Polynomial Chaos Expansion-Based Robust Multi-Objective Control of Hydraulic Turbine-Regulating Systems Under Hydro-Mechanical Uncertainties
- 1Department of Mechanical Engineering, College of Engineering, Qassim University, Buraydah 51452, Saudi Arabia
Abstract
Hydraulic turbine-regulating systems (HTRSs) require advanced robust control strategies to ensure grid frequency stability and prevent mechanical wear under severe non-stationary conditions. This paper presents an original polynomial chaos expansion (PCE)-based robust control framework, designed to address hydro-mechanical uncertainties in HTRSs. Initially, a deterministic multi-objective optimization utilizing the integral of time-weighted absolute error (ITAE) and the integral of the square control signal (ISC) is executed via the multi-objective imperialist competitive algorithm (MOICA) to establish nominal fractional-order proportional-integral-derivative (FOPID) controller parameters. Subsequently, a PCE-based Sobol sensitivity analysis demonstrates that these nominal designs suffer from severe performance variability under real-world parameter drift. To overcome this limitation, a comprehensive four-dimensional robust multi-objective optimization framework is formulated to simultaneously minimize both expected values and standard deviations of tracking errors and control efforts. By integrating generalized PCE surrogate moment evaluations directly within the MOICA optimization loop, the proposed architecture effectively suppresses dynamic performance variance without requiring computationally expensive Monte Carlo simulations. Numerical evaluations confirm that the proposed PCE-MOICA-tuned robust FOPID controller achieves a substantial 57% to 91% reduction in performance variability compared to traditional deterministic baselines and established literature methodologies, including NSGA-II and MOHHO-optimized controllers.
Related
- A connectedness result for convex polynomial multi-objective optimization without boundedness
- Robust Planning with Quadratically Constrained State-Dependent Uncertainties
- Expected Hypervolume Maximization for Multiobjective Optimization under Uncertainties
- A Structure-Preserving Semi-Lagrangian based bi-fidelity method for the Vlasov--Poisson system with uncertainties
- Resilient Cell-Free MIMO ISAC Under Uncertainties: An Impairment-Aware Graph Neural Network Framework
BibTeX
@article{alghafis2026polynomial,
title = {{Polynomial Chaos Expansion-Based Robust Multi-Objective Control of Hydraulic Turbine-Regulating Systems Under Hydro-Mechanical Uncertainties}},
author = {Abdullah Alghafis and Mohamed Nejlaoui},
journal = {Mathematics},
year = {2026},
month = oct,
doi = {10.3390/math14213873},
url = {https://www.mdpi.com/2227-7390/14/21/3873},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Polynomial Chaos Expansion-Based Robust Multi-Objective Control of Hydraulic Turbine-Regulating Systems Under Hydro-Mechanical Uncertainties
AU - Abdullah Alghafis
AU - Mohamed Nejlaoui
PY - 2026
DA - 2026/10/23
JO - Mathematics
DO - 10.3390/math14213873
UR - https://www.mdpi.com/2227-7390/14/21/3873
AB - Hydraulic turbine-regulating systems (HTRSs) require advanced robust control strategies to ensure grid frequency stability and prevent mechanical wear under severe non-stationary conditions. This paper presents an original polynomial chaos expansion (PCE)-based robust control framework, designed to address hydro-mechanical uncertainties in HTRSs. Initially, a deterministic multi-objective optimization utilizing the integral of time-weighted absolute error (ITAE) and the integral of the square control signal (ISC) is executed via the multi-objective imperialist competitive algorithm (MOICA) to establish nominal fractional-order proportional-integral-derivative (FOPID) controller parameters. Subsequently, a PCE-based Sobol sensitivity analysis demonstrates that these nominal designs suffer from severe performance variability under real-world parameter drift. To overcome this limitation, a comprehensive four-dimensional robust multi-objective optimization framework is formulated to simultaneously minimize both expected values and standard deviations of tracking errors and control efforts. By integrating generalized PCE surrogate moment evaluations directly within the MOICA optimization loop, the proposed architecture effectively suppresses dynamic performance variance without requiring computationally expensive Monte Carlo simulations. Numerical evaluations confirm that the proposed PCE-MOICA-tuned robust FOPID controller achieves a substantial 57% to 91% reduction in performance variability compared to traditional deterministic baselines and established literature methodologies, including NSGA-II and MOHHO-optimized controllers.
SN - 2227-7390
ER -