The Janus approach starts by assessing your organization’s readiness to implement a predictive maintenance approach. Predictive Analytics alone do not address resource risk but require adoption of a proactive maintenance culture, and a robust workflow that moves analytic insights along the process of detection, validation, investigation, action, and lessons learned.
We have been through large scale development and deployment of machine learning models for common industry assets and can guide you through the software sales pitch to what is meaningful to your business. We have been there and done that and know both the significant opportunity as well as the classic pitfalls. We recommend selecting an industry proven software platform as opposed to any bespoke data scientist models that come with long term sustainment challenges. Janus can help you pull back the curtains of machine learning tools, to remove any ‘black box’ feel. Its important that algorithms are fully understood by your equipment experts in order to have trust in any alarm or insight.
We can help you select the right equipment to monitor, build your business case, and assess your sensor data gap. We can even build the models using your selected platform by mapping tags to your asset specific model, cleaning your historical data, running the algorithms and fine tuning models on real time data. We will minimize false positives while ensuring your key failure modes have the best chance of early detection. Although we do not currently offer monitoring services we can ensure your monitoring team is set up for success and the change management and training on your new workflows are covered.
Predictive Analytics are mostly justified on reducing asset risk, by detecting defects on the safety and business critical equipment where functional failure results in large impacts. In most cases, predictive analytics programs pay for themselves in one or two significant finds that caught a critical defect before it became a critical business loss .
Maintenance cost reductions are typically the amazing fringe benefit with improved planning and reduced equipment damage, as well as a new way to validate the accuracy of your measurement devices in the predictive models.
Although less quantifiable, the establishment of a predictive workflow can help move the needle in your organization’s reliability culture. We have witnessed data analytics bringing a renewed sense of intrigue and passion for maintenance and reliability practices in organizations that may have struggled in the past.