Using advanced digital technology to predict potential incidents

Back

The Challenge

At Avove, we wanted to improve the efficiency of incident diagnosis and understand the potential impact and added value that advanced digital technologies – including smart data analysis and improved visualisation of data – can have on key business metrics. Process optimisation within incident diagnosis and management leads to improved operator data assimilation, decision-making and the quality of service provided to clients.

The Approach

We worked with the Virtual Engineering Centre (VEC) to look at ways in which simulation combined with large data analytics could be used in process development to predict potential incidents. Our research showed that this approach could be 30% more effective than conventional methods of diagnosing the urgency of a customer-raised order.

Benefits

Using an intuitive digital platform, we have increased flexibility and resilience and gained resource planning and productivity benefits across the repair and maintenance function.

  • Developed a robust predictive incident planning tool, that will improve the customer experience across the water sector
  • VEC overcame challenges with algorithms being used in a commercial solution to deliver productivity savings and benefits
  • We undertook a trial with sample data and the VEC tool; this resulted in a robust job scheduling system based on priority with improved SLA performance. The results included:
    • Job volume and location prediction
    • Quicker intervention times
    • Increased customer satisfaction
    • Diagnosis accuracy increased by more than 13%
    • Vehicle allocation accuracy increased by over 35%

Clients can apply methodologies to network infrastructures and run remotely, if necessary, for the analysis and prediction of behaviour and incidents.

The Solution

We worked with VEC to predict potential incidents and undertook a successful trial in Nottinghamshire to prove the concept worked. Clients are now using this system.

We created a web-based platform to improve incident diagnosis and resource allocation accuracy with real-time updates of customer incidents – adding strategic value to the process

  • The platform provided a patio-temporal forecast of job volumes using an Autoregressive Moving Average eXogenous (ARMAX) model, based on historical job trends and prevailing weather.This allowed the tool to accurately predict (50%) of the weather’s influence on incidents and improve how we predict the job volume of work.
  • We carried out a high-fidelity simulation considering the actual vehicle schedule, resource planning allocation, traffic and random job failure to estimate the key performance indicators (KPIs) of the customer decision support tool

Going above and beyond – an innovative approach to how data analytics can be used in process development to predict potential incidents

Contact us

We’re here to answer any questions you might have or to discuss your project requirements. Please select a subject related to your enquiry, complete the contact form and we will be in touch with you.