en.Wedoany.com Reported - U.S.-based RLCore has launched the RLTune real-time continuous optimization platform, which adds a continuous learning intelligent layer on top of existing control systems in water and wastewater facilities. By leveraging constrained reinforcement learning to dynamically optimize operational performance, it reduces chemical and energy consumption by 15-25%, improves response time by 95%, and achieves process efficiency exceeding 90%. The company showcased this product at the American Water Works Association (AWWA) ACE26 conference.
Industrial systems operate in dynamic environments with constantly changing conditions, including fluctuations in energy prices and chemical costs, variations in influent water, equipment wear, and personnel constraints. Traditional control methods often rely on fixed gains or models that cannot learn from the environment, requiring operators to manually manage the resulting gaps, leading to operational inefficiencies and lost optimization opportunities. It is estimated that controllable inefficiencies in all industrial processes result in annual losses exceeding $1 trillion.
RLTune sits on top of a plant's existing control stack, applying constrained reinforcement learning to continuously improve control decisions under real operating conditions. The platform learns from the real-time plant environment, continuously optimizing industrial processes to achieve operator-defined plant-level KPIs. Shelley Terry, General Manager of Drayton Valley Infrastructure, stated that after the collaboration, unexpected events in the plant control room decreased, chemical usage dropped, and water savings increased, allowing operators to focus on higher-value work. Frank Mannarino, Senior Vice President of EPCOR Water Services, noted that this approach demonstrates how to introduce advanced control and artificial intelligence in a manner consistent with utility operations.
Key features of the platform include: continuous learning and adaptation, automatically adjusting to seasonal changes, influent variations, equipment wear, and process disturbances without the need for manual retuning loops; no requirement for digital twins or complex physical models, learning directly from the actual plant environment; optimization with guardrails, allowing operators to set boundaries and define progressive autonomy while retaining override authority; data logging to improve operational visibility and quantify variability and operating patterns; vendor independence, compatible with SCADA, DCS, PLC, historians, and IoT gateways via OPC-UA; and support for on-premises deployment, keeping data on-site.
RLCore categorizes this as Real-Time Autonomous Optimization (RTAO). Founded in October 2024, the company comprises internationally recognized reinforcement learning experts and product and technology leaders with experience in scaling. The RLTune platform is deployed in municipal, wastewater, and industrial facilities, helping operators build resilience, reduce chemical and energy consumption, improve process stability, and adapt to changes in real time.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
