.png){kind=icon}
返回
Offset-free Lyapunov-based stabilising robust nonlinear control using data-driven model: A nonlinear multi-model computationally efficient approach
.png){kind=icon}
PRE
AI
DOI：10.1016/j.conengprac.2026.107013.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Robust NMPC with data-driven surrogates remains challenging due to structural misspecification and unreliable extrapolation, which can compromise constraint satisfaction. This work proposes an offset-free robust NMPC scheme based on symbolic regression (SR). Using compact NARX models that expose epistemic uncertainty at the operating-zone level, robustness is enforced by embedding multiple SR surrogates as hard constraints. Two configurations are evaluated: (i) RNMPCSZ, which embeds SR models from a single operating zone, and (ii) RNMPCMZ, which jointly enforces zone-specific SR models from several zones, re-scheduling the nominal predictor at each set-point change. In both cases, robustness arises from the intersection of admissible sets defined by the enforced SR models, without modifying the nominal cost. The method was validated on a simulated pilot-scale ESP system. Both controllers preserved tracking while eliminating constraint violations, especially near upthrust, with only minor settling-time increases. Real-time feasibility was maintained even when enforcing up to four models per zone, with RNMPCMZ achieving the largest safety margins.
Keyword:
Robust NMPC
Symbolic Regression
NARX models
Constraint satisfaction
Multi-model approach
期刊
IF:
4.6
论文数:
5.7K
被引数:
1.1W
机构
暂无机构信息
引用论文
Nonlinear Model Predictive Control of Electrical Submersible Pumps based on Echo State Networks
PREAI
分享收藏Data-Driven Non-Linear Current Controller Based on Deep Symbolic Regression for SPMSM.png)
SENSORS
OAAI
SENSORSIF3.5
OAAI分享收藏Efficient hybrid modeling and sorption model discovery for non-linear advection-diffusion-sorption systems: A systematic scientific machine learning approach非线性对流-扩散-吸附系统的高效混合建模和吸附模型发现: 一种系统的科学机器学习方法OAAI
OAAI分享收藏A comparison of strategies to embed physics-informed neural networks in nonlinear model predictive control formulations solved via direct transcription比较将物理信息神经网络嵌入通过直接转述求解的非线性模型预测控制策略的方案OAAI
OAAI分享收藏分享收藏分享收藏
查看更多内容