Ready to Optimize Your Maintenance? 🚀 Try It Free
Hello everyone, I wanted to discuss the best maintenance strategy for our two pumps - one on duty and the other on standby. The first option is to use preventive maintenance (PM) or predictive maintenance (PDM) for the duty pump and reactive maintenance for the standby pump. The second option is to switch the pumps (change over) and apply the same maintenance strategy for both, either PM or PDM. Which option do you think is the most effective strategy? Additionally, when it comes to regular oil changes, which maintenance strategy should it fall under (some may refer to it as predetermine maintenance)? Thank you.
In a duty-standby scenario, it is important to consider the consequences of pump failure. What occurs if the duty pump fails? Understanding the potential repercussions of the duty pump failing is crucial for justifying pre-emptive maintenance. Similarly, what if the standby pump fails to activate? In such cases, action must be taken to ensure it starts and operates at full capacity. Instead of seeking answers from others, it is beneficial to attempt to solve these scenarios independently to gain a better understanding of the logic behind maintenance decisions.
Hello Vee, thank you for your prompt response. The scenario mentioned regarding the EX is purely hypothetical. My intention in asking you was not just to solve a problem, but to share experiences and engage in lively discussions within the group. I recommend keeping the duty pump in continuous operation and conducting regular functional tests on the standby pump to ensure its readiness for actual use. Some plants use a proactive approach of swapping pumps, which has proven effective. I am curious to know which approach is more efficient. While I personally prefer the first approach, the second approach also seems reasonable. If we opt for the first approach, should we conduct maintenance activities on the standby pumps such as oil replacement at set intervals? If so, what specific maintenance tasks should be performed? Thank you.
In my experience, running pumps in duty/standby mode can increase MTBF by approximately 10%, leading to a reduction in costs by 12-15%. This decrease in costs is mainly attributed to a significant drop in seal failures, which are directly impacted by the number of starts. Regardless of the operating policy chosen, certain maintenance activities such as oil top-ups or changes, cleaning, and taking vibration readings are essential for pump maintenance. These tasks are considered housekeeping or PM activities and are relatively low in cost. However, more invasive maintenance work is only necessary if a fault is detected during a test start.
quote: Seal failures can be significantly reduced due to the frequency of starts. This is especially applicable to pumps equipped with mechanical seals. Conversely, pumps with packing glands typically require lubrication from the process fluid. In my professional experience, the material costs for preventative maintenance (PM) outlined in our equipment budgets generally amount to less than 10% of the total equipment value, varying based on the equipment type.
When it comes to pump maintenance, it is important to note the differences between pumps with mechanical seals and those with packing glands. While pumps with packing glands require lubrication typically provided by the process fluid, pumps with mechanical seals operate differently. It is crucial to start packed glands frequently to ensure they are properly lubricated with the process liquid. This is the basis of the alternate running philosophy. Despite the introduction of mechanical seals, many individuals continue to follow old maintenance practices. Mechanical seals have a unique failure mechanism and are best kept running continuously for as long as feasible.
Both maintenance strategies have their merits, but I'd lean toward the second option of applying the same strategy to both pumps. This ensures consistency and allows you to monitor their performance on the same metrics, which can help in identifying trends or potential issues more effectively. As for oil changes, they typically fall under preventive maintenance since you’re performing them at regular intervals to prevent failure. However, integrating sensor data could elevate this to predictive maintenance, as it would allow you to determine the optimal time for changes based on actual performance rather than just a schedule. Ultimately, a blend of proactive and predictive strategies seems to be the most comprehensive approach!
I think going with the first option of using PM or PDM for the duty pump and reactive maintenance for the standby pump makes sense, especially if the standby pump isn't used as frequently; this way, you can keep your primary pump in optimal condition while minimizing costs for the inactive one. However, regularly switching the pumps could help ensure that both are at peak performance when needed, so that's something to consider. As for oil changes, I'd categorize them under preventive maintenance, since they are part of a regular schedule designed to prevent failures and prolong the life of the pumps. Regular oil changes are critical in maintaining performance, even for standby equipment, so they shouldn't be overlooked!
I strongly believe in the concept of rotation for optimal equipment performance. By allocating 50% of service time to P1, 30% to P2, and 20% to P3, you ensure that no two units are at risk of failure simultaneously. This rotation allows for continuous analysis of all three units, ultimately providing three reliable units available 95% of the time. With this approach, you can still achieve 100% availability and successfully execute planned maintenance schedules.
I have to disagree with Sam - from my perspective, implementing rotation could actually decrease reliability, raise expenses, and add to the workload. Rotation can inhibit reliability, lead to higher costs, and result in increased workload.
In some cases, a rotation strategy skewed towards 60/40 or 70/30 can be the optimal choice, despite the fact that a 50% rotation may decrease long-term performance reliability. The operating context always plays a crucial role in determining the best approach.
When it comes to maintaining items that are not gland-packed, opting for a duty-standby situation is the most effective way to minimize failure modes for a set of spared items. Unlike a 70:30 setup, which replicates failure modes for all spare units and increases maintenance efforts, a duty-standby approach offers lower failure rates overall. There are specific scenarios where a 90:10 or even 70:30 setup may be justified, especially for complex items with lengthy start-up procedures. One benefit of a 90:10 or similar policy is the elimination of 'fail-to-start' as a failure mode. In terms of reliability, availability, and cost, the duty/standby system is the top choice, followed by a 90:10 setup. While this approach may not always be feasible, it is preferred when possible for optimal results.
When it comes to choosing between the 70/30 and 60/40 split, it's important to consider more than just the pump itself. Factors such as the type of media being pumped and the likelihood of blockages should also be taken into account. Ultimately, the operating context of the equipment should always guide decision-making, as there are no set rules that apply in every situation. While running things in a duty stand by situation is ideal for reliability, it may not always be feasible.
It would be beneficial for someone with multiple parallel pump sets to conduct a trial over a period of one to two years to determine the most effective operating strategy. My preference is to primarily run a duty pump with occasional tests of the standby pump/s, following a 90/10 ratio. It is likely that the duty pump will eventually require maintenance, at which point the standby pump will take over. A helpful guide for determining when maintenance is necessary can be found in the attached paper. As many respondents have mentioned their book titles, I would like to share mine as well: Ray Beebe, author of numerous papers including "Predictive Maintenance of Pumps Using Condition Monitoring" (ISBN 1856174085). The attached document provides further insight on this topic.
Ray suggests conducting a trial over a year or two with multiple sets of pumps running in parallel to determine their performance. Based on his own trials and findings, he offers advice. Regards.
I am unsure why a 50/50 distribution is considered negative. Some people assume this means more frequent starts, but it really depends on what motivates you to start and operate the machine. Personally, I believe there is a limit to how long a machine can sit idle before reliability is compromised. For this reason, in a 90/10 scenario, I would reduce shifting intervals to avoid idle time and similarly have fewer shifts per unit time in a 50/50 scenario. Could someone point out if I am misunderstanding something?
Hi Pete, using a 50:50 ratio in production leads to an increase in failure modes that need to be addressed, making it a suboptimal choice. The same issue arises with a 60:40 ratio, which I also consider to be a poor choice. On the other hand, a ratio of 90:10 or higher is preferable due to fewer failure modes that need resolution. The number of starts does not impact the 50:50 operating mode; instead, it is influenced by the frequency of starts and duty/standby operations. In comparison, a 90:10 ratio is more favorable than duty/standby operations. However, duty/standby operations offer numerous advantages, including establishing a disciplined structure that facilitates regular PdM readings for both units. Otherwise, monitoring equipment can be inconsistent. My mention of the number of starts is solely to illustrate the deterioration rate of seals. I trust this explanation clarifies things for you. Best regards.
In order to clarify my understanding of this issue, correct me if I'm wrong (CMIIW). When selecting between duty and standby modes for pumps, it is important to designate one pump as the main duty pump that will handle the majority of operations, with another pump standing by to take over in case of any issues. In a 50:50 setup, both pumps are active simultaneously, making it different from a traditional duty and standby arrangement. In a 90:10 scenario, the duty pump operates for 90% of the time while the standby pump is tested periodically to ensure its readiness for emergencies. While a 100:0 setup is an option, it still requires regular functional checks on the standby pump, incurring additional costs. Many operators may lean towards the 90:10 setup for its cost-effectiveness. Regards, Rosmana.
In my opinion, achieving the ideal pump rotation strategy is not always feasible due to various factors, including the nature of the media being pumped. It is important to consider the likelihood of blockages and other related factors when determining the best rotation ratio, such as 70/30 or 60/40, beyond just focusing on the pump itself. Operating context plays a crucial role in the analysis of pumping applications, and setting rigid rules may not always be applicable. Through our trials on different pumping scenarios, we have discovered that a rotation ratio of 70/30 works best for pumping Lime to prevent clogging, while a ratio of 90/10 or higher is more suitable for water or similar media. The optimal rotation ratio can vary significantly depending on the properties of the liquid being pumped. It is important to consider these factors in addition to the pump's performance.
What is the specific gravity of lime and does the duty-standby policy apply to crude oil and slurry applications?
Josh, a common question that many people ask is about the specific gravity of slurry. Based on my experience working with slurry and crude oil, I highly recommend implementing a duty stand-by regime to prevent blockages in the lines. While it may not always be a perfect 90-10 ratio, practical considerations play a role in how quickly slurry can cause blockages if left stagnant. Despite this, a duty stand-by regime remains the most effective option for ensuring smooth operations.
Josh, duty/standby operation is suitable for crude oil service. The policy is not dependent on the specific gravity of the fluid being pumped. Slurries create a more serious fouling issue. Similarly, if the crude oil contains a high amount of sludge, it can lead to similar problems. Specifically, when the pump is not in use, solid particles or sludge may accumulate and block the suction line or the pump itself. This is a unique circumstance that should not be used as a reason to deviate from a duty/standby setup in most cases.
I have been closely following the emails discussing the importance of standby equipment in the petrochemical industry, where continuous processing and series lines of equipment are crucial. I support the idea of having standby equipment ready for immediate use for failure finding purposes, following the 90:10 rule. In our plant, we took it a step further by marking the standby equipment with a bright color for easy identification. This visual cue prompts questions from anyone passing through the plant, ensuring awareness of the plant's condition. In isolated mining sites, the need for reliable standby equipment is equally crucial due to long repair lead times. Having standby equipment ready is essential for maintaining plant operations. - Regards, Neil Blom, Reliability practitioner.
The concept of color-coding equipment is a great way to eliminate confusion between duty and standby items. Having three pumps in a row when only one is necessary can lead to inefficiencies and potential equipment cannibalization. This simple visual cue helps streamline operations and prevent unnecessary redundancy.
In this setup, we have three pumps in place. The first pump is the primary one that takes on the main workload, essentially cannibalizing the other two pumps when necessary. The second pump serves as a standby option, cannibalizing only from the third pump. The third pump is essentially a storage location positioned beside the other two pumps, with its sole purpose being to keep spare parts readily available for the first and second pumps. Its operational hours are minimal, serving as a backup support system for the primary and standby pumps.
I believe that cannibalization may occur as the process of replenishing spare parts is unable to meet the maintenance demands. It is crucial to ensure an efficient spare part acquisition process in order to uphold the duty-standby policy effectively.
I'm perplexed by the assumption that a 50/50 runtime is less reliable than uneven splits like 90/10. Can someone clarify this for me?
Hello electricpete, Typically, utilizing a 50/50 runtime strategy results in both pumps experiencing wear at a similar pace. In a properly planned and monitored system, this can lead to both pumps failing simultaneously. However, in situations where there are extended lead times for spare parts or all available spares are undergoing repairs, it can result in production downtime. This is not a matter of reliability, but rather a question of how reliability is handled and maintained.
Consider implementing a 90/10 strategy for one year, then during the yearly maintenance shutdown, transition the pump settings. Opt for running Pump B at 90% capacity and Pump A at 10% capacity to optimize efficiency and performance.
Starting a pump can put it in the worst condition, as seal failures are more likely to occur with frequent starts. To ensure system availability, it is important to have at least one pump working efficiently at all times. By minimizing the running hours on standby pumps, we can reduce wear and maintain their pristine condition. This proactive approach guarantees that the pump will function effectively when needed, similar to a new one. However, idle standby pumps have a hidden failure mode where they may struggle to start or operate at full capacity. To address this, it is essential to conduct test starts or implement a 90:10 policy, but reversing this ratio can compromise the pump's condition. A duty standby policy proves to be more effective in reducing failure modes and improving system availability compared to other policies. The 90:10 ratio closely resembles a duty-standby setup, offering optimal pump performance. For further details, refer to page 168 in my book.
Demonstrating the effectiveness of a duty standby policy involves examining the proof and underlying assumptions. The concept of starting duty remains consistent regardless of whether the mode is 50/50 or 10/90, with potential issues arising in the 10/90 mode. Considerations include the reliability of machines nearing the end of their refurbishment period. Factors such as timing and runtime play a role in scheduling refurbishments, with the option to manage overhaul schedules for optimal efficiency. In order to minimize the need for frequent refurbishments and avoid having two motors reach end-of-life simultaneously, staggered intervals among pumps in the 50/50 scenario may be more manageable. The planning of refurbishment intervals in a 90/10 interval setup may present challenges. Additionally, the concept of a "pristine" standby machine with minimal runtime may not be ideal for early detection of potential failures before the machine is needed for standby operation.
Dear Pete, I encourage you to explore the section on duty-standby detailed on pages 168-170 of my book. Additionally, I recommend referring to John Moubray's RCM2 book, specifically pages 28-29 and 105-110 for further insight. If any uncertainties persist, we can discuss further. It's worth noting that these concepts are not merely theoretical - I have put them into practice and noticed a significant impact! The operational approach greatly influences system uptime and maintenance expenses, as discussed in the article on 'Operating Philosophy' on the primary Reliabilitycom website. Best regards.
From a practical standpoint, a new plant with new operators and equipment operates 24/7. In the initial 3 years, our approach was to run the main pump continuously, especially with hot products in a refinery, and only run the standby pump for one shift per week (8 hours). This led to an unusually high number of mechanical seal failures. However, after changing our policy to only switch to the standby pump during preventive maintenance and when absolutely necessary, the failure rate decreased significantly. We believe in the 90-10 approach and acknowledge that we made mistakes due to our lack of experience. I appreciate your book, though I wish the Bhopal incident received the same level of documentation as the Alpha Piper disaster. Your feedback on how you conducted your experiments would be valuable. Regards.
I conducted a search on operational philosophy (I spelled it correctly) and I am looking for any relevant links. Can you provide me with some resources on this topic?
Svanels mentioned that they experienced an unusually high rate of mechanical seal failures, which decreased after changing their operating policy. By only using the standby pump during preventative maintenance and when necessary, they saw a decrease in failures. They theorize that frequent starts and operator errors may have contributed to the failures. They suggest a 90-10 approach for pump maintenance to minimize failures. The importance of operating policies on pump performance is highlighted, with suggestions for reducing failure rates. Implementing a new operating philosophy can lead to reduced costs and increased pump availability. For more information on operating philosophies, visit http://www.reliabilityweb.com/art04/operating_philosophy.htm.
Hey Vee, just wanted to update you on our maintenance schedule for the pumps. It took some convincing, but we successfully shifted from weekly 8-hour runs to bi-monthly PM-checks. We explained to operations that running the pumps at full capacity during the short hours was more effective. It's similar to a generator operating at only 60% of its capacity, causing more issues than productivity. Can't wait to read your upcoming book!
How frequently should you conduct a test run on the spare pump? Weekly may seem excessive, but applying the RCM philosophy can provide valuable insight. The key goal is to ensure that the spare pump can effectively take over the duties of the primary pump in case of a malfunction. A leading global manufacturer adopted this approach for their pumps and determined that most pumps only need to be test run once every three years to achieve a 97% availability target. This frequency is driven by the risk of seal binding after long periods of standby (MTBF = 50 years). Consider applying the same analysis to your pumps to determine the optimal testing interval.
Vee, you just made my day! We are currently in the process of implementing a vibration monitoring program in our maintenance department. We have been discussing how to efficiently gather vibration data on our standby pump, especially when we have limited time for our preventive maintenance checks. As part of our routine, we also conduct a 6-month alignment check. Allocating a week for this task would allow us to thoroughly evaluate the standby pump under normal operating conditions and in a stable state. One challenge we face is convincing the plant owner of the importance of running the standby pump for a full week. It sometimes feels like we are dictating how they should manage their plant. Our schedule of 8 weeks on and 1 week off results in a 90-10% split, with 6 starts for the main pump and 5 for the standby pump, achieving an overall availability of 97%. Though I struggle with the calculations at times, particularly in Probability and Statistics, I am committed to working through your example in the textbook to better understand the concepts. My classroom experiences with dice throwing probabilities did not spark my interest, as I am not one for gambling.
Testing the standby pump every three years is essential to prevent potential issues such as canibalization or false brinelling of the bearings due to surrounding vibrations. Failure to conduct regular maintenance on the standby pump could result in early wear and damage to critical components. It is important to prioritize routine testing and inspection of standby equipment to ensure reliable performance when needed.
According to David, a leading global manufacturer found that most pumps need to be test run every three years to maintain a 97% availability target. The main reason for this is seal binding, which occurs after a prolonged period of standby, with a Mean Time Between Failures (MTBF) of 50 years. However, in my experience, pump MTBFs typically range from 12 to 60 months when in operation. When it comes to standby, it is essential to focus on the MTBF for 'fail-to-start'. This can be calculated by dividing the number of failed starts by the total number of starts to determine the percentage of start unreliability. From this percentage, we can derive the start MTBF assuming an exponential distribution. It is unlikely to have a 50-year MTBF based on this calculation, as data from sources like OREDA suggests test starts should occur every 1-3 months. It would be beneficial to understand the data source used by the global producer. Additionally, seal binding is not the only factor to consider for 'fail-to-start', as there are other potential reasons for this issue.
In a recent conversation, Svanels mentioned the challenge of persuading the plant "owner" to run the standby pump for a week. It seems there is a disconnect between "us" and "them" in terms of plant management. However, it is important to remember that we all share ownership of the plant and that establishing a duty-standby policy can benefit everyone by increasing availability and reducing costs. By running the standby pump for a week, we can gather valuable condition monitoring readings on the backup units, ensuring regular maintenance and avoiding unexpected breakdowns. Implementing this policy also encourages operators to report failures promptly, leading to quicker response times and improved uptime. Ultimately, this transparency and cooperation between maintenance and operations can help break down barriers and enhance overall performance. Good luck in dismantling the silos and fostering a collaborative work environment.
Vee recommended referring to pages 168-170 of their book for information on duty-standby, as well as pages 28-29 and 105-110 of John Moubray's RCM2 book. Despite the lack of clarity on why 90/10 is favored over 50/50, the discussion can continue with additional insights. Can anyone provide a concise explanation for this belief? Thank you.
When conducting RCM analysis, it is important to be aware of the potential for brinelling in bearings. While this may not be a common issue in well-designed plants like the one I am currently discussing, it is crucial to consider the possibility of this failure mode occurring in your own plant.
Proper shaft rotation is essential in distributing weight evenly on bearings to prevent excess pressure on one specific point. This principle also applies to long-term storage of pumps and electric motors, ensuring even weight distribution and preventing damage.
When we implemented the 75:25 policy, with the 'A' units given a higher priority, I observed 'B' units running during morning checks. Upon investigating, I discovered that the 'A' units had issues in previous shifts without any breakdown work order issued yet. This highlighted a common issue - operators not reporting failures promptly. This revelation led to a discussion among my team, sparking a debate on the importance of mechanical seals in preventing pump failures. As noted on page 169, mechanical seals play a critical role in pump performance, with wear mainly occurring during start-up when the seal faces lack proper lubrication. Frequent starts can accelerate wear, leading to seal failures and ultimately pump malfunctions. Adopting a 90-10 approach for pumps with mechanical seals can help minimize these issues and reduce maintenance costs. However, the best approach may vary depending on the equipment type. While mechanical seals work well for some pumps, others with bearings or forced lubrication systems may be better suited for different solutions. For instance, water pumps with packing glands may benefit from a switch to mechanical seals to address persistent leaks. In light of these considerations, I seek advice on the most suitable solution for this specific equipment type. Your insights and recommendations are highly appreciated.
For improving the performance of packed glands, I recommend the following steps: create a wooden dummy shaft and a compression mold for packing rings. The mold should be made of steel with two studs and nuts for tightening the packing ring. Ensure that the surfaces in contact with the packing ring are polished to a near-mirror finish. Use pipe sleeves on the studs to limit the compression of the packing ring to a specific thickness. Measure the axial thickness of the new packing ring and compress it fully in the mold. Cut the pipe sleeves to allow the packing pusher to compress the ring halfway. Cut the packing rings on the wooden dummy at a 45-degree angle using a sharp knife and lubricate them before installation. Before working on the pump, check the condition of the shaft sleeve and replace it if damaged. Clean out the stuffing box bore and lubricate the gland studs and nuts. Prepare a metal packing ring pusher and insert the rings one by one using cylindrical pushers. Tighten the gland pusher with a shortened spanner and adjust it as needed to stop any leakage. This process has been proven to significantly extend the life of the packing in pumps, as demonstrated in Catalytic Cracker Bottoms Pumps. In addition, there is a common perception that operators may delay reporting failures. Implementing a duty/standby or 90:10 policy can help address this issue and promote prompt reporting of any problems.
Does a 90/10 ratio result in fewer occurrences compared to a 50/50 split? It's difficult to see how that would be the case.
Pete, the frequency of pump starts is important when considering maintenance and failure modes. Duty-standby systems require less maintenance and experience less downtime compared to systems with a 50:50 pump ratio. Seals may need more attention in systems with frequent starts, leading to potential bearing failures. Operators often prefer more frequent changeovers, but less frequent duty changes can be sufficient. Overhauls based on performance degradation are justified, while time-based overhauls may be too frequent. Implementing a 50:50 policy can increase downtime and maintenance costs. It is worth considering a different approach to maintenance to prevent unnecessary failures and optimize system performance. Why not give it a try and see the positive results for yourself?
quote: As shared by svanels, this topic was recently brought up with my team. One of my colleagues asked if the talented Vee was a part of our plant workforce. It seems many plants would be lucky to have Vee on board.
After analyzing the data using Excel, I have come up with a new approach for pump maintenance scheduling. My proposal involves running the main pump for 8 weeks followed by a week of rest, resulting in 6 starts for the main pump and 5 starts for the backup pump. This schedule ensures a true 90-10 ratio. We have now adopted a different method where we remove the main pump during scheduled maintenance. This adjustment has led to 6 starts for the main pump and 6 starts for the backup pump. By eliminating weekly switches, we are preventing unnecessary wear and tear on the pumps, resulting in a more efficient 99-1 approach. It is crucial to avoid excessive switching between pumps as it can lead to seal problems. Our operations team understands the importance of minimizing pump manipulation to prevent leaks. Taking into account various factors such as power outages, mechanical failures, and cavitation, it is evident that a pump with a mechanical seal can only handle a limited number of starts/stops in a year. For instance, conducting a 2-month preventive maintenance with weekly switches can result in a total of 58 starts per pump, exceeding the pump's capacity.
During my morning rounds, I observed that several 'B' units were in operation. Upon further investigation with the Control room operator, it was discovered that the corresponding 'A' units had encountered issues in the previous shift(s), yet no breakdown work order had been issued. Interestingly, they mistakenly assumed that our plant was the subject of discussion, as the behavior seemed familiar.
Thank you Eugene and Svanels for your kind compliments. I am grateful that you can benefit from the lessons learned from my past mistakes. Best regards.
Hey Vee, I'd love to stay in touch via email. Please send me a message at savanels_at_cq-link.sr. Let's connect digitally!
The MTBF I mentioned for fail-to-start issues due to seal binding is 50 years, as determined through RCM analysis. While other potential failure modes were considered, they were all found to have a higher MTBF. It is likely that the figures used were based on internal failure rate data. However, it is important to note that blindly implementing a three-year test run frequency for standby equipment may not be suitable for every plant, as each situation is unique. The company in question follows robust design procedures, such as using API610 pumps and ensuring foundations are at least 5 times the mass of pump sets. They also have a winterizing program to prevent product solidification in cold weather and primarily deal with clean products. It is crucial to utilize tools like RCM and other reliability modeling techniques to determine the best operation strategy for duty/standby pumps, considering factors such as local design practices, operational considerations, maintenance needs, and equipment criticality. Remember, there is no one-size-fits-all solution, and the resulting strategy will be influenced by a variety of factors.
David, I am struggling to achieve a reliability of at least 97% when using a 3-year test interval for a 50-year MTBF. Can you please help me pinpoint where I may be making errors in my calculations? Even if the calculations are accurate, I am hesitant to adopt such lengthy test intervals due to the numerous uncertainties involved. Additionally, I am unsure about how the MTBF value was determined. It is important to distinguish between fail to start and multiple failed attempts, as 'fake data' is a common issue in reliability analysis. While I am skeptical of the effectiveness of a 3-year test interval, I appreciate your perspective and will consider it with caution before making any decisions.
Could it be 50 months or a aquatic plant?
In analyzing the data provided by David, it is clear that the figures mentioned pertain to a specific population, as evident from the MTBF value given. A 50-year MTBF with a 3-year test interval results in less than 97% reliability. This management strategy is prudent, especially considering that companies often operate multiple pumps under similar conditions. However, crucial details such as the frequency of standby pump usage, acceptable risk levels, and failure costs must be factored in for a comprehensive assessment. Without these key metrics, any attempt to replicate the calculation will yield different results. It is essential to shift the focus from merely extending the lifespan of a single component to managing a range of potential failure modes across various activities. While the decision-making process may seem straightforward based on longevity, it is crucial to consider the broader context and other risk mitigation strategies in place. Each situation is unique, and what may seem like a sound approach for one plant may not necessarily apply universally. There is no one-size-fits-all solution, despite the allure of quick fixes.
It is important to consider other factors beyond just the number of starts when evaluating the operational needs of pumps. Factors such as potential blockages or corrosive effects must also be taken into account, as these can impact the operating philosophy and context in which pumps are used. In situations where standby pumps require frequent starting intervals to prevent damage, there may be a need for design changes. However, it is also important to address the current challenges and manage them effectively. Cheers!
After thoroughly reviewing the discussion in this thread, I gained valuable insights and decided to create a simulation model to compensate for my limited real-world experience with duty/stand-by scenarios. By running a trial simulation, I aimed to acquire "virtual experience" and address any shortcomings. The results of the simulation conducted over a specific virtual period are presented below. I welcome your feedback to make necessary adjustments and enhancements to the model. In this scenario, we are dealing with two pumps operating in parallel. One pump serves as the duty pump, while the other acts as the standby pump. Although both pumps are similar, they fail at different rates - the duty pump fails at a rate of 0.005 times/hour, while the standby pump fails at a rate of 0.001 times/hour. The standby pump has a 0.05 chance of failing to start (one out of every 20 attempts). The repair times for both pumps follow a lognormal distribution with Mean = 2 and SD = 0.5. After running the simulation for 100 repetitions and achieving a 95% confidence level, the system metrics are as follows: - Reliability (for a mission of 8,760 hours) = 0.354965 - MTTF = 8,733 +/- 304 hours - Availability = 99.93% - Duty pump utilization = 95.98% - Standby pump utilization = 3.86% In a revised scenario where both pumps are interchangeable, fail at the same rate of 0.005 times/hour, and have similar failure modes, the system metrics are as follows: - Reliability (for a mission of 8,760 hours) = 0.201455 - MTTF = 5,384 +/- 164 hours - Availability = 99.92% - Duty pump utilization = 49.85% - Standby pump utilization = 50.07% The reliability of the system is higher in the 90/10 case, as expected, with availability remaining relatively consistent. Screenshots of the simulation are included in the attached files. Feel free to suggest any modifications to the experiment. Regards, Rui.
Thank you, Rui, I will investigate further into this matter.
Rui, I found your information to be quite helpful. I am particularly interested in the availability numbers provided, which may involve certain assumptions that require further examination. For instance, increased unreliability usually leads to more failures. How many events occur for each case? Is there a downtime associated with each failure, and if so, what is the Mean Time To Repair (MTTR) being utilized? I would recommend considering a 6-hour window with a variability of ±2 hours for each failure event as a starting point. Additionally, since you already have a simulator in place, have you considered running it for scenarios with a 75:25 and 60:40 distribution? Your insights on these situations would be greatly appreciated. Thank you.
I apologize for forgetting to translate the screens into English. I have added new ones to this post to address the inconvenience. I will respond to Vee's questions shortly, as I am currently leaving the office. Regards, Rui. Download the attachment(s): 50-50_90-10.zip (252 KB, 1 version).
Thank you, Vee, for your feedback. The 90:10 model involves simulating 1,000 failures of the duty pump. Each iteration varies in the time span between the system starting and the 1,000th failure of the duty pump being resolved. One iteration produced the following results: Simulation horizon = 199,879 hours Total system failures = 26 Failures before mission completion = 17 Reliability = 0.346153846 MTTF = 7,688 hours MTTR = 6 hours Availability = 99.92% On the other hand, the 50:50 model simulates 1,000 failures of both pumps, A and B. The interval between system start and recovery from the 1,000th failure varies for each iteration. Results for one iteration were: Simulation horizon = 192,915 hours Total system failures = 36 Failures before mission completion = 29 Reliability = 0.194444 MTTF = 5,359 hours MTTR = 5 hours Availability = 99.92% For the 75:25 (or 60:40) situation, does this mean pump A operates for 75% of the time and pump B for 25% (9 months and 3 months in a year, respectively)? The possibility of simulating these scenarios depends on your response. Any combination is possible, the challenge being the required coding time. Best regards, Rui.
Thank you, Rui. I am unsure why you are assuming 9 months of continuous running for the 75:25 ratio. What assumptions did you make for the 90:10 or 50:50 situations? It is important to consider multiple iterations to gain more meaningful results. For instance, one iteration may show a MTTR of 6 hours while another shows 5 hours. After performing 100 or 1000 simulations, patterns may start to emerge. To ensure accurate comparisons among scenarios, I recommend setting the time for activities A and B per cycle to 8 weeks across all scenarios. This will establish a consistent baseline for analysis.
In a discussion about reliability simulations, Vee and Rui are seeking clarity on different duty cycle scenarios like 75:25 and 90:10. Rui seeks clarification on what the regimen entails to ensure accuracy in their simulator. They discuss the importance of running multiple simulations to get reliable results. Rui acknowledges a mistake in using the term MTTR incorrectly and corrects it to MDT. After running 50 simulations, they find the MDT to be around 4.45 hours in the 50:50 scenario and 6.02 hours in the 90:10 scenario. Rui suggests standardizing the cycle time for better comparisons and expresses readiness to address further inquiries on duty & standby scenarios.
I appreciate your efforts, Rui, and apologize if I have caused any confusion. To clarify my request for simulation on the model you've prepared, I would like you to analyze various scenarios: 50:50, 60:40, 75:25, 90:10, and a proper duty/standby setup (with the standby operating for 8 hours each time) with an 8-week cycle. Assuming an MTBF of 20000 hours for 'stopped while running' and Mean Time to Restore (MDT) of 6 hours +/- 2 hours, with a start reliability of 0.98 for each pump and a 95% confidence level over a two-year period, I seek answers to the following questions: 1. What is the system availability? 2. How many failures occur with the running pump? 3. How many failures to start are observed with the standby pump? Additionally, if the MTBF is changed to 10000 hours while keeping the MDT constant, what impact does it have on the results? Moreover, if the start reliability is adjusted to 97% with the MTBF remaining at 20000 hours, what are the implications? I hope this specification is clear, as I believe this analysis holds significant value. Thank you for your assistance.
Hello Vee, thank you for your time. Just to clarify, when you mention a 75:25 ratio, are you indicating that pump A operates as the primary pump for 75% of the 8-week period (equivalent to around 1005 hours), with pump B serving as the backup? And then for the remaining time (approximately 335 hours) within the 8-week timeframe, pump B becomes the primary pump while pump A switches to standby mode? Kindly confirm or provide any necessary corrections. Thank you for your assistance. Rui
Indeed, Rui, agreed.
In my opinion, the second choice stands out as the top pick.
For those of you following the results of the duty/standby simulation, please bear with me for a couple more days. The coding and validation process has been challenging, and I need a bit more time to ensure accuracy. Thank you for your understanding. - Rui
After putting in a lot of effort, I have managed to obtain some results that I would like to share with you. The data suggested by Vee for simulation tests showed a highly favorable situation with virtually no failures over a 100,000-hour operation period. I intentionally selected data that could potentially lead to frequent failures to ensure the simulation model returned meaningful results for comparing different performance scenarios. While some may find the data unrealistic, it serves its purpose well for scenario comparison. Assumptions: 1. The system consists of one duty pump and one standby pump, with the standby pump replacing the duty pump only when necessary. 2. Failure modes, whether random or due to wear and tear, follow an exponential distribution. 3. The duty pump is more likely to experience failures than the standby pump, with failure rates of 0.05 failures/hour for the duty pump and 0.01 failures/hour for the standby pump. 4. There is a static probability of standby pump failure upon request, with probabilities of 0.1 and 0.05 for different MTTF values. 5. The standby pump only operates for the duration needed for the duty pump to be repaired. 6. The simulator will later incorporate Vee's suggestion of the standby pump operating for 8 hours each time. 7. A cycle (x:y) involves alternating duty and standby pump roles over a 12-week period. 8. Pump restoration follows a normal distribution with an average of 6 hours and standard deviation of 2 hours. 9. The assumed mission duration is 200 hours. 10. No preventive maintenance is considered to maintain simplicity for scenario comparisons. Based on test results, scenarios with differing duty and standby pump reliability showed minor variations, suggesting that regimen variations have minimal impact on reliability and availability. Further details and simulation results can be found in the attached files. If you have any questions or recommendations for data, please feel free to reach out.
We greatly appreciate Rui for the outstanding work done.
Hello Rui, I apologize for my absence last week and am currently catching up on my emails. I have not had a chance to properly review the detailed note you sent me, but I will make sure to do so soon. Thank you for your substantial input in the meantime.
Engaging discussion! I've caught up on the conversation and need to do some reading to fully grasp the calculations, simulations, and theory being discussed. I wholeheartedly agree that the efficiency depends heavily on factors such as the pump, moving medium, and operating conditions. I have a practical example from a steel mill's Billet Caster spray water system. The system consists of three spray water pumps (2 operating, 1 standby) with 200 hp split case horizontal motors and an average output of 3200 gpm at 160 psi. Additionally, there are three filter water pumps (2 operating, 1 standby) with hazelton submersible pumps, 100 hp, and an output of 3400 gpm at 30 to 40 psi. The maintenance strategy in place may not be optimal, but it involves monitoring run time hours for both types of pumps. The spray water pump strategy includes one standby pump that takes over when a primary pump fails, while the filter water pump strategy involves pulling pumps for rebuild when approaching 9000 operating hours to prevent costly failures. The current approach results in approximately 3 rebuilds per year, costing $30,000 for each rebuild by an outside shop. To improve efficiency and reduce maintenance costs, a suggestion is to install two larger vertical pumps controlled by a VFD to eliminate the need for a bypass valve. This modification could lead to energy savings and decreased maintenance expenses. Any feedback on this proposed solution would be greatly appreciated, as well as recommendations for further reading on the topic. Thanks!
Apologies for the error in point 7 of my previous post. I have corrected it below, and here it is again: 7. A 12-week cycle (x:y) designates pump A as the primary pump and pump B as the backup pump for x% of the time and pump B as the primary pump and pump A as the backup pump for y% of the time. Kind regards, Rui.
By implementing a duty standby policy, you can reduce the number of failure modes to address and increase system availability compared to other policies. When operating under a 90:10 ratio, it is similar to a duty-standby setup. It is recommended to have a company conduct vibration analysis on both pumps quarterly. How should we schedule pump operation to ensure both pumps are surveyed during this process?
Hello Brighton! When it comes to ensuring the reliability of your standby pumps, there is a specific order of operations to follow. First, you need to determine the failure rate of your standby pumps when they fail to start. Next, you must establish the required system availability, which is influenced by the level of risks you are willing to tolerate. These factors will ultimately dictate how often you should conduct test starts. In some cases, it may be challenging to obtain precise answers to these questions. In such instances, default values and a system availability target, such as 97%, can be used as a starting point. For example, if you are comfortable with a 95% availability, quarterly test starts may be sufficient. Once you have decided on a testing frequency, you can coordinate with your team to ensure that all necessary personnel are available during the testing period. It is important to take readings on the standby pump before switching it on, and to monitor its performance once it is operational. By staggering the test starts to accommodate operator schedules, you can ensure that both the vibe technicians and operators work together seamlessly. Adjustments to the availability of vibe technicians may be necessary depending on the chosen testing frequency. Remember, the key is to align the cart (availability of vibe technicians) with the horse (testing frequency) to effectively manage the reliability of your standby pumps. This approach can also be applied to other policies, such as the 90:10 rule, within your operation.
Hello Vee, I lack expertise in plant maintenance but I am eager to develop a solid maintenance plan for the plant I work at, with your guidance. I plan to have the vibration technician monitor the duty pumps and have the technicians operate the standby pumps overnight so the vibration technician can check them the next day. My goal is to switch back to the duty pump after running the standby pump for a day. Can you confirm if this aligns with your recommendation? Additionally, I would like to know your maintenance strategy for critical equipment such as compressors and compressor after coolers. Your advice would be greatly appreciated.
We are coordinating to have both the thermography specialist and vibration analysis expert on site together. They previously utilized the thermography specialist for electrical board readings. Could you please provide information on the mechanical services he offers to enhance equipment performance? Thank you for your assistance.
Apologies for all the questions, but we are gearing up for a shutdown followed by three years of continued plant operations. Will this impact the duty-standby approach to operating pumps? Thank you.
If you're looking for more detailed feedback on your topic, it may be best to begin a new discussion thread. Rest assured, we will definitely provide a response. However, this current thread has become quite lengthy, with a total of 77 comments (it's our megapost!). For newcomers, starting from the beginning and reading through all the posts can be quite the endeavor, akin to reading a novel. Thank you for understanding and happy posting!
Seigga, it may be beneficial for you to create a new post with your question in order to receive updated and diverse feedback. This can help you gain fresh insights and perspectives on your query.
What are the acceptable vibration levels for equipment when not in use? A forum user, svanels, shared their experience with testing a standby pump every three years. Unfortunately, running the pump could lead to premature wear and damage due to excessive vibration. This highlights the importance of setting appropriate vibration limits for equipment during standby periods to prevent issues like false brinelling on bearings.
✅ Work Order Management
✅ Asset Tracking
✅ Preventive Maintenance
✅ Inspection Report
We have received your information. We will share Schedule Demo details on your Mail Id.
Solve Your Operational Challenges with Oxmaint → Get Free Demo
Answer: - Preventive maintenance involves scheduled inspections and tasks to prevent equipment failure. - Predictive maintenance uses data and analytics to predict when maintenance is needed based on the actual condition of the equipment. - Reactive maintenance is performed after a breakdown occurs, often resulting in unplanned downtime.
Answer: - The effectiveness of the maintenance strategy can vary based on factors such as equipment criticality, cost considerations, and resource availability. - Some may argue that applying predictive maintenance to the duty pump and reactive maintenance to the standby pump could be a balanced approach.
Answer: - Standardizing the maintenance strategy for both pumps can simplify maintenance planning and resource allocation. - However, individual pump requirements and operational conditions should also be considered when deciding on a maintenance approach.
Answer: - Regular oil changes can fall under preventive maintenance (PM) as part of a routine maintenance schedule to ensure optimal pump performance. - In some cases, oil analysis and condition monitoring may be used to implement predictive maintenance practices for oil changes.
Join hundreds of satisfied customers who have transformed their maintenance processes.
Sign up today and start optimizing your workflow.