
ValveStic is an advanced automated control valve health monitoring system that uses convolutional neural networks (CNN) and principal component analysis (PCA) to detect and measure the severity of stiction in control valves. This system allows for real-time, non-invasive monitoring of valve conditions using routine operating data, eliminating the need for external expertise or invasive procedures. ValveStic's flexible design makes it applicable across various industrial processes, regardless of valve brands, providing an efficient solution for reducing operational costs, optimising maintenance, and improving overall plant productivity.
Control valves are critical components in processing plants, responsible for regulating flows and ensuring smooth operations. However, wear and stiction—where valves stick and cause process oscillations—pose significant challenges, reducing valve lifespan and leading to poorer product quality. These issues can negatively impact profitability, with 30% of oscillatory problems attributed to stiction and 70% of a valve's life cycle costs dedicated to maintenance. Traditional OEM-based detection methods are often invasive and specific to particular brands, limiting their applicability and efficiency in industrial processes.
ValveStic stands out for its innovative use of convolutional neural networks and principal component analysis to monitor and predict valve health in real-time. Unlike traditional methods, it is entirely non-invasive, requiring no physical disruption of valves or complex brand-specific procedures. It extracts time-series data and visualises valve health, enabling early detection of stiction issues and accurate classification of their severity. This flexibility and automation make ValveStic suitable for a wide range of industrial processes and control valves, regardless of brand, providing a seamless integration with existing plant operations and maintenance strategies.
ValveStic offers a revolutionary solution by employing CNN and PCA to monitor valve health through a fully non-invasive approach. By analysing routine process data, ValveStic continuously detects stiction in real-time and assesses its severity, categorising it as either weak or strong. The system eliminates the need for brand-specific or invasive interventions, providing automated diagnostics that can be understood without external expert input. This approach allows plants to optimise maintenance schedules and extend the operational life of valves while reducing maintenance costs and improving process efficiency across multiple industries.
ValveStic's competitive advantage lies in its non-invasive, real-time monitoring capabilities, which set it apart from traditional, invasive OEM-based methods. The system's ability to function across different valve types and brands makes it universally applicable, reducing dependency on brand-specific diagnostics. ValveStic also requires no external expertise, offering easy-to-understand automated diagnostics that streamline maintenance decisions. By enabling early detection of valve stiction and optimising maintenance efforts, ValveStic reduces operational costs and enhances the longevity and performance of control valves, giving industries a cost-effective and efficient solution.