After reviewing the queries and comments made by Manou, I will address them accordingly. In Manou's text, I have included reference numbers, such as "[1]", to direct to specific responses.
MANOU'S COMMENT: 1- When making condition-based maintenance decisions on equipment, such as monitoring vibration levels, EXAKT examines a reasonable number of similar equipment with a good history of replacement and vibration data. Through the analysis of age replacement data and vibration readings, EXAKT estimates the Weibull parameters for age and covariate parameters for various vibration readings. Subsequently, EXAKT creates a statistical decision graph to guide future decision-making regarding the equipment's maintenance. The Hazard function is represented as h=h0 e**Z, suggesting that the equipment's age is reduced by multiplying the base age hazard h0 by Z. An advantage of CBM strategies over Age replacement is the focus on extending equipment lifespan and avoiding premature replacements.
DAMING'S RESPONSE:
1. The term "guess" is not commonly used in statistics for estimating parameters. Maximum Likelihood Estimation, utilized in EXAKT, is well-documented in statistics literature.
2. "Age" and "various vibration readings" are variables, not parameters, as explained in textbooks on PHM.
3. The age of equipment is not accelerated; the hazard rate depends on age and Z (vibration reading).
MANOU'S COMMENT: 2- Could you explain the Decision Graph, which represents the boundary between green and red areas and helps determine when to replace equipment based on covariants like velocity and acceleration?
DAMING'S RESPONSE: The Decision Graph in EXAKT provides recommendations for replacement policies without dictating specific actions.
MANOU'S COMMENT: 3- What is the decision criteria in EXAKT?
DAMING'S RESPONSE: The decision criterion in EXAKT is primarily based on expected costs associated with risks.
MANOU'S COMMENT: 4. The cost of long-term equipment replacement is determined by an optimal hazard level, influencing decisions based on cost rather than perceived risks.
DAMING'S RESPONSE: The optimal decision in EXAKT considers both cost and risk, aiming to minimize expected costs associated with different maintenance strategies.
MANOU'S COMMENT: 5. Let's examine three points from the Decision Graph to determine the optimal hazard levels for cost-effective replacements at different ages and Z values.
DAMING'S RESPONSE: Considering the optimal hazard levels at various points in the Decision Graph requires accurate calculations and an understanding of the underlying modeling principles in EXAKT.
MANOU'S COMMENT: Expressing concerns about the accuracy of the calculations and decisions made in EXAKT, particularly in relation to maintenance cost and replacement strategies.
DAMING'S RESPONSE: Addressing misunderstandings and providing clarification on the methodology and rationale behind EXAKT's decision-making process.
In conclusion, it is essential to thoroughly understand the principles and calculations involved in maintenance decision-making tools like EXAKT to ensure effective and reliable outcomes.
Are you looking for the best substance to use right now? This forum remains open for discussion. Want to know if the EXAKT program is good to go? Your input is appreciated.
I recently reviewed the presentation attached below and found it highly informative for anyone interested in physical asset management. The experts discuss the integration of RCM, CMMS, and CBM, which is crucial for optimizing asset performance and reliability. Have you had a chance to read it? What are your thoughts - do you find it too technical or incredibly valuable? Personally, I believe there is a wealth of insights and valuable information included. Are there any technical experts or organizations who have evaluated this work? Check out the presentation here: http://www.omdec.com/articles/p_ABBCBM21June2005.html.
Josh, don't be too quick to judge maintenance workers. I have found that many of them possess a high level of intelligence based on my experience training them over the years. Regards,
- 01-12-2024
- Jessica Freeman
I agree with Josh that we should avoid making broad generalizations, but I also agree with the essence of his point. Within any organization, there will be individuals who gravitate towards a more abstract problem-solving approach and others who lean towards a more concrete one. Neither approach is inherently superior to the other. Sometimes those with a more abstract mindset may take longer to make decisions or overlook important factors that others would consider.
Personally, I tend to lean towards a more abstract approach. However, after reviewing the material, I find myself wondering if this tool is too abstract for a simple task. Is its main purpose simply to determine if a trend is increasing or decreasing? Why not use straightforward automated alarms based on levels or rates of change to alert analysts, who can then use a more subjective approach to assess the trend and potential risks? It's important to consider that the analyst must not only analyze the data but also rely on their knowledge and expertise in the specific area.
For example, in vibration data analysis, why do we need to transfer data from one program to another just to analyze trends? This results in multiple programs sending us alerts and complicates the process unnecessarily. I fail to see how the added costs, complexity, and effort are justified in this scenario.
- 01-12-2024
- Shawn Thompson
Josh, I must say I have some reservations about the company you have been linking to. The 'P-F' curve, often attributed to John Moubray, actually originated from the Nowlan & Heap concept. The integration of Condition-Based Maintenance (CBM) in a Reliability-Centered Maintenance (RCM) program is standard practice, alongside Preventive Maintenance (PM) and other strategies. The definitions provided in the article seem misleading, appearing to be a haphazard compilation of information from various sources without a thorough understanding. Relying solely on human perception for anticipating equipment failures can lead to inaccuracies, as demonstrated by the disparity between expected and actual failure types. Some recommendations offered also raise questions about the accuracy of the information provided. It is not groundbreaking that the results of RCM are utilized in Computerized Maintenance Management Systems (CMMS) programs. Overall, my concerns align with those expressed by ElectricPete. - Howard
- 01-12-2024
- Heather Coleman
Dear maintenance professionals, I have simplified the criteria to just "fantastic or too technical" to better cater to our audience. It's clear that we have a group of highly skilled individuals in maintenance. Nevertheless, continuous improvement is key. I hope the project team can address the feedback we've shared. Thank you.
Josh, it's crucial to remember that in order to sell, vendors need to advertise their products or services. ABB once had a maintenance website known as maintenance-tv.com, but it unfortunately fell victim to the dot-com bust, along with numerous other businesses.
I am in need of expert systems, fuzzy logic, and neural networks for maintenance and reliability tasks. Can you recommend any reputable sources for these cutting-edge technologies? Thank you.
Murray's discussion on Weibull-PHM may seem complex, but it is crucial not to overlook its significance, as seen by the extensive research and validation conducted by prestigious institutions such as The University of Toronto, University of Texas, and University of Maryland. The US Navy is actively implementing proportional hazard models (PHM) to forecast equipment performance and streamline maintenance operations on ships. By incorporating sensors to monitor critical equipment conditions and utilizing reliability block diagram models with various distributions including Weibull-PHM, future Navy vessels will have a system in place to predict equipment health throughout missions. This initiative, set to span several years, aims to have a fully operational equipment health monitoring system in place on the first DD(X) ship in a decade. The proven tools and methodologies employed in this project can also benefit industrial facilities seeking to enhance their maintenance practices.
Isn't this debate just a bunch of academic jargon that even self-proclaimed 'experts' can't seem to agree on (often seen as competing vendors trying to justify their own existence)? While P-F curves are useful for teaching the basics of failure and reliability, applying this theory to real-world situations (such as in the navy) may seem questionable. Let me elaborate on my skepticism: I am in charge of maintaining gas compression trains that consist of large aeroderivative gas turbines, agricultural power turbines, and centrifugal compressors - all of which contribute significantly to our revenue. Each component requires a tailored maintenance approach. While some opt to follow manufacturers' recommendations for overhauling aeroderivatives (resulting in more spare parts sales for the OEM), we prefer a combination of performance monitoring and borescope inspections to determine when an overhaul is necessary, often extending intervals beyond OEM suggestions without increasing costs. So, sales experts, how would your software benefit our unique situation? In our experience, one out of six power turbines (up to 25 years old) encountered an uncommon exhaust diffuser failure - how would your software and weibull analyses address this issue with an 18-month lead time and a high cost of $1.8 million? As for the centrifugal compressors, which are mechanically straightforward, we rely on simple vibration and performance monitoring to guide maintenance decisions. In the rare event of catastrophic failure (usually due to a design flaw causing impeller resonance fatigue), our proactive approach of changing bearings and dry gas seals as needed has proven effective. Ultimately, what additional value does advanced software/weibull analysis/P-F curves offer that practical common sense cannot provide in our specific case?
- 01-12-2024
- Victor Thompson
Dear Buzz, I appreciate your skeptical perspective, as healthy criticism is important. However, resorting to personal attacks and insults is unnecessary. In my professional opinion, I believe that The New Weibul Handbook by Robert Abernathy could provide valuable insights on this subject. Abernathy, a leading expert in the field who has designed numerous jet engines, may have influenced the systems you mentioned. Having worked in various real-world scenarios and with multiple navys, I have observed that the challenges faced are often similar, though varying in complexity. I encourage you to explore Abernathy's work for a deeper understanding. Best regards,
Your initial feedback serves as the foundation: "A combination of simple vibration analysis and performance monitoring is used to determine necessary actions." Transitioning from manual data collection to continuous online monitoring eliminates human error and ensures accurate data collection. By implementing software that recognizes failure patterns, potential issues can be identified and addressed before they become critical. Additionally, scheduling maintenance based on performance metrics can optimize production schedules and equipment uptime.
In the past, predictive technologies like vibration analysis were dismissed as "snake oil" by traditionalists. However, it is essential to acknowledge the expertise of professionals in the field and consider the benefits of adopting new maintenance approaches. Embracing advanced technology can lead to more cost-effective and accurate maintenance solutions compared to relying solely on subjective judgment. It is crucial to remain open-minded and explore innovative ways to improve equipment reliability and performance.
Hello, are you the Mr. Machinery expert? Is it true that speaking kindly to your machines helps them run smoothly? Just kidding! Are you interested in this field? How much experience do you have with it? Thank you.
Attached is my presentation on the state of condition monitoring in the Steel Industry, which I presented at a conference in 2005. As a consultant specializing in asset management, I strive to help companies enhance their practices by leveraging data on equipment performance. I am dedicated to assisting firms in optimizing their operations by utilizing collected data effectively. If you are interested in knowing more about the future of condition monitoring, I am available for a discussion. Contact me at sshores@dei-group.com. Steve Shores, The DEI Group.
Thank you, Sshores, for the presentation. I appreciate the high-level overview, but I am seeking more detailed examples of how expert systems, fuzzy logic, and neural networks are applied in maintenance and reliability. I am eager to delve into real-world applications for future references. Thank you!
Switching from traditional manual data collection methods to real-time online monitoring eliminates the need for human involvement and reduces the risk of errors. While proponents of online systems often emphasize this point, it's important to recognize that manual monitoring offers unique benefits. Beyond just collecting data, manual monitoring allows operators to physically inspect equipment, using all their senses to detect abnormalities such as leaks or loose connections.
Potential errors in manual monitoring may arise from placing transducers incorrectly or entering incorrect values. Some may try to create a "fear factor" to promote online systems as superior, but in reality, industry professionals often rely on instrument technicians to investigate and verify any changes detected. It's common practice to check multiple parameters before taking action based on a single reading.
Online monitoring systems can be valuable for predicting equipment failures and scheduling maintenance based on predefined rules. However, it's essential to consider whether the software has the necessary data inputs to make accurate assessments. For example, can the system predict the deterioration of specific components like dry gas seals in a centrifugal compressor train?
In scenarios where equipment degradation is gradual and not immediately impactful on performance, manual inspection and expertise are still crucial. A computer may not easily detect issues like minor erosion on turbine blades, which can have significant cost implications depending on the timing of repairs.
When considering the implementation of new technology for maintenance decisions, it's important to respect the expertise of industry professionals and not solely rely on automated systems. While technology can offer insights and guidance, factors such as market conditions, contractual obligations, and resource availability also play a significant role in decision-making.
Experience in handling complex systems and equipment maintenance should be valued, regardless of whether it comes from the military or commercial sectors. Success stories are important, but it's equally important to acknowledge and learn from instances where technology may have failed to identify issues or provided inaccurate assessments, leading to costly consequences.
While online condition monitoring may not always be the optimal solution, it often proves to be highly effective. Not every organization operates a traditional "plant," and not all have the resources for manual checks by operators. Relying solely on human judgment for condition monitoring in large-scale electricity distribution networks would be inefficient. Many companies now rely on computer indications to remove capital equipment from operation swiftly, especially in industries like mining, power generation, and gas distribution. The automation of engineering judgment is increasingly common, despite some reservations. In my military experience, bureaucratic processes can lead to wasted resources, but the focus remains on managing highly technical equipment in diverse operating scenarios. Dismissing military asset management knowledge outright would be shortsighted, akin to stereotyping all recently privatized companies as outdated and inefficient. Machines would likely advise prioritizing strategic asset management goals, regardless of the source of the advice.
Do you view your machines as tools rather than friends in this community, Buzz? For those who value human expertise, consider the following questions: How do you ensure that all your decisions are sound? If you possess flawless judgment, we'd love to tap into your brilliant mind and digitize that knowledge. And if you've made mistakes, we're eager to learn from them in order to prevent future errors and incorporate them into a software program. Would you be interested in a 5% royalty on software sales? Do you have a succession plan in place for your company post-retirement to ensure its longevity? How many team members do you have, and have you chosen a successor? How much knowledge are you willing to impart to them, and what has been the success rate of this transfer? Is your company overly reliant on one person to the extent that you doubt the benefits of computerization over human expertise? As someone who values efficiency, I believe that incorporating technology will streamline my work. With my expertise in machinery, I am confident that I can master software more effectively than anyone else, including the programmer. Do you believe that human intelligence can outperform computer software? Warren Buffett once said, "Risk comes from not knowing what one is doing." Do you think you can succeed alone without external resources such as universities, industry organizations, OEM vendors, and consultants? I rely on various sources for knowledge and always welcome input from consultants. Did you personally design and build your machinery, or did you utilize a consulting firm before joining the company? What is the longest interval for overhauling your gas turbine? Can a gas turbine operate indefinitely without maintenance, and under what conditions could it run smoothly for up to 10 years? Thank you.
1) How do you ensure the correctness of your decisions? We are interested in learning from your expertise and possibly creating a computer program based on your knowledge. Mistakes are also valuable for learning to avoid repetition. We are willing to offer a royalty of 5% for any software sales. Thank you for the compliment, but I do not claim to have a genius mind. Decision-making is a collaborative effort involving inputs from various departments. Improvement is a constant process, and while computers have their uses, they are not the ultimate solution. In terms of contributions, perhaps sharing information on maintenance boards would be beneficial.
2) Do you have a succession plan in place for your company post-retirement? How many subordinates do you have, and is there an anointed successor? I believe in sharing information openly within the organization to facilitate ease of access. All relevant details are stored on a central server for accessibility.
3) Do you believe that computerizing processes will enhance efficiency compared to human expertise? I am open to utilizing technology to improve my job and am willing to provide feedback for software enhancement.
4) Warren Buffet once said, "Risk comes from not knowing what one is doing." It is essential to seek external inputs from universities, organizations, and consultants to enhance knowledge. While OEMs and consultants play a role in providing information, it is essential to verify and analyze the data. Overhauls are necessary for gas turbines depending on operating conditions and quality of fuel and air. Investigation into machine failures is crucial to determine the root cause. Further information on gas turbines can be found in reputable sources such as the ASME IGTI shop.
- 01-12-2024
- Vanessa Carter
Buzz makes a valid point regarding the influence of computers on our beliefs. In today's digital age, many automatically trust information obtained online without considering the human element. This is evident in the recent trend of people uncritically accepting emails claiming to be from Santa Claus in Nigeria. It is important to remember that the Internet can both inform and deceive, and critical thinking is essential when evaluating online content.
Buzz and Josh, I appreciate your emphasis on the importance of a strong defense and not solely relying on "magic" software solutions. While software can streamline processes and aid decision-making with features like what-if scenarios, it is important to remember that people and their ideas are equally crucial. It's essential not to get too caught up in technology, as it is not a one-size-fits-all solution. Ultimately, thoughtware is more valuable than software. Let's focus on constructive discussions without biases or sarcasm, and be open to learning from each other even when we disagree. Let's ensure everyone feels welcome to participate and contribute their thoughts. V.Narayan.
I am currently delving into the comparison between human expertise and technology. Thank you for the valuable details on aeroderivative gas turbines from ASME IGTI. Are you a member of IGTI as well? We are dealing with aeroderivative GTs that have a low initial cost but high maintenance expenses. We are seeking long-term solutions and would appreciate advice on fuel and air systems, as well as contracting options. Thank you for your insights.
Key Points: 1. Neural Networks and Fuzzy Logic are utilized in reliability and maintenance as self-learning systems. This innovative concept involves learning from both successes and mistakes. I have firsthand experience working on a project with EPRI and TXU involving a similar system for rotating machinery. 2. I have created an 'expert system' for motor circuit analysis (specifically, ALL-TEST Pro's EMCAT) and have contributed to the development of other expert systems. These software applications prioritize providing reliable recommendations rather than being infallible. It is important to use these recommendations as guidelines, especially with broad-based systems. 3. My recommendations often emphasize a combination of continuous monitoring and human data collection for assessing condition. Continuous systems are preferred for rapid fault detection in situations with limited manpower or safety concerns. While electronic monitoring is valuable, I still advocate for human involvement when feasible and cost-effective. Maintaining a skilled workforce is essential, and support may be necessary in the form of training, centralization, instrumentation, or monitoring equipment. The military is already implementing advanced technology, such as a nearly autonomous vessel with minimal crew and automated fault detection capabilities. As warfare evolves, the military is adapting by streamlining processes and reducing waste to effectively manage shrinking budgets. Overall, the optimization and reduction of preventive maintenance are progressing, demonstrating the military's commitment to efficiency and responsiveness in changing environments.