top of page
  • Writer's pictureAndrew Lewis

14 Risk Factors that Cause ROI Losses in Your PdM Program

Updated: Feb 23, 2023

High ROI doesn't come easy... Focus on these key risk factors to prevent losses.

Done well, an Industry 4.0 Predictive Maintenance (PdM) program will transform the performance of your industrial operation. But these do not come easy. We outline 14 risk factors that will reduce project success. In our previous blog post we outlined 10 ways to boost your ROI. These two post contrast both sides of the same coin - optimizing ROI in AI projects.

WHY IT MATTERS: We have demonstrated that a well planned and executed PdM program can produce ROIs as high as 60x across 234 assets. But this takes careful planning an execution and advoidance of the pitfalls outlined below.

BACKGROUND: Return on Investment (ROI) is calculated by dividing the profit earned on an investment (or cost) by the cost of that investment. This is the key KPI to determine the effectiveness of your POC and deployments, for a single asset or a group.

ROI = Profit/Cost

THE DETAILS: Your target ROI should be at least 10x to make the project worthwhile. As your PdM skills grow you can easily double that. But these factors will lower your ROI. They are best managed in the POC and deployment of each asset, but they can also crop up at anytime.

Decrease Profits, Lower ROI:

  1. Improper Asset Selection – Infrequent alarms, common in high availability operations. Look for excessive predictive maintenance (PM) cycles that prevent assets from failing. Look for asset selection that takes these factors into consideration before the sensors are attached.

  2. Weak AI Models – These models have a very high percentage of false positives and false negatives. Many factors cause weak models, including: sparse historical data to build models, high variability of operation not related to asset failure, failure modes not detectable with attached sensors. Look for asset selection that takes these factors into consideration before the sensors are attached.

  3. Failed Battery – This scenario prevents a failing asset from alarming. Look for platforms that will alarm a failing battery before it stops working so a replacement can be scheduled. Look for PdM platforms that will alarm degraded battery voltage.

  4. Sensor Drops Off Line – sensors can get pulled off for several reasons including accidental dislodging, sensor intentionally moved to another asset, or the sensor-Bluetooth connection is lost. Look for PdM platforms that will alarm sensor drop-off.

  5. Wins (Profits) Not Reported – No structured system to document Wins from the PdM program. Contributing to this: technicians may feel threatened by the AI system. They will take credit for wins through existing methods. Look for a simple documentation process, better training, education that shows how jobs will become more interesting solving fundamental problems.

  6. Poorly Equipped Maintenance Staff – technicians must know how to use and maintain predictive maintenance tools can be costly in terms of staff time and training costs.

  7. Premature PdM Project Cancellation – If returns are low will leadership can abandon the project before results can be developed. Look for informed leadership committed and well briefed for the journey ahead.

  8. Short Term Need for PdM Evaporates – Excessive plant outages, extended planned downtime, reduces adverse impact of unplanned downtime because many unplanned repairs will spill into planned downtime. Look for informed leadership committed to building PdM solutions for long term ROI.

Increase Cost, Lower ROI:

  1. Data Science Staff – the PdM platform requires human curated machine learning (ML) model building. Staffing scales as more sensors are added. Look for next generation PdM platforms that have ML automation to virtually eliminate the requirement.

  2. Maintenance Staff – Each alarm requires inspection by a technician to make complex decision if asset actually failing (alarm valid) and estimate how long before failure. As sensors are scaled inside a site, alarms will scale up placing a heavy work load on maintenance staffing. Look for PdM platforms that provide alarm diagnostics. These will make inspections more productive and help less experienced technicians

  3. Multi-disciplinary team – A requirement for a complex, highly customizable PdM platform. The team is unable to work as a cohesive unit, solve problems effectively and make deliverables. Look for team leadership that helps each discipline not understanding limitations of their role in the whole project. Look for next generation PdM platforms that have eliminated this requirement.

  4. Missed Failures – Sensors do not cover every kind of failure, eg, failure of a pressure line. PM cycles are still required to inspect for these. Look for new sensors types, eg, acoustic, thermal, to be added to the PdM platform.

  5. Missed Critical Failures – Catastrophic failures that result in significant costs. These can be hidden in a relaibiltiy program that shows no corrective or maintnenace work orders, no unplanned downtime. Look for a better asset selection process to screen out these assets for PdM.

  6. High Platform Costs – Sensor, platform and cloud storage costs. Look for a next generation PdM platform that has lowered these costs by 10x and more.

ZOOM OUT: In each of the above ROI reducing factors, Look for work arounds to each of these to minimize or eliminate these factors. These factors must be monitored for the long term.AI-based deployments are not static. Results change over time due to variability of human behavior and equipment aging making the AI models less relevant.

THE BOTTOM LINE: AI-based predictive maintenance is already having a significant impact on plant operations. It is imperative to manage your PdM platform on an on-going basis to ensure continuous high ROI. The more ROI you can demonstrate, the more budget you will secure for scaling the PdM at multiple sites.

NEXT POST: Your Monitron POC, How to Execute for Maximum ROI is a blog about building AI solutions for augmenting decision making and empowering people that make them. It's authored by the engineers at DecisionIQ.

ABOUT DECISIONIQ: We are "boots on the ground" factory engineers expert at the adoption of AI and machine learning into operations. As consultants and system integrators we bring our experience in a mix of a well structured programs that enable our clients to produce winning results and maximum ROI. You will learn to use AI to build competitive advantage by increasing plant uptime, quality and yield. With 24 POCs and 16 deployments we have achieved for our client’s ROIs as high as 60x and a cumulative $22 million in savings.

AN AWS PARTNER: We provide capability in design of AI solutions for industrial applications. We work with the AWS Industrial IoT services stack, cloud migrations and Amazon Monitron POCs and deployments. For qualifying customers, many of these capabilities are eligible for AWS subsidized funding.

©Copyright 2023, DecisionIQ, Inc.

56 views0 comments


bottom of page