Scholarlink

<p>Model predictive controller has firmly established itself as a powerful control strategy for controlling dynamic systems subject to constraints. However, the heavy computational burden for solving the optimization problem at each sampling instant, especially in real-time scenarios of fast systems, remains a current challenge. Therefore, this work presents a simple model predictive controller (MPC) with an optimized gain-scheduling strategy to allow the relaxation of the constraint formulation on MPC with the goal of expanding its practical applicability in time-variable systems. The developed control algorithm uses a gain-scheduling strategy with optimized parameters considering several harsh conditions for the control system in its offline optimization procedure. The controller parametrization is performed using the weight chimp optimizer (WCO) driven by a set of developed rules that consider the system constraints and its closed-loop performance. The advantage of this approach is reducing the computational burden of MPC by relaxing the constraints’ formulation and incorporating them directly in the offline WCO-based controller tuning procedure. The result is a simpler controller, but with high performance. A case study of a renewable energy system with an LCL filter is presented, where the proposed controller regulates the grid-injected current in around half to at most two grid cycles and maintains the stability of the closed-loop system even when the grid inductance varies five times its nominal value, which is a considerable challenge. A comparison with other three optimizers is also presented to demonstrate the superiority of the proposed control approach.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Relaxing constraint formulation on MPC using weight chimp optimizer with gain-scheduling strategy applied on renewable energy systems

  • 期刊论文

  • 发表日期：

    2025年6月2日

查看全文
  • Paulo Jefferson Dias de Oliveira Evald,
  • Matheus Schramm Dall’asta,
  • Lenon Schmitz,
  • Jéssika Melo de Andrade,
  • Telles Brunelli Lazzarin

摘要

Model predictive controller has firmly established itself as a powerful control strategy for controlling dynamic systems subject to constraints. However, the heavy computational burden for solving the optimization problem at each sampling instant, especially in real-time scenarios of fast systems, remains a current challenge. Therefore, this work presents a simple model predictive controller (MPC) with an optimized gain-scheduling strategy to allow the relaxation of the constraint formulation on MPC with the goal of expanding its practical applicability in time-variable systems. The developed control algorithm uses a gain-scheduling strategy with optimized parameters considering several harsh conditions for the control system in its offline optimization procedure. The controller parametrization is performed using the weight chimp optimizer (WCO) driven by a set of developed rules that consider the system constraints and its closed-loop performance. The advantage of this approach is reducing the computational burden of MPC by relaxing the constraints’ formulation and incorporating them directly in the offline WCO-based controller tuning procedure. The result is a simpler controller, but with high performance. A case study of a renewable energy system with an LCL filter is presented, where the proposed controller regulates the grid-injected current in around half to at most two grid cycles and maintains the stability of the closed-loop system even when the grid inductance varies five times its nominal value, which is a considerable challenge. A comparison with other three optimizers is also presented to demonstrate the superiority of the proposed control approach.