Abstract
Enterprise information construction is the most effective way to improve the efficiency of process enterprises. Information integration technology is the key to implementing enterprise information. At the same time, fault diagnosis is also the technical focus of enterprise information. The fault diagnosis technology based on artificial intelligence method and the remote fault diagnosis technology based on B/S mode open up a new field for the research and application of fault diagnosis technology. When a certain mode, the output maximum value F n of the neural network is close to 0, and other values Smaller, both close to 0.1, then the system can be considered F n failure. Practice has proved that if there are enough training samples, the network has a strong diagnostic function to determine the correct rate of 96% or more, indicating that the network training effect is good and meets the diagnostic requirements.
Introduction
Human beings have entered the threshold of an unprecedented information age. The basic characteristics of the information age are the information technology revolution involving all fields and the information explosion caused by it, and the network is the technological core of the information age. In just a few years, the Internet has brought about tremendous changes in society’s business, government, and society, changing people’s perceptions and lifestyles [1]. The emergence of various new information needs has opened up new connotations and service spaces for the network. The country’s strategy of “informatization-driven industrialization” and vigorously promoting the transformation of traditional industries with information technology have already made enterprises feel that they will face a new round of important competition. In the era of knowledge economy, the market environment in which enterprises operate is fundamental. Sexual changes, market competition is becoming increasingly fierce. Petrochemical and chemical enterprises are capital- and technology-intensive industrial sectors. They are process industries with continuous production conditions. The production processes are interdependent. The production process depends on the adjustment of process parameters. To achieve, the control information requirements are timely, stable and reliable, and there must be strict, accurate and timely management plans in management. According to the above characteristics, experts pointed out that the entry point of informationization construction in the chemical industry should be reflected in the advanced control of advanced production equipment. Optimized process flow, efficient production scheduling, high-quality equipment management, accurate and timely energy and material measurement to achieve the goal of safe, stable and long-term production process. At the same time, accelerate the technological innovation of enterprises through the application of information technology. Progress in management innovation and institutional innovation [2].
The “10th Five-Year Plan” of the chemical industry pointed out that the technical focus of computer chemical applications is computer production control and optimization technology, integrated manufacturing technology, chemical fault diagnosis technology, monitoring and safety system technology, engineering design technology, molecular design technology, simulation technology, etc. In order to improve the core competitiveness of enterprises, to accelerate the pace of development of traditional chemical industry by informationization, it is necessary to use computer and network technology in real-time information systems in scientific research, production, management, etc., and real-time data display using Web technology is An indispensable component, which is remotely monitored by industrial field data, is of special significance for remote scheduling at the production site [3].
The innovation of this paper is based on the network structure and system operation principle of B/S remote fault diagnosis system, and points out several key technical problems to be solved by implementing this technology.
Literature review
Pang discusses the development of factory real-time information systems under Web technology and its nature. The various data information of the industrial control system is centralized into the Web database, and the corresponding data is transmitted to the client’s Web browser through the Web server, which is based on a popular development mode [4] on the B/S structure mode. Dhimish believes that remote fault diagnosis is inseparable from the network, and it is the medium for monitoring the communication between the site and the diagnostic center. In order to improve the accuracy of fault diagnosis, it is necessary to improve the performance, reliability and fault tolerance technology, communication protocol and communication service quality assurance technology [5]. Guangning pointed out that fault diagnosis often has many uncertain factors, so it is necessary to adopt an expert system that supports fuzzy knowledge and deterministic knowledge synthesis and reasoning. Fuzzy knowledge reasoning can better simulate the ideas and methods of experts to solve problems when the initial information is incomplete or not very accurate. The essence is to infer the credibility of the conclusions based on the preconditions and the uncertainty of the rules themselves. Use less sophisticated knowledge to get the most accurate answer possible. The difficulty and key of the expert system based on fuzzy and determined knowledge comprehensive reasoning is the expression and acquisition of knowledge and the management of knowledge base [6].
Experimental methods
Concentrating various data information of the industrial control system into the Web database, and transmitting the corresponding data to the client’s Web browser through the Web server is based on a development mode on the current popular B/S structure mode [7].
ASP dynamic server page
ASP is a server-side script execution environment that belongs to Active X. It is a server-side scripting environment developed by Microsoft. It is included in 15 (PW)S. The system provides two scripting languages: VB script and Java script, while VB script is used as the default scripting language for the system. The ASP program exists on the Web server in plain text with the extension .asp. It can be opened with any text editor. The ASP program can contain plain text, HTML mark up, and script commands [8].
Browser/server mode (B/S)
The three-layer B/S architecture has many advantages that traditional C/S architecture does not have, and it has a combination of Internet/Intranet technology, which is the general trend of technology development [9].
B/S/ three-tier architecture
The three-tier structure of the B/S mode Modern enterprise network is Web-centric, adopts TCPPIP technology, uses HTTP as the transmission protocol, and the client accesses the Web and the back-end database connected to the Web through the browser. This is called B/S mode. It evolved from the CPS (Customer P Server) model. In CPS mode, the corresponding application must be installed on the foreground client workstation, and the server is a pure database server. Once the application changes, all client workstations using the application must reinstall the application, causing a large inconvenient. The B/S mode simplifies the client software, and only needs to install the browser as the running platform of the client application, and concentrate all development, maintenance and upgrade work on the server side, which is easy to maintain. The user uses the browser to access the Internet, submits the request to the web server, the web server processes the request, queries the database, and organizes the result of the query into an HTML page and sends it to the client, which is displayed on the user’s browser, and constitutes a three-layer B/S mode structure [10].
Neural network based fault diagnosis method
At present, research on neural network fault diagnosis systems has been very active. Because the neural network has complex multi-mode processing and association, speculation and memory functions, it is very suitable for fault diagnosis systems. It has the self-organizing self-learning ability and can overcome the shortcomings of the traditional expert system when the heuristic rules are not considered [11]. Therefore, the application of neural networks to process monitoring systems has become a very active area of research. Using neural networks for fault diagnosis of control systems, there are two ways to perform offline diagnosis and online diagnosis.
Key technologies of B/S based remote fault diagnosis system
Remote monitoring and diagnosis based on B/S is a new development direction in the field of equipment fault diagnosis. Therefore, there are still many theoretical and technical problems that need further research. In order for it to have a broad application prospect, the following key technical issues must also be addressed. Active access to the expert system knowledge base through Web pages is a key issue that must be addressed in building remote fault diagnosis expert systems. ASP technology and ADO technology can be combined to achieve access to the database [12]. The following Java Script statement first creates an ADO database access object.objConnectiontion, this object is linked to the ODBC data source, and then use this object to open the database named “remote user diagnosis”. Once you open the database, you can execute powerful SQL standard database manipulation statements (DML).
Experimental results and analysis
This paper uses a 3-layer BP neural network for fault diagnosis. The number of input layer nodes of the network is set to five, corresponding to five different characteristic frequency bands; the number of output layer nodes is set to three, corresponding to three common fault types of the rotor. The fault diagnosis program calls the neural network toolbox of the language a, and calls the triabnpx command. The triabnx P adopts the momentum method and the learning rate adaptive adjustment strategy, thereby improving the learning speed and increasing the reliability of the algorithm. The function trian lm uses the Lveenbegr-Marquardr optimization algorithm to make learning time shorter. The relationship between training learning error, learning rate and step size is shown in Fig. 1. It can be seen that when the BP algorithm is used, the error precision meets the requirements when the training reaches 987 steps. When the average training error is required to be smaller, the longer the training time, the more the number of training steps, and sometimes it does not even converge. When the requirements are met, the accuracy should be minimized to shorten the training time.
Relationship between learning error, learning rate and number of learning steps of B-P algorithm.
The BP optimization algorithm has higher accuracy in training results. In the fault diagnosis of high-speed rotating mechanical systems, the neural network determines the output data of the neural network to the fault conversion rule: when the output maximum value F n of the neural network in a certain mode is close to 0.9 and the other values are small, both are close to 0.1, then it can be considered that the system has an F n fault. Practice has proved that if there are enough training samples, the network has strong diagnostic function.
Statistics of all fault types diagnosis
Relationship between learning error and learning steps of improved BP algorithm.
According to 130 the above analysis results, it can be seen that in the diagnosis of each fault type (Shoulder, Carrier frequency offset, Low frequency offset, Spectral interference), the diagnostic errors of the three faults of offset and low frequency offset are lower than those of the actual low-frequency offset. In addition to the first verification point, the difference from the target value does not exceed, and basically meets the target value. The diagnostic accuracy rate of the four faults is greater than or equal to the above, indicating that the network training effect is good and the diagnostic requirements are met.
The fault diagnosis and accident warning system is still in progress. Since the whole research involves many aspects such as computer technology, network technology, and expert system, the research time is limited, and the system still needs further development and improvement. To sum it up, there are the following aspects:
Further expansion of the knowledge base the industrial production process is extremely complex, and the fault forms and mechanisms are diverse. To form a complete and accurate knowledge base of fault diagnosis, it is necessary to continuously communicate with the on-site technicians and optimize the various diagnostic rules to form. A relatively complete set of expert systems. The further improvement of the data acquisition system is difficult to debug due to the small probability of failures and accidents. Therefore, in order to verify the accuracy of the knowledge base, it is necessary to develop a simulation data generator of an object to generate a batch of data for various faults. Send to the fault diagnosis system. Integration with other enterprise subsystems The Intermec system is a fault diagnosis and accident warning system. The ultimate purpose of its establishment is to provide comprehensive management services for enterprises. It should be integrated into the enterprise B/S system as a decision system. Share resources with other systems such as fixed assets, budgets and final accounts, so as to achieve more effective management of the entire enterprise resources. The Internet-based intelligent monitoring system has the ability to collect display information and intelligent monitoring functions, and can use knowledge analysis information to propose improvements. On the one hand, it not only provides the operator with a normal operational reference, assists the operator in supervising the production operation, but also tells the operator the process and cause of the failure and accident formation, and proposes treatment suggestions so that the operator can quickly respond to the fault and accident. Make a correct reflection. On the other hand, it also guides the management to develop production plans, arrange equipment maintenance schedules, etc., which will become an important guarantee for factory safety production.
The growing Internet network is expanding to all corners of the world in an exponential relationship. The Internet technology has been increasingly matured and has been applied in most companies and companies around the world. In recent years, we have been note that Internet technology and Web technology are being introduced into the field of industrial automation. How to improve the operational efficiency of enterprises, reduce production costs, enhance the competitiveness of enterprises, and become the key to the successful entry of enterprises into the market, and enterprise information construction is the most effective way to improve the efficiency of process-oriented enterprises. At the same time, remote monitoring and fault diagnosis also make full use of the network to enable enterprises to obtain technical support in a timely manner, communicate information between various departments, and acquire, accumulate and integrate various aspects of knowledge, so that enterprises have stronger competitiveness and its application. With the rapid development of technology, equipment fault diagnosis technology must be combined with the Internet, and adopt an open architecture to have strong vitality and broad application prospects. The Web-based browser server model is most likely to become the computing model of the era of network computing. Therefore, combining the fault diagnosis technology with this model will inevitably greatly improve the level of fault diagnosis, which has important practical significance.