In November 2014 the Business and Industry Section (BIS) and the Institute of Asset Management ran a joint event, 'Statistical challenges in asset management', focusing on case studies from the water industry.
David Smallbone (BIS) provided an introduction on the role of statistics in good practice asset management, including an overview of the new international standard ISO 55000 and guidance set out in the Institute of Asset Management’s ‘Asset Management – An Anatomy’. This was used to set the context for the two speakers, both of whom had worked closely with water industry clients to develop statistical methods for modeling and predicting asset degradation to support planning of interventions.
Dr Atai Winkler (Overbeck Analitica) presented on ‘Using Survival Analysis to Model Asset Failure and Replacement’. This covered how survival analysis, a method usually used in medical statistics, has been successfully applied to model asset failure and replacement. The case study took intervention and asset data covering a 12-year period from a major UK water company. The talk discussed the deterioration curves that had been fitted for different types of asset descriptor, such as manufacturer. Dr Winkler then demonstrated how the Cox proportional hazards model was used firstly to identify assets at greatest risk of immediate failure (and therefore in greatest need of immediate attention), and secondly to simulate the financial consequences of various asset maintenance and replacement policies.
Following the talk, a number of questions were raised about the quality of the data available and the time taken to cleanse data before applying modeling approaches. The consensus in the room was that setting up data capture systems and obtaining sufficient, good quality data to populate these models was a common challenge across all industries.
Professor Jake Ansell (Edinburgh University) followed with a talk on ‘Optimal Asset Management’, based on the work he has covered in his career supporting several of the main water companies. In the water industry, regulation means that companies have to develop well-laid plans for asset management both in the short term (five years) and long term (25 years). The problem is the number of assets that each water company has and the modelling of the performance of assets over protracted periods. Professor Ansell gave an overview of the issues and described how the Whittle Index had been employed to achieve a £1bn saving for a water company in its asset management.
Following Professor Ansell’s talk questions were raised on how to best model deterioration and degradation. Professor Ansell cited a number of papers and also the Water Statistics User Group’s work in this area. The audience also discussed the methods used to estimate and assess the impacts and costs of asset failure, another common challenge across all industries when predicting the amount of funds required to manage assets and where this spend should be prioritised under competing constraints.
The event was well attended by people from both academia and business, with representation from a wide range of industries and asset management roles.