How to bolster asset management efficiency using analytics platforms

Asset management insights

• Subject matter experts are tasked with monitoring equipment, and artificial intelligence can be added to improve asset management.

• This article will explore common barriers that hold manufacturers back, and how to navigate and progress past the barriers to realize artificial intelligence benefits.

The enterprise industrial landscape is digitizing and evolving rapidly and asset management is no area of exception. Reminiscing with any seasoned process engineer is likely to recount how they used to “feel the vibration of the pump” in their feet, “hear the hum of the compressor” to assess operating efficiency or “waft the air in the mechanical room” to detect overheating motors.

While these sensory — and often acutely accurate, depending on the engineer’s experience and expertise — methods of monitoring equipment health were relevant within recent decades, they are outdated in a world that is undergoing a tremendous digital transformation. Industry has installed sensors everywhere, built data warehouses and established remote technical facilities for centrally locating engineers and other personnel, collectively subject matter experts (SMEs).

Additionally, the intersection of information technology (IT) and operational technology is becoming increasingly important and complex. Now, instead of relying on human senses to determine process health, artificial intelligence (AI) is tasked with continuously inspecting incoming field data. The promise is faster, more reliable, earlier and more accurate anomaly and upset detection.

Furthermore, the unceasing global increase in automation, paired with the ongoing struggle to hire enough skilled labor to meet demand, is necessitating that SMEs maintain responsibility for more industrial assets than in the past. The only way to successfully accomplish this is to rely on automated technologies, such as AI, to operate efficiently and maintain industrial equipment effectiveness.

Leveraging AI is dominating the conversation of process analytics and manufacturing optimization and major manufacturing corporations are eager to understand and leverage AI to operate in a more environmentally friendly manner, enhance safety and increase profitability. These companies are investing millions and even billions of dollars in the race to realize these benefits.

However, for most organizations, the promise of AI has yet to be fully realized.

Asset management jobs

Recently at a major multinational oil and gas company, teams conducted numerous efforts to derive business value from increasingly digitalized data. Operating in a large global corporation with hundreds of thousands of assets, the primary objective was to optimize plant operations worldwide.

Historically, experts were stationed at individual plant sites. While this was highly beneficial for the immediate plant site — ensuring an experienced SME was always on hand to address issues and teach plant personnel directly — it limited the sharing of knowledge across similar sites regionally or globally. Leadership teams recognized this shortcoming and, leveraging newfound digitalization trends, began creating a centralized team of experts to collaborate and support operations globally.

The immediate challenge was sharing learnings throughout the enterprise, answering, “How can we leverage our expertise and scale it worldwide?” Energetic and expert engineering and IT teams were eager to tackle this knowledge-scaling challenge. Instead of stationing an SME at each site, the team combined them at one location, relying on data connectivity, video-conferencing and other technologies to bridge the gaps with the sites.

However, the field data itself was not structured in the way working teams needed it to be. Correcting all the imbalances and asymmetries in the data frequently required human intervention and these data-cleansing efforts took years in several cases to remediate. In the end, the company’s grand idea of co-locating key SMEs panned out, but it unearthed a rampant issue in the interim regarding how data was collected and managed.

Problem with asset structures

Before delving into the major challenges, it is important to briefly describe asset structures, which are foundational tools used to harness the full analytic capabilities of software platforms. They organize physical locations, pieces of equipment and data on said equipment into a hierarchical structure.

When structuring assets, several issues can impede teams’ abilities to scale and derive value from their data. The five main challenges are:

Figure 1: Seeq’s Asset Group Editor provides an intuitive interface for organizing and adding calculations in analyses. Courtesy: Seeq

1. Differing equipment ages

This is a problem in many plants regardless of age, ranging from brand-new facilities to those more than a century old. Management often thinks, “a pump is a pump is a pump,” but unfortunately, engineering design and equipment evolve over time, so one-size-fits-all approaches do not work. Some equipment is fully “sensored,” while other similar assets have no instrumentation onboard.

Notably, one offshore plant consisted of only five data historian-connected sensors. Plant personnel were responsible for measuring all other recorded parameters with handheld pressure gauges and other instruments.

Even at new largescale facilities with modernized network infrastructure and data connectivity, project design and construction periods sometimes take place over extensive periods and/or designs change multiple times. These and other factors often result in different variations of similar-type assets.

2. Incorrectly mapped sensors

Major manufacturing facilities can have tens of thousands of sensors, making it challenging to connect and centralize them correctly. Over time, instead of correcting errors, teams sometimes develop simple organizational workarounds, such as proceeding with the knowledge that “sensor XX-1A actually means XX-A1,” which can cause inefficiencies and potential inaccuracies and other errors.

3. Sensor failures

One new cryogenic facility was built with the expectation not to shut down ever in the first five years. When a sensor broke shortly into the plant’s operation, engineers used other process data to calculate a “soft” sensor reading in place of the failed physical sensor to avoid an operational halt, without intending to replace the field sensor until the next planned shutdown. While evading downtime, this calculation did not provide the same accuracy of the original sensor, inherently decreasing operational efficiency potential. Additionally, were the team to “blindly map” the broken sensor into an asset tree, it would provide inaccurate data.

4. Different design choices

There is persistent tension between project and operations teams within industrial organizations. Project divisions strive to meet requirements with maximum cost efficiency, while operations groups desire the extras that enable improved situational awareness and simplified optimization. This can result in corner-cutting during design and commissioning to save costs, such as omitting contractually nonrequired sensors that operations teams may view as essential. These types of decisions and miscommunications can have long-term operational efficiency and data accuracy implications.

5. Use case dependency

Asset trees might be structured by equipment type, process flow or plant area, but different analyses require different structures. For instance, control loop performance monitoring may require one structure, while pump health analysis requires another. In one case, a manufacturing company’s IT department attempted to build one asset hierarchy to rule them all, but it failed to recognize the need for different structures for different analyses. As any industrial SME could predict, this effort ultimately failed.

These are some of the leading difficulties for rapidly deploying and scaling asset structures. Fortunately, modern advanced analytics platforms are helping manufacturers overcome these challenges.

Advanced analytics tools for asset management

When ill-equipped, there is a temptation to give up after struggling through constructing one asset hierarchy, but this leaves significant value on the table.

Figure 2: Seeq enables superimposing process data and customized calculations to help plant personnel derive a holistic picture of plant performance with context. Courtesy: Seeq

Another occasional pitfall is deploying Python or another programming language to structure data, but this skillset is typically confined to data science groups. Because it is not readily available to all teams, this can lead to maintenance and sustainability challenges of the code as time goes on.

The solution is enabling personnel to build asset hierarchies using point-and-click tools or by ingesting a CSV file, making the procedure accessible to almost anyone capable of operating a computer. AI-equipped advanced analytics platforms facilitate this, empowering SMEs and other users to experiment and iterate freely with asset groups.

Organizing data into an asset tree in these types of software platforms empowers users to:

• Use asset swapping to rapidly create identical visualizations for different pieces of equipment.

• Write high-value calculations for components, then scale them across all similar components in the tree.

• Automatically generate scalable content and custom analyses.

• Reference the tree as a starting point for roll-ups, calculations, displays, dashboards and reports.

Point-and-click tools: As mentioned previously, there are many challenges to creating successful asset structures and SMEs working with structures not fit for their purpose can experience significant hindrances. Point-and-click tools, however, enable SMEs to build and iterate quickly without IT involvement, which can be time-consuming and require numerous approvals.

By empowering plant personnel to create asset structures at the operations level, teams can iterate rapidly on what works and distribute this to the entire operating unit efficiently. This method prioritizes business value over striving for the “one size fits all” asset structure.

When reviewing and selecting tools, teams should consider the ease at which calculations can be added. In many analyses, signal data alone is not enough. Effective tools provide simple ways for adding calculations that can be replicated at scale. Figures 1 and 2 show a tool that empowers the operations team — without IT — to build out calculations quickly and efficiently to enhance analyses.

Figure 3: With just a few lines of Python code, subject matter experts can build out massive asset structures within Seeq, based on entry in a CSV or other spreadsheet-based text file. Courtesy: Seeq

Template files (CSV): For use cases that are too large to manage efficiently with point-and-click tools, using a spreadsheet program to create asset structures can be beneficial. With a few lines of Python code, advanced analytics platforms can ingest a spreadsheet-based text file — such as a CSV — and build the asset structure within the platform.

This approach provides significant benefits for iteration at scale, enabling manufacturers to quickly build large asset structures, deploy them, evaluate their effectiveness and make necessary adjustments (see Figure 3).

Monitoring solutions

Customizable asset structures within analytics platforms provide business value for scaling analyses across similar assets enterprisewide. This also facilitates faster creation of customized calculations when applicable, which enhance process efficiency. Outputs, such as dashboards, enrich monitoring and streamline decision-making processes, especially combined with AI-based anomaly detection and triggers (see Figure 4).

Figure 4: Dashboards provide at-a-glance asset health monitoring, fostering informed operations and maintenance decision-making. Courtesy: Seeq

By prioritizing ease of use and scalability, AI-equipped advanced analytics platforms empower organizations to leverage their data effectively, driving smarter decision-making and maximizing process efficiency.

Innovate asset management to drive productivity

In the dynamic landscape of manufacturing and industrial analytics, managing complex asset hierarchies effectively is paramount for optimizing plant operations and deriving actionable insights. From the varying ages of equipment to the need for flexible solutions that empower engineers at the operations level, the complexities of asset structures demand innovative approaches.

Point-and-click tools enable rapid iteration and deployment, placing business value at the forefront of decision-making. Additionally, platforms integrated with Python scripting capabilities provide efficiency at scale, enabling quick creation of large asset structures.

As industry continues to digitize, leveraging AI-equipped advanced analytics platforms to support overextended SMEs is critical for revealing data-driven insights, increasing operational safety and maximizing process efficiency and productivity. By prioritizing ease of use, scalability and the ability to adapt to diverse use cases, these software platforms empower manufacturing leaders and plant personnel to navigate the complexities of asset hierarchies with confidence.

Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our WTWH Media editorial team and getting the recognition you and your company deserve. Click here to start this process.