industrial-control-room-getty Getty Images

How to Bridge Asset Performance Data with Key Business Metrics using AI

Managing asset data for large industrial companies can be a clunky, error prone process. Involving artificial intelligence can proactively cut waste and reduce your data-induced headaches.

Most industrial service providers today have process analysts monitoring their assets in the field. Each process analyst is often responsible for maintaining assets across 12 to 15 different plants using a variety of software systems that house data in silos. These analysts and service representatives might have years of on-the-job experience at their disposal, but they struggle to monitor the performance of those assets with the dual challenges of limited human resources and clunky legacy tools. Because of that, they rely on spreadsheets and phone calls to get the information they need. This manual process is error prone and leads to many inefficiencies. Not only that, it also takes them longer to get to the root cause of an event, like an energy spike or unexpected downtime. This leads to higher costs and lower productivity.

 As a result, analysts are unable to link all of the data together, leaving unanswered questions that impact their revenue metrics. These questions include:

●      What’s the percent change in asset energy consumption month to month?

●      How does a particular asset parameter impact the revenue generated from that asset?

●      How does each operational activity affect the life of the asset?

●      How can I increase my throughput efficiency?

●      How can I reduce my downtime?

Industry executives and operators face daily challenges and pressure to do more with less and reduce resource consumption. There is immense potential for artificial intelligence (AI) technologies to help companies achieve those goals and accurately answer the above questions to mitigate rising costs and increase profit margins.

 Uncovering blind spots with real-time and historical data

Industrial processes are managed using multiple software platforms -- such as SCADA, ERP and CMMS -- which are generally hosted on a mix of aging, on-premises infrastructure and legacy asset performance management (APM) systems. Multiply this scenario by numerous facilities across regional, national, or even international footprints, and you get an idea of the challenge facing today’s process manufacturing entities – even those that have already embraced automation.

For example, beverage companies have to rely on different sources of water to produce their products. In order to ready that water to produce the beverage, they have to treat it to the required level before they can use it for production. The quality of the feed water depends on the location of the plant. Even within the same plant, the quality of the feed water can change over the year, depending on how much water is drawn from each source. In order to effectively manage the cost of treating feed water, you have to carefully track the performance of the assets that are used in the process. Legacy systems make it difficult to access the data and track performance. These systems often cause blind spots in industrial processes that are major contributors to waste and inefficiency, leading to revenue attrition.

While there are standardized best practices for leveraging asset data, APM is still in its early days, and every organization seems to have its own approach. Some companies rely on reactive APM, running to fix assets when they fail, but otherwise not managing their usage or health. Others are proactive, in the sense that they manage assets based on usage; these companies service equipment on set time tables, whether the assets need attention or not. More sophisticated teams listen to asset signals and respond accordingly.

Listening to asset signals involves keeping track of data and understanding what actions need to be taken. For example, a food manufacturing company needs to maintain their throughput efficiency even when their assets consume more energy than required. If energy consumption spikes too much, they have to shut down the asset and fix it before resuming production. Energy is the most expensive part of operations, so organizations have to make sure they can predict energy spikes and take necessary precautions to prevent them. This will enable them to maintain cost efficiency of the throughput and maximize uptime as well.

By leveraging historical and real-time data within these systems, AI can uncover blind spots by connecting data and business metrics to deliver crucial insights automatically and continuously. AI can help build a bridge between asset behavior and future revenue by creating KPIs that link to key business metrics. With cloud-based APM solutions, analysts can make more informed decisions to achieve higher throughput efficiency, increase revenue retention, and reduce total cost of operations over the asset lifecycle.

Creating proactive industrial plants

The inability to access data from industrial processes forces executives to rely on experience or rigid assessment procedures. Both approaches can lead to misjudgments that result in wasteful spending and disrupt the manufacturing process. If they rely on experience, they will often fall back on intuition as to how to run a particular asset. This usually results in conservative settings leading to excessive resource consumption. On the other hand, right assessment procedures employ static rules to evaluate a given situation, which leads to suboptimal judgments about what needs to be accomplished.

AI offers the ability to identify and correct this issue. AI-driven Industrial Internet of Things (IIoT) platforms provide process analysts with real-time data from the manufacturing process, which allows them to be proactive instead of reactive. Managers can head off real issues rather than throw money at problems that may not actually exist. Additionally, plants will be better able to craft new products tailored to specific customer demands, all of which drives increased brand loyalty and market differentiation. This will hugely benefit the companies in this field that are rapidly transforming into software-centric, recurring revenue businesses.

Ultimately, decision makers can employ data-driven approaches instead of unreliable tests and static rules of thumb. Sensor data generated at industrial plants will help predict trends and measure KPIs, connecting it all back to business metrics, which can be used to increase efficiency and maintain profit margins.

Prateek Joshi is an artificial intelligence researcher, an author of eight published books, and a TEDx speaker. He is chief executive of Plutoshift, which provides asset performance management for industrial process facilities to increase energy efficiency while maximizing throughput.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish