Intelligent generator monitoring disc based on machine learning and mechanism model

Abstract

The invention relates to a generator intelligent monitoring disk based on machine learning and a mechanism model, which comprises an intelligent decision-making layer, an intelligent analysis layer and a fusion perception layer, and is characterized in that the intelligent decision-making layer comprises a key quality module, a quality defect module, a fault diagnosis module, a maintenance strategy module and a scheduling optimization module; the intelligent analysis layer comprises a big data analysis technology module, a process monitoring technology module and a big data acquisition technology module; the fusion perception layer comprises a business demand consultation module, a process flow carding module, an equipment operation analysis module, a process monitoring design module and a monitoring data acquisition module. The system has the advantages that response is timely, and faults can be timely processed.

Description

Technical Field

[0001] The present invention relates to the technical field of operation supervision systems, and in particular to an intelligent monitoring panel for generators based on machine learning and mechanism models. Background Art

[0002] The study of electrical operation failures and solutions in thermal power plants is an important basis for ensuring the stable and safe operation of thermal power plants and giving full play to their value in the power system. In the production and operation of thermal power plants, power generation, transmission, and use are carried out simultaneously, and all links must operate at the same time. If a problem occurs in a certain link, the production of the thermal power plant will face the possibility of suspension. The electrical system is a key component of the operation system of a thermal power plant. Because the electrical equipment of the thermal power plant is complex and has high and low interactive distribution, it increases the difficulty of locating and repairing electrical operation failures. The storage of electric energy in thermal power plants is relatively difficult, and strict requirements are placed on the capacity of electrical equipment. At the same time, backup preparations need to be made. A complex power grid operation framework is formed under the joint operation of multiple units. In this case, greater pressure is brought to the stable operation of the electrical system of thermal power plants. Based on this, we will increase the research on electrical operation failures in thermal power plants, continuously refine fault solving experience, and provide assistance for electrical operation and hidden danger elimination.

[0003] The occurrence of electrical operation failures in thermal power plants threatens the normal production of thermal power plants and affects the social benefits of thermal power plants. Electrical operation runs through the entire production process of thermal power plants and is of great significance to the safe production of thermal power plants.

[0004] The electrical operation system involves a large number of electrical equipment, and all processes from equipment selection, introduction to operation must strictly require quality. However, because each process involves a large number of influencing factors, if the electrical operation failure cannot be eliminated in time, it will not only increase the risk of the electrical operation system, but also bring additional costs. Because of this, thermal power plants must attach importance to solving electrical operation failures, do a good job in daily inspection and maintenance, build a safety barrier for the operation of the electrical system, do a good job in hidden danger inspection and risk warning, ensure that the equipment inspection and maintenance strategy is implemented, do a good job in electrical system inspection and fault resolution records, and continuously improve the level of electrical system fault resolution.

[0005] In recent years, large steam turbine generators have frequently participated in deep peak regulation. The different expansion differences of the materials of the generator components cause relative displacement and wear, resulting in frequent stator and rotor failures, long maintenance cycles and high costs, which seriously affect the safe and stable operation of the unit. At present, because the real-time state quantities of the generator system are in a decentralized state, lacking reasonable alarm thresholds and internal correlations, manual monitoring alone is slightly insufficient when potential failures may occur during the operation of the generator, and there are still problems with the existing generator monitoring technology.

[0006] As the main work of power plants, operation monitoring is still carried out in the traditional way. Monitoring the operating parameters of equipment and timely discovering equipment abnormalities are the main contents of the monitoring work of operators. With the traditional DCS fixed limit alarm setting, the situation is often quite bad when the alarm is triggered, which usually causes equipment shutdown and affects the safety and economy of the unit. In addition, the energy of operators is limited, and it is inevitable that there will be problems such as inadequate monitoring and delayed perception of parameter trend changes. Therefore, it is particularly important to build a real-time intelligent operation monitoring system, fully combine the experience advantages of personnel and the computing advantages of computers, and change the operation monitoring from people looking for data to data looking for people, which represents the future development trend. Summary of the invention

[0007] 1. Technical Problems Solved

[0008] The technical problem to be solved by the present invention is that currently, abnormalities in power plant equipment require operators to monitor the panel, but the response is not timely and the fault cannot be handled in time.

[0009] 2. Technical Solution

[0010] In order to solve the above technical problems, the present invention provides the following technical solutions: a generator intelligent monitoring panel based on machine learning and mechanism model, including an intelligent decision-making layer, an intelligent analysis layer and a fusion perception layer.

[0011] The smart decision-making layer includes a key quality module, a quality defect module, a fault diagnosis module, a maintenance strategy module and a scheduling optimization module;

[0012] The intelligent analysis layer includes a big data analysis technology module, a process monitoring technology module and a big data acquisition technology module;

[0013] The fusion perception layer includes a business demand consultation module, a process flow combing module, an equipment operation analysis module, a process monitoring design module and a monitoring data acquisition module.

[0014] Furthermore, the upper layer of the intelligent analysis layer is a smart decision-making layer, and the lower layer of the intelligent analysis layer is a fusion perception layer.

[0015] Furthermore, the intelligent analysis layer also includes a visualization module and a fault prediction module.

[0016] Furthermore, the big data analysis technology module includes a mechanism analysis submodule, a graph theory submodule, a qualitative simulation submodule, an expert system submodule, a propagation path identification submodule, a data classification and mining submodule and a predictive reasoning submodule.

[0017] Furthermore, the process monitoring technology module includes an analytical module submodule, a knowledge submodule, a data driven submodule and a multivariate statistics submodule.

[0018] Furthermore, the big data acquisition technology module adopts a distributed parameter system modeling and control module.

[0019] Furthermore, the business demand consultation module is signal-connected with the process flow combing module, the process flow combing module is signal-connected with the equipment operation analysis module, the equipment operation analysis module is signal-connected with the process monitoring design module, and the process monitoring design module is signal-connected with the monitoring data acquisition module.

[0020] Furthermore, the fusion perception layer also includes a predictive maintenance module.

[0021] Furthermore, the big data acquisition technology module contains a stator core temperature high fault correlation analysis and judgment model, a stator core temperature measuring point abnormal damage fault correlation analysis and judgment model, a stator winding interlayer temperature high fault correlation analysis and judgment model, a stator winding interlayer temperature measurement point abnormal damage fault correlation analysis and judgment model, a stator core end local looseness analysis and judgment model based on time-space correlation, a stator winding water outlet temperature high fault correlation analysis and judgment model, a stator winding water outlet temperature abnormal damage fault correlation analysis and judgment model, and a stator wire rod water outlet pipe blockage correlation analysis and judgment model based on multi-dimensional scaling correlation.

[0022] 3. Beneficial Effects

[0023] The advantages of the present invention compared with the prior art are:

[0024] 1. Through the core model algorithm, the automatic modeling tool is used to dynamically identify various working conditions and operation modes of the unit, and then the monitoring model is established. The massive real-time production data is automatically scanned regularly by the computer, and the abnormal parameters are captured in combination with the alarm logic;

[0025] 2. System learning can be achieved by collecting on-site production parameter measurement points, accumulating historical data, and then using a large amount of data to train the model to form a normal historical estimate of the measurement point, thereby achieving the purpose of predicting a certain technical indicator. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is an application architecture diagram of the present invention;

[0027] Figure 2 This is a diagram of the early warning module of the present invention;

[0028] Figure 3 This is a diagram of the architecture of the remote expert diagnosis system of the present invention;

[0029] Figure 4 This is the architecture diagram of the digital twin system of the present invention;

[0030] Figure 5 This is a communication architecture diagram between the remote engineering station of the present invention group and the production unit;

[0031] Figure 6 This is an application architecture diagram of the smart wearable device of the present invention; DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] It should be noted that all directional indications such as up, down, left, right, front, back, etc. in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture as shown in the accompanying drawings. If the specific posture changes, the directional indication will also change accordingly.

[0034] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0035] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the invention.

[0036] The present invention will now be further described with reference to the accompanying drawings.

[0037] Example

[0038] Generators are a key component of the operation of the electrical system of a thermal power plant. Because the electrical system works for a long time, the load of the generator increases, and heating is inevitable. However, if it is not handled in time, the problem of generator heating will be aggravated, threatening the safe operation of the generator and shortening the service life of the equipment. Generator heating failure is a common problem in the electrical system of a thermal power plant, and the causes of generator heating are diverse. For example, the wear of the internal components of the generator increases under long-term operation, and the influence of excessive friction will inevitably cause the generator to heat up. When the generator is in the state of heating, the iron loss and copper loss increase significantly, and the current fluctuates accordingly, continuously increasing the operating load of the generator. The generation of heat energy during the operation of the current is an inevitable process, but if the heat energy cannot be evacuated in time, the speed of heat dissipation will increase, threatening the safe operation of the generator. In addition to mechanical heating during the operation of the generator, if the bearing lacks oil, it will also cause the generator to heat up. Therefore, for the heating failure of the generator, it is necessary to objectively analyze the cause of the failure and then formulate targeted solutions.

[0039] Build a generator status big data platform, establish an efficient and stable generator intelligent fault analysis and expert knowledge base system, make full use of big data collection, big data preprocessing, big data storage and management, big data analysis and mining, big data presentation and application technologies, form equipment fault warning models, evaluation and warning indicators, health status samples and all-operating state characteristic parameter warning thresholds, and realize generator status fault warning based on big data.

[0040] The system is based on the generator temperature and electrical data obtained from the ICS platform, and is mainly composed of stator core temperature, stator winding water outlet temperature and stator winding interlayer temperature fault intelligent early warning modules.

[0041] Each module deeply mines and analyzes the historical stator core temperature, stator winding water outlet temperature and stator winding interlayer temperature data of the generator, summarizes the typical fault characteristics related to the generator stator temperature, and constructs a fault judgment model with the following typical characteristics: (1) Stator core temperature high fault correlation analysis and judgment model, stator core temperature measurement point abnormal damage fault correlation analysis and judgment model; (2) Stator winding interlayer temperature high fault correlation analysis and judgment model, stator winding interlayer temperature measurement point abnormal damage fault correlation analysis and judgment model and stator core end local looseness analysis and judgment model based on time-space correlation; (3) Stator winding water outlet temperature high fault correlation analysis and judgment model, stator winding water outlet temperature measurement point abnormal damage fault correlation analysis and judgment model and stator wire rod water outlet pipe blockage correlation analysis and judgment model based on multi-dimensional scaling correlation.

[0042] The system can quickly identify fault signs such as high generator temperature and damaged measuring points. It can also accurately locate blockage in the wire rod water outlet pipe and loose core end faults, and issue first-level warnings, second-level warnings and adjustment suggestions to operating personnel to avoid generator burning accidents.

[0043] A visual analysis function is designed. The generator temperature data is analyzed in the form of broken lines and charts on the ICS platform, which can present data characteristics and temperature change trends of fault measurement points in multiple dimensions. In order to further confirm the change trend of the fault point, the target shooting of each temperature state quantity is retained in the visual chart trend for 5 minutes. During normal operation, the target shooting analysis state quantity display points are relatively concentrated, and the deviation changes little. If the stator end is loose and the water outlet rod is blocked, the value of the corresponding time and space temperature state quantity has a large change in the relative deviation, and the target shooting analysis state quantity display point is obviously larger.

[0044] Through this visual analysis function, operators can timely observe and analyze the relationship between the causes of stator end loosening and water outlet bar blockage and the changes and trends in temperature data, and then discover the early signs of typical stator failures, reducing the inspection pressure on operators.

[0045] Based on the SIS system, the real-time production data and equipment measurement point alarm are visualized and integrated, which is responsible for the real-time online monitoring of the equipment parameters of the whole plant. It has complete parameter trend change alarm and equipment operation status prediction functions, combined with measurement point alarm technology and parameter

[0046] The lower limit automatic approximation technology, online parameter function technology and online parameter curve technology can realize visual alarms such as parameter over-limit (high), over-limit (low), severity and comprehensive abnormalities. It has basic functions such as real-time tracking of abnormal parameters, intelligent meter reading, limit function technology, system debugging, batch import and export, and hierarchical authority management.

[0047] The failure mechanism of the unit is complex, the types of failures are numerous, and there is a complex mapping relationship between the fault symptoms. Therefore, in order to realize the intelligent early warning of hydropower unit failures, it is necessary to build a model of the operating parameters of the key points of the equipment, extract the fault characteristics, and compare them with the fault model, and then make diagnostic decisions to achieve intelligent and automated early and accurate warning.

[0048] Equipment fault diagnosis technology plays an important role in ensuring the safety of mechanical equipment, improving product quality, saving maintenance costs and preventing environmental pollution. The use of modern equipment fault diagnosis technology in production can bring huge economic benefits to enterprises. Although the economic significance of this technology is multifaceted, it can be summarized in two aspects, namely, the reduction of sudden failures and the reduction of maintenance costs. Predictive maintenance based on condition monitoring does not stipulate the maintenance cycle, but it is necessary to monitor the condition of the equipment regularly or continuously, and find out whether the equipment has deterioration or signs of failure based on the results of condition monitoring and fault diagnosis, and then arrange repairs when necessary. It can detect and diagnose the existing faults before the equipment fails, and can more accurately estimate the reliable time of continuous operation, thereby extending the service life of the equipment, reducing unexpected downtime accidents, and controlling excessive maintenance, thereby reducing spare parts consumption and maintenance workload, and preventing human failures caused by maintenance, effectively reducing maintenance costs.

[0049] The architecture diagram of the remote expert diagnosis system is as follows: Figure 3 As shown in the figure, it is mainly composed of a digital twin system, a remote engineer station, smart wearable devices, and an expert pool. It builds a bridge between the production unit and the group's expert diagnosis center, and directly empowers the front-line production units with expert knowledge.

[0050] The digital twin system architecture is shown in the figure below: Figure 4 As shown. The modeling engine of the digital twin enabling platform should be simple, easy to use and powerful, support multi-dimensional modeling of equipment, scenarios and mechanisms, provide a high-concurrency operating environment of at least one million points, and establish a digital twin model of the unit / equipment based on the actual physical objects of the production unit and the group's data lake data. Based on the digital twin model, real-time monitoring of the production process, alarms for abnormalities / faults, optimization of production energy consumption, and prediction of operating status can be achieved, and the digital twin system can complete information exchange with the group cloud.

[0051] The communication architecture between the group remote engineering station and the production unit is as follows: Figure 5 The remote engineering station sends the control logic, screen and corresponding data points of the distributed control system (DCS) in the production area 1 to the group remote engineering station terminal through the industrial Internet or the control dedicated line network by means of encryption and packaging.

[0052] The remote engineer station terminal translates and analyzes the received information, restores the received control logic, images and data points, and on this basis, traces the root cause of the fault problem.

[0053] For urgent or difficult problems on site, operators can connect with remote experts anytime and anywhere through smart wearable devices, and can fix on-site problems without waiting for experts to come to the site. Through the high-definition camera of smart glasses, real-time video or pictures are sent back to accurately describe the problem. Experts mark through the management background or mobile terminal, and direct on-site engineers more directly and intuitively through text, voice intercom, etc., to improve the efficiency of problem solving.

[0054] The real-time database of the power enterprise reads real-time data from the control system, and the intelligent monitoring system reads real-time data of system configuration related points from the real-time database. The background calculation engine parses the formula configured by the user to start the real-time index calculation and the dynamic calculation of the monitoring parameter benchmark value. The monitoring alarm calculation service calls the result calculated by the previous calculation engine to start the calculation of the alarm information. Users can use the monitoring configuration and debugging in this process, and view the data and alarm information through the monitoring interface. The automatic meter reading service obtains data from the real-time database and displays it on the intelligent monitoring user interface.

[0055] The data fusion intelligent monitoring system includes functional modules such as data fusion from multiple sources, curve thresholds, template creation, instantiation, rules, trends and processes; it uses artificial intelligence algorithms to calculate the baseline values ​​and normal variation ranges of major operating parameters in real time; it monitors the unit operating parameters online in real time, while reducing workload and improving the intelligence and comprehensiveness of the operation monitoring; it improves predictive capabilities through dynamic monitoring bands and intelligent settings.

[0056] Establish interfaces with SIS database, indicator calculation and other data, and use the intelligent monitoring configuration model to comprehensively calculate, analyze and uniformly display data from multiple sources; configure the monitoring data model, sort out the operation judgment rules of the key parameters of the unit, and use them to capture and judge operation anomalies, and continuously accumulate and improve them during the daily use of the system; systematize the experience and knowledge of operators through the expert system, and apply it to intelligent monitoring; configure display interfaces, permissions and other content.

[0057] Build a supporting guarantee system, mainly including data sources, management systems and infrastructure guarantees, to ensure the normal operation of the system and allow the intelligent monitoring system to better serve the daily operation.

[0058] The implementation of this system will effectively prevent major accidents in the production process of power plants, provide reliable technical means for the safe and stable production of power plants, and continuously strengthen safety supervision at the production site. It will play an important role in ensuring the safe operation of electricity and improve the productivity of power generation enterprises, generating good social benefits. It will strengthen the stability of power plant operation, thereby promoting the improvement of production reliability. It will reduce the probability of accidents in the production process, which is conducive to the harmonious and stable development of society. It will improve the work efficiency of the production process, so that the liberated staff will have more time and energy to create more value for society.

[0059] The invention and its implementation methods are described, and such description is not restrictive. The drawings show only one implementation method of the invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and implementation method similar to the technical solution without creativity, which should fall within the protection scope of the invention.

Claims

1. Intelligent monitoring panel of generator based on machine learning and mechanism model, including intelligent decision-making layer, intelligent analysis layer and fusion perception layer, characterized by: The smart decision-making layer includes a key quality module, a quality defect module, a fault diagnosis module, a maintenance strategy module and a scheduling optimization module; The intelligent analysis layer includes a big data analysis technology module, a process monitoring technology module and a big data acquisition technology module; The fusion perception layer includes a business demand consultation module, a process flow combing module, an equipment operation analysis module, a process monitoring design module and a monitoring data acquisition module.

2. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The upper layer of the intelligent analysis layer is the intelligent decision-making layer, and the lower layer of the intelligent analysis layer is the fusion perception layer.

3. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to any one of claims 1 or 2, characterized in that: The intelligent analysis layer also includes a visualization module and a fault prediction module.

4. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The big data analysis technology module includes a mechanism analysis submodule, a graph theory submodule, a qualitative simulation submodule, an expert system submodule, a propagation path identification submodule, a data classification and mining submodule and a predictive reasoning submodule.

5. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The process monitoring technology module includes an analytical module submodule, a knowledge submodule, a data driven submodule and a multivariate statistics submodule.

6. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The big data acquisition technology module adopts a distributed parameter system modeling and control module.

7. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The business demand consultation module is signal-connected to the process flow combing module, the process flow combing module is signal-connected to the equipment operation analysis module, the equipment operation analysis module is signal-connected to the process monitoring design module, and the process monitoring design module is signal-connected to the monitoring data acquisition module.

8. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The fusion perception layer also includes a predictive maintenance module.

9. The intelligent monitoring panel of a generator based on machine learning and mechanism model according to claim 1 is characterized in that: The big data acquisition technology module contains a stator core temperature high fault correlation analysis and judgment model, a stator core temperature measuring point abnormal damage fault correlation analysis and judgment model, a stator winding interlayer temperature high fault correlation analysis and judgment model, a stator winding interlayer temperature measurement point abnormal damage fault correlation analysis and judgment model, a stator core end local looseness analysis and judgment model based on time-space correlation, a stator winding water outlet temperature high fault correlation analysis and judgment model, a stator winding water outlet temperature abnormal damage fault correlation analysis and judgment model, and a stator wire rod water outlet pipe blockage correlation analysis and judgment model based on multi-dimensional scaling correlation.

Images