Open Access

ARTICLE

Optimized Pilot Hydraulic Valves for Urban Water Systems via Enhanced BP-Coati Algorithms

Shuxun Li^1,2, Xinhao Liu^1,2,*, Yu Zhang^1,2, Yu Zhao^1,2

1 School of Petrochemical Technology, Lanzhou University of Technology, Lanzhou, 730050, China
2 Machinery Industry Pump and Special Valve Engineering Research Center, Lanzhou University of Technology, Lanzhou, 730050, China

* Corresponding Author: Xinhao Liu. Email:

Fluid Dynamics & Materials Processing 2025, 21(10), 2495-2526. https://doi.org/10.32604/fdmp.2025.068674

Received 03 June 2025; Accepted 29 August 2025; Issue published 30 October 2025

Abstract

Hydraulic control valves, positioned at the terminus of pipe networks, are critical for regulating flow and pressure, thereby ensuring the operational safety and efficiency of pipeline systems. However, conventional valve designs often struggle to maintain effective regulation across a wide range of system pressures. To address this limitation, this study introduces a novel Pilot hydraulic valves specifically engineered for enhanced dynamic performance and precise regulation under variable pressure conditions. Building upon prior experimental findings, the proposed design integrates a high-fidelity simulation framework and a surrogate model-based optimization strategy. The study begins by formulating a comprehensive mathematical model of the pipeline system using electro-hydraulic simulation techniques, capturing the dynamic behavior of both the pilot valve and the broader urban water distribution network. A coupled simulation platform is then developed, leveraging both one-dimensional (1D) and three-dimensional (3D) software tools to accurately analyze the valve’s transient response and operational characteristics. To achieve optimal valve performance, a multi-objective optimization approach is proposed. This approach employs a Levy-based Improved Tuna-Inspired Wake-Up Optimization Algorithm (L-TIWOA) to refine a Backpropagation (BP) neural network, thereby constructing a highly accurate surrogate model. Compared to the conventional BP neural network, the improved model demonstrates significantly reduced mean absolute error (MAE) and mean squared error (MSE), underscoring its superior predictive capability. The surrogate model serves as the objective function within an Improved Multi-Objective Mother Lode Optimization Algorithm (IMOMLOA), which is then used to fine-tune the key design parameters of the control valve. Validation through experimental testing reveals that the optimized valve achieves a maximum flow deviation of just 1.11 t/h, corresponding to a control accuracy of 3.7%, at a target flow rate of 30 t/h. Moreover, substantial improvements in dynamic response are observed, confirming the effectiveness of the proposed design and optimization strategy.

---

Keywords

---