Variant Of Model Predictive Control Research Articles

Model predictive complex system control from observational and interventional data

Complex systems, characterized by intricate interactions among numerous entities, give rise to emergent behaviors whose data-driven modeling and control are of utmost significance, especially when there is abundant observational data but the intervention cost is high. Traditional methods rely on precise dynamical models or require extensive intervention data, often falling short in real-world applications. To bridge this gap, we consider a specific setting of the complex systems control problem: how to control complex systems through a few online interactions on some intervenable nodes when abundant observational data from natural evolution is available. We introduce a two-stage model predictive complex system control framework, comprising an offline pre-training phase that leverages rich observational data to capture spontaneous evolutionary dynamics and an online fine-tuning phase that uses a variant of model predictive control to implement intervention actions. To address the high-dimensional nature of the state-action space in complex systems, we propose a novel approach employing action-extended graph neural networks to model the Markov decision process of complex systems and design a hierarchical action space for learning intervention actions. This approach performs well in three complex system control environments: Boids, Kuramoto, and Susceptible-Infectious-Susceptible (SIS) metapopulation. It offers accelerated convergence, robust generalization, and reduced intervention costs compared to the baseline algorithm. This work provides valuable insights into controlling complex systems with high-dimensional state-action spaces and limited intervention data, presenting promising applications for real-world challenges.

Microsecond Enhanced Indirect Model Predictive Control for Dynamic Power Management in MMC Units

The multi-modular converter (MMC) technology is becoming the preferred option for the increased deployment of variable renewable energy sources (RES) into electrical power systems. MMC is known for its reliability and modularity. The fast adjustment of the MMC’s active/reactive powers, within a few milliseconds, constitutes a major research challenge. The solution to this challenge will allow accelerated integration of RES, without creating undesirable stability issues in the future power system. This paper presents a variant of model predictive control (MPC) for the grid-connected MMC. MPC is defined using a Laguerre function to reduce the computational burden. This is achieved by reducing the number of parameters of the MMC cost function. The feasibility and effectiveness of the proposed MPC is verified in the real-time digital simulations. Additionally, in this paper, a comparison between an accurate mathematical and real-time simulation (RSCAD) model of an MMC is given. The comparison is done on the level of small-signal disturbance and a Mean Absolute Error (MAE). In the MMC, active and reactive power controls, AC voltage control, output current control, and circulating current controls are implemented, both using PI and MPC controllers. The MPC’s performance is tested by the small and large disturbance in active and reactive powers, both in an offline and online simulation. In addition, a sensitivity study is performed for different variables of MPC in the offline simulation. Results obtained in the simulations show good correspondence between mathematical and real-time analytical models during the transient and steady-state conditions with low MAE. The results also indicate the superiority of the proposed MPC with the stable and fast active/reactive power support in real-time simulation.

Control pathway for an advanced divertor on ITER

Abstract This paper presents the development of a coupled shape and divertor controller for ITER with capabilities to control the flux expansion and an advanced divertor configuration, the x-divertor (XD), in which a secondary x-point is placed in the downstream scrape-off layer. Due to the high-performance nature of ITER and its relatively few shaping coils, satisfying constraints on the coil currents, power supplies, and plasma shape is a challenge for this configuration. To meet these constraints the controller uses the constrained linear quadratic regulator (CLQR) framework [1] , a variant of model predictive control (MPC). Previous work [2] has shown the existence of XD equilibria on ITER and in this study, we identify a control pathway for achieving the “pure” XD where the secondary x-point is placed at the outer strike point, representing a maximally flux-expanding scenario. Constraints are observed throughout the transition. One limitation is that at high flux expansion, the shallow angle of incidence of the magnetic field on the divertor results in a reduced plasma wetted area via tile shadowing. On the other hand, the high flux expansion and flaring characteristic of the XD facilitate enhanced detachment, which could negate the shadowing effect to some degree. In this case it would be desirable to operate as close to the flux expansion limit as possible. To this end we also demonstrate the controller's ability to track a trajectory for the flux expansion near the accepted angle of incidence limit.

Model Predictive Control of two‐phase flow using a diffuse interface approach

AbstractWe propose a feedback control strategy for the stabilization of two‐phase flow with variable densities, where we use an inexact variant of Model Predictive Control, called Instantaneous Control. The physical system is governed by the Cahn–Hilliard Navier–Stokes system with different densities proposed in [1]. (© 2014 Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Instantaneous closed loop control of the radiative transfer equations with applications in radiotherapy

AbstractIn this paper we consider a time‐dependent Boltzmann equation for particle transport in biological tissues, which can be used for radiation dose calculation. We assume that the dose should be delivered to a target volume which moves over time. We formulate this as an optimal control problem and derive closed‐loop control laws for the dose delivery problem using boundary and distributed control. Optimality conditions are derived. For the construction of the closed‐loop control laws we use an inexact variant of model predictive control called instantaneous control. We compare numerical results obtained with instantaneous control to those obtained by optimal open‐loop control, and present numerical simulations in one and two spatial dimensions. This work could be applied to image‐guided radiation therapy, where patient motion during treatment is one of the future challenges.

Multiplexed MPC for Multizone Thermal Processing in Semiconductor Manufacturing

It is known that a faster sampling rate allows a controller to react to unmeasurable disturbances faster. Adding feed- forward could decrease the effect of disturbance significantly, but feedback control is still required to cope with modelling errors and unknown disturbances. A variant of model predictive control (MPC), called multiplexed MPC (MMPC), has been proposed recently. The motivation for MMPC was to reduced real-time computational load and hence allow faster sampling on multivariable systems. In this brief, we report the first successful application of MMPC on a semiconductor manufacturing process. MMPC with feedforward has been demonstrated experimentally on a multizone bake plate application. While most of the temperature drop will be compensated by feedforward control, feedback provided by MMPC will cope with the errors due to warped wafers. This result is important for the semiconductor wafer baking process, because temperature nonuniformity will affect critical dimension of the wafer.

Calculation of optimal thermal load of intermittently heated buildings

Abstract A building is permanently in thermodynamic non-equilibrium due to changing weather, free gains and indoor temperature set-point. Load calculation in dynamic conditions is an essential goal of building energy simulation. This paper demonstrates that the load calculation is a control problem. Supposing that the thermal model of the building is linear and that the model of the building, the weather conditions and occupational program are known in the design stage, the paper proposes an unconstrained optimal control algorithm which uses feed-forward to compensate the weather conditions and model predictive programming (MPP) for set-point tracking. MPP is obtained by modifying the dynamic matrix control (DMC), a variant of model predictive control (MPC). The peak load depends on the set-back time of the indoor temperature: smaller the set-back time, larger the peak load, but smaller energy consumption. Then, the choice of the weighting coefficients in the model predictive programming may be done on economical considerations.