency control solution using adaptive control techniques and model predictive control (MPC) based on the Multi-Objective Practical Swarm Optimization Algorithm MOPSO. To solve the problems of parameter tuning in Load Frequency Control, the suggested adaptive control approach is utilized to accomplish on-line adjustment of the Load Frequency Control parameters. During the electrical grid's integration, the system under investigation is a hybrid Wind/PV/FC/Battery smart grid with variable demand load. To achieve optimal outcomes, all of the controller settings for various units in power grids are determined by means of a customized objective function and a particle swarm optimization method rather than a regular objective function with fluctuating restrictions. To suppress the consumption and generation balance, MPCs were designed for each of the Storage Battery, Wind Turbine Generation, and the model Photovoltaic Generation. In addition, demand response (real-time pricing) was used in this scheme to reduce the load frequency by adjusting the controlled loads. The suggested control strategy is evaluated in the Simulink /MATLAB environment in order to analyse the suggested approach's working in the power system, as well as its effectiveness, reliability, robustness, and stability. The simulation findings show that the proposed control method generally converges to an optimal operating point that minimises total user disutility, restores normal frequency and planned tie-line power flows, and maintains transmission line thermal restrictions. The simulation results further indicate that the convergence holds even when the control algorithm uses inaccurate system parameters. Finally, numerical simulations are used to illustrate the proposed algorithm's robustness, optimality, and effectiveness. In compared to previous methodologies, the system frequency recovers effectively and efficiently in the event of a power demand disturbance, as demonstrated. A sensitivity test is also performed to assess the suggested technique's effectiveness.

Optimal Frequency Control Management of Grid Integration PV/Wind/FC/Storage Battery Based Smart Grid Using Multi Objective Particle Swarm Optimization MOPSO and Model Predictive Control (MPC)

This article forecasts the performance of smart-grid electrical transmission systems and integrated battery/FC/Wind/PV storage system renewable power sources in the context of unpredictable solar and wind power supplies. The research provided a hybrid renewable energy sources smart grid power system electrical frequency control solution using adaptive control techniques and model predictive control (MPC) based on the Multi-Objective Practical Swarm Optimization Algorithm MOPSO. To solve the problems of parameter tuning in Load Frequency Control, the suggested adaptive control approach is utilized to accomplish on-line adjustment of the Load Frequency Control parameters. During the electrical grid's integration, the system under investigation is a hybrid Wind/PV/FC/Battery smart grid with variable demand load. To achieve optimal outcomes, all of the controller settings for various units in power grids are determined by means of a customized objective function and a particle swarm optimization method rather than a regular objective function with fluctuating restrictions. To suppress the consumption and generation balance, MPCs were designed for each of the Storage Battery, Wind Turbine Generation, and the model Photovoltaic Generation. In addition, demand response (real-time pricing) was used in this scheme to reduce the load frequency by adjusting the controlled loads. The suggested control strategy is evaluated in the Simulink /MATLAB environment in order to analyse the suggested approach's working in the power system, as well as its effectiveness, reliability, robustness, and stability. The simulation findings show that the proposed control method generally converges to an optimal operating point that minimises total user disutility, restores normal frequency and planned tie-line power flows, and maintains transmission line thermal restrictions. The simulation results further indicate that the convergence holds even when the control algorithm uses inaccurate system parameters. Finally, numerical simulations are used to illustrate the proposed algorithm's robustness, optimality, and effectiveness. In compared to previous methodologies, the system frequency recovers effectively and efficiently in the event of a power demand disturbance, as demonstrated. A sensitivity test is also performed to assess the suggested technique's effectiveness.

In-depth characteristic analysis and wide range optimal operation of fuel cell using multi-model predictive control

Humidity and voltage are the critical parameters that have a significant influence on the thermal stability and electrical efficiency of the proton exchange membrane fuel cells (PEMFCs). However, the inherent coupling between the two parameters makes the controller design challenging from a system's perspective. Furthermore, frequent load fluctuations during fuel cell operation give rise to considerable fluctuations in voltage and humidity. To mitigate this undesirable variation, the system coupling of voltage and humidity and nonlinearity resulting from varying load are analyzed in this paper, where the coupled dynamics of relative humidity and output voltage are manipulated by the humidifier electrical power and compressor voltage. Based on the analysis, a multi-model predictive control (MMPC) scheme is formulated for the coupled plant dynamics, where the prediction model utilized is developed based on the subspace identification method for local models integrated with fuzzy membership weightings. The MMPC formulation has taken advantage of considering the actuator saturation and the allowable output range as input and output constraints. Simulation results show that the proposed MMPC can improve the tracking performance with a flexible trade-off between the relative humidity and output voltage.

Multiple model predictive control of perching maneuver based on guardian maps

Considering the strong nonlinearity of Unmanned Aerial Vehicles (UAVs) resulting from high Angle of Attack (AOA) and fast maneuvering, we present a multi-model predictive control strategy for UAV maneuvering, which has a small amount of online calculation. Firstly, we divide the maneuver envelope of UAV into several sub-regions on the basis of the gap metric theory. A novel algorithm is then developed to determine the ploytopic model for each sub-region. According to this, a Robust Model Predictive Control based on the Idea of Comprehensive optimization (ICE-RMPC) is proposed. The control law is designed offline and optimized online to reduce the computational expense. Then, the ICE-RMPC method is applied to design the controllers of sub-regions. In addition, to guarantee the stability of whole closed-loop system, a multi-model switching control strategy based on guardian maps is put forward. Finally, the tracking performance of proposed control strategy is demonstrated by an illustrative example.

Research on optimization of primary frequency regulation of thermal power units based on multi-model predictive control

The nonlinear problem of system cannot be well solved using traditional power regulation system, which causes regulation lag and insufficient output and poor accuracy of the actual performance of the primary frequency regulation (PFR) on thermal power units. In order to further improve PFR performance of thermal power units, a multi-model predictive control method is adopted to optimize and improve the traditional steam turbine power control system. A PFR optimal control strategy is proposed, which improves the regulation performance and robustness of the power control system. The simulation results show that the proposed optimization algorithm can overcome the nonlinear problem of the steam turbine governing valve, and the regulating process is stable and fast without overshoot. This research is of great significance to improve the contribution rate of PFR of thermal power units and to better satisfy the assessment of power grid PFR, thereby to stabilize the frequency of power grid.

Application of feedforward predictive control in DC furnace coordination system

In order to solve the control difficulties such as large dynamic delay, serious coupling between machine and furnace and strong nonlinear, the feedforward and feedback predictive control is proposed to control the coordinated system on the basis of multi-model predictive control, and the static feedforward part designs the feed line feedforward of coal quantity against coal quality disturbance on the basis of the accurate balance of energy. Thus, the reference total fuel quantity under the target load instruction is determined, and the dynamic part fully considers the dynamic compensation of the unit energy storage, gives a certain overshoot fuel quantity to the boiler regulation, speeds up its response, and ensures the rapidity of the unit. The real-time simulation results of virtual DPU (Distributed Processing Unit) show that the designed feedforward-feedback predictive control can quickly respond to load change, improve the resistance to fuel disturbance, reduce the fluctuation of main steam pressure, and better ensure the safe, economic and stable operation of the coordination system.

Fractional-order multi-model predictive control for nonlinear processes

Fractional-order multi-model predictive control for nonlinear processes