Scaling Asset Performance Management Across the Enterprise

Overview

ARC Advisory Group recently conducted a global survey of asset performance management (APM) best practices. Survey results included 180 respondents from industry practitioners and several one-on-one interviews with industry end users from the energy, chemicals, food and beverage, and mining industries. This research was conducted to provide a better understanding of how industry leaders address APM scalability. An APM system is considered scalable when it doesn’t need to be redesigned to maintain effectiveness during or after deployment across a variety of unique assets, or across multiple plants. ARC found that 17 percent of survey respondents cannot scale APM across different assets at the same plant, regardless of resources. A staggering 53 percent of respondents also indicated that their APM deployment was scalable, but not without tremendous effort and resources. Lack of data science alignment was also cited as a top challenge.

Clearly, there is need to address the resource-intensive nature of APM projects and heavy-customization of APM solutions. Considering the fact that most of our respondents already have an APM program, there was a surprisingly low adoption of machine learning (ML) and artificial intelligence (AI) tools. To help improve the scale out of APM, leading industrial users are:

• Addressing the resource-intensiveness of many APM solutions by minimizing the customization of APM implementations.
• Creating better harmony between APM project teams, the maintenance organization, and with the data science organization.

What Is Asset Performance Management?

Asset performance management (APM) is an approach to managing assets that prioritizes achieving business goals in addition to traditional asset reliability and availability goals. Industrial companies rely on APM as one of the primary enablers of digital transformation. APM optimizes the performance of physical assets in their operating ecosystem, typically employing a digital thread throughout the asset lifecycle, supporting digital twins for assets or asset groups, supporting connected workers, and maintaining a network of parts and service providers. By leveraging an ecosystem of data from connected assets and applying digital models, advanced analytics, and machine learning, modern APM solutions seek to support market or customer-driven production intelligently and sustainably by improving asset reliability and availability while reducing risk and cost.

Business goals: Through digital transformation in asset and risk management, and improvements in asset availability and uptime, users achieve higher revenue and profitability while improving customer satisfaction with on-time delivery and quality.

Asset ecosystem: The ecosystem for assets extends beyond the plant floor and facilities to include a wide range of applications across asset-intensive industries that leverage sophisticated OT, IT and engineering systems and related production, maintenance, and engineering personnel. It also encompasses third party partners and other providers of parts and services.

Digital tools: Apply modern technologies across a range of tools like smart devices, augmented reality, and mobility to improve business processes and create new methods for asset management.

Data & analysis: Enable greater depth of collaboration across the asset ecosystem by using digital twin, digital thread, and other modern information assimilation and management approaches.

Practices and apps: Traditional practices and applications become more effective when enhanced with data, digital tools, and support for business goals.

Challenges Companies Face in Rapidly Scaling APM Solutions

Given the breadth and complexity of APM solutions, it is not surprising to find the lack of resources, data science alignment, customization, and domain expertise to be top challenges when scaling APM solutions. With the recent sharp rise in machine learning (ML) and artificial intelligence (AI) solutions in the industrial market, asset owners and APM practitioners face the challenge of addressing the vast number of vendors going to market without industry expertise. Asset owners expect that APM ML and AI solutions are proven and adaptable to various asset classes and processes - and these solutions will not be used as a training or test project to prove some generic AI solution.

APM vendors often go to market with generic asset profiles or templates. While these are helpful to look at the asset performance in a testing or OEM facility, they don’t help much to help bridge the gap between a discrete asset (pump, chiller, exchanger, column) and how the asset operates in the context of the process it supports. A strong connection between process models and asset profiles is required to reduce the time needed to build models and copy-paste them to other asset classes.

According to APM practitioners, it is still quite difficult to copy and paste models from one plant to another plant or site. There are many unique characteristics to the assets at each location. For example, even process licensors have subtle variances in processing technology. Each process has differences in constraints, limitations, and operating procedures and is subject to many different process disturbances, making it tricky to adapt and scale an APM solution. It is a resource-intensive activity to gather the tags manually associated with an asset. Most of this work is a manual process, which must be done before any ML is applied. When the asset is a pump, each pump has a unique signature depending on the environment and its operation. The chart below illustrates the inability of APM practitioners to scale out APM effectively. Seventeen percent cannot scale APM across different assets at the same plant and a staggering 53 percent can scale APM but require tremendous effort and resources.

ARC Advisory Group clients can view the complete report at  ARC Client Portal

If you would like to buy this report or obtain information about how to become a client, please  Contact Us

Keywords: Scaling Asset Performance Management, Machine Learning, Data Science, ARC Advisory Group.