Best Tools for Determining Optimal Preventive Maintenance Intervals

Michael, I want to clarify my previous comment as I may have caused some confusion. In my opinion, the selection of optimization software should be tailored to the specific industry it will be used in. Personally, I believe that RCM, RBI, and IPF software are the most effective optimization processes available. I wouldn't recommend using any other software for optimization. These tools can provide valuable insights into inventory levels, maintenance schedules, and other important quantitative data. While a CMMS may not be equipped to handle all of these functions, data from RCM software can be seamlessly integrated into a CMMS if they both use open databases and their fields can be mapped accordingly. This integration can be achieved through a transfer software. These considerations are crucial when considering factors such as industry requirements, company culture, and workforce skills. There may not be a simple solution to solve all optimization challenges, but a knowledgeable IT professional familiar with computer systems can address these concerns efficiently.

• 19-08-2024
• Kevin Griffin

Thank you, Vee, for your feedback. It might be misleading to discuss PM optimization, as it is often confused with maintenance optimization, which is primarily based on reliability analysis and mathematical calculations.

• 20-08-2024
• Chad Cook

Michael stated that he believes maintenance optimization should heavily utilize reliability analysis and mathematical models. He believes that both parties are in agreement on this particular aspect.

• 20-08-2024
• Laura West

Hello Michael, Achieving your desired outcomes in EXCEL can be straightforward without the need for expensive software. By implementing the two provided formulas (utilizing a numerical method to calculate the integral), you can determine the optimal time intervals for PM tasks that result in either minimizing maintenance costs or maximizing availability. Since the formulas cannot be differentiated to identify minima or maxima, a numerical approach is essential. Attached is an example in EXCEL showcasing the results, where the minimum cost occurs at t = 825 hours and the maximum availability at t = 1,050 hours. Best regards, Attachment: Example_of_OTI_in_PM.pdf (212 KB) - 1 version

• 20-08-2024
• Oscar Hayes

In determining the best time intervals for preventive maintenance tasks, the goal is to achieve the optimal balance between minimizing maintenance costs and maximizing equipment availability. While some may argue that the focus should be on managing tolerable risks rather than strictly minimizing costs or maximizing availability, it is important to consider the cost curve, which often shows little sensitivity to small changes in intervals. This indicates that the maintenance strategy should prioritize cost-efficiency and risk management, rather than solely focusing on minimizing costs or maximizing availability.

• 20-08-2024
• Quincy Brooks

Hello Vee, I appreciate your input on this topic. When considering the most cost-effective approach for a combination of corrective maintenance (CM) tasks and preventive maintenance (PM) tasks over an extended period, it is important to strike a balance. The ALARP concept, which stands for "As Low As Reasonably Practicable," focuses on minimizing economic costs as well as other potential costs such as harm to people and the environment. Every action has associated costs, so it is crucial to weigh all factors and determine the most suitable course of action. The cost curve tends to plateau in the middle range, indicating that total cost is not significantly impacted by minor adjustments around the minimum value. This flexibility can be advantageous for decision-makers, as they are not confined to a specific "optimal point" and can choose from a broader range of options. It seems that we may have slightly different interpretations on this issue. However, I believe my previous post addresses the queries raised by Michael. Warm regards,

• 20-08-2024
• Quinn Bell

Thank you for joining Tui. I agree that Excel is a powerful tool for calculations, but it does have its limitations. One drawback is that Excel sheets can be difficult for others to understand, aside from the creator. Additionally, Excel doesn't assist in building or selecting the best model for the problem at hand, leaving users somewhat isolated. Inputting numbers blindly can lead to errors, highlighting the importance of expertise when using Excel. While I appreciate your age replacement model sheet, I'm curious about other models you may have. I've created basic models in Excel myself, such as age, block, and minimal repair models. On the topic of maintenance optimization, I believe that the key consideration should be tolerable risk, rather than solely focusing on minimum maintenance cost or maximum availability. It's essential to prioritize safety and not solely rely on maintenance optimization when developing a maintenance program. Safety should always be the primary concern, with cost considerations coming after. Quantifying safety in terms of costs can be tricky, as some decisions may have ethical implications, such as risking a machine operator's life to save money. It's crucial to use mathematical models to assess the consequences of decisions objectively. The flat cost curve mentioned in Tui's calculations is specific to that particular scenario.

• 20-08-2024
• Cameron Price

Michael, I found your insights on EXCEL to be quite intriguing! Personally, I have a different perspective on the matter. While I do have access to software tools like Reliasoft's Weibull 6++ and Block-Sim, I prefer using my own custom tools. These tools are ones that I have developed myself, and I have a deep understanding of their results and limitations. I always approach data with careful consideration and rely on my extensive experience, rather than blindly inputting numbers into an Excel sheet. I have a strong command of EXCEL and have been using it in various capacities such as consultancy, training, and lectures for a long time. If you're interested, you can visit my website www.rassis.com, where you'll find a wealth of information in Portuguese. You'll find published papers, presentations, and numerous EXCEL files covering topics like investment project evaluation, reliability, statistics, production management, and performance monitoring. Many of my students and professionals who use my tools have found them to be incredibly helpful in solving real-world problems. I have also developed custom software over the years, particularly for cases requiring Monte-Carlo simulation due to complexity, always starting with EXCEL as a foundation. In regard to cost and availability curves, I often observe a distinctive flat shape in the mid-region, with variations in amplitude. I have a file available on my website (www.rassis.com/manutencao.html) that demonstrates the calculations mentioned in my previous post, focusing on a Weibull distribution with 3 parameters. I am currently working on an application involving a Boolean Table representing a system with up to 6 repairable components, tailored to specific cases like yours involving complex systems. This task becomes more challenging as failures are represented by Weibull distributions rather than exponential ones, requiring consideration of degradation levels when components are replaced. If you have a file that closely matches this model and are willing to share, I would greatly appreciate it. Thank you for your insights and sharing your experiences. Regards,

• 20-08-2024
• Ruby Cooper

It's always great to hear about your positive experience using Excel for this particular purpose. One interesting point you made is your preference for using tools that you have personally created and fully understand. While that ideal scenario works well for you, in a consulting role, it's important to consider that the individuals using your tools may not have the same level of expertise or be familiar with the tools' construction. Personally, I find it a bit challenging to interpret other people's Excel spreadsheets, and I can imagine they feel the same way about mine. However, I acknowledge the value in your customers embracing Excel in their work. Regarding our differences on utilizing Excel, I may not have effectively communicated my point about assisting in building or selecting models. It seems you may have a higher expectation of user knowledge, which could result in more complex and extensive spreadsheets. I want to clarify that my mention of "blindly punching in numbers" was not meant to imply that Excel users do this, as I also use Excel. I only meant to highlight that Excel doesn't impose restrictions on data entry. While Excel is a powerful tool, there may be specialized software options that can better serve non-experts. I am looking forward to exploring your website as I'm sure it offers a wealth of interesting content. Unfortunately, I don't have experience with the specific issue you mentioned.

• 20-08-2024
• Andrew Clark

Rui, please take a look at the documents on tolerable risk at www.nfrmp.us/ncls/docs/Eric_Halpin_Tolerable_Risk.pdf and/or www2.emersonprocess.com/.../ART_TolerableRisk_Sept07.pdf. It is important to understand that what constitutes tolerable risk is determined by society and can vary over time and across different societies. While there are no universal norms, there are widely accepted guidelines, as indicated in the provided URLs. In my opinion, it is crucial to prioritize managing the drivers of performance, such as reliability and productivity, rather than focusing solely on the consequences like cost and availability.

• 20-08-2024
• Nina Sanders

Rui mentioned that in the event of a system intervention for component replacement, it is crucial to take into account the level of degradation experienced by the other components. Could you further explain the concept of considering the "memory of degradation levels" and also outline the necessary outputs of the model?

• 20-08-2024
• Penelope James

I greatly appreciate your input on risk analysis and the provided URLs (unfortunately, I encountered an issue with downloading the second one - could you please check for any errors in the link?). I want to clarify that I do not advocate for everything to be based solely on monetary factors in mathematical scenarios. My focus is specifically on maintenance policies, where the outcomes are purely economic costs with no implications on human safety or environmental impact. In such cases, it is crucial to apply concepts of risk assessment and techniques for reducing risks as a priority. To ensure clarity in my Excel sheets for others, I have implemented a color scheme to differentiate between data, intermediate calculations, and results. Data cells are highlighted in blue with numbers in dark blue, intermediate calculation cells are in green with numbers in dark blue, and results cells are in yellow with numbers in red. This system can be seen in any of the sheets on my website to facilitate easy interpretation. In response to your questions, Sean, the key outputs of the model include MTBF, MTTR, and system availability. When considering degradation levels, the system states are defined and classified as "operational" or "non-operational" for each component within a redundant system. Utilizing Monte Carlo discrete events simulation, the system's progression is simulated based on Weibull distribution failure modes. This approach considers the memory effect in degradation modes such as wear and tear, corrosion, fatigue, and fluency. The simulation determines component failure moments and system states, allowing for timely replacements and continued system operation. If you require further clarification or examples, please do not hesitate to reach out. Thank you for engaging in the discussion.

• 20-08-2024
• Vince Myers

Rui, thank you for the information, it has greatly clarified things for me. I refer to the process as "ageing" rather than "degrading" and prefer the term "remaining life" over "memory", as you mentioned in your explanation. To input into the model, you will need: 1) a Boolean truth table (consider using a fault tree instead) to illustrate the system reliability structure based on component states. 2) Weibull parameters for each component, including shape and scale. 3) Information on the replacement time for failed components (is it fixed or sampled from a distribution like normal?). Also, will components continue to age and possibly fail when: a) the system is offline b) maintenance (component replacement) is being carried out? Once you confirm the details above, I will work on creating a model to share later this week.

• 20-08-2024
• Diane Kelly

Upon reviewing the enclosed document, Rui mentioned, "I was unable to access the second article attached." In the context of discussing maintenance policies, it is emphasized that economic costs are a direct result of poor reliability and productivity, rather than the drivers of performance. By addressing these root causes, costs will naturally decrease, similar to how risks are managed. This highlights the primary concern with the strategy of solely focusing on cost optimization. For more insights, please refer to the attachment titled ChemEng_TolerableRisk_Sept07.pdf (187 KB).

• 20-08-2024
• Jessica Freeman

Attached please find the article, Rui. In the article, we discuss the importance of addressing maintenance policies and how economic costs are a direct consequence of poor reliability and productivity. These factors, which I refer to as "drivers", are the root causes of issues that must be managed effectively. It is crucial to focus on treating these drivers rather than solely on the outcomes or consequences, such as costs or risks including health, safety, and environment (HSE). By addressing the drivers, we can ensure that the right consequences will follow. Concentrating on a single consequence, such as cost, may lead to unintended consequences like HSE issues. While we may not share the same perspective, it is important to acknowledge and respect each other's views. Please see the attached document for further details.

• 20-08-2024
• Faith Perry

Dear Sean, Thank you for your post. I am attaching an example to help illustrate the case I am currently working on. Please see the attached file, "Repairable Complex System.pdf," which provides a clearer explanation of my project. I hope this helps clarify things for you. Regards, [Your Name]

• 20-08-2024
• Jack Parker

Hello Rui, when it comes to modeling the situation, does the repair of a component start immediately upon failure, or does it begin when the whole system fails? While in the example provided both scenarios are the same for a series system, they may differ in a more intricate system. Best regards, Sean.

• 20-08-2024
• Shawn Thompson

Rui, you primarily utilize this as a tool for budgeting or maintenance planning. It could work well for both purposes if a sufficient amount of data is accessible.

• 20-08-2024
• Ulysses Ross

Hello Sean, I appreciate your dedication to addressing this issue. As soon as a component fails, the repair process (including replacement) starts promptly. Thank you for your attention to this matter. Best regards.

• 20-08-2024
• Nathan Scott

I utilize predictive models to anticipate the performance of a future system comprised of electronic and electro-mechanical equipment, aligning with RAMS criteria specified by clients. These models serve as valuable decision-making tools, allowing users to input data and explore various "What if..." scenarios. Best regards,

• 20-08-2024
• Ian Butler

In a quote by Vee, it is stated that reliability and productivity are crucial factors in driving overall performance. Instead of focusing on the consequences, it is suggested to manage these drivers for better results. Vee, could you clarify if productivity refers to human productivity, plant productivity, or both? Please provide examples for better understanding.

• 20-08-2024
• Helen Diaz

In our discussions, we typically prioritize human productivity over the Plant's volume efficiency. Productivity, in this context, is the ratio of output to input, where input encompasses all resources like personnel, equipment, and machinery. Our primary focus is on optimizing critical resources, especially manpower, to minimize downtime and improve overall efficiency. This downtime, measured by Mean Time To Repair (MTTR), accounts for all the time from equipment stoppage to production resumption, including necessary preparations and safety procedures. MTTR is a crucial indicator of productivity, as it reflects the average downtime experienced. By enhancing work preparation processes, we can increase hands-on-tools time and boost productivity levels. Improving MTTR not only enhances plant availability but also maximizes the utilization of workforce effectively.

• 20-08-2024
• Bob Smith