I am still talking about Emerson Exchange, and have a few more to go. This post is about analytics. Jonas Berge, Senior Director, Applied Technology, Plantweb, Emerson Automation Solutions, has often supplied me with great insight usually about networks in the past. We chatted briefly at Exchange and then followed up with email conversations. In this one, he talked about analytics.
Digital Transformation has a foundation in data. Data is useless without a formal way of thinking about it. There are two kinds of analytics tools.
We are left with two tasks. We must first understand the two types, how they are derived and their strengths and weaknesses.Then we choose the right analytics tool for the problem.
There are principles-driven tools and data-driven tools.
One must remember that advanced predictive techniques can only be practically applied to a subset of use cases.
An over-emphasis on one approach means companies won’t position themselves to capture all the potential benefits.
When factoring the effort and expertise required to develop accurate machine-learning models, remember most organizations already have systems in place to record maintenance- and reliability-related data, but the effectiveness of such systems can be undermined by poor housekeeping. The same assets or issues may be described in different ways in different systems, for example, making integration difficult. Companies may use free-text fields to record issues or maintenance actions, making automated search or data analysis harder. Or critical data may be inaccessible, locked away in spreadsheets or on paper notes.
The application of machine-learning techniques to monitor asset condition has already received considerable attention, even though their cost and complexity may ultimately limit their application.
When a machine is prone to a narrow range of well-understood failure modes, it is often possible to address a potential problem in a simpler way, for example by monitoring the temperature or vibration of a component against a set threshold.
Model-based predictive maintenance becomes a breakthrough way to solve selected high-value problems. This approach has the most potential where there are well-documented failure modes with high associated downtime impact, for example in a critical machine on a larger production line.
Root-cause problem solving, using approaches such as fault-tree analysis as well as cause-and-effect or failure-modes-and-effects analysis (FMEA), is a fundamental part of any organization’s maintenance and reliability strategy.
Not all condition-monitoring techniques require elaborate algorithms or complex models, however. Data-driven condition-monitoring approaches use simple queries that are run periodically or in real time against time-series data generated by machines and external sensors. If threshold conditions are passed, these systems can trigger investigative or corrective action in the digital-reliability-engineering workflow, or directly to maintenance execution.