EAM implementation practice. The interview with the head of Smart.EAM
![EAM implementation practice. The interview with the head of Smart.EAM EAM implementation practice. The interview with the head of Smart.EAM](/uploads/thumb/0/media_single/1109-2.jpg)
– Kirill, why did the EAM project start in INTERPIPE?
The main initiator of the project was the director of finance and economics. His task was to achieve transparency in understanding the structure of working capital and optimize its size. Summarizing, I will say that this is a typical problem of the procurement process of any company. Namely, how to optimize applications for material resources to ensure the smooth operation of production. Such as maintaining critical components in stock (for example, bearings), and consumables (like oil). Procurement and finance services are always trying to optimize purchases in terms of finance. Some funny cases describe what the potential for development is. I remember one state enterprise where the procurement department run a model. They deleted everything from the procurement plan, and didn’t buy anything for a month. Surprisingly, the plant worked, as if nothing had happened. This example shows what the "urgent need" of goods means, which can come from manufacturers.
Returning to the project at INTERPIPE, I would say that similar problems arose here, and the company's management wondered how to manage this process. "Here we come to the warehouse, we look, there are bearings. We ask: "Why do we need 2 more bearings, if we already have 3 bearings?" – "Well ..." There is no clear answer. "And for what equipment is this spare part or part intended?", "What is the service life of this machine or part?", "When was the last time it changed?", "What kind of regulation... ", etc. The answers to all such questions were clearly not convincing. That is when they finally got the idea of a management system.
– Did you already have implemented systems at that time?
Yes. The main production assets management system has been working in IT-Enterprise since 2002 and has already been implemented at dozens of enterprises. By this time, the EAM direction became progressive, and we were already finishing the first modern project at the Poltava Mining and Processing Plant. The new solution showed itself perfectly in practical implementation. INTERPIPE NIKO TUBE was the first factory of the company, which began to introduce our EAM. Then there was INTERPIPE Nizhnedneprovsk Pipe Rolling Plant. In Nikopol, they started from one workshop, then they replicated to the whole enterprise.
– And then INTERPIPE STEEL?
Yes. At INTERPIPE STEEL, we saw completely different management options. Here there were already developed APCS and MES. In other words, there was information about how equipment worked before. This information was used for production management. For us, this was an opportunity to manage equipment maintenance processes much better. We were able to contact the Q-MET database (MES from Danieli) and get access to information from 23 thousand sensors. This allowed evaluating more than 1 million events on the equipment for 1 day. For example, the system automatically recorded the operating hours of more than 100 pumps of the energy service.
– What was the state of the regulatory reference information (RRI) before the project started?
If we talk about INTERPIPE STEEL, then, of course, a part of the RRI was implemented at the enterprise. You know that it was a very big project and the whole team that implemented it worked well. However, something was present, something was absent, something was translated, something was not. It was obvious to us that documentation necessary for the subsequent maintenance was secondary to the priorities of the project. After all, the main thing was to get the plant up and running
Therefore, our task for the RRI was precisely to describe the many different types of equipment from suppliers. More precisely, it is more than 6 thousand technical places and almost 1,5 thousand objects of repair. For example, a bucket consists of eleven elements. Some of them need to be checked once every six months, others every month, etc. Something needs to be changed periodically. Thus, the descriptions become rules in the system. The system itself will then remind, prompt, and guide. For example, this week "you need to change the gasket in the site X of the unit Y".
– Did you make these descriptions by yourself or the customer did?
In our opinion, the customer should always make the data and regulations. If this does not happen, the system will not work. The customer will depend too much on the performer and/or will not learn how to work. In addition, customers better understand their own equipment and know it better than the performer.
– Besides integration with data from MES, what else have you implemented?
So that information can be used in business processes of service, it must be prepared. We usually introduce 3 levels of changes:
1. Collection of information. We do not need information from the sensor every second. It is much more important to understand the long-term trend (changes) in the behavior of a particular piece of equipment. And these accumulated volumes of information grow constantly.
2. Preparation of information for storage and analysis. We use a distributed architecture, where information is stored on different servers. Information should be stored in a special way in order to take up less space so that it can be analyzed faster. After all, there are dozens, and in the near future, hundreds of terabytes of data. This information must be very carefully processed. For example, when starting or stopping equipment, they go through large fluctuations of various physical parameters. The trend of these parameters (temperature, current, etc.) can lead to false conclusions and false alarms. These are all pretreatment tasks. They are rather nontrivial given the large amount of equipment throughout the plant as a whole and the various possible states. All such situations, the system should clearly distinguish and not allow false conclusions.
3. Data processing – building trends, using statistics, recommendations and visualization.
![](/uploads/thumb/0/platform_content/1684-.jpg)
Kirill Kostanetsky at the 2017 Kiev International Economic Forum, where IT-Enterprise presented the Smart.Factory and Smart.EAM cases for the first time.
– How are you doing today with preventive and predictive diagnostics?
Let us start with the preventive one. 80% of the problems can be solved by introducing the so-called "yellow zone". If the indicator in it is approaching the border of the "red zone", this means that it’s time to intervene, the system should warn about the upcoming breakdown. This is how preventive diagnostics works.
There is a second part of the tasks, where the boundary between the yellow and red zones floats. These are about 18% of situations. It "floats" for various reasons. It can be different types of dependencies, linear or exponential. If this is exponential one, the problem may occur tomorrow. If linear one – in a week. It depends on how the indicator changed in the past. To understand this, we use statistics. Trends show when equipment enters the red zone. Now this is a forecast using mathematical apparatus and statistics.
There are 2%, where we can not use mathematical apparatus. There is the need for machine learning methods. This is cases where we cannot describe these links. The system itself learns and builds these connections. The human mind cannot grasp this. We cannot talk about the 100% development of these competencies yet, but we are moving quickly in this direction.
– How did you describe these situations with the red and yellow zones in the regulations?
The employees of the customer company set the parameters of the zones. These are technologists, first of all. However, there is also documentation from manufacturers. The parameters of normal operation and anomalies, as a rule, are indicated there.
– How are EAM systems developing in the world?
EAM is a category of software for managing a manufacturing enterprise. All EAM systems work according to close rules. It can be said that they work according to the logic of common sense. The main planning method has always been the schedules of scheduled preventive maintenance. Resources and materials must be adjusted for them. Therefore, at first, all the plans were considered scheduled preventive maintenance, and the systems were called CMMS.
What is new now? First of all, there is an emphasis on diagnostics. Any decisions to stop the equipment and conduct maintenance should be based on specific measurable performance indicators. Therefore, there is a rapid development of diagnostic devices (sounds, vibrations, temperature, electrical indicators and even smells). Diagnostics can be collected by staff or automatically ... The starting point of the maintenance process is an automated information collection system. Previously, many parameters were collected manually. Now everything is automated. This eliminates the notorious "human factor" and many errors associated with this. We ourselves underestimated the accuracy of planning and the conditions under which a lot of information automatically got into the system. Today we see that this radically changes the accuracy of planning, allows tracking micro trends, freeing up a lot of people, etc. This is a fundamental difference. That is, the starting point for these processes in EAM today is information from the process control system.
If we talk more broadly about changes, then with the advent of preventive and predictive diagnostics, the associated functionality changes.
– What indicators or results do you consider the main ones in this project?
First, let us consider the economic indicators. At INTERPIPE, this is a reduction in the stock of spare parts for current and capital repairs and a decrease in the share of low-turning parts in them. We have achieved a 10% reduction, and that is millions of dollars. However, the project will bring the main effect in the future, when all systems will be configured and statistics will be accumulated.
If we talk about trends in the world, then everyone goes to operational statistics. Business executives began to very carefully count the money. It is impossible to remain cost-effective and at the same time allow yourself to use classical scheduled preventive maintenance.
– And how do we perform scheduled preventive maintenance?
In most enterprises, the human factor dominates. That is, uncle Vanya comes in and, using his many years of experience, says: "This will work for another six months, this – 3 months, this already has to be replaced", etc. Interestingly, such an approach is no longer a classic SPM. The classic one is like a car – after 10-15 thousand kilometers, we change the oil and filters. Such classic SPMs exist at few enterprises, while the majority of the others are dominated by the "uncle Vanya effect". However, you understand the shortcomings of this approach. After all, everything now depends on the qualifications, and maybe even purely subjective factors peculiar to a particular "uncle Vanya".
– It is clear with economic indicators. However, the middle manager often asks the question: "How will this change at my work and in my services? For example, should I now reduce my staff by, let us say, 20%?"
This is a very delicate question. In practice, the leaders of the level of the chief engineer want to keep, and not reduce the number of their people.
We usually give such an answer: "Specialists should stay. However, the profile of their work is changing – they have to do more analytical work". Modern EAM significantly increases the requirements for specialists! Previously, it was a mechanic, who first ran through the incident and was able to make high-quality welding. Now it is a mechanic, who knows modern materials, understands lubrication, causes of pressure drop, instrumentation and who can analyze and diagnose cause-and-effect relationships using a computer. Therefore, people need to relearn.
![](/uploads/thumb/0/platform_content/1685-.jpg)
Employees of INTERPIPE STEEL and other plants are equipped with mobile devices that allow to quickly obtain the necessary data and transfer them to a centralized ERP system.
– Nevertheless, do you see connection with the requirements of the leadership in your EAM practice? For example, we heard about two growing trends from metallurgists: requirements for reducing the number of people as well as massive independent dismissals of specialists – they are leaving the country. Both are related to the pace of implementation of systems such as EAM.
Trends that we see in enterprises at their level rather go from management. In some places, there is pressure to reduce the number of people. However, in my opinion, this may lead to the fact that there will be no one to service the equipment. Whatever you say, the rate of obsolescence of production assets as a whole is higher than the pace of modernization in our country. Therefore, it seems to me that the output should consist in the parallel implementation of such systems. Nevertheless, at the same time, people must be trained. There should not be a situation when the reasons for stopping the machine are diagnosed within 6-8 hours, this should take minutes. Of course, at the same time, such specialists should be adequately paid. Moreover, such changes are underway at a number of enterprises.
However, in some enterprises, in my opinion, there are no clear communication policies with respect to such systems. I repeatedly heard "We will implement this system and they will fire me. Why do I need it?" On the other hand, if you look objectively, the information system allows seeing the real load of people, their employment and efficiency. By experience, we see a completely different situation in the service at different enterprises of the same industry. For example, at one enterprise, the equipment works without stopping, at the other one it stops every month.
– You mentioned that a lot of statistical data is being processed. Are any regularities already visible that allow optimizing and improving maintenance and repair management system work?
There are many examples of improvements. For example, at INTERPIPE Nizhnedneprovsk Pipe Rolling Plant and INTERPIPE NIKO TUBE, there was previously only a service regulation and scheduled preventive maintenance, all were following it. Now they make measurements and see the situation in real time and this is already a preventive diagnosis. People really use it. Previously it was impossible to prove that there was a problem with the equipment if it was not obvious. There were external contractors who came with their expensive equipment, measured something, made conclusions and evaluated the condition. Now all this is available to the specialists of the enterprise, and they can pre-empt negative scenarios.
The second example is about changing culture, and it is very representative, in my opinion. There is always a period between the appearance of a service request in the system (automatically generated) and the actual response to this application. The system allows to clearly see this reaction time, and this statistic is available to management. So, according to the results of the first such analysis, many managers had a shock. They saw that the service staff responded to requests in a completely different way than was supposed by the regulations. For example, there is an urgent request for the repair of critical equipment, but nothing happens for 3 days.
– That is, services do not react to signals and recommendations of the system?
They react, but in their own order and in their understanding of convenience and priority. This happens not only because of negligence. People often have other priorities, and in general, there can be significant differences between them. Analysis of these deviations, understanding of priorities, the way management sees it and the way plant services do leads to optimization of work and better service efficiency.
– But how are these priorities aligned in practice?
In general, this is solved by organizational and programmatic methods. What are organizational methods? It is the rules, procedures and maintenance regulations. Programmatic methods are used when we set priorities for different types of equipment in process algorithms. Everyone can have different causes of breakdowns, from cosmetic ones to the safety of people. So priorities are distributed and it is necessary to act according to them.
– Are there any dashboards on KPI and what is their level?
One of the main indicators in the world is OEE (overall equipment effectiveness). OEE is often measured in manual mode. Somewhere it is not measured at all. Only a small part of our enterprises use automated tools. Let me explain the importance of this by the example of downtime – one of the 3 main components of OEE. Usually the sensors do not say why the equipment is stopped. They just say it is stopped. Therefore, a person diagnoses the cause. For this, we have a separate business process. For example, at INTERPIPE NIKO TUBE, there is a process that allows identifying the causes of downtime. Often this is not obvious. For example, maintenance personnel says: "You are not operating the equipment properly." And operational personnel says: "No, that was you who badly repaired it the last time." There are conflicts. Therefore, we need rules and processes that unequivocally interpret certain situations. At INTERPIPE NIKO TUBE, at the end of the month, the chief engineer examines such disputes and uses all the information from the system.
– It's great ... Is there an automatic measurement of OEE at INTERPIPE?
Yes, the company has switched to a single OEE measurement system with uniform rules for calculating OEE elements including downtime. As a result, we know that OEE has improved by 10-15%. The premium part of the specialists is tied to the OEE results.
– Do you analyze the reliability of the equipment by the method of RCM? What is RCM analysis?
RCM analysis is a process (technique) aimed at establishing the root causes of why equipment fails. Any downtime that is longer than the standard time generates a number of questions. The system automatically asks a group of specialists in reliability specifying questions such as "what is the reason for downtime", "did downtime affect people’s safety, ecology, production processes, etc.", "how could it be possible to preempt or eliminate it faster", etc. Questions are built as simply as possible and with unambiguous answers like "yes" and "no". As a result, a recommendation is issued. For example, "it is necessary to increase the frequency of checking the oil pressure". The system introduces this into plans, and the regulations change further. Thus, there is a constant improvement of the service system.
– How have your own views on EAM and the development of similar systems in Ukraine changed after this implementation?
This is a modern approach. Companies can no longer do without this. It is impossible to collect statistics on the equipment without automating business processes for maintenance, ideally, with automatic input of information from the process control system and MES.
Of course, we are also completely moving away from standard scheduled preventive maintenance and are switching to preventive and then more and more predictive diagnostics, where data processing plays an increasingly important role. Another trend is high-quality data visualization. We already use many features so that the information is reflected in the best way.
The future is in these trends, we are actively working on it, and we are already implementing it.
– How much easier is it now to convince the customer with economic arguments like "look here, how the cost of service at plant X has decreased?" That is, how simpler is it than, say, 5 years ago? And can you guarantee improvements in these indicators?
In systems of similar complexity, the result can only be guaranteed if the project activity is properly organized. At the project review stage, we always discuss with the customer which key indicators will affect the results of implementation. We usually give our own and world statistics and indicate the conditions under which such improvements can be obtained. Therefore, for example, improvement in the OEE indicator is important yesterday, today and it will be important tomorrow, but technology is moving forward, and tomorrow we will introduce changes to what is best today. Earlier, we talked about automating business processes of scheduled preventive maintenance. Today we say that everyone in the world is switching to predictive diagnostics, but this requires information and processing in the data warehouse.
– Thanks for the interview! We wish you success and good projects!
The interview was taken by Yurchak Alexander (AIAU).
Based on: Industry 4.0 in Ukraine