Case Study: How Creating Video Game Hacks Helped Me Create Asset Failure Prediction Models

Abdul Basit Bashir, Reliability Engineer


Estimating the time left before an upcoming failure based on the asset’s current condition is desirable in many industries. Failure prediction models can be created to help make such estimations systematically for your assets. These models are datasets describing relationships between variables in asset’s operation, maintenance and condition data.

One of the approaches of creating hacks for video games which is based on data mining can also be used when creating failure prediction models for industrial assets. The technique involves identifying hidden patterns in the dynamic data produced when the game is running. The hacking approach starts from having no information about the data to making sense of how the variables relate with each other. The technique is used to intervene the game’s normal operations by actions such as renewing a game character’s life repeatedly. This objective to favour a game’s player has a similarity to the desire of an asset owner which is to intervene an asset’s typical degradation process. The purpose however is that instead of a game character’s life, it is to repeatedly renew an asset’s life.

This presentation will explain how the approach works. It will be described how the many condition data types can be taken advantage of using the same approach, including vibration, oil condition, ultrasound, electrical signatures and thermography.

Learning Takeaways:

  • An approach for creating asset failure prediction models
  • Importance of the data exploration required before any machine learning and artificial intelligence based models are created
  • Importance of integrating condition data from multiple technologies for predictive maintenance
  • Some of the prerequisites to be met for automating predictive maintenance
  • Some tips on how to meet those prerequisites



Abdul-Basit Bashir is a Mechanical Engineer specializing in predictive maintenance. He has studied through Master of Maintenance and Reliability Engineering program from Federation University, Australia. He is ISO certified in Vibration Analysis (Cat II) and Lubricant Analysis (Cat II). He likes to apply the most worthwhile and effective solutions against problems regardless of their popularity or complexity. Abdul-Basit lives in Perth and works at an alumina refinery as a Reliability Engineer.