skip to main content

Explore Research Products in the NSF-PAR

  • NSF Public Access
  • Search Results
  • Learning‐based adaptive‐scenario‐tree model predictive control with improved probabilistic safety using robust Bayesian neural networks
Title: Learning‐based adaptive‐scenario‐tree model predictive control with improved probabilistic safety using robust Bayesian neural networks
Summary

Scenario‐based model predictive control (MPC) methods can mitigate the conservativeness inherent to open‐loop robust MPC. Yet, the scenarios are often generated offline based on worst‐case uncertainty descriptions obtaineda priori, which can in turn limit the improvements in the robust control performance. To this end, this paper presents a learning‐based, adaptive‐scenario‐tree model predictive control approach for uncertain nonlinear systems with time‐varying and/or hard‐to‐model dynamics. Bayesian neural networks (BNNs) are used to learn a state‐ and input‐dependent description of model uncertainty, namely the mismatch between a nominal (physics‐based or data‐driven) model of a system and its actual dynamics. We first present a new approach for training robust BNNs (RBNNs) using probabilistic Lipschitz bounds to provide a less conservative uncertainty quantification. Then, we present an approach to evaluate the credible intervals of RBNN predictions and determine the number of samples required for estimating the credible intervals given a credible level. The performance of RBNNs is evaluated with respect to that of standard BNNs and Gaussian process (GP) as a basis of comparison. The RBNN description of plant‐model mismatch with verified accurate credible intervals is employed to generate adaptive scenarios online for scenario‐based MPC (sMPC). The proposed sMPC approach with adaptive scenario tree can improve the robust control performance with respect to sMPC with a fixed, worst‐case scenario tree and with respect to an adaptive‐scenario‐based MPC (asMPC) using GP regression on a cold atmospheric plasma system. Furthermore, closed‐loop simulation results illustrate that robust model uncertainty learning via RBNNs can enhance the probability of constraint satisfaction of asMPC.

  more » « less
Award ID(s):
2302219
NSF-PAR ID:
10442867
Author(s) / Creator(s):
 ;  ;  ;  
Publisher / Repository:
Wiley Blackwell (John Wiley & Sons)
Date Published:
Journal Name:
International Journal of Robust and Nonlinear Control
Volume:
33
Issue:
5
ISSN:
1049-8923
Page Range / eLocation ID:
p. 3312-3333
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ( , Energies)
    null (Ed.)
    Stability and reliability are of the most important concern for isolated microgrid systems that have no support from the utility grid. Interval predictions are often applied to ensure the system stability of isolated microgrids as they cover more uncertainties and robust control can be achieved based on more sufficient information. In this paper, we propose a probabilistic microgrid energy exchange method based on the Model Predictive Control (MPC) approach to make better use of the prediction intervals so that the system stability and cost efficiency of isolated microgrids are improved simultaneously. Appropriate scenarios are selected from the predictions according to the evaluation of future trends and system capacity. In the meantime, a two-stage adaptive reserve strategy is adopted to further utilize the potential of interval predictions and maintain the system security adaptively. Reserves are determined at the optimization stage to prepare some extra capacity for the fluctuations in the renewable generation and load demand at the operation stage based on the aggressive and conservative level of the system, which is automatically updated at each step. The optimal dispatch problem is finally formulated using the mixed-integer linear programming model and the MPC is formulated as an optimization problem with a discount factor introduced to adjust the weights. Case studies show that the proposed method could effectively guarantee the stability of the system and improve economic performance. 
    more » « less
  2. ; ; ; ( , American Control Conference)
    We propose a learning-based robust predictive control algorithm that compensates for significant uncertainty in the dynamics for a class of discrete-time systems that are nominally linear with an additive nonlinear component. Such systems commonly model the nonlinear effects of an unknown environment on a nominal system. We optimize over a class of nonlinear feedback policies inspired by certainty equivalent "estimate-and-cancel" control laws pioneered in classical adaptive control to achieve significant performance improvements in the presence of uncertainties of large magnitude, a setting in which existing learning-based predictive control algorithms often struggle to guarantee safety. In contrast to previous work in robust adaptive MPC, our approach allows us to take advantage of structure (i.e., the numerical predictions) in the a priori unknown dynamics learned online through function approximation. Our approach also extends typical nonlinear adaptive control methods to systems with state and input constraints even when we cannot directly cancel the additive uncertain function from the dynamics. Moreover, we apply contemporary statistical estimation techniques to certify the system’s safety through persistent constraint satisfaction with high probability. Finally, we show in simulation that our method can accommodate more significant unknown dynamics terms than existing methods. 
    more » « less
  3. ; ; ( , AIChE Journal)
    Abstract

    This article proposes a novel distributionally robust optimization (DRO)‐based soft‐constrained model predictive control (MPC) framework to explicitly hedge against unknown external input terms in a linear state‐space system. Without a priori knowledge of the exact uncertainty distribution, this framework works with a lifted ambiguity set constructed using machine learning to incorporate the first‐order moment information. By adopting a linear performance measure and considering input and state constraints robustly with respect to a lifted support set, the DRO‐based MPC is reformulated as a robust optimization problem. The constraints are softened to ensure recursive feasibility. Theoretical results on optimality, feasibility, and stability are further discussed. Performance and computational efficiency of the proposed method are illustrated through motion control and building energy control systems, showing 18.3% less cost and 78.8% less constraint violations, respectively, while requiring one third of the CPU time compared to multi‐stage scenario based stochastic MPC.

     
    more » « less
  4. ; ; ( , Proceedings of 2022 American Control Conference (ACC))
    This paper introduces a computationally efficient approach for solving Model Predictive Control (MPC) reference tracking problems with state and control constraints. The approach consists of three key components: First, a log-domain interior-point quadratic programming method that forms the basis of the overall approach; second, a method of warm-starting this optimizer by using the MPC solution from the previous timestep; and third, a computational governor that bounds the suboptimality of the warm-start by altering the reference command provided to the MPC problem. As a result, the closed-loop system is altered in a manner so that MPC solutions can be computed using fewer optimizer iterations per timestep. In a numerical experiment, the computational governor reduces the worst-case computation time of a standard MPC implementation by 90%, while maintaining good closed-loop performance. 
    more » « less
  5. ( , Proceedings of the 52nd Frontiers of Power Conference)
    null (Ed.)
    Increased capacity associated with renewable energy sources has created a need for improved methods for controlling power flows from inverter-based generation. This research provides a comparative study of finite-control-set model predictive current control (FCS-MPC-based) with respect to conventional proportional-integral-based (PI-based) synchronous current control for a three-phase voltage source inverter (VSI). The inverter is accompanied by an inductive-capacitive-inductive (LCL) filter to attenuate pulse width modulation (PWM) switching harmonics. However, an LCL filter introduces a resonance near to the control stability boundary, giving rise to substantial complexity from a control perspective. In order to avoid potential instability caused by the resonance, active damping can be included in the PI-based current control. Though properly designed active damping can improve inverter stability, in practice the robustness of standard PI control is not attainable due to variability in the grid inductance at the point of common coupling (PCC). This is due to impedance variations causing large shifts in the LCL resonance frequency. Weak grid conditions (i.e., a low short-circuit ratio) and a correspondingly high line impedance are particularly susceptible to LCL induced resonance instabilities. As an approach to operate with grid impedance variations and weak grid conditions, FCS-MPC has the potential to produce superior performance compared to PI-based current control methods. This comparative study indicates that FCS-MPC has improved resonance damping and fast dynamic capability in a system with renewable energy sources under weak grid conditions. Detailed results from MATLAB/SimPower are presented to validate the suggested FCS-MPC method where it is robust to uncertainty in the grid impedance variations. Overall results indicate an improvement over conventional PI-based current control methods. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email