"Specialized and innovative" is the core asset of small and medium-sized enterprises, which not only have outstanding growth, but also have strong profitability. In the future, these companies will not only enjoy the dividends of policy support, but also gradually grow into "giant" companies with development potential under the strong bullishness of market funds.
Industrial operations
From "clairvoyance" and "tailwind ear" monitoring to equipment abnormalities, to intelligent monitoring software data processing and push alarms, to the intelligent diagnostic platform's algorithm model processing data to generate preliminary diagnostic information, and finally combined with the evaluation and analysis of diagnostic analysts to provide fault diagnosis services.
PHM (Prognostics and health management), that is, fault prediction and health management, originated in the field of aviation and was used to reduce the accident rate of military aircraft, the most representative case is that the US F-35 fighter using the PHM system after the maintenance manpower is reduced by 20-40%, the sorties rate is increased by 25%, and the total logistics cost is reduced by 50%.
- System principles
Predictive maintenance uses hardware pre-filtering, and then uploads the filtered signal, and after the server collects the signal, it adopts the analysis method based on time domain analysis to obtain the mean, standard deviation, peak (the difference between the maximum value and the minimum value) and the steep value of a period of sine vibration wave. Because it is a circular rotation of vibration, so regardless of the size of the vibration, the total mean is not much change, and into a normal distribution, only after the sensor abnormal movement, due to gravity caused by the component change, will make the mean offset larger, so that the mean can be defined as a reference line, and analyze the long-term change of the mean to obtain its normal fluctuation range (according to the abnormal standard of normal distribution, 3 times its standard deviation), according to the degree of its out-of-range, to determine the abnormal alarm. The total standard deviation and peak gradually increase with the use and deterioration of the equipment, so multiple continuous standard deviations and peaks are normally distributed based on an ascent curve, so we can determine its abnormality based on the speed of its rise. The total steep value is a relatively fixed value, regardless of the vibration size, as long as the equipment material strength is normal, the steep value is relatively fixed, and it is easy to get its abnormality, when the analysis value is abnormal, and then through the FFT (fast Fourier change) frequency domain spectrum analysis, so as to finally get the abnormal problem point (energy frequency distribution and energy size, different fault points of energy frequency and size and part are not the same) and causes, and then through the industry big data analysis to match the corresponding solution.
- PHM Intelligent O&M: Realize the dual value of security + economy for customers
PHM uses sensors and other methods to obtain various types of data such as equipment working conditions, surrounding environment, online or historical operating status, etc., through feature extraction, signal analysis, data fusion modeling, to achieve equipment operating state monitoring, failure model modeling, residual life prediction and reliability evaluation, etc., is a set of mechanical, electrical, sensing, artificial intelligence, communications, network and other multidisciplinary interdisciplinary high-end technology.
The big data fault analysis and prediction system is equipped with a brain-like algorithm suitable for high robustness in complex environments, and establishes a knowledge graph based on situational awareness through a multi-scale feature interception model, which is applied to the fault analysis and prediction system. At present, the status quo of the enterprise monitoring system is decentralized and independent operation, and the management personnel cannot understand the implementation of the production equipment in time; due to the small number of fault diagnosis experts in the enterprise, it is not enough to meet the needs of a large number of equipment fault diagnosis and prediction. The traditional method is to perform maintenance programs by regularly shutting down key equipment; by deploying a predictive maintenance system, saving historical data of important equipment, setting data cleaning rules, and regularly diagnosing equipment faults based on various matching analysis algorithms and fault diagnosis rules; combined with real-time monitoring of equipment operating status data, predictive maintenance through big data modeling can detect potential failures in advance and reduce the frequency of unplanned downtime.
Artificial intelligence powers the development of predictive maintenance technology
(Source SMEE)
In recent years, the rapid development momentum of the industrial Internet, in the intelligent manufacturing, Industry 4.0 and other policies to promote, 5G, big data, artificial intelligence and other new technologies are an important means to release the value of industrial data elements, in the National Industrial Information Security Development Research Center released the "Industrial Internet Innovation and Development White Paper", the main application scenarios include equipment / and product management, business and operation optimization, social resource collaboration three categories. The equipment and product management scenarios are the most widely used in the industry, including condition detection, fault diagnosis, predictive maintenance, and remote O&M.
The overall situation of the application of the industrial Internet platform
In equipment and product management, condition monitoring and alarm are the most important application scenarios. For the vast majority of industrial enterprises, the normal operation of equipment is the premise of production, through the status of equipment detection can effectively avoid non-fault downtime, thereby reducing the production loss and maintenance costs caused by equipment failure.
However, condition monitoring is the passive collection of equipment operating data, failing to predict equipment operation failures proactively, and with the introduction of the concept of "Industry 4.0", predictive maintenance has become one of the important innovation points of this concept. More sophisticated sensors, faster communication networks, and more powerful data computing platforms also pave the way for the development of predictive maintenance.
By the end of 2017, more than 80% of the 153 companies were actively addressing issues related to predictive maintenance. The types of companies surveyed include power transmission and transformation engineering and hydraulic power, electrical automation and robotics, machine tools and manufacturing systems, software and digital enterprises, etc.
Understanding and mastery of predictive maintenance
Predictive maintenance market situation
- Market size
In the 2019-2024 Forecast Maintenance Market Report, the global maintenance market size is forecast to reach $3.3 billion in 2018, and its compound annual growth rate is expected to exceed 39% by 2024 to reach $23.5 billion.
- Global industrial chain
180+ predictive maintenance enterprise market
According to the classification of the Industrial Internet Industry Alliance, industrial intelligence problems are divided into four categories according to influencing factors and complexity: one is multi-factor complex problems, the second is multi-factor simple problems, the third is the simple problems with few factors, and the fourth is the complex problems with few factors.
Artificial intelligence in the two major technical directions of industry
Among them, deep learning and knowledge graph are the two major technical directions of the current industrial implementation of artificial intelligence: deep learning focuses on solving complex problems with fewer influencing factors but higher degree of computation; knowledge graph focuses on solving problems with more influencing factors but relatively simple mechanisms.
An overall view of industrial intelligence applications
The concept of predictive maintenance is developed from the concept of "condition monitoring", according to the overall view of industrial intelligence applications, the development of machine learning and deep learning has greatly promoted the application of predictive maintenance. In the predictive maintenance scenario of the equipment/system, the machine learning method fits the complex nonlinear relationship of the equipment operation, which can improve the prediction accuracy and reduce the cost and failure rate. Its commonly used methods are time series model prediction method, gray model prediction method and neural network prediction method.
- Lightweight real-time detection of edges
By equipping the edge device, holographic information modeling obtains the spatio-temporal state information of the device from multiple channels and multiple perspectives, making the device information more reliable. Multi-perception fusion from multiple levels, multi-space to the device data to complement and optimize the combination of information, so that the device information is more comprehensive. The fault monitoring and prediction system based on AI algorithm and situational awareness makes the predictive maintenance response more accurate and fast.
Application areas: pneumatic actuators, stepper motors, servo motors, conveyor belts, photoelectric switches, visual inspection and other industrial automation detection lines and production lines, according to the equipment status data, so that the equipment in real time to maintain the optimal state, and complete real-time fault monitoring and prediction.
Application areas: machinery industry, electronic power industry, steel industry, petrochemical industry, etc.
In short, with the intelligent manufacturing, the popularity of Industry 4.0, the promotion of the industrial Internet by the state and many other factors, the future development of predictive maintenance is expected, but how to truly make it play the greatest value in the industry still needs to be continuously explored by all parties to make it blossom.
Major intelligent O&M enterprises in China
01 Yung Chi Ri Xin Stock Code[688768]
Founded in 2007, Anhui Rongzhi Rixin Technology Co., Ltd. is a high-tech enterprise in the field of industrial Internet. Provide customers with equipment intelligent operation and maintenance cloud platform solutions and dynamic equipment predictive maintenance products and services.
02 Donghua Test Stock Code[300354]
Jiangsu Donghua Test Technology Co., Ltd. was established in 1993, is a "one-stop" test system solution composed of sensors, conditioning amplifiers, data acquisition instruments, analysis software, engineering application software and professional services, widely used in scientific research, testing, teaching, equipment manufacturing and other fields.
03 Ward (Tianjin) Intelligent Technology Co., Ltd
(Tianjin) Intelligent Technology Co., Ltd. is a wholly-owned subsidiary of Ward Transmission (Tianjin) Co., Ltd. Mainly engaged in intelligent equipment health management cloud platform, the platform integrates modern sensor technology, wireless communication technology, Internet of Things technology, cloud computing and other cutting-edge technology, for industrial machines to wear intelligent monitoring equipment, real-time collection of equipment operation data.
04 Aerospace Intelligent Control (Beijing) Monitoring Technology Co., Ltd
Aerospace Intelligent Control (Beijing) Monitoring Technology Co., Ltd. was established in December 2018 by the overall restructuring of Beijing Aerospace Intelligent Control Monitoring Technology Research Institute, is a well-known equipment condition monitoring and fault diagnosis high-tech enterprises in mainland China, the company's intelligent operation and maintenance of industrial Internet cloud platform with fault diagnosis and life prediction as the core includes intelligent point inspection management system, online status monitoring system, equipment fault AI diagnosis system, equipment asset management system and other complete functions.
05 Shuo Orange (Xiamen) Technology Co., Ltd
Founded in 2016, Shuo Orange Technology is a high-tech enterprise that combines machine learning and equipment noise analysis with mechanical noise recognition in the field of predictive maintenance. The company's main product machine auscultation master, through the non-contact way to collect machine noise and the application of the original noise feature set system to standardize, systematic identification, analysis and processing, for equipment predictive maintenance, product automation quality inspection, environmental anomaly alarm, equipment remote monitoring.
06 Chengdu Anerfa Intelligent Control Technology
The company was officially operated in early 2018, and is now a participating unit of Chengdu science and technology-based small and medium-sized enterprises, "General Technical Specifications for the Construction of Intelligent Coal Preparation Plants" and other standards, and the Anfa Equipment Predictive Maintenance System (Intelligent Platform for Industrial Equipment Operation and Maintenance) was listed in the "Recommended Catalogue of Intelligent Technologies and Equipment for Coal Preparation Plants".
Chengdu Anerfa Intelligent Control Technology Co., Ltd. is a state-level high-tech enterprise in the field of industrial Internet. Intelligent O&M and security early warning service provider in the resource industry. Main products: intelligent equipment health management and fault early warning, operation and maintenance data services, microseismic early warning, mill load grinding sound temperature integrated sensor, mining production and operation big data SAAS, etc. Products and services are exported to Brazil, Morocco, Russia, the Philippines, Mongolia and other major resource countries.
Comparison of major O&M manufacturers in the market
Intelligent O&M is distributed among the main manufacturers in the market
The above is the fourth batch of noteworthy "specialized and special new" companies in the mining field excavated by Zhongying Bedrock for everyone in this period.
END