Week 10 - Lecture 47 : State-of-the-Art and Requirements of PHM

So friends uh with a background in the previous lecture uh now we go to the next stage um we discuss the state-of-the-art in PHM and requirement. ments of PHM. U this is something we are trying to understand. Um what is the road ahead of us if you want to implement uh PHM at systems level. Uh so so we should know which are the area where it has been implemented and then first of all uh for understanding any uh aspect of any system we have to have a reference system. So we will uh we will work on a uh reference system uh which is uh which is a candidate for consideration of uh PHM. Okay. Um so let us see uh what will be part of the introduction of this particular uh lecture that is 10 oblique 2. Um selection of reference domain um for uh for for application area. um we should select a domain which is a maintaining the highest level of safety and record and only one name comes to us that is nuclear systems. Um as I mentioned more than 430 plants and this number is going to grow grow exponentially in in in u this decade and next decade. Why? because there is a idea of small modular reactor. Okay. So it is better uh and probably for for PHM for uh small modular reactor is going to going to play a very very critical role. Uh because uh small mod modular reactor means uh they are systems where the design is modular like you assemble and then they have a power ranging from almost like 5 megawatt to third 300 megawatt and virtually you can say they are not in public domain but then there will be number can be there for like any business or you know any uh industry can have there uh this kind of reactor with all due safety norms uh uh you know addressing uh there in their backyard to meet their like railways can have um their own small modular reactors, steel industry can have their own small. So now nuclear is entering at least in India it is entering into the public domain and uh that is where the number of nuclear reactors should go and uh as a researcher and as a uh scientist in the area of reliability and risk I see a huge role uh the the PHM is going to play uh there because uh they will have a different modus apprenty uh uh there are five SMRs located in a district and they are buying being maintained by a central control room by a team of pooling engineers. So that uh you know first of all they will be highly reliable uh so like um need of personal attention will be less compared to the present systems and then if if at all anything situation comes uh um there can be a team visit the plant and come back. Okay. uh for simple example these type of uh systems will have a fueling requirement which might range from 2 years to 10 years to 15 years. So that itself is there are system where maximum 1 year 1.5 years and then for some plant it is a continuous fueling is required. So we are it is a it is a a different paradigm and when it is a different par paradigm it will require a different approach and improved approach and probably by the time if PHM is matured it is still matured like I I'll show you in this uh in this uh lecture only how it is being used uh in safety critical and complex systems. So it is already crossed that initial threshold. Then overview of some monitoring and prediction and what is the requirement of or what are the requirements of PHM when we talk at the system level. Now we are okay uh on component level the technology is matured at 10% of science and then efforts and resources are required and PHM can be implemented for a pump for Bing monitoring for lithium ion ion batteries uh for a gear gear health monitoring. These are the component they were creating problem. If you look into the history of uh operation of complex engineering system they were and they have been attacked uh like capacitor uh is a hazard also at the same time it's a failure whether it is in micro electronics or power electronics it affects adversely uh the safety and uh reliability also. um then SSC's categorization how it should be done because you know uh it cannot be as I mentioned it cannot be me PHM cannot be implemented uh for all the component but definitely and effectively it can be implemented for some 10% or 20% of the component and that is where we draw the maximum benefit remaining all trickles down so it is not required also even routine maintenance procedure also can suffice there actually so I would say why we have chosen uh nuclear plant First of all, they are the leaders in maintaining the safety of uh the plant. Okay. Uh and of course there is it is in print that the nuclear uh plants uh they maintain the highest safety records. So um the domain which understand safety or risk reduction will only embrace it like for the case especially when we talk about the SMR and all that because SM PHM brings new possibilities uh for SMR small modular reactor because it's a uh the operation maintenance paradigm will be much different than what we have now where dedicated staff is put for operation and maintenance and you know um even in the odd hours the the help is available maybe in 1 hour or so or so when we can have this kind of thing for a normal system existing system why not we have a PHM where there is a uh centralized control room and there are 10 SMRs being managed each of 50 mega megawatt or 20 megawatt or whatever or even 5 megawatt you know so uh PHM is a resource intensive option this is one thing which is holding us that is point number one is that it is a resource inensive operation operation and second thing is reliability of PHM itself should be demonstrated. So uh so at present it is being installed in many system and probably I think very uh risk critical systems it is being used like aviation and uh you know uh space and all that. So let us see uh domain specific studies followed by identification and prioritization and since P ph PHM is resource intensive it cannot be applied anywhere and everywhere it has to be 20% of the component which are falling on the top uh from the point of view of risk or reliability level. Now background nuclear plants are categorized as complex engineering system. Yes we know we have discussed a lot on this. Traditionally if events and failures are uh are the unreliable situation uh then to a great extent there is a provision in nuclear plant to provide advanced burning of the instrument. The point I'm trying to tell you is that nuclear plants uh have a system where if if risk is the one of the metrics then before the risky situation comes nuclear plants by design itself they provide redundancy diversity and then in monitoring also certain signals which come in advance uh and then they'll let you know about some transients and all that. So those provisions are there though they are not in the name of PHM and like I gave you one example in in olden days 60s or 50s the pipelines health is used to be monitored by having coupons are there they are underground pipelines and there are coupons over the years you remove the coupon and see the line how how is the condition of the pipeline how much thickness reduction has happened. Okay. So these concepts were there but now we are having a very uh very uh you can say resource inensive environment in terms of uh AI and ML. It can provide us a better opportunity for indication right from starting right in the control room of the plant uh uh and you know online signature and uh one can imagine a scenario like there is a there is a panel in the control room which is giving rul of all the components. uh it could be like a like a video vis video display unit and wherever it is re reaching um almost like u um very close to the I I will not use any any quantitative indicator here very close to the failure it will give alarm so that you have time for management and for since it is been implemented in a plant the plant people know when it is required to give a signal so that a management action can be uh can be implemented that can be done only at the plant level because plant design and operation document they define these kind of criteria. So those scripts and alarms will flow in and they will tell you preventive maintenance is one of the way which is a proactive way of maintaining it. Condition based maintenance is almost very close to the PHM and it has become one of the popular technique any leakage or any anything there it will be the humidity uh a sensor humidity uh humidity sensor will tell us that this is the situation is going to happen. Um or you know for a bearing uh if the temperature goes beyond certain limits it may not give the remaining useful time but it definitely wants you you take action. Actually in service inspection and maintenance again it is a proactive approach where periodically it is done above provisions predict prevent and proactive features are having it. So but of course they are not in terms of a modern approach where it sitting in a control room you can tell and in fact they are all very resource consu consuming activities. So PHM will solve them and it will bring closer to uh and the accuracy level or effectiveness level will be much higher compared to what we have today. But these capabilities are there like for example the in the design principle defends in depth in uh in in design strategy. This itself ensures a lot of protection of the system where redundancy diversity fails criteria fault tolerant systems. Fault tolerant miss if any fault is there it there should be having it could be redundancy or it could be some extra provisions for replacement or repair so that the fault itself does not cause any problem single failure criteria no single failure of the component should result into compromise on risk and uh reliability. So these are the features and then defense in depth is implemented like this probably you'll appreciate what kind of safety that uh that's why it is a one of the safest industry like you have a fuel fuel the the first barrier is the fuel sheet itself second barrier will be the boundary and then third barrier and fourth barrier you can see here uh the third barrier is uh this thing pipeline and a design basis and like that The so so these are we are talking about barriers uh first barrier, second barrier, third barrier, fourth barrier and all then level of protections are also there. So level level of protection means action or provisions that are there actually. Okay. So uh so these all so up to here this is a normal operation then uh safety actions comes into into the play and then general correct criteri. So this whole approach is called you can open any um IA book and you can find diagram and study it. But it is a robust approach and this has this has got the nuclear industry going into developing a system which are robust, reliable and giving the service. uh so um so that is what we talk about the defense in depth the major criteria multiple barriers and multiple level of protection that that's what I have discussed and then when it comes to the operational uh specifications the technical specific safety limit limiting condition for operation because all these things should translate into when plant goes into the operation. So these are and there are some document it's called technical specification uh that is a sort of a u document uh which tells uh the condition uh limits for under which the plant should operate and it should not be violated. If any of these things get violated it becomes a uh case for regulator regulatory intervention. Okay. And then maintenance inservice inspection maintenance and all. So with this robust provisions and maybe many more it is not uh it is not possible to cover in one one or two slides there are provisions uh but at the outset the these are demonstrating as a best but only a limitation was these are all very conservative criteria. uh in 50s or 60s when the plants were developed uh the material property the designs the computational uh uh features thermal hydraulic features and you know power of computers was not there actually but now the time has come that we come out of the conservative domain and uh try to have like risk informed approach inter at international level it is and in risk informed approach it is the probabistic probabistic risk assessment is playing a key role and in fact now we have come to come to a uh stage where we can call riskbased approach uh because um in uh in traditional deterministic approach factor of safety was one criteria and uh that was uh uh like you know some experience uh experts uh they worked out 2.5 5.5 it was based on the experience and all that but now the materials are uh properties are known well so we can say always the uncertainty associated with the strength uncertainty associated with the stress and the overlapping thing will indicate the probability of failure. So when we know the material some way how it fails and all that fractures and all those things. So we can we can because you know once we have overconervative system becomes uh overburdened actually you know. So those kind of ages can be can be removed and um probabistic risk assessment and deterministic safety assessment they can overlap complement supplement each other and better risk information disbase approach uh can be uh can be had in place actually you know so that is what the objective is and PHM is one of the things which is intervention as part of proistic riskbased approach actually there is a growing trend in the exploring the application of prognostics And uh already a lot of research is going on actually you know and people are interested because it has got benefits. Okay. Um before we go ahead uh even though in my other slides it is there but now say uh risk means safety risk and uh and the security risk. So security risk is a relatively new paradigm and that has to be addressed. Even if I ensure safety I need to address the security provisions also. So this is a new uh technology or area or domain being developed unless until you start both safety and security risk and there are many more I we'll be discussing about them then only it becomes a holistic safety actually okay so already condition based maintenance is there and we we have talked about it again the technology has moved uh you know risk informed from hardcore conservative deterministic approach to risk informed approach especially in regulatory domain. Okay. And then p plant reliability that is un uninterrupted operation higher availability targets can be realized through PHM. It's again one additional promise that is coming into picture. Uh but all said and done we are 40 30 nuclear plants and strategy. In fact for all these 40 30 plus plants uh risk assessment has been done that is probabistic risk assessment studies. So you you can see how slowly the whole thing is turning and now probability is uh is uh seen as a integrated part part of the safety uh risk and reliability. Um uh it is not a uh you know exhaustive research but the salient feature I have picked up like in aerospace uh uh this PHM uh is playing a role vital role and this is a paper they are probably uh if you refer this book u my book and pet's book uh role of prognostics in support of integrated riskbased engineering uh and this paper is available in open domain Okay, it is a uh society of uh uh prognostics and health management that they have published. I think it was 2012. So from there I have picked up and uh then aircraft engine damage monitoring. This is one of the active areas being pursued uh in uh PHM. Okay. Um then uh we have electronic systems. If I have to tell the electronic systems are the uh you can say leaders in implementation of PHM. So there are many papers and you'll see many component like um right from IGBT to capacitor to uh to metallization uh so many things uh and physics of failure models have been developed. So I think electronics has taken a lead uh in this implementation of uh PHM and then we have a health monitoring of lithium ion battery. probably uh in the last decade you'd have seen uh that the batteries are exploding uh in locations where they had a uh you know fatalities or undesired effects and that's why this field lithium ion battery has taken over and probably you'll find now that this particular thing has reduced still in isolated examples are there uh but but the use is also increasing because now lithium ion batteries are becoming part of uh automobile sector you know so electric vehicle we call so there the number has increased and that's why it is seen more also but to a great extent problem of uh you know health effect has been reduced to a great extent and probably that insight come from PHM uh to a great extent you know and prognostics for B monitoring and gears you'll find almost like every third day one paper is being published so because uh uh pumps, generators uh where rotating devices are there, power electronics um also is of course I have talked about electronics but there are a lot of work is going on in power electronics also and uh then engine turbine conditions and all you know turbines are the biggest churner of power actually and if anything happens in the turbo generator it's a problem you know uh it sends the disturbances in the whole grid So turbine condition monitoring and engine condition monitoring have become a area for research and implementation actually. And then telecommunication and structural systems um like dams and you know many structures they have sensors located on them uh for any u any degradation happening or any loss of strength or stress and this this one you know. So these are the some few things which I have picked up and quoted in my paper. U but there are many now because this paper was written in 201213. So there are many areas. uh it is up to you to open the uh internet and see how PHM is becoming uh uh PHM or it's a similar technology they are becoming popular in identifying or you know the degradation uh not only degradation for but for uh estimation of the remaining useful life also. Now uh um let us if if I have to go at implementation at system level uh what are the requirements the m metrics is there. So if I today if I take a decision or a government or a industry body takes a decision for implementing PHM to what are the metrics they should have with them to go ahead with that. So the first this this will be the first job a report will be prepared where we are what is the stage of the plant and how we can implement it what is the state of art which enables which are the component can be brought in the uh for PHM and which are the component which will be left out which can be managed by condition monitoring or which are the component which will follow the traditional maintenance management approach that can be decided. So first is planned stage and phase. If it is a design stage, there should be a complete mapping on uh the power of PHM at the same time limitation of a PHM so that it is easier to provide PHM provision like for a nuclear plant for a reactor vessel if I want to monitor its health for let's say 100 years. So I should be having installing a sensor there. But that sensor should be replaceable also. We sure we know that they will have a limited life 10, 15, 10 whatever or two. So uh a lot of thing goes into consideration when you provide a sensor. But you once you make a provision there is no problem actually you know. So uh for passive component we should have for active component which are the 10 20 components for which we would like to monitor the health of the bearing. It could be circ recirculation pump uh it could be turbo generator. Uh it could be shutdown cooling pumps. Uh it could be uh fire pumps uh which should fall because from safety point of view they have high and then most important thing which are the barriers that we have to monitor from the point of view of security which are the uh which are the fence uh fence wall that we have to monitor and ensure integrity on those walls uh from the point of view. So we have gone from safety from security both uh this is depends on plant stage. Now suppose if the plant has completed let's say 40 years and we are going for refurbishment then we should explore where the sensor should be provided that we can easily go from 40 years to 60 years okay uh the job will be less compared to the new plant but even if we inst install PHM at least 10 places uh we have a huge assurance uh in terms of reliability and safety actually now what is the objective what we want to do uh whether it was a compromise on safety all across or uh reliability or what suppose if the safety component was less which happens in many industries. So it should be reliability throughput we have to increase. Okay. And that should be the objective of course without compromising the safety. Okay. So uh objective and then scope also comes into the picture and then whatever domain we are talking about what whatever is the state-of-the-art that is available we can tap the market available technologies at the same time if you put for 20% for resource and all that uh then we can uh we can think about implementing prognostics and health management and again this is SS category that we want to we have to create a different category wherein like a ready rocker it should be available and a list I should get from from top top 10% to uh 20% which are the component list we should have and that scope and objective also should be clear and then implementation level what what level we should implement it actually you know so um and you know uh it is some there are many decisions have to be taken in the plant uh which will decide the scope and objective of PHM and that will come during implementation level only. What is available access to that? What is not available? We have to work out a way and uh wherever it is not possible can we have some other other approach something like that and a risk assessment approach before implementation of PHM after implementation of PHM so that we can realize the net benefit in terms of safety and here also in terms of the uh you know uh um in terms of the u plant deliverables what advantage we'll have are we reducing number of uh trips on the plant. Uh are we reducing that is increasing the reliability? Are we reducing the shutdown periods of the plant that is plant availability we are increasing? So throughput will increase and then existing maintenance approaches. I mean um these things together they should reduce the maintenance of the routine burden and that uh reduction whatever uh gain we have made whether it is called return on investment for PHM whether we can uh we can uh use that uh money for implementing PHM and to get the so it is a vicious circle positive circle wherein gain will go on increasing without compromising plant safety and reliability. implementation approach probably we have uh we have talked about it this was a level and it was a what kind of approach we want to use it and all uh this is a topic uh we can go on talking uh and it is domain specific actually you know um it should be in the face manner it should be proof of concept first and then uh propagate it to some selected component or just in one go Because the similar technology is operating in uh other plants and then what kind of provision you are going to make it so that it is uh involving the control room staff uh to take in fact management component also should be part of implementment approach and that will require a special effort because we should know plant documents and you know so and this thing so this is a complete uh matrix for implementing the uh PHM or PHM design I gave you one example that which categories you should select for monitoring and all that. So first category you can say it is a pive component um you know which is like a which is like uh it is a core of the plant and they it is replacing them is difficult task. So the and in other words lifelimiting component. So what kind of the uh status that is available uh the M stand for monitoring offline ISI offline condition based maintenance diagnostic prognostics online diagnostic online prognostics so what are what is the matrix what is the state of AR that is available for uh reactor structure nonstructure mechanical components pump electrical power pump then power electronics uh uninterrupted power supply and all that micro electron ICS digital card IC PLC programmable logic controller field programmable gets interconnects and controls cables and then process. So this table probably you have to go through and understand what is the state-of-the-art as on today and remarks I have given from IAA or references that I have with me and you will get the complete comprehension of what is the state-of-the-art available and if you want up to generation three if you go for generation four nuclear power plants what advancement you you have to do uh in terms of implementation and categorization of SSC Um okay so and then we have um prognostic design requirement. So design requirement will differ uh the plant stages. So plant stage could be under design by uh operating plant uh useful life period and then uh aged per plant uh shutdown uh refurbishment period and operating uh operation after refurbishment. So uh objective here will be uh drain stage improve safety. Okay. Using PHM operating plant improve safety and availability both. Okay. And then uh aged plant um monitoring remaining useful life. Okay. And shutdown under refurbishment followup later on. What happens if we do refurbishment? How what are whether we are getting the benefits on that? So refurbishment, extended maintenance, uh where is the plant is uh plant is renewed. Okay, not all components but major component which are aging they are replaced and then uh the followup and retrofitting is required uh operating after referment uh reduction in uh operational cost. Okay. So after refurbment mean this is a uh not a new technology. uh what happens the if plant has operated for say design life was uh that time when it was built plant was built it was told it is 40 year life now you have come to 40 years life but you see that with little extended maintenance we can extend the life of the plant to another 20 years okay and that is where the this uh plant level and its requirement uh uh prognostic requirement reduction in operational cost it can be done okay so for state-of-the-art enablers online monitoring Then online monitoring and PHM this these are the comments over here subject micro electronics you know here if we talk about the under design and digital control uh channels power electronics electrical so like that if you see all these metrics parameters if you see and different stages of the plant you'll understand the requirement how where PHM can fit into the design uh design of PHM can be done accordingly its scope objective it can be uh it can be decided. So in this lecture we saw uh nuclear we have selected a domain because uh safety demonstration is happening all there uh and then deterministic approach it it is a blessing but then now with the state-of-the-art and modern technology that are available there is a time to have a complimentary risk based approach or risk informed approach where it uses deterministic as well as probabilistic both uh and uh the complement of supplement each other to uh to realize um higher level of safety and availability. Advanced senses systems approaches we have discussed role of PHM environment scope uh major pronostic design requirement the metrics we have discussed and categorization of SS that also we have discussed uh categorization of SS we'll discuss in the uh fourth lecture when we talk about the uh in detail methodical model we'll provide uh when we will talk about the importance measure risk importance measure uh fuzzle basically risk reduction worth risk achievement worth that time this discussion will be complete These are some references over here. Um uh this is my book on uh not book this is a one paper in uh journal uh international journal of prognostics and health management. Uh I think the year is 2012 that I had published after that lot of water has flown and I have been adding up everywhere when I went and this uh PET and MGK prognostics and in fact I'll tell you at the outset there are not many uh books in available in open domain. So whatever available in open domain uh you can buy and you can um you can learn from there. Uh what we are we discuss under prognostics and health management. This is going to be the new technology um for maintenance management or support of maintenance management to reduce the resources uh that we uh invest in maintenance management. Thank you.