USING ANN-BASED MODELS IN SIZING AIR CHAMBER FOR PROTECTING TRANSMISSION LINES FROM WATER HAMMER EFFECTS
Mohamed Soliman Mohamed Soliman;
Abstract
Abstract
Hydraulic transients caused by sudden power failure of pump motor can cause cavitation and water column separation leading to collapse of pipeline. These negative pressure transients can be effectively controlled or eliminated by installation of air vessels (hydropneumatic or accumulators). In such cases, the vessel optimum size is needed for effective design and economic implementation. Since vessel sizing involves various parameters with complex relationships and an iterative solution of the differential water hammer equations, no simple and reliable sizing tool is applicable. In this study, a new and simple ANN-based model for sizing surge vessels for protecting pumping transmission lines has been developed. A wide range of flow, pipeline and pump parameters are chosen to cover many possible practical problems for a simple rising pipeline profile. The water hammer equations are solved in an iterative manner with the method of characteristics, and the Monte Carlo Simulation (MCS) is utilized to generate 2,000 realizations for the optimum vessel size for various operational cases with downsurge and upsurge transient conditions. Two approximate ANN-based models were developed for prediction of the initial and the final gas volume in the required vessel. The developed models outperformed the existing design charts yielding better scores in various statistical indicators, e.g., high coefficient of determination of 0.98 and four order magnitude less error. The introduced models have the advantage of being in simple form with clear relation between various influencing problem parameters. These developed models can serve as basis for preliminary surge protection selection and conceptual design, with no iterations and can substitute complex commercial modeling software, especially for practitioners and decision makers in feasibility study phases. This can be done with less error and higher degree of accuracy compared to existing design charts. The uncertainty analysis revealed that the predicted models had low uncertainty band of +/-0.16 compared to +/-0.78 for previous design charts.
Hydraulic transients caused by sudden power failure of pump motor can cause cavitation and water column separation leading to collapse of pipeline. These negative pressure transients can be effectively controlled or eliminated by installation of air vessels (hydropneumatic or accumulators). In such cases, the vessel optimum size is needed for effective design and economic implementation. Since vessel sizing involves various parameters with complex relationships and an iterative solution of the differential water hammer equations, no simple and reliable sizing tool is applicable. In this study, a new and simple ANN-based model for sizing surge vessels for protecting pumping transmission lines has been developed. A wide range of flow, pipeline and pump parameters are chosen to cover many possible practical problems for a simple rising pipeline profile. The water hammer equations are solved in an iterative manner with the method of characteristics, and the Monte Carlo Simulation (MCS) is utilized to generate 2,000 realizations for the optimum vessel size for various operational cases with downsurge and upsurge transient conditions. Two approximate ANN-based models were developed for prediction of the initial and the final gas volume in the required vessel. The developed models outperformed the existing design charts yielding better scores in various statistical indicators, e.g., high coefficient of determination of 0.98 and four order magnitude less error. The introduced models have the advantage of being in simple form with clear relation between various influencing problem parameters. These developed models can serve as basis for preliminary surge protection selection and conceptual design, with no iterations and can substitute complex commercial modeling software, especially for practitioners and decision makers in feasibility study phases. This can be done with less error and higher degree of accuracy compared to existing design charts. The uncertainty analysis revealed that the predicted models had low uncertainty band of +/-0.16 compared to +/-0.78 for previous design charts.
Other data
| Title | USING ANN-BASED MODELS IN SIZING AIR CHAMBER FOR PROTECTING TRANSMISSION LINES FROM WATER HAMMER EFFECTS | Authors | Mohamed Soliman Mohamed Soliman | Issue Date | 2018 |
Recommend this item
Similar Items from Core Recommender Database
Items in Ain Shams Scholar are protected by copyright, with all rights reserved, unless otherwise indicated.