Our pump sets boast an impressive 16-year Mean Time Between Failures (MTBF) record. As Vee pointed out, numerous factors must align for this level of reliability to be achieved. It is crucial for pumps to be specifically engineered for longevity to ensure a high MTBF. Beware of cheaper, lower-quality pumps on the market that are unlikely to offer the same level of durability and reliability.
Richard, are you referring to each individual pump set or a collective group of them, which consists of all installed pumps, regardless of their operational status, typically located on the higher end?
In maintenance analysis, it's important to note that a standby pump does not accumulate operating hours when calculating Mean Time Between Failure (MTBF). MTBF is determined by dividing cumulative operating hours by the total number of failures, which includes all stoppages caused by breakdowns, CBM analysis, and PM rework (i.e., all corrective maintenance work). It's essential to exclude stoppages for operational reasons, preventive maintenance (PM), or project work when calculating MTBF.
If you are looking to improve the MTBF of your equipment, here are some steps to consider. First, conduct a detailed analysis of the process associated with the problematic equipment to identify any contributing factors beyond the equipment itself. Next, thoroughly review the Pump-Set to identify any design flaws or issues that may be affecting performance. Make sure to involve your operations and maintenance teams in this process. Then, examine the installation and operation of the equipment in question, paying close attention to any changes that may have occurred over time. Lastly, review the maintenance history to identify any recurring issues or failures. It's important to follow engineering principles and refer to manuals or information related to the equipment's design, installation, operation, and maintenance. By following these steps, many issues can be prevented. Remember to always strive to be knowledgeable about the equipment you are working with. Practice these guidelines to improve equipment reliability.
Thank you, RM, for your valuable advice. We are able to carry out those tasks as suggested. Have you noticed any improvements in equipment Mean Time Between Failures (MTBF) following the analysis?
- 14-09-2024
- Wesley Jenkins
I concur with Vee's explanation of estimating MTBF for equipment. It seems that the MTBF values mentioned in HP BlocH's article are applicable to the entire pump class, rather than for each specific pump set. This led me to inquire with Richard about whether the MTBF of his pump sets, which can reach up to 16 years, are calculated based on individual units.
Implementing a pump switching schedule is essential for effectively managing standby pumps in CMMS. Without this, it can be challenging to accurately account for non-operating hours. Manual entry of downtime may be required in order to estimate MTBF directly in CMMS. How are others addressing this issue? One solution is to use the breakdown field when submitting a work notification, indicating equipment failure with specified downtime dates, object parts, damage, and cause codes. This ensures accurate tracking and maintenance of equipment.
Josh originally inquired about whether MTBF calculations were performed for each individual pump, each type of pump, and the overall pump class. The MTBF was only calculated for centrifugal pumps as a type, not for each specific pump. However, the engineer did not agree to perform these calculations.
Have you identified the cause of the high MTBF in pumps that have failed, as inquired by operations? Why have estimates for the MTBF of each pump not been provided? Is it possible for the SAP PMIS to conduct this standard analysis? Have non-operating hours of pumps during standby been taken into account?
- 14-09-2024
- Victor Thompson
quote: Originally posted by Josh: Are you referring to each specific pump set or a collection of them that includes all installed pumps, regardless of whether they are in use or not, that typically lean towards the high side? In this case, I am referring to individual pump sets, measured by the total running time of each pump in the set divided by the number of CM work orders for the set. I exclude regular PM work that involves shutting down the pump, such as changing oil in bearings, or inspecting line shaft bearings with a borescope. I also do not include performance testing or other types of condition analysis that require the units to be shut down (I am unsure if this is what Vee means by "CBM analysis" or not).
Richard, that sounds promising. When evaluating pumpsets, how many do you typically assess? Is the MTBF determined through CMMS data or by manually calculating failures documented in CMMS? Have you accounted for non-operating hours when pumpsets are on standby? Is there a structured pumpset switching schedule in place? How do you monitor the total operating hours of each pumpset, including standby units? Are these hours consistently recorded in CMMS?
quote: According to Josh, have you identified why there is a high MTBF for certain pumps as questioned by operations? Why did you not agree to calculate the MTBF for each individual pump? Can your SAP PMIS software perform this type of standard analysis? Did you remember to exclude non-operating hours of pumps while they were on standby? Apologies for my lack of clarity, I did not agree with calculating the MTBF for a group of pumps instead of each individual pump. I have previously calculated the MTBF for each pump and presented the findings to our engineer, however, they insisted on only calculating the MTBF for the pump group. Additionally, they failed to exclude pumps in standby mode from the calculations. Is this a mistake on their part?
Panuphan, have you considered calculating the Mean Time Between Failures (MTBF) for each pump individually, rather than as a group? What was the response from your maintenance manager? Did they align with your engineer or have a different perspective? Have they addressed the issue of reconciling the gap between the perceived high MTBF and the actual performance of the pumps as witnessed by operations?
Richard, let me further explain the point you brought up: Condition-based maintenance (CBM) is focused on identifying items that are already showing signs of failure. This means that the failure is inevitable, and the signs of degradation are becoming unacceptable, whether it be through vibration levels, TAN numbers, flow reductions, or any other parameter. By recognizing these indicators, we can estimate when the item will eventually fail. As a result, all CBM-based condition monitoring activities are aimed at addressing failures that are already in progress.
On the other hand, preventive maintenance (PM) actions are taken proactively, either requiring the machine to be stopped or not. These actions are taken before the actual failure occurs, sometimes well in advance or just before it happens. PM is based on predictions of an item's lifespan, typically determined by historical data or expert judgment. This approach is generally conservative, ensuring that the item is unlikely to be unusable at the time of the PM activity. The decision to perform PM is economic, aimed at minimizing production losses by replacing or repairing the item early, even if it means sacrificing some remaining useful life. In these scenarios, the items have not yet failed and may not do so for some time. Therefore, all PM activities can be excluded from the list of failures.
It is important to differentiate between CMs carried out post-breakdown or following rework after a PM, as these instances can be considered clear failures. Additionally, modifications or temporary fixes implemented to aid operations may be categorized as CMs in your system, but these should be recognized as breakdowns and should not be included in the failure analysis.
Hi Josh, here are the responses to the questions you asked:
- We have a total of 49 pump sets in the plant, with 16 considered critical for monitoring based on size and process criticality. MTBF data is tracked for these 16 sets, while the others are either rarely used or small and therefore do not have MTBF data maintained.
- The MTBF for larger pump sets is estimated from individual run hours recorded in the DCS, while others are tracked manually. The number of failures is extracted from the work order system.
- Non-operating hours of pump sets on standby are excluded, with only operating hours considered. Some pumps have individual run hours tracked, while others are tracked as a pair. A formal pump set switching schedule is implemented on a monthly basis.
- The cumulative operating hours of each pump set, especially those with standby units, are not recorded in the CMMS. The MTBF measure is seen as a benchmarking tool and part of a suite of KPIs. The rolling 12-month calculation is used for the 16 critical sets, but it may not be as meaningful for smaller plants with fewer pumps and breakdowns.
- Other tools such as pareto analysis are used to focus improvement efforts, with the MTBF measure serving as a benchmark for comparisons and trend analysis over time.
Overall, the MTBF measure is useful as a benchmark and to track changes over time, but may not be directly comparable to others without knowledge of what is included in the calculation. If the MTBF number is low, it can help prioritize improvement efforts.
- 14-09-2024
- Quentin Foster
Check out our Pump MTBF graph to observe the step change resolution caused by the small sample size. Explore the detailed analysis of our pump's Mean Time Between Failures to see how it can impact your operations.
Thank you, Vee, for the clarification. In my calculations, I generally follow the same methods. I have previously factored in pump overhauls based on vibration data, but one area where I have not typically accounted for breakdowns or failures is oil changes guided by sample results.
Thank you for your response, Richard. I have taken note of some key points:
1) It is beneficial to extract running hours from DCS. Currently, we manually input running hours from equipment meters into CMMS to schedule preventive maintenance tasks based on hours of operation. I am exploring if running hours can be automatically exported from DCS to CMMS periodically.
2) Calculating equipment MTBF using a 12-month rolling average may result in infinity when there are no failures. Should this calculation be cumulative over the equipment's lifespan or age so far? It would be helpful to get input from others on this matter as well. Monitoring and comparing MTBF trends can help evaluate the effectiveness of equipment maintenance strategies.
3) Does your CMMS include a breakdown field for users to indicate equipment failures when submitting work requests? This feature would automatically track the number of equipment failures unless the field is not completed by the user. This streamlines the process of calculating equipment MTBF within the CMMS system.
- 14-09-2024
- Frances Fisher
Richard, if all your pumps are equipped with mechanical seals, you have an opportunity to enhance their reliability by adjusting the operating schedule. Consider operating pump 'A' for 7 weeks and pump 'B' for 1 week, or even running pump 'A' for 12 weeks and pump 'B' for 1 week. By implementing this alternating schedule, you can potentially increase your Mean Time Between Failures (MTBF) by approximately 10%, leading to improved uptime and cost savings. Unlike pumps with gland packing, pumps with seals do not require alternate running. For more information on the impact of pump starts on seal wear, refer to Chapter 40, page 303 of the book "Case Studies in Maintenance" (previously known as 100 Years in Maintenance).
- 14-09-2024
- Shawn Thompson
Thank you for your insights on the rotation schedule, Vee. Our pump sets have a combination of mechanical seal and packed glands. Surprisingly, mechanical seal replacements make up a small percentage of failures despite the 50/50 rotation. On the other hand, packed gland leaks and failures result in a much higher number of corrective work orders. I will incorporate this information for the team's evaluation. Additionally, could you please clarify the rationale behind changing the title of the book?
quote: Shared by Josh: 1) Extracting running hours from the DCS system is a beneficial practice. In addition to this, we manually input running hours from equipment meters into our CMMS system to schedule preventive maintenance tasks based on usage. I am uncertain if running hours can be regularly transferred from DCS to CMMS. Personally, I import data into Excel from PI and then perform a batch import into Maximo on a monthly basis. Currently, we have not taken the electronic integration route yet. quote: 2) When calculating equipment Mean Time Between Failures (MTBF) using a 12-month rolling average, the result turns into infinity in case of no failures. Should this calculation be cumulative and based on the equipment's overall lifespan or age up to that point? It would be helpful to hear input from others on this matter. While I calculate this for individual pumps and sets, it may not provide a clear indication of performance improvement over time. I still encounter situations where this metric yields infinity for two pump sets. quote: 3) Does your CMMS system include a breakdown field that users can mark when raising a work request? By ticking this field, the system can automatically track equipment failures, unless left unchecked by the user. This eliminates the need to manually review work orders for each equipment to determine the number of failures. Our goal is to automate the calculation of equipment MTBF through CMMS. I concur with this approach, although we have not utilized this field for unknown reasons. While this can be changed, it is not our current priority.
I agree that calculating MTBF (Mean Time Between Failures) for individual pumps and sets is important, but it may not provide a comprehensive measure of performance improvement or decline over time. I have encountered two pump sets with an MTBF measurement of infinity. This raises questions about the reliability of using MTBF as a sole metric. How can a pump set have an MTBF of infinity? Could it mean that there have been no failures since the equipment was installed?
- 14-09-2024
- Gregory Hughes
Vee, if we follow your recommendation to replace pumps with mechanical seals every 9 or 12 weeks, is it still necessary to manually rotate the shafts to avoid false brinelling?
- 14-09-2024
- Yvonne Mitchell
Richard, addressing packed glands is crucial for preventing failures. Have you considered pre-moulding the packing rings? Utilizing a shaft fummy and mould box to pre-compress the rings may require an investment of time and money, but I have seen a significant improvement in performance as a result. Alternatively, replacing the packed glands with mechanical seals could be a viable option, as there have been many advancements in this area that could meet your needs. Regarding "100 Years," a new print run was necessary, as the Publisher believed that the new title better reflected the content of the book.
Despite the opinions of many experts, I believe that false brinelling is often exaggerated. In my experience, what is often perceived as false brinelling is actually caused by stray currents, requiring a different solution altogether. The idea that shafts need frequent turning only applies when they are very slender and rotors are extremely heavy. There is also a misconception about motor freeze hardening and the need to warm up bearings. Despite working in maintenance for 40 years, I have yet to see any evidence supporting these claims. Vishal shared a story about how unnecessary maintenance work was created by sparkies replacing perfectly functioning motor bearings simply because they were considered "old". As a result, the new bearings are showing significantly higher levels of vibration. It's important to consider the real cause of issues before creating unnecessary work for ourselves.
Are you referring to shaft currents and bearing defects when mentioning stray currents? These issues have been a concern since the early 1900s, with research detailing how shaft voltages and currents can impact bearing lifespan. Factors such as accidental potential on motor frames, electrostatic charging, shaft magnetization, electric dissymmetry, and common mode voltage can all contribute to these damaging effects. Induced currents flow through the rotor, bearings, and frame, creating a path across the bearings to ground. To mitigate these risks, solutions like isolated bearings, shaft grounding, and non-conductive grease can be implemented. Other options include Faraday shields, individually tuned filters, and specialized inverter designs. Ensure proper maintenance and monitoring to prevent potential damage. For more information, visit http://reliabilityweb.com/inde...and_bearing_defects/ from PdMA Corporation.
Vee, are you saying that even for pumps packed with glands, there is no need for manual rotation of their shafts?
One common misconception in the automotive industry is the idea of motor grease hardening. Many mechanics, or "sparkies," believe that it is necessary to warm up bearings that have not been in use for a while. However, there is no definitive proof that motor grease hardens when left idle. It is important to clarify this myth and debunk any false beliefs. Additionally, it is surprising that there is no auto spelling check feature available to prevent confusion.
I am currently experiencing some eye issues, so please forgive any typos in my response. To clarify, grease is used to lubricate, and sparkies refer to electricians. In regards to gland-packed pumps, is manual rotation of the shafts necessary? Gland-packed pumps require periodic running to prevent the packing from drying out and becoming rigid. Rotors with slim shafts can potentially bend under their weight when not in use, so they need to be turned 270Λ regularly. This is particularly important for heavy rotors found in multistage pumps or compressors. Most pumps operate at low speeds, typically below 3600 rpm, and have sturdy shafts.
- 14-09-2024
- Jasmine Howard
In response to Josh's query about stray currents, these may refer to shaft currents and bearing defects. This list is inclusive, but one crucial addition would be Welding return currents. Negligence from welders, particularly during shutdowns, in connecting their 'earth' lines properly can result in return currents passing through bearings, such as those in blowers.
What are your thoughts on the argument that Mean Time Between Failures (MTBF) may not be an effective indicator of reliability? Some critics suggest that calculating MTBF using a 12-month rolling average can result in infinity when there are no failures. Should this calculation be done cumulatively over the equipment's lifecycle or based on its age so far? It would be valuable to hear from others on this topic. I personally agree with this perspective, as I have encountered situations where MTBF is infinite for certain pump sets. Additionally, I am curious about opinions on the relationship between MTBF and conditions like cataracts or conjunctivitis.
Josh, while MTBF can be useful for measuring reliability improvements and computing test intervals for hidden failures, it may not be the best indicator for determining maintenance timing. It is commonly believed that MTBF is reliable, but there can be errors in its calculation, especially with small sample sizes. This can result in very poor results, making single-item MTBFs unreliable. Grouping items together to increase sample size can present other challenges, as items in the group may not be identical or independent. Despite these issues, trending MTBFs such as for pumps remains the preferred method for benchmarking progress and making intersite comparisons. It is important to consider fundamental issues such as the definitions of failure and MTBF, as these can vary from site to site. In personal news, I am currently experiencing complications from glaucoma in both eyes, which is affecting my ability to read and write. I hope to have this resolved in the coming weeks.
We utilize MTBF data for long-term budget forecasting. What is particularly intriguing is uncovering the original design flaws. However, it can be frustrating when management changes before a costly design change is approved. To maintain equipment performance, we have developed an Excel tracking system that monitors machine run time for timely oil changes. By analyzing the run time data, we can identify when parts are operating above optimal temperatures, indicating they are in service. While this system is efficient for tracking machine components over a 2 to 3 year period, we found it challenging to apply the same method to calculate MTBF due to the file size and lengthy processing time. Our equipment history file is categorized, listing incidents such as feed pump malfunctions and required repairs. Despite having a large number of equipment pieces since 1984, the average MTBF is around 5 years. Interestingly, the intake screens, although not critical to production, become a top priority when multiple failures occur during heavy debris seasons. It is fascinating to observe the evolution of repair costs over the years, highlighting the importance of regular maintenance practices.
Are you utilizing a CMMS system at CBMV?
Vee, I'm saddened to hear about your eye issues. I wish for a speedy recovery for you.
- 14-09-2024
- Heather Coleman
According to Josh, approximately 70% of data is managed through SAP, while the remaining 30% is scattered across shared drives, intranet protocols, and old project documents dating back to 1985. This includes manual entries in file cabinets with dates and initials covered in "white out." The transition from WPTS to SAP CBMN occurred in 1999 and has been in effect ever since.
Is it possible to retrieve Mean Time Between Failures (MTBF) data from your SAP PM system by generating a standard report?
I find it intriguing that I have yet to set up a standard report in SAP to track MTBF. It's uncertain if such a feature exists within the system. There are various functionalities in SAP that we haven't utilized, and the report you mentioned might be one of them. It's simple to determine actual costs and order dates for parts. However, distinguishing between a routine yearly inspection with component replacements and a repair following a failure can be challenging. MTBF is a fascinating subject that can provide valuable insights based on data from SAP and field assessments. It remains unclear how MTBF can benefit us, aside from the possibility of extending the pump's lifespan between maintenance. Just like rotating bald tires to the back can reduce the frequency of flat tires over time. In real life, I recently helped an elderly neighbor change a tire and noticed the need for air in the remaining ones. She confessed to neglecting tire checks and enjoying the sound of squealing tires while cornering. I reminded her of the importance of proper tire maintenance and promised to return to check on her tires. CBMN
When it comes to distinguishing between routine maintenance and emergency repairs in a facility, the proper utilization of a Computerized Maintenance Management System (CMMS) can play a crucial role. By categorizing work orders as either Preventive Maintenance (PM) or Breakdown (BD), organizations can effectively track whether components were replaced during a scheduled yearly inspection or in response to a sudden failure. Recognizing the importance of this distinction can enhance overall maintenance efficiency and cost-effectiveness. Would you agree with this approach?
Exploring MTBF (Mean Time Between Failures) is a fascinating subject that involves generating statistics using data from SAP and field analysis. However, the true value of MTBF may not always be clear. As one individual noted, did the system changes lead to an extended pump lifespan between overhauls? When discussing MTBF, it is important to consider which system is being referenced - the maintenance management system or CMMS. MTBF essentially represents the average runtime of equipment, but how do you assess the reliability of your equipment without relying on MTBF?
It is fascinating to analyze the original design flaws that often create obstacles in projects. Despite thorough investigation and convincing management of the need for a costly design change, the challenge arises when management turnover occurs, leading to the need to restart the process. It is crucial to establish a standard procedure for the design change process, including a designated form for approving signatures. This form should include a list of all previously exhausted causes for reference by both current employees and incoming management. Have you implemented such a procedure in your organization? How long is the typical timeframe for addressing design issues like these?
Are you looking for a way to efficiently track machine run time for timely oil changes? Our Excel file does just that by recording hours of operation and temperatures to assess maintenance needs. With minimal manual inputs required, the file effectively monitors components critical to production within a 2 to 3 year timeframe. Wondering if CMMS can handle this data? Absolutely! CMMS can seamlessly integrate with the Excel file to trigger preventive maintenance tasks like oil changes based on runtime. Additionally, CMMS can also monitor oil temperatures for trend analysis, complementing any existing DCS systems. Don't let valuable maintenance data go to waste - optimize your operations with CMMS integration today.
Quote: I'm unsure of the benefits of MTBF, apart from what was mentioned about CBMN. I'm a little confused because I thought you mentioned tracking MTBF for multi-year budget planning. Prior to that, Vee highlighted the importance of MTBF and strategies for boosting confidence levels: MTBF is valuable for assessing reliability enhancements and determining test intervals for latent failures. However, it is not reliable for scheduling maintenance timing, contrary to popular belief. Calculating MTBF can be prone to errors, especially with small sample sizes, and a sample size of 1 may yield inaccurate results. In my opinion, relying on single-item MTBF is unreliable. Grouping items together to increase sample size may present challenges as the items may not be identical or independent. Nevertheless, I believe this is a better approach than calculating MTBF for individual items. There are also fundamental issues related to defining failures and MTBF, as these definitions may vary between sites. Overall, trend analysis of MTBFs, such as for pumps, remains the most effective method for benchmarking progress and making comparisons between different sites.
It seems like your plant is making good progress by tracking MTBF, whether through manual methods or in a CMMS system. Some sites don't include MTBF in their KPI reporting or monthly maintenance reports at all, so you're already ahead of the game. Keep up the effort in monitoring and charting MTBF trends to stay on the right track.