Enhancing Pressure Estimation in Water Distribution Systems Utilizing Optimized Scene Context Aware Graph Convolutional Network for Improved Predictive Performance

Water Distribution System (WDS) is an essential component of urban development and plays an important part in the normal functioning of the city. Some significant obstacles are faced due to the complex and periodic environment of pressure variations in WDS. Therefore, this research proposes a novel Enhancing Pressure Estimation in Water Distribution Systems utilizing Optimized Scene Context Aware Graph Convolutional Network for Improved Predictive Performance (PEWDS–SCAGCN–IPP). The data for this work is taken from large-scale water distribution networks in the area of Oosterbeek, Netherlands. Then, the input data is pre-processed using Adaptive Multiscale Gaussian Co-Occurrence Filtering (AMGCoF) to remove outlier, replace missing values, and normalization from input data. After that, the preprocessing data is given into the feature extraction stage using the Fractional Fourier Jacobi Wavelet Transform (FFJWT) to extract nuanced and localized features. The extracted features are given to the Scene Context Aware Graph Convolutional Network (SCAGCN) to estimate the pressure in WDS. Finally, Educational Competition Optimizer (ECO) is employed to optimize the SCAGCN to accurately estimate the pressure in WDS. The proposed PEWDS–SCAGCN–IPP is implemented, and the performance metrics like Accuracy, Mean Absolute Error, Root Mean Square Error, Mean Absolute Percentage Error, Nash and Sutcliffe Coefficient of Efficiency are analyzed. Finally, the performance of the proposed PEWDS–SCAGCN–IPP method attains 16.20%, 26.81%, and 30.18% higher accuracy, 26.28%, 16.11%, and 32.52% lower RMSE, and 16.52%, 26.85%, and 31.62% lower MAE when compared with the existing methods, respectively.