Enterprise Asset Management
Blog: Capgemini CTO Blog
Let’s dive deeper into asset reliability
Previously, we shared notable differences between asset reliability and asset integrity and why differentiation is essential. This point of view will dive deeper into the realm of asset reliability, where we will share our insights on how to differentiate and identify your critical assets and the necessary data they generate. This POV will strive to ensure you can monitor and analyze your assets to predict failures before they affect your business-critical operations. As we shared previously, the study asset reliability focuses on rotating assets, typically machinery that has a “heartbeat” or machinery in motion. We believe a successful automated reliability program starts by identifying your critical assets using proven methodologies, determining their essential data points and parameters, and selecting the right amount and frequency of data to collect.
Identify your critical assets/separating the vital from the trivial
Not all equipment, and the unique data generated, needs to be connected, monitored, and analyzed within an asset reliability program. We say this because the cost to monitor ALL equipment and ALL heartbeats would be far too high, and the value never realized. The total cost of ownership (TCO) for such an endeavor would include installing sensors, fine-tuning the sensors, network connectivity needs, data strategy and storage needs, systems to make sense of the data, and the sheer maintenance of this ecosystem. So how do you focus your efforts on your critical assets? We recommend you perform a thorough review of all your equipment. This initial review will help you select the “right” assets to be included in your program. Your organization may already have an established, reliability-centered maintenance
(RCM) program that can be the starting point for the identification triage. We believe the asset selection team optimally consists of individuals from process controls, mechanical and reliability engineers, operations and maintenance personnel, and OEMs all working together.
Effectively done, this group will identify the most critical equipment based on your organization’s predefined business and HSE criteria. This review will yield the equipment listing and the components at the highest risk of failure, providing the most value by systematically being connected, monitored, and analyzed. We suggest performing this review periodically so that additional equipment pieces are included in this program as your organization progresses along with its connected reality.
Your assets’ relevant data
Now that you have identified your critical assets and their components, it is time to focus on their data. Machinery complexity has grown in the last few decades – more types and more specialization. Pumps and compressors must operate within tolerable vibration to achieve throughput. Fundamental to this is the reliance on condition monitoring, assessing the performance to deliver expected outcomes, which requires data.
Data is, or should be, the valuable feedstock that drives any useful analysis. Sensors today measure variables such as pressure, throughput volume, vibration, and temperature and produce data many times per second. These sensing devices can be hardwired into field automation or be connected to field wireless networks creating an internet of things (IoT). The use of methods, such as RCM, will adequately guide these devices’ installation nature, defining components, the physical variables, and the precision and frequency of which these measurements are to be collected.
Over the last ten years, access to data has dramatically improved. With cheaper and greater access to data, deeply rooted data numbness or skepticism has also set in.
Many engineers or control room operators (CROs) routinely question the data streaming towards them, as the automation equipment can be unreliable or unmaintained. But, overwhelmingly, those concerns about the data are unfounded, as most of it is accurate with time.
With so much data at our fingertips, using and leveraging it becomes the next valuable challenge. There are hundreds of analysis tools available that help make sense of this data deluge. Some is fit for a specific purpose while some may have broader breadth but may lack fit-for-purpose intricacies. These tools can create user notifications in the most basic form, such as when a threshold value is breached. Basic business rules can provide CROs real-time assistance as they monitor hundreds of pieces of equipment and dozens of processes. Most tools today allow more complex algorithms and calculations, but more subtle, hidden problems usually go undetected when just using this business-rules approach. Artificial intelligence (AI) provides a significantly improved system to evaluate this data, identifying previously hidden correlations in the variables. Now, suddenly, we can see the unseen!
In addition to identifying curious anomalies, the most useful aspect of AI, and the related machine learning, is to predict failure at a point far before manifesting real impacts. The goal of these tools is to “find a lemon before it becomes a lemon.”
Using the data for action/Integrated towards action
Using the newer analysis tools has never been easier, as they are built with users in mind – tools that provide all the required functionality but integrate with other sources for a seamless experience. Now, master data can move into the monitoring apps to use the exact equipment data tags of interest. Analysis recommendations and work order creations are only a step away, with just the touch of a button.
But there are many companies or disciplines within the organization that aren’t ready to embrace these technologies. Data gathering feels slip shot, and analysis is highly dependent on the engineer or CRO. Outcomes are highly variable. Starting anew feels like a burden. But worry not, those that have suffered before you can provide a roadmap to improvement. And it begins, like in many other situations, with an honest assessment of the current state, coupled with a bold view of the future. Consider using an assessment tool, and skilled guides and steady hands to help along the journey. Here is where we can help.
WHO WE ARE:
The act of monitoring an asset is not new. We monitor, analyze, and predict most days without thinking about these acts. The simple surveillance of reading the gauges on our car’s dashboard or checking oil levels, or feeling how the vehicle’s suspension system reacts over potholes and tight turns is a data-gathering exercise brains act as the analysis and prediction engines. But with today’s ever-increasing equipment complexity, our minds can no longer do all the work, and we must rely upon more sophisticated technologies to help us “find a lemon before it becomes a lemon.”
When I was a young engineer, in a small central California oilfield, we performed all of the analysis in “plotbooks” populated with field oil well production data from our fleet of technicians. I would get a single production value each month, and a dozen for the full year. A dozen data points on valuable assets seem unheard of by today’s standards. Measurement devices are much less costly, as are communications and data management components. How did we ever get the job done way back then?