How to reduce the cost of maintenance of equipment by 40%
What is the predictive maintenance of industrial equipment and how does this methodology help to save
One of the key KPIs is the reduction of equipment maintenance costs by up to 40% due to increasing its efficiency, reducing downtime and increasing the productivity of technical specialists. But if the management of equipment and its indicators are extremely important for the world’s largest enterprises, these processes are traditionally underestimated by us.
Predictive maintenance and RCM methodology
Untimely maintenance of industrial equipment is not only direct loss from breakage in the form of costs for new parts. This is also a reduction in production volumes, the costs of reorganization and re-planning processes. In addition, disruptions in supply plans, reputation losses among customers and partners, which leads to even greater costs in the future.
To ensure the maximum level of reliability, it is not enough to apply the traditional approach to maintenance – to keep the schedule of work, to control the purchase of spare parts, to retain experienced specialists. Methodology Industry 4.0 provides a completely new approach – remote monitoring and control and predictive maintenance – predictive maintenance.
Until 2022, the market and the request for solutions for predictive services will increase seven-fold, or up to $ 10 billion in monetary terms, this will become one of the main trends of technological modernization.
The very technology of predictive maintenance is based on the methodology of service based on reliability (RCM).
Let’s say that we have a pump at the enterprise that pumps the liquid. At the beginning of operation, he performs the work perfectly, but over time, the filter gets clogged, the indicators fall. At one point, the indicators fall to a critical point, when the pump does not pump enough water. In this case, we are talking about equipment failure. To prevent such a scenario, the classical maintenance methodology involves replacing this pump after a certain time – usually the average performance of the site.
But breakdowns do not always happen after the same time. Sometimes the node can continue to work in the regular mode much longer, and sometimes fails before the deadline. The methodology of RCM is not to calculate the average lifetime, but to find the point at which performance begins to fall. When we see a deterioration in performance, we understand that we need to prepare for the failure of equipment, and we can carry out repairs.
RCM underlies predictive maintenance, the essence of which is in the periodic or continuous assessment of the condition of the equipment. With this approach, the ultimate goal is maintenance at the time when it is most cost-effective. Performing repairs on the calendar, industrial enterprises spend millions to purchase spare parts and carry out work where they could be avoided. Today, all companies and enterprises are going to improve the way equipment is serviced based on the forecast.
From RCM to Smart Predictive Maintenance
The development of this methodology and its implementation at the enterprise is consistent.
The first step is to display data, monitor equipment status, and manually monitor parameters.
The second step is a clear understanding of how the change in parameters affects the operation of the equipment, and the development of boundary values, for which it is necessary to correct the equipment. In addition to the benefits of preventing breakdowns, monitoring and setting indicators help determine the conditions under which the equipment performs its work most effectively. Once these boundaries are broken, you can interact with the equipment to restore the most effective mode.
Performing repairs on the calendar, companies spend millions on the purchase of parts and work where it could be avoided
The next step is to create an automatic notification system. In the previous stages, indicators could be monitored manually or using bar codes that are read by mobile devices and show the status of equipment in real time. This step involves the integration of all data into a single network and the installation of a program that automatically generates alerts to management personnel.
The last step is fully automated control of equipment: the system itself guides the entire process from monitoring indicators to processing requests for repairs and ordering the necessary components.
What is built up in the end is called smart predictive maintenance – intelligent, preventive maintenance of equipment. Constant monitoring and automation of maintenance not only prevent failures almost with 100% probability, but also save millions of dollars a year for enterprises.
What else will change the predictive service?
In addition to the obvious superiority in preventing breakdowns and rational use of resources, this approach gives a few more pleasant bonuses.
First, the generation and processing of a huge array of data on the operation of equipment allows you to simulate realistic scenarios of its operation. We can view the equipment indicators for yesterday, for the past week or for the whole year, and based on this data it is better to understand what conditions improve the overall performance of the company, and which – are harmful. Predictive maintenance makes it possible to build these forecasts and, based on a clear calculation, manage the entire production process.
Secondly, with such a scheme of service, the maximum person falls out of the process. And although job cuts due to robotics are not the best trend, it is difficult to overestimate the benefits for production. Software control allows you to instantly exchange data and approve decisions instantly, which saves time and protects production from downtime. Errors that a person might make inadvertently or accidentally can not be committed by a program that has clear settings. This makes the work of a manufacturing enterprise many times more efficient and coherent.
That use these techniques, it is not necessary to have the most up-to-date equipment and equip it with sensors, methodological approaches work on any equipment
Systems that carry out such maintenance will be developed with the help of artificial intelligence technologies based on machine learning.
Imagine that we have a complex setup that produces hundreds of indicators: temperature, pressure, speed, currents and so on. When this equipment fails, machine learning is used to predict non-obvious failures. We may not know what parameters have affected, what has changed, it is difficult for a person to determine. The computer and the machine can analyze huge data streams and learn from the indicators of the past, in order to prevent the breakdown of even the most complex systems in the future.
This is only a small part of what modern production control systems are. In general, they consist of equipment catalogs, document archives, equipment and node traffic records, repair cards, application and job management, admission orders, incident analysis, downtime accounting and analysis, warehouse management, procurement and write-offs, personnel management, e. All these systems exist and are used today, constantly improving not only internal systems, but also the shell. Developers from year to year make the interface easier and more understandable for work, and a platform for using such solutions can be a regular smartphone or tablet.
This approach to servicing shows significant economic results, quickly pays off, demonstrates good indicators of lower costs for equipment maintenance and improved technical readiness. But most importantly – in order to use these techniques, it is not necessary to have the most up-to-date equipment and constantly equip it with sensors. Methodological approaches work on any equipment, which is especially important for manufacturing enterprises in Ukraine.
Автор: Кирилл Костанецкий, руководитель проекта SmartEAM
По материалам : Новое время – https://nv.ua/techno/technoblogs/chto-takoe-prediktivnoe-obsluzhivanie-2476568.html
Accidents and breakdowns of the main and auxiliary equipment cause long downtime of the equipment, interruptions in the production process, and in some cases severe injuries to service personnel. To avoid any accident or breakdown, you need to warn her…
Since 2016, the INTERPIPE plants have been introducing an automatic system for maintenance and repair of equipment (PM). The system allows you to monitor equipment operation online, reduce planned and emergency downtime due to preventive diagnostics. To date, data on…
Artificial intelligence is the property of computer systems to perform creative, intellectual tasks. Machine learning is a set of algorithms and statistical models with which computers can solve the problem without using detailed instructions, using templates and logical conclusions. Anomaly…