Anomaly Detection: How Uniper has Implemented this Technique Alongside Existing CM Techniques

Omoruyi Onobhamiukor, CEng MIMechE, Lead Technical Consultant, Uniper Technologies Limited


Uniper’s Condition Monitoring Team has a long history in the vibration analysis of large turbine generator assets, with remote vibration support provided to over 100 assets over the past 25 years.  Uniper has taken a hands-on approach when it comes to developing condition monitoring solutions; often developing in-house software tools to meet our exact needs. These solutions originally focused on enhancing the vibration analysis of critical rotating assets, but has moved heavily into analysis of plant process data over the past 10 years.

Generation assets have invested in plant data historians which provide a business with safe secure access to all plant datasets.  Through successive market challenges the industry now operates with far less staff, resulting in the available data being used predominantly reactively instead of proactively.  Uniper chose to develop an anomaly detection solution for this dataset as the tools on the market at the time either didn’t provide the complete toolset or where deemed too costly for large scale deployment.  The solution, SpheriCAL and the Advanced Condition Monitoring Service it supports are often seen as a replacement to conventional condition monitoring techniques, but Uniper has always recognized the importance of ‘the right tool for the right job’ and has focused on integrating this alongside the other CM techniques.

The insights gained from monitoring plant process data have proven valuable in recognizing the onset of faults earlier than previously possible, often preventing the previously normal consequential damage which may be detected through other monitoring techniques.  This capability has forced Uniper’s Condition Monitoring team to rethink the combined overall condition monitoring philosophy, significantly expanding the range of assets and systems covered and ultimately improving the overall asset health assessments.  This presentation will include case studies highlighting the importance of integrating condition information from different techniques to provide the asset owners with the most clear and confident picture of their plant.


  • Insights into how anomaly detection can be implemented
  • Importance of integrating new techniques alongside existing monitoring solutions
  • Real life case studies


Omoruyi Onobhamiukor is a Mechanical Engineer with a Master’s degree from The University of Nottingham and a Chartered member of the Institution of Mechanical Engineers. For the past eleven years, he has been involved in developing condition monitoring techniques and strategies at Uniper and is currently technical lead of Uniper’s Condition Monitoring Service which provides remote on-line and in the field support to international power and industrial clients by incorporating data from monitoring systems like the Beran and historian data from PI in diagnosing and assessing the condition of plant machinery. Data driven techniques are at the centre of his team’s operations, but maintaining a strong focus on delivering actionable insights remains a clear and core strategy.