Acquirium: Toward Interoperable Data Driven Applications in Water Treatment Systems

Water treatment systems face increasing pressure for automation driven by decentralization trends and regulatory requirements, yet deploying the necessary data-driven processes remains difficult due to fragmented data sources, unstructured metadata, and lack of interoperability across facilities. We present WaTr, an ontology designed for water treatment operations that captures fine-grained treatment processes, substance flows, and external references to heterogeneous data sources, and Acquirium, a co-designed data management platform that enables metadata exploration without requiring expertise in semantic query languages. WaTr recasts and significantly extends existing building ontologies to support a wide array of water treatment technologies and concepts. Acquirium supports a novel dataframe-style API for progressive graph query refinement and a lightweight “soft-sensor” abstraction for deploying portable analytics. We demonstrate that Acquirium reduces application code by up to 4 times compared to existing platforms, while scaling to hundreds of concurrent processes and complex soft-sensor compositions with sub-second end-to-end latency. These contributions lower the barrier to developing interoperable, data-driven applications for water treatment facilities.