ARIKI

Automated remote inspection of critical infrastructure using intelligent camera systems

ARIKI is developing and testing a system for automated remote monitoring of water supply, treatment and disposal plants At the forefront of this project are intelligent camera systems that independently capture defined areas of the critical infrastructure and use edge computing to perform on-site AI-supported analysis of the data collected. Conditions and damage are detected and the persons in charge notified in real time.

Market perspectives and product claims
In Germany, around 6,000 companies are responsible for the continuous transport, treatment, storage and supply of drinking water, rainwater and wastewater. They operate many thousands of plants for these purposes, including wastewater treatment plants, pumping stations and special structures, such as stormwater overflow basins. In future, it will be possible to automate the monitoring of these plants using the system developed by ARIKI. Thanks to the intelligent camera and the integration of an edge-based AI module into the system, the need to travel by car to plant inspection sites is minimised. Furthermore, the need for manual visual inspection of monitoring images is eliminated.

The automation of routine inspections decreases workload for highly qualified staff. Plant operating costs can be significantly reduced as a result. At the same time, eliminating car journeys cuts the CO2 emissions resulting from operation of the infrastructure. Continuous monitoring is another benefit of automated remote inspection supported by AI. It also avoids risks to staff in hazardous situations, such as heavy rainfall events. Incident prevention is improved and public protection optimised.
The system developed in this project is being piloted in a water management plant but could potentially be transferred to other fields of application within the area of critical infrastructure (e.g., wind turbines, substations or power plants). ARIKI is therefore developing a digital operating concept aimed at enabling centralised remote monitoring of decentralised plants with a reference architecture for Germany and Europe.

Challenge and innovation
Modern societies rely on functioning water management systems. This requires continuous monitoring and maintenance of various plants for water supply, treatment and disposal. At present, site visits are required to inspect this large number of plants, most of which are in remote locations. The planned ARIKI inspection system integrates intelligent devices and machines and is thus part of the Internet of Things (IoT) ecosystem. One feature of an IoT system is that it produces large volumes of data. Huge computing capacities are required to process this data centrally. The transfer of data from decentralised sources as in the use case described (i.e. inspection of plants that form part of the critical infrastructure) reduces work speeds and presents significant challenges in terms of IT security and data protection.
ARIKI uses an edge-based IoT solution to overcome these challenges. It enables direct on-site analysis of large volumes of data (e.g. from cameras). For the solution, ARIKI developed an edge computing module for cameras and an AI model capable of analysing image data in accordance with inspection requirements. The AI model has to be trained for various special tasks, depending on requirements. For this purpose, ARIKI uses AI training methods such as transfer learning, which enables transfer of the knowledge acquired in solving one specific problem to the solving of a different problem. When connected to a monitoring system designed for the specific application, the solution enables the definition of focal points for the inspection of each plant and continuous automated monitoring by the camera technology.

Use cases
The end-to-end solution is being piloted and tested at a wastewater and water treatment plant operated by an associated partner. A particularly important aim of the pilot phase is to demonstrate how the camera system, AI module and monitoring platform interact with one another and to determine whether any adjustments are needed.

Consortium
Wahtari GmbH, Schirra IT (consortium leader), Ingenieurbüro für Wasserwirtschaft und Ressourcenmanagement GmbH, August-Wilhelm Scheer Institut für digitale Produkte und Prozesse gGmbH; Austrian partner: Software Competence Center Hagenberg GmbH

Benefits