A smart alarm management project led by Northumbrian Water and funded by the Ofwat Innovation Fund has completed its architectural design, its AI roadmap and the first draft of industry alarm management guidelines. The partners describe these as the project’s first three major milestones.
The Smart Alarm Management (SAM) project is a collaborative Catalyst project. It aims to improve how control room alarms are triaged, prioritised and actioned using artificial intelligence (AI), machine learning and data modelling. According to the partners, the tool will help controllers identify root causes more quickly and make faster, better-informed decisions. This is intended to help prevent incidents such as flooding, pollution and water supply interruptions.
Why water control rooms need smart alarm management
Water company control rooms monitor networks round the clock and act as central hubs for operational incidents and crisis events. Every day, control rooms across the UK receive thousands of alarms. These alert controllers to issues including pressure changes, equipment faults, unusual flow rates, rising sewer levels, water quality anomalies and power failures.
Regulatory requirements have increased the need for quicker responses, and the technology has advanced. As a result, the number of sensors feeding data into control rooms has grown. The project partners said this can make it difficult for controllers to identify which alarms signal the most serious risk and need the most urgent attention.
H2O Global News has previously reported on smart infrastructure monitoring of water storage tanks.
Grouping alarms by root cause
SAM will use machine learning and data modelling to prioritise the most important alerts and group them by root cause. It will also gather and present relevant insights and recommend what action to take, based on historical data.
Under this approach, an event that might once have generated 20 separate alarms could be presented as a single, consolidated item. That item would include a priority ranking, historical insights and recommended actions. An earlier proof-of-concept reduced alarm volumes by up to 75%, according to the project.
Technical design and draft guidelines completed
The project has two workstreams. The Technical workstream is building an intelligence layer from existing alarm data. It has now completed its architectural designs and AI roadmap. These set out how the SAM tool will be developed, its technical requirements and a plan for features such as grouping alarms by root cause and automatically recommending actions in response to an alarm.
The Alarm Management Guidelines workstream is establishing standardised ways for the industry to manage alarms and setting out how AI can be used most effectively. Its work is based on alarm standards developed in the oil and gas sectors. The first draft of the guidelines will be shared with partners for review, then with the Water Control Room Forum and the wider industry for comment.
Partners and funding
Northumbrian Water leads the project in partnership with technology consultancy Softwire, Wessex Water, Anglian Water and the Engineering Equipment & Materials Users’ Association (EEMUA). The Ofwat Innovation Fund awarded the project £1.6 million in spring 2026.
The partners-in-kind are United Utilities, Scottish Water, Welsh Water, Affinity Water and South West Water, and they will be the first to roll out the technology. Spring Innovation is the knowledge-sharing partner. Spring was launched in 2021 to support delivery of the sector’s 2050 Water Innovation Strategy and is backed by all UK and Irish water and wastewater companies.
Controllers from the partner water companies will support development of the tool. They will provide access to their alarm data and give feedback on the guidelines as they are drafted.
Angela MacOscar, head of innovation at Northumbrian Water Group, said: “This collaborative project is set to make life easier for control room teams by turning complex alarms into clear, helpful insights. It helps teams quickly understand what’s happening, why it matters and what to do next – allowing faster actions to be taken through an approach that’s designed to be consistent across the sector.”
She added: “July’s launch event really helped shape the project’s delivery approach. Since then, the team has made great progress in achieving its first major milestones, moving us closer to creating a system that can be adopted by all water companies, helping them work more efficiently to deliver better outcomes for customers and the environment.”
Model development in the next phase
The project launched at the Northumbrian Water Innovation Festival in July 2026. It is now entering its next phase, which will focus on developing the AI and machine learning models. The partners will also refine the first draft of the Alarm Management Guidelines with the wider industry.
Spring will host an interactive workshop for the project in February.
Frequently asked questions
What is the Smart Alarm Management (SAM) project?
SAM is a collaborative, Ofwat Innovation Fund-backed Catalyst project led by Northumbrian Water. It uses artificial intelligence, machine learning and data modelling to improve how water control room alarms are triaged, prioritised and actioned.
How much funding did the SAM project receive?
The project was awarded £1.6 million by the Ofwat Innovation Fund in spring 2026.
Which water companies are involved in SAM?
Northumbrian Water leads the project with Softwire, Wessex Water, Anglian Water and EEMUA. United Utilities, Scottish Water, Welsh Water, Affinity Water and South West Water are partners-in-kind and will be the first to roll out the technology, with Spring Innovation as the knowledge-sharing partner.
How much can AI reduce control room alarm volumes?
An earlier proof-of-concept reduced alarm volumes by up to 75%. An event that once generated 20 separate alarms could be shown as a single consolidated item with priority ranking, historical insights and recommended actions.
What happens next in the SAM project?
The next phase will focus on developing the AI and machine learning models and on refining the draft Alarm Management Guidelines with the wider industry. Spring will host an interactive workshop in February.
