A Second Systemic Look at Preventive Maintenance 

Preventive maintenance remains the preferred maintenance type for mining and metallurgical organizations even today. It is well-known as a proactive strategy for industrial asset management aimed at preventing downtime. It relies on a logical sequence of planned and scheduled maintenance tasks based on equipment usage, operating hours, or the calendar. Tasks may include inspection, cleaning, lubrication, adjustment, replacement of parts, testing, and other routine activities designed to keep the equipment in good working condition and extend its useful life. This method, usually applied to critical production assets, is considered cost-effective, safer, and more efficient when the program is properly followed.

However, the discourse among researchers, experts, and technology companies on various platforms is exclusively focused on predictive and even prescriptive maintenance based on the real condition of assets. The claim of this discourse primarily revolves around better equipment availability for improved annual performance. Preventive maintenance methods based on the real condition of the asset, powered by suitable instrumentation, could reduce the frequency of maintenance activities by 10 to 20% from the schedule. Additionally, preventive maintenance is undoubtedly less costly in terms of manpower and materials.

A legitimate concern arises regarding the risk of undertaking unnecessary interventions. Predictive maintenance, despite requiring a higher initial investment in monitoring technology, can often contribute to long-term cost reduction by avoiding unnecessary interventions and enabling more efficient use of maintenance resources. This efficiency depends on several factors, including the type of equipment, the complexity of operations, available resources, and how each approach is implemented.

At the moment, this argument does not seem to resonate widely. Companies typically choose preventive maintenance due to its simplicity of implementation and management, considering the specific characteristics of the equipment and operations to determine the most suitable approach. Thus, most organizations strive to adhere to their production plan at all costs. Many prefer to shutdown more frequently for planned maintenance, preventing unforeseen downtimes that could cost hundreds of thousands of dollars per hour in revenue. These frequent shutdowns are calculated in yields deemed satisfactory, providing a relative sense of security and better control.

Nevertheless, there is a broad consensus among superintendents and maintenance managers that a broader adoption of condition-based and risk-based maintenance methods is inevitable to systematize data analysis and have the agility needed to adjust the maintenance program based on actual needs. The adoption of technology-assisted techniques and methodologies is expected to grow to enhance productivity and address major industry challenges.

Moreover, many managers have been dreaming of artificial intelligence (AI) since the current technological boom. To achieve this, several actions must be taken to centralize, structure, and categorize data to give this dream a chance to materialize. It should be noted that there is often a significant capital investment required to enable the use of AI for maintenance purposes. Equipment is not always equipped with the necessary sensors to gather a sufficient amount of data to change the maintenance mode.

In this text, we explore three lines of thought in favor of adopting technology-assisted methods:

• A second look at maintenance.
• Digital inspection: helping to prioritize actions.
• Creating collective intelligence.

Maintenance Maturity 

There is a significant disparity between mature organizations and those striving for maturity when strictly comparing them from a maintenance perspective in mining and metallurgical organizations.

In a mature organization, the maintenance program is well-established and strictly adhered to. Maintenance managers are familiar with their equipment, emergencies are less frequent, and production exhibits better predictability regarding annual tonnage. In most cases, operations and maintenance teams work collaboratively, greatly facilitating production.

In contrast, in an organization seeking maturity in the mining sector, the maintenance program is not as rigorously followed. Maintenance activities related to the operating hours of certain assets may be stretched or set aside due to emergencies elsewhere on the site. Despite significant efforts to regain relative stability, maintenance teams often act as firefighters, addressing issues as they arise. It is challenging to regain control in such situations, and the risk of catastrophic breakdowns and accidents may be higher.

Clearly, the motivations to integrate modern computer-assisted practices with mathematical models are not the same. Let’s explore the nuances.

A Second Look at Maintenance 

Undoubtedly, the vast majority of industrial maintenance personnel are highly skilled. They must design and follow a robust, efficient, and comprehensive maintenance program in a context where every minute of downtime is extremely costly.

Being human, individuals may overlook certain adjustments to the program, which can be crucial or catastrophic depending on the perspective. Humans cannot analyze anomalies in real-time or detect every degradation trend in an asset portfolio. There is too much data to track, digest, and process, especially as this data is scattered across multiple software or systems. So, can we really blame them? No, obviously not.

In this context, digitized analysis methods based on best engineering practices in integrity and reliability are necessary additions to determine the health condition of a fixed, rotating, or rolling asset. Tools like APM+ enable the detection of situations, anomalies, or micro-events that would otherwise go unnoticed in preventive maintenance. Most importantly, these detections occur at a faster pace, providing more time to adjust the maintenance program and mitigate abnormal wear.

To a lesser extent, the detection of anomalies or signs of aging may not generate additional actions in the preventive maintenance program. However, during regular maintenance activities, personnel can pay special attention to additional or simply different elements. Thus, equipment instrumentation and comprehensive asset documentation provide vital assurance to the maintenance team. By taking a different and systemic approach to the condition and health of the asset, it is possible to cover blind spots in so-called traditional maintenance. This allows for the consolidation of information from various data sources.

For organizations striving for maturity, technological assistance is even more necessary. Analyzing trends or events reported by sensor or inspection data helps prevent unperformed maintenance from turning into bigger fires to extinguish.

There are numerous stories of systems capturing situations that could have resulted in major issues. In various organizations at different levels of maturity, where projects have been initiated, the detection of anomalies has proven to be critical. Serious integrity issues were discovered when the program did not specifically address certain aspects. This was the case for critical rotating assets like a mill but also for fixed assets such as highly damaging sulfuric acid pipelines. The averted incidents saved workers and prevented significant losses for the organizations.

The second digital look enhances the maintenance program. The results of mathematical analyses can indeed become a crucial source of information for daily and operational decision-making, translating into tangible gains. While sometimes challenging to calculate for an external firm, the return on investment from avoided accidents and the assurance of annual production are undoubtedly real.

Digital Inspection: Helping to Prioritize Actions  

The instrumentation of assets is undoubtedly an important factor in monitoring the vital signs of assets. However, to effectively guide the maintenance program and ensure effective asset management, inspections are of paramount importance.

What is confirmed is that the digitization of inspections is not yet a given for all mining sites, even for mature maintenance organizations. However, it could serve as a gateway to digitizing operations, a transition that is relatively easy to cross. Those who have crossed it have mostly done so using form tools that are not adapted to the complexity of engineering of asset integrity and reliability.

A generic form creation tool, even if the results are transferable to CMMS, Power BI, or an ERP, is far from sufficient. This can be easily explained. On the one hand, there is significant pressure on employees to create comprehensive and effective forms, and this task requires a lot of time. On the other hand, there will always be a significant lack of vital features for tracking data over time.

It is necessary to implement forms dedicated to each type of asset, created from an analysis of failure modes, their effects, and their criticality, including a severity grid to qualify defects and an evaluation of the risk associated with each asset. Itis important that these forms be backed by an inspection plan structured by an engineer to guide the work of inspectors. This plan directs the work of both internal resources and subcontractors, ensuring comparable data during computer-assisted or non-computer-assisted evaluations.

The chosen product must also be capable of reading, displaying, and analyzing data of various types, such as photos, quantitative and qualitative data, in addition to thickness measurements (NDT). But above all, this type of tool should not be an empty box where all filling efforts have to be carried out by the organization. Forms and health index calculations must be included. The latter must be based on engineering expertise in asset integrity and reliability assessment. It must demonstrate a certain level of intelligence to reduce the workload of existing resources rather than perpetuating a trial-and-error approach.

Finally, a tool of this kind must provide data traceability, enabling the comparison of results with past data and the observation of trends. These results should be displayed on adynamic dashboard for better visibility into the health of assets and the risks they pose, leading to improved prioritization of interventions. Even mature organizations can expect improved performance. Better predictability results in better cost control. By focusing on the right priorities, the risk of unforeseen emergencies is reduced, and unnecessary investments in lower-priority tasks are avoided.

For organizations striving for maturity, unfortunately, emergencies often dictate daily tasks. Attention to static assets is practically nonexistent. Nevertheless, plants are aging, and static assets such as tanks, pipelines, chimneys, and buildings approach the end of their lifespans. These assets are frequently inspected and analyzed by external firms that must inform the organization about the health status of assets and the associated risks. These expert reports are often shelved due to attention being focused elsewhere, but also because of the challenge in cross-referencing information and prioritizing the interventions to be implemented. This lack of visibility inevitably creates latent emergencies that will sooner or later jeopardize production.

Yet, these are the organizations that are least involved in the implementation of assistance tools. The reason is simple: even if they wanted such a tool, time is a constraint. They are aware that implementation cannot happen magically without the involvement of internal resources. The consideration is valid, but many APM+ integration teams, like those of Stelar, bring together teams of engineers, reliability specialists, maintenance technicians, and IT professionals. This collaborative effort allows them to achieve gains without the perceived mountain of efforts, especially since they can benefit from the support of an engineering firm with the required knowledge to implement the models.

Creating a Collective Intelligence

Over the past few months (2023), we have witnessed implausible situations in organizations that are otherwise mature in maintenance. To adequately analyze the integrity of critical assets within an organization, external integrity and reliability experts had to reach out to retirees for information regarding construction plans, materials used, installation dates, Tmin, etc. What is most disconcerting is that these situations are not isolated occurrences.

The metallurgical plants were built in the 1950s and 1960s. Many critical systems are naturally reaching the end of their lifespan. We addressed this in the article “The Relationship Between CMMS and Stelar”: investments to replace or rebuild certain critical and major assets are imposing and sometimes illogical from an economic and ecological point of view.

But how can we hope to extend the life of assets if we don’t have basic information about them? How can we expect to prolong the life of assets if we don’t have records of studies, inspections, failures, repairs, expansions, or any other events throughout their lifecycle? Asking the question provides its own answer, and it creates headaches.

Over the past decades, organizations could rely on the knowledge of their best employees who remained with the organization for years. They knew the site inside out, and the assets held no secrets for them, especially as the equipment was always in the later stages of their lifecycle curve. This briefly explains why there was hardly any urgency to digitize information about the life of assets. The situation is catching up with organizations.

The challenge is exacerbated by the workforce situation. This is well-known; the available and skilled workforce is becoming scarce, and employees stay for much shorter periods. They are required to juggle multiple tasks simultaneously due to the shortage, and many retire. The loss of information is vast, and the associated training costs are enormous, with trial and error due to lack of knowledge, and prolonged downtime due to inexperience. Additionally, there are growing risks of breakdown saffecting health, environment, and production of the organization.

It would be unwarranted to sound the alarm, but it is essential to recognize and be aware of the risk. Organizations must regain control of their data to reclaim the intelligence of key individuals. They must be able to collect new relevant information about the life of assets in a centralized and structured manner. This information is crucial for ensuring operations run as smoothly as possible, independent of the talent or knowledge of a handful of people. We can call this organizational intelligence.

Furthermore, younger workers expect new methods. They have grown accustomed to technology filling certain aspects of their lives, especially their knowledge. If they have previously worked for modern companies, they have had access to tools that assist them in their work. Failing to support young workers with relevant tools hinders their integration into a world where shutdowns are avoidable at all costs.

In addition, mining organizations benefit from organizational intelligence for the accelerated implementation of mining projects based on historical data; quicker learning for younger workers; smoother transition between rotating teams streamlining corporate service operations; and facilitating information sharing among different divisions within an organization for data analysis.

Next Steps

The transition to a predictive maintenance approach focused on asset condition, with the accumulation of centralized and structured information in an "asset medical record," has been a significant step in supporting the business objectives of companies.

However, while preventive and prescriptive maintenance strategies have been particularly successful for production equipment such as rotating machinery, predicting the most at-risk assets remains a challenge. These assets are not always limited to rotating equipment but may also include fixed assets such as a tank or a pressure vessel.

Therefore, the next step is to implement a risk-based maintenance strategy to identify the most critical assets to address. This approach relies not only on operational data but al soon their actual condition and the consequences associated with a failure.

‍