Research on a multivariate measurement system for polyethylene gas pipelines utilizing a variable weight UMAP model

In producing gas-polyethylene pipelines, the five major quality indicators-wall thickness, inner diameter, outer diameter, concentricity, and ovality-exhibit complex interactions, making it challenging for traditional methods to comprehensively evaluate the measurement system's performance. To address this issue, this study introduces a multivariate measurement system analysis model based on Variable-Weight Uniform Manifold Approximation and Projection (VUMAP) dimensionality reduction. First, a dynamic weighting method is proposed, which adaptively adjusts the weights of quality characteristics using K-Means + + clustering and the Particle Swarm Optimization (PSO) algorithm, thereby overcoming the limitation of traditional uniform weighting in accurately reflecting the contributions of different quality characteristics. Second, the VUMAP dimensionality reduction algorithm is refined to enhance its capability to preserve the nonlinear relationships within high-dimensional data, ensuring the integrity and reliability of the dimensionality reduction results. Finally, an integrated multivariate measurement system analysis model is developed to comprehensively assess the measurement system's performance, effectively overcoming the limitations of traditional methods in data structure exploration and capability evaluation. The model's effectiveness is validated using a gas-polyethylene pipeline production process dataset.