Remote condition monitoring and predictive diagnostics of the technical condition of expensive equipment, critical structures and structures (predictive maintenance)
Specialists of the monitoring and data analysis department of our Center carry out remote monitoring of data from any dynamic systems to maintain stability, identify signals of failures and system changes in the functioning of the observed objects, predictive diagnostics of the technical condition of expensive equipment, critical structures and structures.
The applications of our monitoring and data analysis methods are endless. They are used especially effectively in controlling processes with many factors that are not even differentiated by observers and under conditions of uncertainty.
Refusal to monitor the technical condition of structures and structures led to an environmental disaster due to the depressurization of a diesel fuel tank that occurred on May 29, 2020 at a thermal power plant affiliated with Norilsk Nickel.
On August 17, 2009, in the turbine room of the Sayano-Shushenskaya HPP, water was released from the turbine crater as a result of the destruction of hydraulic unit No. 2, which was put into operation in 1979. Hydraulic unit No. 2 was damaged due to the destruction of the studs securing the turbine cover to its stator, which led to to tear off the cover. The hydraulic unit rotor with the turbine cover and the upper cross was thrown upward by the flow of water. Water filled the turbine shaft and flooded the turbine room. Power and auxiliary equipment were damaged, and building structures of the turbine hall building collapsed. All 10 hydraulic units failed. Analysis of vibration data using Shewhart control charts would have avoided this disaster.
Contamination of oil in the Transneft Druzhba oil pipeline with organochlorine compounds in 2019 led to reputational losses on an international scale and significant direct financial losses to the Russian oil industry.
In all projects, we help our customers solve current problems and identify new opportunities (knowledge) for improving processes and safe operation of equipment, and develop clear practical recommendations taking into account economic feasibility.
We invest 30% of the company's income in scientific and research activities.
Goal: increasing economic efficiency
As a result of monitoring, analyzing and interpreting data, you will gain capabilities previously unavailable to your team:
- identify signals of process failures;
- stabilize unpredictable (statistically uncontrollable) processes to be able to predict and achieve the best state that your system can provide;
- adjust processes to target values (center in the tolerance, standard, specification);
- improve forecasting accuracy;
- make an objective assessment of the results of any changes “before” and “after”;
- develop procedures for optimal control of technological processes to minimize losses from errors of the first and second kind (with a significant reduction in both false alarms and missed signals);
- predict the occurrence of emergency situations and take actions in advance to eliminate them or mitigate adverse consequences;
- maintain the performance characteristics of expensive and critical equipment, structures and structures at the required level, promptly identifying signs and the degree of degradation of control parameters;
- minimize the consumption of expensive components in technological processes while improving the quality and uniformity of the products produced;
- implement engineering innovations.
Types of data analyzed
For remote predictive diagnostics, monitoring data coming from equipment and technological sensors, laboratory tests, and also collected manually can be used, for example:
- vibration data (vibration diagnostics), gap, flow rate, liquid level, pressure, temperature, humidity, gas analysis, position, speed, force, viscosity, density, hardness, radioactivity, size, illumination, dust content, concentration, presence and amount of impurities, tribodiagnostics , acidity, contamination, acoustic measurements, electrical measurements and other measurements;
- data on the frequency of events, incidents, any quantitative data (counting data.
Predictive diagnostic technology
Real multifactorial processes have uncertainty due to the nature of the variability of known measurable, numerically immeasurable and even non-differentiable factors. This uncertainty and the technology staff's lack of knowledge of how to manage it creates enormous waste.
Our solutions are based on the Statistical Process Control (SPC) methodology, the main tool being Shewhart control charts. Despite the apparent simplicity of control charts developed by Professor Walter A. Shewhart, their correct practical use and interpretation still remains the domain of a few specialists.
Our work with data includes the entire range of operations, starting with the organization of the procedure for collecting it.
We examine data inextricably from the context of its production. Therefore, to develop recommendations for process improvement, we rely on scientific knowledge and work with subject matter experts.
Finding the causes of failures, anomalies and inefficiencies in processes with the movement of research “upstream” gives an understanding of what, how and at what level needs to be done for continuous improvement of processes:
Drawing. Cause-and-effect relationship of the resulting variability of controlled parameters
Benefits of diagnosing using statistical process control (SPC) methods
Low data requirements
Data monitoring and diagnostics using SPC methods require an incomparably smaller amount of data than diagnostics using machine learning (ML) methods. For the purposes of diagnosis and forecasting by SPC methods, it does not matter whether the data represents 100% of the data series or part of it, but the data must be organized rationally and with respect to distribution over time. Manually collected data will do. As a result, there are minor requirements for the computing power of the equipment and the storage of historical data (continuous data collection is not required).
See Donald Wheeler's article: Experiments, big data or Shewhart control charts? Various approaches to process improvement. Does your approach do what you need?
Not reactive, but knowledge-based actions
Using SPC, analyzing the cause-and-effect relationships of failures adds new information to our customers’ knowledge base about the system of process functioning, rather than asking the dispatcher to take reactive actions without explaining “why,” as is the case with machine learning (ML).
Relies on scientific and expert knowledge of processes
In the improvement work, scientific knowledge about the nature of the phenomena under study is used and competent employees of the customer company are necessarily involved. Exploring the context of data production and working with experts in the field under study are our fundamental principles.
Takes into account the statistical state of a process to make predictions
Takes into account the statistical state of the process under study (stable or unpredictable).
Minimizes the consequences of making errors of the first and second types
Allows you to minimize the consequences of making errors of the first and second types in attempts to improve, indicating specific operational rules for influencing the process, namely, whether intervention in the work of the analyzed process is required or vice versa, only systemic changes will help.
See article The concept of variability and process control and open solution APCS. Errors of the first and second kind .
Takes into account the rapid loss of relevance of data accumulated over the past period
Everything flows, everything changes.
Creates a customized failure signature
Our methodology operationally accurately separates noise from signals, revealing the individual signature of equipment and process failures rather than requiring fictitious boundary values to be set.
For example, a problem indicating a lack of such knowledge was reported by OSyS (Optimized Systems and Solutions, a subsidiary of Rolls-Royce), which specializes in providing DSS (Decision Support System, DSS) for monitoring and optimization high value assets:
"When analyzing performance parameters, OSyS specialists were faced with the fact that they either had to set the threshold values too narrow, and then the number of alerts would sharply increase, or the values were too wide, and then it turned out that the values in the alert did not meet the required limits. However, technical tools, knowledge and analysis skills helped them create a failure signature for all kinds of data streams that can come from the equipment."
The world classification of types of maintenance is divided into the following types:
- reactive maintenance (repair or replacement upon failure, Reactive Maintenance);
- preventive scheduled maintenance (PPR, maintenance, Preventive Maintenance);
- Predictive maintenance based on signs of failures and the onset of degradation;
- maintenance focused on economic feasibility and safety (rationally combines all previous types of maintenance, Reliability Centered Maintenance).
We specialize in the third type, namely predictive maintenance, carried out based on the actual condition of critical and highly valuable structures, structures and equipment, in order to increase the degree of their safe operation, increase their useful life and reduce operating costs.
The effect of using predictive maintenance:
“Independent reports* show that launching a predictive maintenance program provides, on average, the following economic benefits:
Reduced maintenance costs: 25% - 30%;
• Reduction in the number of failures: 70% - 75%;
• Reduced downtime: 35% - 45%, including through advance ordering of spare parts;
• Increase in productivity by 20-25%.
By receiving prompt messages from us about the presence of special reasons that take the process of functioning of the observed object out of a stable state, as soon as this has become statistically significant, you can easily identify these reasons and start managing them (eliminate or, if possible, reduce their influence). Signals about the beginning of degradation of the monitored parameters will allow timely measures to be taken to eliminate or slow it down even before the onset of significant wear of the component or equipment.
Types of companies interested
1. Companies engaged in design, production, construction, warranty and post-warranty maintenance of expensive equipment, critical structures and structures.
2. Operating companies.
Monitoring mode
Monitoring can be carried out in the mode required by the customer, from one update per week on weekdays to hourly around the clock. One update can include either one data point or a new series of historical data points for the entire time period from the previous update.
The information presented on this page is for informational purposes only and does not constitute an offer or public offer. Prices, terms and availability of products or services are subject to change without notice. For detailed information about prices, conditions and ordering, please contact us.