New ✨ Introducing Oxmaint Asset Hub for Machine Builders and OEMs. Explore Now

Q&A Community

Optimizing Spare Parts Inventory for Preventive Maintenance

Question:

In many companies, I often come across maintenance departments that perform preventive maintenance (PM) but fail to differentiate between demand stemming from wear-out failures and demand resulting from random failures. This leads to a mixture of demand, causing average stock levels and holding costs to increase beyond their optimal levels. It is crucial to correctly identify and address these separate demands in order to justify holding certain spare parts in stock permanently. I am interested in hearing about your experiences and insights on this issue. Share your thoughts with me, Rui.

Top Replies

Looking to differentiate between demand for spare parts caused by wear out failures and random failures? What criteria or guidelines can be used for this distinction? Should stock levels be based solely on demand resulting from random failures?

Hey Josh, I totally agree. Implementing a just-in-time ordering system for parts is essential for preventive maintenance tasks, while maintaining a small inventory for unexpected failures is also key. What are your thoughts on this approach? Let me know. - Rui

While Just-In-Time (JIT) is preferred, obtaining spare parts only when needed may not be feasible for small quantities. It is necessary to maintain a stock of materials, especially in remote locations, to avoid disruptions in operations. One possible solution is to negotiate a price agreement with a vendor who can stock the parts for multiple customers, ensuring fast delivery and eliminating the need for multiple quotes. Should we also consider stocking parts for unforeseen failures, or would a price agreement suffice in such cases?

Identifying worn-out spare parts like gaskets, impellers, wear rings, bearings, and lubricants is fairly simple. However, determining which components are prone to random failures can be more challenging. Complex machinery, systems, or assemblies are susceptible to failures caused by a variety of factors. This raises the question of which parts should be kept in stock. Additionally, many may not have analyzed Reliability-Centered Maintenance (RCM) curves to pinpoint equipment that is prone to random failures at specific times.

While I may not be an expert in spare parts or RCM, I enjoy asking questions to dig deeper into the topic. Do you keep stock of your entire industrial plant or just a portion? Do you store assemblies or individual parts? Utilizing an RCM analysis can be a valuable starting point to understand the impact of failures at an industrial site and to establish maintenance strategies for equipment. When it comes to failures, it's not always black and white - there are often various factors at play. By conducting an RCM analysis, we can determine the consequences of component failures and assess the potential impacts on health, safety, the environment, company reputation, and production. Understanding the frequency and causes of failures is crucial in determining the maintenance strategy to adopt. While some failures may be time-based, most are not, leading to the need for condition-based monitoring or run-to-failure approaches. Each site may need to make individual decisions based on their specific circumstances. Forecasting failures through condition-based monitoring can be complex, requiring thorough monitoring, clear criteria, and confidence in the workforce to act upon identified conditions. Without a solid understanding of potential failure scenarios, probabilities, and consequences, identifying which parts to stock becomes challenging. In conclusion, developing a maintenance strategy based on potential failure modes, probabilities, and monitoring processes is essential for effective spare parts management. While discussing these concepts may seem complex, it is crucial for optimizing plant maintenance and operations.

More Replies

A reciprocating gas compressor was in operation for over five years with regular maintenance, including valve replacements every one and a half years. However, a catastrophic failure occurred due to a fouled 1st stage discharge valve, leading to piston and bore damage. The fault tree analysis identified various contributing factors to the failure, prompting the implementation of condition-based monitoring and age replacements for critical components. Although valve replacement on condition is cost-effective, it slightly increases the risk of failure over time. To optimize maintenance cost and prevent downtime, it is advised to replace time-based parts and valves every one and a half years. By stocking spare parts, the risk of not having components during an emergency is mitigated. It is crucial to proactively assess equipment for maintenance strategies and spare part management to avoid costly failures. Understanding the consequences of failure is key to effective equipment maintenance.

Rui, it seems like your initial query pertains to the increase in inventory holdings stemming from the failure to distinguish between parts used for random failures versus those used on a time-based schedule. Is that accurate? If so, the key lies in determining the additional investment required and whether it is justified to address the issue. (To clarify, I am not referring to choosing between planned or run-to-failure strategies, but rather segregating spare parts usage.) If the items in question are low-cost, it may not be worth the trouble. The solution involves evaluating each situation individually to determine the extra investment and effort needed to reduce it. It's important to note that I have a bias as I have created a process for evaluating MRO inventory investment and lowering it without affecting operational capabilities. For more information, please visit www.InitiateAction.com/inventory_management.htm Feel free to let me know if I understood your question correctly.

This is how I address this issue, and I am quite satisfied with the results so far. In a CMMS, work orders are typically traced back to either the production department or the maintenance department. Parts replaced due to maintenance interventions are logged in the database for tracking purposes. However, parts replaced after inspections due to defects should also be included in the maintenance group. For example, discharge valves on gas compressors that fail periodically should be categorized as maintenance-related. By tracking the average number of units per day requiring maintenance and predictive maintenance, it is possible to determine the optimal reorder point for spare parts efficiently. This approach relies on a Poisson probability distribution to calculate reorder points accurately. Factors such as unit price, holding costs, and opportunity costs should also be considered when determining order quantities. Utilizing the MRP algorithm in the CMMS ensures timely ordering of parts for preventive maintenance tasks. While classifications may sometimes be challenging, using a scientific method is more effective than relying on general rules. Overall, this strategy combines just-in-case and just-in-time principles for efficient spare parts management. Your insights on economics are appreciated, and I look forward to exploring the link you shared. Feel free to share your thoughts on the points I have discussed. Thank you. Regards, Rui.

Jaz's case involving the discharge valves of a reciprocating gas compressor raises the question of whether it is cost-effective to keep expensive items in stock. Specifically, when considering the crankshaft, would it be more economical in the long run to maintain one unit in inventory or none at all and risk potential damages from a failure in the future? To better assess this, I kindly ask for specific numerical data such as the unit purchase cost, lost production cost per hour, corporate cost of capital, warehouse cost, failure rate, mean running hours per year, and the expected lifespan of the crankshaft or any other part of your choosing. Your input will greatly aid in making an informed decision. Thank you, Rui.

When considering the impact of equipment failure on your processes, it is crucial to weigh the costs and effects involved. If downtime is a concern and you have a robust planning and scheduling system in place, it is advisable to focus on preventive measures or predictive maintenance rather than relying solely on contingency planning for potential failures. This approach can help minimize disruptions and optimize productivity by addressing issues before they escalate.

In an effort to provide comprehensive information, Rui Assis presents an analysis of the production impacts of stopping a compressor. During the outage and 7 days post-service restoration, there is an estimated loss of 0.5-1% of 850 tons of production, with each ton valued at $1100 and costing $700 to produce. The cost breakdown includes variable costs tied to raw materials. The unit purchase cost and lost production cost are factors to consider, along with corporate cost of capital and warehouse expenses. The failure rates of valves and crankshafts are projected based on specific patterns, with the compressor operating around the clock. Lead times for compressor parts and potential downtime scenarios are outlined, including estimated costs for part replacements. The text also discusses factors contributing to crankshaft failures and the importance of proactive maintenance strategies. In the event of a catastrophic compressor failure, the projected outage duration and associated service costs are detailed. Additionally, considerations for managing raw materials during a compressor stoppage are highlighted. Emphasis is placed on the maintenance of spare valve sets to minimize risks associated with unexpected failures.

For those participating in the discussion, it is crucial to reassess the essence of the conversation. The concept revolves around the importance of segregating the necessity to keep parts in stock based on anticipated or unplanned replacements. Parts that require replacement on a scheduled basis should be procured through Just-In-Time (JIT) ordering, while parts needed unexpectedly (with lead time exceeding the forecasted time) should be readily available in stock. Drawing from my experience in the aluminum industry overseeing a fleet of over 500 mobile equipment units, a Reliability-Centered Maintenance (RCM) pilot project was undertaken. Initially, 10% of part replacements were time-based (e.g., oil and filters), and the remaining 90% were based on conditions. However, post RCM analysis, it was revealed that 50% of replacements were time-based, and the other 50% were condition-based. Implementing JIT ordering for time-based part replacements for preventive maintenance (PM) led to the preparation of PM packages up to 40 days before the scheduled PM activities commenced. While this duration may seem extensive, the processes were still in their developmental stage. Parts were also stocked for breakdowns in alignment with the discussion. The responsibility of ordering parts through JIT fell on the Planner and the Computerized Maintenance Management System (CMMS). The specifics of how this ordering was executed, whether manual or automated, are critical points for consideration. Transitioning to my current role in a fixed plant within the oil industry, there is a greater scope for condition-based monitoring. The culture within fixed plant operations facilitates such monitoring, unlike mobile equipment environments. A fixed plant is likely to have more parts replaced based on conditions, thereby influencing stocking and ordering protocols. Prior to delving into decisions regarding stocking and JIT ordering of parts, it is advisable to assess the bigger picture. Analyzing equipment failure consequences, devising Equipment Maintenance Strategies, and identifying tasks that are vulnerable albeit necessary are foundational steps. Spare parts should be in place for high-risk components. A compressor case study sheds light on the intricacies of managing parts for condition-based actions. Analyzing failures through RCM helps in optimizing Equipment Maintenance Strategies. For instance, implementing condition-based monitoring on compressor valves and pre-planned replacements based on abnormal temperature readings resulted in substantial cost savings for the company. In conclusion, strategic decision-making is paramount in determining when maintenance tasks should be conducted. Balancing operational hours against associated risks is key to efficient maintenance practices.

Thank you, Jaz, for taking the time to share your valuable experience and insights with us. I appreciate the detailed information you provided, including the numbers. Regarding your inquiries, the Unit Purchase Cost pertains to the cost of specific parts, such as a crankshaft. Could you please specify this cost? The Lost Production Cost refers to the financial impact of one hour of production loss, calculated by the mean contribution margin of the product mix (average unit selling price minus average variable costs) multiplied by the average production hour rate. How much does your company lose for each hour that production halts due to compressor failure? I eagerly await your responses to these critical questions. In the meantime, I will thoroughly evaluate your situation to offer my insights. - Rui

Thank you for your insights, Rui Assis. I believe there are two key issues addressed in my previous posts - one pertaining to valves and the other concerning the crankshaft. A valve malfunction can potentially result in a crankshaft failure, with valve failures occurring about once every five years if caught early, and leading to catastrophic consequences if left unattended. Crankshaft failures, on the other hand, are condition-based and typically require replacement every 20 to 25 years. These failures occur once or twice in a century, including the risk associated with catastrophic valve failures. The replacement and maintenance schedules for valves and crankshafts are essential to minimizing downtime. Valve parts have a lead-time of 14 to 21 days, while crankshaft parts require a similar timeline for delivery. The costs associated with lost production due to downtime can be significant, with lost production costs totaling $100 per hour for each day of outage, plus an additional 7 days for recovery. Flaring costs during downtime amount to $833 per hour. Proper monitoring and maintenance of valves and crankshafts are crucial to avoid costly and extended unplanned outages.

It's amazing how active things get when you take a short break! Last week, Rui, you mentioned this quote: Phill, I believe I have addressed your query. Thank you for your insights on economics and the link you shared, which I will explore with great interest. Would you like to provide feedback on the points I shared? I trust you found the free articles at the link enjoyable, Rui. In response to your comments (and other posts from the past week), it's clear that the calculation approach can provide a scientifically sound solution. However, two key issues arise from this. Firstly, is it the best solution? Secondly, what problem are you actually trying to solve? Let me address these one by one. During my time at university, I learned about statistical analysis of spare parts and similar topics, and I sometimes rely on that knowledge. In my experience, a 'good enough' answer can often be reached through a simple review. For instance, in the case of deciding whether to hold onto a crankshaft and its parts or not, the key factor is usually the lead time of supply versus the compressor's requirement. Generally, the cost of production loss far outweighs the cost of the part. If the lead time is 2-3 weeks, halting production for that period (at a cost of $15,000 per day for gas flaring) is not a favorable option. It would be more beneficial to work with the supplier to reduce the lead time to a more manageable timeframe. Addressing delays during their summer shutdown could also be tackled by creating consignment stock for two weeks. When it comes to the fundamental question you're trying to answer, my experience suggests it often revolves around two goals - ensuring availability and reducing working capital. These objectives are often mistakenly seen as conflicting, but they can be achieved simultaneously. However, traditional optimization methods may not always be effective in achieving both these goals. I recommend referring to two resources for further insight. Firstly, check out the article 'When Optimization Doesn't Optimize' in the May/June issue of Reliability Magazine. If you can't access the magazine, I can share the article on my website upon request. Additionally, download the paper '5 Myths of Inventory Reduction' from www.InitiateAction.com/5_myths.htm for a deeper understanding of inventory management concepts. Finally, I acknowledge that my approach may challenge the commonly favored 'calculation' method in inventory management. However, problem-solving in this field involves understanding the factors influencing decision-making and the constraints impacting those decisions. Sometimes, starting from scratch isn't necessary. I hope this provides some clarity and guidance.

Hello Jaz, I have reviewed the document attached regarding the valves related case. I found your explanation to be a bit unclear, but I attempted to come up with a solution by making some assumptions. It appears that the opportunity cost of failure is underestimated in this scenario. If you believe the data I used is incorrect or if you would like to discuss a different case, please feel free to adjust the values in the equations. Alternatively, you can provide me with the correct numbers, and I will gladly recalculate. It is important to note that conducting calculations using scientific methods should only be done after applying RCM analysis to the equipment and determining that it is both technically feasible and financially viable, as John Mowbray pointed out. The decision to move forward with the solution ultimately rests with you. Thank you, Rui Attachment: Valve_failure.zip (30 KB) - Version 1

Rui, thank you for the opportunity to do a trial. Let's address the costs associated with valve failures. When the compressor is not running, the costs amount to around $1000 per hour, in addition to a one-time cost of approximately $17,000 every time the compressor stops (resulting in process disruptions even after it is restarted, regardless of whether it was off for 4 days or 20 days). Regarding downtime, it will be 4 days if valves are readily available in stock, but it can extend to 17 to 24 days if valves need to be ordered. This information is crucial for understanding the financial implications of valve failures and the associated downtime.

Check out the attached document, Jaz. The latest data reveals a shift in the situation, highlighting the necessity of having valves in stock without a doubt. This method can also be applied to various other scenarios. Best regards, Rui. Download the file here: Valve_failure_1.zip (30 KB, 1 version).

Apologies for the oversight, but I realized I neglected to include the $17,000 cost per compressor stoppage when spare parts are available in inventory. Please update the calculation for the recovery cost to 4 days x 24 hours/day x $1,000/hour + $17,000 = $113,000, rather than $96,000, and the total cost of the "no stock strategy" to $70,649 instead of $60,957. The conclusion remains unchanged. Thank you, Rui.

During the 1980s, as a freelance consultant engineer, I honed my expertise in flow (or lean) production while collaborating with renowned automobile and electronic companies to assist them in transitioning to lean production techniques. Addressing inventory-related challenges was a common occurrence during this period. Your article on '5 Myths of Inventory Reduction' caught my attention, although it appears to be more geared towards individuals lacking in-depth knowledge of cost and operations management, including FMECA analysis principles and cause analysis for assessing redesign economics. I am puzzled by your apparent aversion to optimization. When faced with constraints that necessitate spending to achieve measurable benefits, utilizing mathematical models for decision-making should take precedence over relying solely on rules of thumb or common sense, which can often fall short. For instance, in the case of a compressor, determining the optimal maintenance schedule using math can reveal cost-saving opportunities like reducing the recovery period to 4.5 days. Without employing math, it is challenging to assess the feasibility and worthiness of such decisions. Qualitative methods like RCM are essential precursors to quantitative analyses, fostering better understanding and engagement towards solutions. While quantitative techniques are valuable in materials management, your '5 Myths of Inventory Reduction' piece appears devoid of such approaches based on the examples provided. Apologies if I have misunderstood your perspective. Best regards, Rui.

Thank you for your reply, Rui. Let's consider another scenario: a cooling tower system consisting of 8 cooling cells, each equipped with its own electric motor, gear box, and fan. The primary concern lies with the gear boxes. The cost of each gear box is $40,000, with an estimated lifespan of 30 to 40 years and a characteristic life Beta of 1.5. Unfortunately, there are no condition-based tasks available to accurately predict failures, mainly due to the unique installation and location of the gear boxes. Out of the 8 gear boxes, 6 are 24 years old and 2 are 6 years old (added when the plant was expanded). In case of a failure, the cost is incurred only during the Summer season, with no additional costs in the Spring, Fall, or Winter. The cost of failure during Summer can range from $1000/hr to $1500/hr per failed gearbox, depending on the ambient temperature. The repair time with the part in hand is 2 days, but it increases to 40 days if the part needs to be ordered. The question now arises: should we keep a stock of gear boxes, and if so, how many? It is important to note that this issue is specific to the Summer season, as the gear boxes operate continuously throughout the year, 24 hours a day.

Hi Jaz, I had assumed you would be able to handle it on your own by now. Estimated life refers to the projected timeframe during which equipment is expected to remain functional before being disposed of or replaced due to technological advancements, decreased efficiency, or other factors. This is commonly referred to as "useful life". Is this what you meant by stating 30 to 40 years? I have my doubts. When you mention a Characteristic life Beta of 1.5, are you referring to 1.5 years? I have to go now, but let's discuss this further later. - Rui

Rui, I have a challenge for you regarding the maintenance of a cooling tower system with 8 gearboxes. During the summer months, any gearbox failure would result in decreased production rates due to lack of spare cooling capacity. However, if a gearbox fails in the Fall, Winter, or Spring, there would be no consequences. This scenario is interesting because it involves calculating the probability of a gearbox failure during the summer months out of the total of eight gearboxes. The estimated remaining life of the cooling tower system is between 40 to 50 years. I interpret Beta = 1.5 as indicating a slight increase in the probability of gearbox failure over time, rather than a constant failure rate.

I gladly accept the challenge. To approach the problem differently from the compressor case, I plan to create a simulation model, which should make finding a solution easier this time. However, I will need some additional time to do so. Regarding your data, it seems that when you mention beta = 1.5, you are referring to the shape parameter of a Weibull distribution rather than the characteristic life. A shape parameter of 1.5 indicates a slight increase in the hazard rate over time. In order to proceed, it is crucial to estimate the scale parameter, which represents the time at which the probability of failure is 0.632 in a 2-parameter Weibull distribution. Without this information, progress will be difficult. Rui

Rui, we have estimated the scale parameter to be between 30 and 40 years based on a probability of failure of 0.632, with a shape parameter of 1.5. Does this information provide assistance?

Do you believe that engineering students at a bachelor's degree level may not receive instruction on numerical methods or the Weibull distribution unless they take specific mathematics courses? Furthermore, university subjects are often taught in a theoretical manner without demonstrating their practical applications, potentially lacking the motivation for students to fully engage with the material.

Numerical methods are a crucial aspect of Engineering courses in the first year at the University. However, the focus on Reliability is limited to certain branches of Mechanical Engineering courses in PT, such as Product Engineering and Industrial Engineering. Reliability training is also offered at institutions like the Air Force Academy and Navy Academy. Currently, Process engineering courses do not include Reliability in their curriculum, which is why I often provide professional training on this subject at companies' sites or at specialized institutes like ISQ. I strongly believe that all professionals in operations management should be well-versed in reliability and maintainability concepts. Understanding these concepts provides a comprehensive understanding of the technical and management aspects, aiding in making informed decisions to enhance operational efficiency and safety. Ultimately, this knowledge leads to higher levels of job satisfaction.

Jaz, I am committed to addressing the cooling tower issue over the weekend. You have my word on it. Best regards, Rui

Dear All, I will be posting this information as a new thread shortly. Since the discussion is focused on inventory stock levels, I believe this document will be relevant. I hope you find it helpful. Best regards, Attachment: inventory.xls (443 KB, version 1)

Thank you once more, Rui, for your keen interest in these case studies. I am eagerly anticipating the outcomes and am particularly intrigued by the insights gained from the compressor valve case. I am hopeful that delving into the intricacies of the cooling water tower gearbox case will provide further understanding. It is my belief that initiating an "RCM" type process is essential to pinpoint areas requiring focused spare part analysis.

In response to Daryls's initiative, I am also sharing a helpful tool on this forum—a simple yet effective calculator for determining the optimal time to order a critical component and the appropriate order quantity to maintain in stock. This Excel file, which I often distribute during my material management courses, can be a valuable resource for your inventory planning needs. Feel free to download the attachment titled "Spares.xls" to take advantage of this tool. Regards, Rui.

Hi Jaz, Please find attached my response to your query. Apologies for the delay. If the data doesn't meet your requirements, feel free to inform me. I hope it proves helpful. Best regards, Rui PS: I have made some edits to the English phrasing and included a revised document. Sorry for any inconvenience. Attachment(s): Cooling Tower Report PDF (50 KB) Version 1

Explore the attached EXCEL file showcasing the implementation of the MRP (Materials Requirement Planning) algorithm. MRP enables efficient planning of parts required for preventive maintenance tasks just in time. While typically found in CMMS software, this demonstration offers insight into the underlying logic by examining the formulas encoded in EXCEL. Experiment with altering data in the pale blue cells, observe the outcomes, and analyze the findings. Best regards, Rui Attachment: MRP.xls (59 KB) Version 1

I am truly impressed by the thorough analysis you provided, Rui. Your dedication and expertise are truly appreciated. I am interested in learning more about the process behind the analysis - such as the software you utilized (off-the-shelf or custom-made Excel spreadsheet) and the time it took to complete. Your insights have sparked my curiosity.

Hello Jaz, I'm thrilled that you enjoyed my work. It was a rewarding experience to dedicate approximately 6 hours resolving this intriguing problem. Utilizing my trusty self-created Excel spreadsheet, I delved into the issue with enthusiasm. While at the gym later in the day, I further pondered the case and added additional insight on the transient conditions. I believe the case is now comprehensive. Feel free to let me know if any data adjustments are needed. Great job and best regards, Rui PS: I have updated the document (originally missing the English label for table 5) Attachment(s): Download Transient Conditions PDF (26 KB) - Version 1.

Dear Rui, I apologize for the delayed response. I appreciate your feedback on the '5 Myths of Inventory Reduction' article, but I want to clarify that it is not solely focused on inventory reduction strategies. It is actually a part of the Inventory Cash Release Process, which I discuss further in my book. The '5 Myths' were derived from observations of client behavior that often hindered optimal outcomes for their companies. Regarding optimization, I do not dislike it, but I believe true optimization involves considering various system and behavioral factors, not just relying on calculations. In today's complex business world, a more holistic approach is needed to improve inventory performance effectively. While algorithms have their place, they are not the only solution. I understand your emphasis on applying math for trade-offs rather than relying solely on common sense. However, the Inventory Cash Release Process goes beyond common sense and incorporates scientific methods to drive lasting results within organizations. This process, as opposed to standalone techniques, enables behavioral changes that can be embedded long-term. I am passionate about providing companies with sustainable solutions for inventory management. I invite you (and other readers) to explore more on this topic in the article "When Inventory Optimization Doesn't Optimize" in Reliability Magazine. Feel free to reach out if you wish to delve deeper into this discussion. Thank you for your engagement in this forum.

Hello Phill, I want to express my gratitude for your detailed explanations. Please understand that I greatly value individuals like yourself who possess the expertise and dedication to influence others to improve operational management practices through logical reasoning. While this may not be my strong suit, I have chosen a different path and consciously avoid getting involved in organizational processes and teamwork. I find solace in working independently. As a consultant engineer specializing in operations economics and more recently in reliability and maintainability, as well as an academic professor, I have found that quantitative methods are often the key solution to complex problems, such as the case of Jaz's cooling tower. I strongly believe that theory and practice are complementary to each other, and the challenge lies in finding the right balance between the two. Personally, I find a sense of satisfaction and contentment in the close relationship between mathematics and philosophy. Wishing you continued success, Rui.

Rui, I have accepted your challenge and completed my assessment of the cooling tower issue. The cooling towers play a crucial role during the summer months, with a potential cost of failure ranging from $48,000 to $60,000. Currently, there are 8 cooling towers utilizing a run-to-failure strategy, which unfortunately leads to inevitable breakdowns. The extended delivery time of 38 days seems excessive and may not be easily shortened. This raises the question of whether there are alternative operational methods for the plant, or if the vendor can provide units on consignment. It is important to explore these options before investing in new gearboxes. Having spent a decade maintaining equipment in a similarly seasonal industry, I understand the importance of minimizing summer failures, especially when there is high demand for products. I question the rationale behind employing a run-to-failure strategy for such a critical piece of equipment. This alternative perspective may spark further discussions and considerations. Cheers!

Thank you, Rui and Phillip, for sharing your insights and suggestions regarding this matter. In my role as the reliability coordinator for this plant, I appreciate having multiple perspectives to draw from. It's important to be able to provide diverse answers to the same questions, as different peers will have varying expectations. Ultimately, building confidence in our decision-making abilities is crucial. By being able to defend our positions, we can increase the likelihood of decisions being implemented and resources being allocated accordingly. This will ultimately lead to more successful outcomes.

Hello Phill, I completely understand the skepticism towards numbers - how reliable are they really? This is something that I believe everyone with the proper qualifications and a sense of responsibility always considers when approaching a new issue. In my opinion, if the delivery time cannot be significantly reduced, then it is necessary to keep at least one unit in stock. The analysis time is only 30 seconds. When compared to my 6-hour work process, can you imagine how I must feel? Rather than feeling miserable, I am confident that I have taken the right approach, assuming that the data is accurate and the situation cannot be altered significantly. I have even used this experience to better prepare myself for handling similar cases quickly, in less than 30 seconds. I am puzzled as to why you did not take the opportunity to showcase your superior skills (or intuition?) to Jaz and the rest of us. Why not consider 2 or even 3 units? What assurance do you have that a second failure will not occur shortly after restarting the system? The appropriate course of action depends on the potential consequences and the delivery time. If there is no way to address the issue, then it may be necessary to increase the number of spare units on hand. Making decisions based on solid mathematical principles is key, rather than relying solely on intuition. In conclusion, I agree with your methods, but there are instances where mathematics can offer a more precise and effective solution. I believe we are in accordance on this matter, so further discussion appears unnecessary. Best regards, Rui

For those who haven't read all the posts, consider partnering with a local supplier who carries equipment supplies to streamline your maintenance process. Create kits with all necessary spare parts under one number for easy access during preventive maintenance. It's also important to benchmark each machine in your facility and analyze historical data to identify potential failure rates. Ask yourself what the impact would be on your business if a specific machine were to fail - if it would be catastrophic, ensure you have the necessary spares on hand. Develop a Bill of Materials (BOM) listing suppliers, part numbers, and lead times for each item to better manage your inventory. Prioritize parts with long lead times or high failure rates to prevent operational disruptions. This proactive approach will help you stay prepared and minimize downtime. We hope these tips are helpful for optimizing your maintenance strategy.

Thank you for your valuable input, G. It's essential to recognize that issues are not always simply black and white. There are various shades of gray to take into consideration, including different levels of severity such as negligible, marginal, critical, catastrophic, or everything in between. While I appreciate your wise and commendable approach, it may be beneficial to also conduct a quantitative analysis in certain situations where the cost of failure versus holding costs does not clearly favor a particular solution. - Rui

Rui, I appreciate your insightful response. I agree that there are varying degrees of failure consequences and approaches to determining the need for spare parts. Conducting quantitative analysis, as you mentioned in your post, is crucial in such situations. This concept resonated with me when I made my previous post, and you articulated it effectively here. Your work on the cooling water tower serves as a valuable contribution to making a strong business case for investing $40,000 in a critical spare part. When making such decisions, I must compete for funding against other departments with equally convincing proposals for alternative uses of the capital.

Jaz, I appreciate your recognition of my hard work. I wish you success in your quest to secure $40,000 for a meaningful "charitable cause" investment. Best regards, Rui

Rui, I want to apologize if I hurt your feelings, as that was never my intention. I was simply trying to illustrate that not all problems necessitate a detailed quantitative analysis for resolution. In light of adding an alternative perspective to the discussion, here is some additional insight. I don't recall stating that I don't trust numbers; quite the contrary - I believe that the right numbers can provide the answer. However, the challenge lies in determining the correct set of numbers and the underlying assumptions associated with them. In my view, most calculation errors stem from these assumptions. I presume you would agree based on your reference to different shades of gray. As engineers, we often rely on calculation methods ingrained in us during our university education, which we trust in. Nevertheless, as you mentioned, not everything is clear-cut, and there are various shades of gray to consider. Despite the black-and-white nature of the 'calculation' approach, it is essential to limit the analysis to a logical set of numbers to navigate these nuances - otherwise, every circumstance would need to be reviewed. Every problem-solving scenario, whether related to inventory or any other issue, hinges on key factors that influence the outcome. Instead of scrutinizing every number, focus on the ones that truly matter or could alter the decision. The inventory approach I recommend was formulated using the scientific method of hypothesis falsification, streamlining the process by discerning true and false conditions. This approach, pivotal in scientific progress for centuries, forms the foundation of problem-solving. Returning to the topic of cooling towers, the numbers we can rely on are tangible figures like costs, lead times, repair durations, and downtime expenses - easily verifiable metrics. On the other hand, failure rates are speculative and uncertain, compounded by the inevitability of failure with no maintenance. Therefore, my rationale was as follows: failures are costly in terms of downtime, parts have extended lead times, and failure likelihood is high, necessitating a spare. The crucial step is questioning the constraints: Can lead times be shortened? Can maintenance tasks be expedited? By challenging these limits, an optimized solution can emerge. For further exploration, I recommend researching Double Loop Learning on Google, developed by a Harvard Professor in the 1970s. Additionally, you can refer to my article "Double Loop Your Thinking" on my website www.InitiateAction.com/articles.htm. It appears that perhaps the main issue lies in justifying the spare to management, rather than recognizing the need for it. Convincing management involves a separate problem-solving process, likely requiring calculations to bolster your argument. Rui, your demonstration of this technique has undoubtedly influenced decision-making positively and was not in vain. Keep up the good work. Cheers!

Hello Phill, in retrospect, it appears that the primary challenge lied in justifying the need for a spare part to management rather than recognizing the need itself. Convincing management to approve the purchase involves a different set of skills, which includes making persuasive calculations to support your argument. I appreciate you highlighting this aspect of the issue, as it was something I had not previously considered. Throughout my career in consultancy, my focus has always been on advising management personnel. As a result, I have noticed a tendency among practitioners to overlook the significance of incorporating quantitative methods that integrate financial considerations when addressing technical issues. Money-related matters have always been the focal point of my work, which is why I lean more towards utilizing numerical data rather than qualitative approaches, despite understanding their effectiveness. I find statistics and simulation to be invaluable tools for evaluating reliability concerns. Thank you for your insights and for sharing the links to your articles, which I have found to be insightful and informative. I look forward to exploring "Double Loop Learning" from Harvard, as per your recommendation. Best regards, Rui.

I present you with a challenging scenario to test your problem-solving skills. Imagine your company has 3 pumps, labeled A, B, and C, each with gaskets (2 per pump) that fail according to a 2-parameters Weibull distribution. The shape parameter is consistent across all pumps at 2.6, but the scale parameter varies: Pump A lasts 900 running hours, Pump B lasts 1,200 running hours, and Pump C lasts 1,600 running hours. The average daily working times for the pumps are: Pump A = 16 hours, Pump B = 14 hours, Pump C = 13 hours. Preventive replacements occur at set intervals: Pump A at 500 running hours, Pump B at 700 running hours, and Pump C at 900 running hours. The pumps are operational 365 days a year. The lead time for replacement gaskets from the vendor is 1.5 months. If you aim for a 98% service availability, do you think it's necessary to keep gaskets in stock? If so, how many should you have on hand? Share your insights on how to manage this maintenance challenge effectively. Rui.

Although there is no quantitative method to confirm, my assumption is that all 3 pumps must always be running simultaneously and require the same gaskets. In order to optimize performance, would it be beneficial to set the minimum and maximum levels of gaskets at 1/3 sets? Please provide your answer after considering others' replies.

Thank you, Josh, for sharing your guess. I look forward to hearing from others. Best regards.

Starting with the assumption that there are significant consequences if any of the pumps fail and that management would be greatly concerned about the cost of stocking gasket sets, I conducted my analysis. In my opinion, the re-order point should be set at a minimum of 3 for 98% Availability. Since each pump requires 2 gasket sets, I recommend stocking 4 (ordering in multiples of 2). Drawing from my experience in mobile equipment maintenance, we always kept a minimum of 3 pistons for a 4-cylinder diesel engine to ensure availability. I choose not to disclose my specific approach to this problem at this time.

Jaz, I am extremely interested to learn about the process you used to arrive at those numbers. Let's compare our methodologies directly once we gather more responses, if you're willing. Thank you.

We eagerly welcome responses from additional peers, particularly from individuals like Phill and Daryl who have previously engaged with this topic. Your valuable insights and input would greatly enhance our discussion.

Hello everyone, back in 1990/91 during my graduate studies, I delved into the Weibull distribution. Interestingly enough, it's been 15 years since I last used or even thought about it! To be honest, my knowledge on Weibull is quite rusty. Rui, I believe your analysis will be spot on given the assumptions of the problem. Here's my take on it: I assume gaskets are supplied and used in sets, with our maintenance system efficiently ordering parts for preventative replacement. Starting with two in stock ensures I never run out for PM, sometimes even having up to 3 sets available. If the PM stock is 'reserved,' I manage unexpected failures with just 1 set on hand, occasionally having up to 3 sets. Without more insight on failure history, this stock level seems appropriate. As for the cost of the gaskets, this is crucial information to know. If inexpensive, stocking up is a no-brainer. However, holding excess sets may not be the best solution. Additionally, the significance of the 98% figure is worth pondering. Are 2 failures out of 100 truly acceptable? This could result in substantial downtime if stock isn't available. Rui, we eagerly await your final analysis!

Thank you, Phill, for sharing your insights on this matter. In my view, it is essential to ensure that parts required for preventive maintenance (PM) are ordered in advance to avoid any delays in the process. The parts kept in stock should be designated solely for corrective maintenance (CM) tasks. It is important to note that even with a PM schedule in place, there is still a possibility of a gasket failure occurring unexpectedly. Many assume that maintaining no stock is necessary if PM is followed diligently. However, this assumption may not be entirely accurate, which is why I am raising this question. I have not provided the unit price of the gaskets, but it is relatively low. If the price were higher, a balance between holding costs and opportunity costs would need to be struck, leading to a scenario similar to the one observed with the Jaz cooling tower. Instead, I have focused on the desired level of service (98%), which indicates a 2% chance of encountering a stockout when parts are needed urgently. It is crucial to have a reorder point in place, even if the gaskets are inexpensive. Stay tuned for the exciting conclusion next weekend, as I will reveal the final outcome and explain the methodology used. I hope that this information proves valuable to readers. PS: Rui has already discovered the correct answer.

Please take a look at the attached files containing information on gaskets for three pumps. I believe you will find them insightful and helpful. Thank you, Rui.

It appears that I need to make two separate posts to attach multiple documents. Nonetheless, please find the attached PDF document titled "Gaskets of 3 Pumps." Regards, Rui. Size: 23 KB. Version: 1.

Please find attached an Excel file on the Weibull distribution for your review. This file includes information on the expected mean life of a part that fails within a certain timeframe, the conditional probability of failure after an idle period, and the expected mean life of a part operating after an idle period. By adjusting specific cells, such as N13 and N3, you can obtain valuable insights into the MTTF and other relevant data. Feel free to explore the attached file, "Weibull_distribution.xls", for further details. Best regards, Rui.

Dear Rui, your dedication to answering the question is truly commendable. Thank you for your hard work and thorough analysis. It appears that the quantitative approach has yielded a significantly higher number of predicted gaskets compared to the other inputs. Here is the breakdown: Josh - 1 - 3 sets, Jaz - 4 gaskets (equivalent to 2 sets), Phill - 1 - 3 sets, and Rui - a minimum of 4 sets, with the maximum quantity uncertain. The surprisingly large number of gaskets predicted by the quantitative analysis can likely be attributed to the higher unreliability of gaskets in this specific scenario, exceeding what would be considered normal in practical applications. I believe your assistance may be required here, Rui. Upon reviewing the spreadsheet, it appears that the probability of gasket failure before the next preventive maintenance (PM) for each pump is around 20%. Therefore, if pump A undergoes gasket replacement every 3 weeks (or 500 hours), on average, a failure is expected to occur every fifth cycle (20%) - is my understanding correct? This translates to an average of one failure every 15 weeks, which equates to more than three failures per year. This observation suggests potential flaws in the current PM strategy and/or the design of the pumps. Do you agree with this assessment, and do you validate my logic? Since your initial analysis was primarily illustrative, I would appreciate your insights on the implications of these figures. Thank you for your collaboration.

In this scenario, brute force calculation is utilized without the use of Weibull analysis. The assumption is that regular replenishment for preventive maintenance (PM) is managed by the stores, simplifying the process of determining the minimum stock needed. However, as we know, real life is rarely that simple. The minimum stock required is 2 sets, equivalent to 4 each. The motivation behind this is the operation of 3 pumps, with the client's preference to operate pumps independently. Since one set can potentially malfunction, quick replacement is necessary. In our situation, the vendor is not in close proximity, with a minimum lead time of 2 months unless willing to pay premium costs. Additionally, API pump gaskets are not widely available in hardware stores. In light of these factors, the vendor suggests purchasing 6 sets rather than just 1 for convenience.

Rui, in response to your inquiry about my analysis approach, here is how I conducted it: Initially, I utilized a Weibull spreadsheet to determine the likelihood of gasket failure prior to preventive maintenance, with all three pumps estimated to have around a 20% failure rate. Subsequently, I calculated the Mean Time To Failure (MTTF) for 5 cycles. Next, I employed your spreadsheet to input the MTTF data for each of the three pumps, each equipped with two gaskets, in order to reach my conclusion. This entire process consumed approximately 20 to 30 minutes of my time, which could potentially be expedited with more practice. As noted by Phillip, the pumps exhibit a significant failure rate, especially considering the impact on maintenance budgets and overall throughput. In my opinion, it is crucial to address problematic components that have the potential to cause substantial financial losses and operational disruptions. While some may prioritize addressing known issues, I advocate for a shift towards proactive maintenance strategies to mitigate unforeseen failures that can significantly impact maintenance costs and operational efficiency. In my current role, I frequently encounter inquiries about enhancing the reliability of specific "bad actor" equipment that poses a challenge to our operations. While we excel at managing and repairing known issues, there is a need to focus on preventive measures to avoid unexpected failures that can incur substantial costs. Since joining this site six months ago, we have successfully averted an issue that could have accounted for a significant portion of our maintenance budget and throughput. Moving forward, my objective is to foster a culture that emphasizes proactive maintenance practices to address and prevent costly bad actors effectively. It is essential to not only address immediate concerns but also prepare for potential future challenges that may have a significant impact on overall operations.

Hi Jaz, I'm uncertain about giving advice on addressing your unexpected challenges, but have you looked into the PMO strategy proposed by OMCS? I wanted to discuss the term 'bad actor' commonly used in the engineering field. I am interested in its origins, meaning, and history. Can you provide any insight on this topic? Thanks.

Chronic issues like bad actors are a common occurrence in many industries, while acute events such as fires or explosions are rare but can have a significant impact when they happen. Which type of event incurs higher costs? It is often argued that chronic issues like pump failures caused by bad actors are more costly in the long run, while the costs of acute events can be spread out over time. It might be best to open a new discussion thread to delve deeper into this topic.

Phill, you raised some intriguing questions - it's fascinating to consider the various idioms used in different regions around the world. Josh summarized it well. The term "bad actor" is one I picked up from the oil and gas industry in North America. Building on your earlier point about expecting failures every 15 weeks, I would classify pump A as a "bad actor." Additionally, after calculating the spare parts needed, I would assess the potential consequences of the pump failing as being significant. From a personal standpoint, I would also factor in the cost of the parts, prioritizing efficiency over time spent. This might prompt further analysis and potentially a redesign. I really appreciated Rui's scenario as it presents a solid test for complex business situations that require a high level of scrutiny.

I first learned about PMO during a seminar at the 2006 RCM conference in Las Vegas. While I wouldn't consider myself an expert, the seminar provided me with valuable insights. Neil Bloom's discussion on RCM and its ability to prevent unexpected failures caught my attention, shedding light on the importance of RCM. Many companies have the potential to save on maintenance budgets, but the costs of not preventing failures can outweigh these savings. It is essential to continuously improve maintenance strategies to prevent catastrophic events, as equipment failures can contribute to larger failures. By asking the right questions, one can uncover underlying equipment failures that may lead to catastrophic events. The goal is to reduce the frequency of these failures, with equipment faults potentially playing a larger role than previously thought. This line of thinking can lead to further exploration and improvement in reliability practices.

It's great to see the high level of engagement and interest in this topic! Phil, you make a valid point about interpreting the numbers. When it comes to maintenance tasks, the periodicity of preventive maintenance (PM) was chosen with a probability of gaskets needing replacement under a corrective maintenance (CM) task around 0.2. Many people follow the rule of thumb "90-10", meaning if no other criterion applies, there is a 10% chance for CM and 90% for PM. Cost effectiveness plays a crucial role in decision-making when considering the costs and time associated with CM and PM tasks. For example, if the cost difference for replacing gaskets in pump A during CM and PM tasks is significant, the decision-making process should be based on minimizing costs or maximizing operational availability. It is advisable for the maintenance manager, rather than store personnel, to determine reorder points based on factors like cost and availability of facility. Calculations are essential, especially when vendors offer bulk discounts which may affect the overall cost effectiveness. Jaz, your approach to calculating Mean Time to Failure (MTTF) and demand for pumps is interesting. However, it is important to ensure the correct interpretation of data to avoid overstocking, especially when considering parts needed for CM and PM tasks separately. Josh, I have developed a software tool called INES to assist with cases like these, which was created by my son using VBA. The tool helps in analyzing stock data and determining reorder points efficiently. I look forward to hearing your thoughts and solutions to the problem presented. Your expertise and knowledge in reliability and efficiency are highly appreciated. Regards, Rui

Upon reviewing the attached information, I realized there were some errors in my initial calculations. I used the spreadsheet to determine the reorder point, and in recalculating the Mean Time to Failure (MTTF), I found that the MTTF for each pump can be calculated as 1/(probability of failure before preventive maintenance) multiplied by the maintenance timeframe. The probability of failure before preventive maintenance was determined using a 2-parameter Weibull calculation, with all pumps showing a failure probability of approximately 0.2. Upon reevaluating, I discovered that 4 gaskets are needed to ensure a 99.14% success rate (as opposed to the previously suggested 3 gaskets for a 96.33% success rate). It appears that in my initial calculation attempt, I mistakenly used the scale parameter as the maintenance timeframe. I will double-check the spreadsheet to ensure accuracy. I have yet to review the solution provided by Rui, as I wanted to first share my revised calculations. After posting this update, I will review Rui's response. Additionally, I need to make a correction regarding the conversion of a 2500-hour MTTF for Pump A to weeks: it should be 22 weeks if the pump operates for 16 hours a day, not 15 as previously mentioned. Edit: After realizing I had input the incorrect lead time and MTTF, I have removed my previous answer for accuracy.

Thank you, Rui, for your kind feedback. In regards to my analysis on the gasket reorder point issue, I am confident in my approach and solution. I utilized a unique method to address the initial query, which may differ from yours. Interestingly, it took me three attempts to double-check the numbers on the spreadsheet before posting my response. Hopefully, my manager doesn't stumble upon this!

Hello Jaz, I would like to address a miscalculation in my previous post and have already made the necessary corrections. I have attached an image of the "Spares" Excel file for your reference. It is important to note that this file does not take into account planned maintenance (PM) actions and should only be used for random failure modes. For wear-out failure modes, refer to the "Gaskets of 3 pumps" Excel file shared earlier this month. By separating CM and PM actions, the reorder point is 4 units, but when combined, it doubles to 8 units. This impacts both the average stock levels and holding costs. Thank you, Rui.

Rui makes a valid point regarding the ordering process for PM components. In the past, I assumed that the planner would order gaskets and other parts separately as needed for preventive maintenance. However, this approach may lead to higher administrative costs. When I worked in mobile equipment, we followed a similar method of keeping stock for unexpected events and ordering kits for regular PMs. While this kept planners busy, we had not yet optimized the process for efficiency.

In my view, assigning the responsibility of determining reorder points shouldn't solely fall on store employees. One approach could be to consider lead times and understand the item. When it comes to inventory management, it's important to have a clear understanding of the nuances, especially for stores specializing in beanies.

When determining reorder points, it is crucial to designate someone who is both responsible and accountable for the operational and financial impacts. This alignment is essential for optimizing inventory levels and avoiding unnecessary costs. Unfortunately, many companies overlook this key factor, leading to either overstocked or understocked inventory. By ensuring that the right individuals are making these decisions, businesses can effectively manage their inventory and maximize their resources. Cheers!

Frequently Asked Questions (FAQ)

FAQ: 1. What is the importance of differentiating between demand from wear-out failures and demand from random failures in optimizing spare parts inventory for preventive maintenance?

Answer: - Correctly identifying and addressing these separate demands is crucial to justify holding certain spare parts in stock permanently. Failure to differentiate between them can lead to increased average stock levels and holding costs beyond optimal levels.

FAQ: 2. How can maintenance departments effectively address the mixture of demand resulting from wear-out failures and random failures?

Answer: - Maintenance departments can optimize spare parts inventory for preventive maintenance by understanding the root causes of demand, analyzing historical data, and implementing strategies to manage stock levels based on the specific demands.

FAQ: 3. What are some best practices for optimizing spare parts inventory for preventive maintenance?

Answer: - Best practices include conducting thorough analysis of maintenance data, implementing predictive maintenance techniques, establishing clear demand forecasts, and continuously monitoring and adjusting stock levels based on demand patterns.

FAQ: 4. How can companies ensure they are justifying holding certain spare parts in stock permanently?

Answer: - Companies can justify holding certain spare parts in stock permanently by accurately identifying demand sources, optimizing stock levels based on demand patterns, and continuously evaluating the cost-benefit analysis of stocking specific spare parts for preventive maintenance purposes.

You must be a registered user to add a comment. If you've already registered,
sign in. Otherwise, register and sign in.