Power grid production domain operation monitoring system
Technical Field
The invention relates to the field of power grid information, in particular to a power grid production domain operation monitoring system.
Background
With the development of big data and the development of the internet of things, the application of the big data and the internet of things on a power grid system is more common, and a smart power grid is generated. One of the applications of the smart grid is to realize remote monitoring, so that production and operation conditions are mastered in real time, the production and operation conditions are closely combined with a power grid enterprise operation management strategy, transparency and automation of power grid management are enhanced, and better development of a power grid system is realized.
In the prior art, the internet of things or big data are generally adopted to monitor the equipment of the whole power grid, the monitoring range of the equipment is usually embodied in the monitoring of each node of the whole power grid, and relevant maintenance personnel are dispatched to maintain after a certain node of the power grid is monitored to be in fault. The method is also applied to power grid production equipment, and only monitoring and after-fault alarming can be started, so that corresponding defect production equipment records and defect elimination records in the production field exist only as historical records, cannot be applied to analysis of defect occurrence of the actual power grid production equipment, and can only be used for overall calculation of indexes and projects in power grid management business. The defects of the production equipment are often not all accidental factors, the factors such as severe environment and overlarge load of the line can obviously improve the probability of the defects of the production equipment, so that the defects of part of the production equipment can be regularly and circularly generated. The defect production equipment prediction can be realized after the rule is obtained, and the defect elimination cost and the loss caused by the defects of the production equipment can be obviously reduced after the defect elimination preparation is made in advance. Although the reference rule may be provided by combining the statistics of the defect occurrence times and the defect production equipment distribution of the existing equipment, the accuracy is not high, and prediction errors are easy to occur.
Disclosure of Invention
The invention aims to overcome at least one defect of the prior art, and provides a power grid production domain operation monitoring system, which can not only realize real-time alarm of defective production equipment, but also establish a prediction model according to all production equipment states, production equipment distribution and production equipment connecting lines of the defective production equipment, realize advance preparation of defect elimination, avoid hidden danger of defect occurrence, optimize defect elimination process and provide basis for monitoring and prediction of other aspects of power grid operation.
The technical scheme adopted by the invention is that the power grid production domain operation monitoring system comprises:
the data acquisition module is used for acquiring basic data reflecting the state of the production equipment, distribution data of the production equipment and line data of the production equipment in real time, and judging whether the production equipment is in a defect state, and the defect type and the defect grade under the defect state based on the basic data; the basic data are data reflecting the state of the production equipment in the operation and maintenance stage of the production equipment, and comprise defects of a production equipment body, changes of the distance to the ground of a production equipment connecting circuit, load changes of the production equipment connecting circuit, position changes of the production equipment relative to the ground and the like, the distribution data of the production equipment comprise geographical distribution of the production equipment, the terrain where the production equipment is located, the quantity of the production equipment in the whole production domain and the like, and the data of the production equipment located lines comprise line distribution of the production equipment. The defect types comprise the types of defects of specific equipment in an overhead power transmission line and a power transmission cable line, such as inclination of a tower or damage of a cable support and the like; the defect grades comprise emergency defects, major defects and general defects, namely the defect grades are usually distinguished in the prior art;
the defect management module is used for receiving the data acquired by the data acquisition module, counting the defect number and defect distribution of the production equipment in the management area within specific time so as to provide a defect elimination basis for workers, and counting the defect elimination number and the time from the beginning of defect elimination to the completion of defect elimination of each defect production equipment; the defect management module also stores the statistical data to form historical data for subsequent monitoring. The statistical data of the defect management module can not only monitor the defect condition of the production equipment in real time, but also can be stored in a storage medium to form a continuously updated historical record, thereby being beneficial to providing reference data for analyzing the defect occurrence in the production domain. The data in the specific time can be calculated in response to enterprise management, and enterprise index data calculation such as defect times, defect elimination timeliness rate of production equipment, defect elimination rate of production equipment and the like is facilitated;
the data processing module is used for performing machine learning based on historical record data of the defect management module and establishing a prediction model to predict the occurrence of defects of production equipment in the region; the method can continuously train through updating of the historical records to obtain the prediction model with high accuracy, can predict the defect occurrence of the production equipment in the production domain according to the prediction model, is beneficial to realizing important attention on the predicted defect production equipment, and avoids the occurrence of the defects of the production equipment. When the situation cannot be avoided, the defect elimination advance preparation can be realized based on the prediction result, so that the operation stability of the power grid production domain is improved, and the economic loss caused by the occurrence of the defects is reduced;
the central control client end is used for visualizing the original data and the processed data of the data acquisition module, the defect management module and the data processing module and sending warning information based on the defect management module and prediction warning information based on the data processing module. The central control client acquires and visualizes data of all modules, and is beneficial to manual query and management of personnel at the central control client. And the personnel defect elimination arrangement is carried out through the real-time defect generation warning information sent by the defect management module, and the key attention and the preparation for eliminating defects in advance on the predicted defect production equipment are realized through the predicted warning information based on the data processing module.
Preferably, the data processing module performs machine learning by using a random forest algorithm and/or a convolutional neural network algorithm and/or a logistic regression algorithm based on an artificial intelligence learning system TensorFlow. TensorFlow is a common open-source machine learning platform in the prior art, and is beneficial to realizing more various machine learning by utilizing various algorithms on the basis of the TensorFlow.
Preferably, the data acquisition module comprises a current transformer, a voltage transformer, a positioning sensor, a humidity sensor, a temperature sensor and an image sensor which are arranged on the production equipment and/or the production equipment connecting line, and the current transformer, the voltage transformer, the positioning sensor, the humidity sensor, the temperature sensor and the image sensor are all common detection equipment and can be widely applied to each production equipment of a power grid; the data acquisition module also comprises a data transmission component for transmitting data.
Preferably, the system further comprises a plurality of mobile terminals for manually inputting basic data of the state of the reaction production equipment, production equipment distribution data and line data of the production equipment, wherein the mobile terminals are provided with positioning parts and are internally provided with characteristic information for arranging personnel at the mobile terminals. When the data acquisition module is invalid or needs manual data acquisition, the mobile terminal personnel are arranged for acquisition and recording, so that the situation that data cannot be acquired is avoided. In addition, the mobile terminal is also used as communication equipment for receiving the information of the central control client terminal, so that personnel provided with the mobile terminal are guided to eliminate defects and maintain; the characteristic information comprises personal basic data, technical posts, defect elimination history records, defect elimination rate, defect elimination time for various defects and maintenance range of correspondingly configured mobile terminal personnel.
Preferably, the scheduling module is used for planning missing people, acquires the information of the defective production equipment and the distribution information of the defective production equipment through the central client, acquires the position information of the mobile terminal and the characteristic information of the personnel allocated to the mobile terminal through the mobile terminal, and sends the scheduling information to the mobile terminal which is pre-arranged to be allocated by the missing people according to a specific scheduling strategy. The information of the defect production equipment at least comprises information of defect types, defect production equipment grades, defect production equipment models and the like, and the scheduling module schedules personnel according to all the received information and a specific scheduling strategy, so that the fastest real-time defect elimination is planned. Or after receiving the predicted warning information, arranging personnel and time to realize planning for removing hidden troubles caused by defects or planning for eliminating defects in advance.
Preferably, the data processing module receives all the information acquired by the scheduling module, establishes a defect elimination binary network according to the existing distribution of the production equipment and the position information of the mobile terminal, establishes a partial order relation to a set of the defect equipment according to the distribution of the defect production equipment, and models historical record data of the defect management module and characteristic information of personnel provided with the mobile terminal to obtain a deep neural network model; and determining an optimal specific scheduling strategy and sending the optimal specific scheduling strategy to a scheduling module according to an effect strengthening or weakening deep learning model generated after the mobile terminal personnel are eliminated and configured in the historical elimination record. In the process of establishing the deep neural network model, bipartite graphs are used for matching production equipment and personnel at a mobile terminal, the personnel and the production equipment are used as nodes and connected, and the weights on the connecting sides are continuously changed along with the change of missing data in a historical record.
Preferably, after the scheduling module acquires the specific scheduling strategy, route planning is performed according to the existing map, the information of the defective production equipment, the distribution of the defective production equipment, the position information of the mobile terminal and the characteristic information of personnel provided with the mobile terminal, and the specific equipment required for eliminating the defect is prompted.
Preferably, the data processing module is further combined with a prediction model and a deep neural network simulation for planning and scheduling to predict an optimal specific scheduling strategy in the defect occurrence state. And storing the predicted defect occurrence as simulated historical data, so that an optimal specific scheduling strategy is formulated when a predicted result occurs by using a deep neural network for planning and scheduling, and the simulated historical data is deleted after the optimal specific scheduling strategy is obtained, thereby reducing the load.
Preferably, the defect management system further comprises an index monitoring module, wherein the index monitoring module is used for counting the defect occurrence frequency, the defect elimination rate and the ratio of the defect elimination rate to a preset threshold of the index monitoring module in the defect management module management area. The index monitoring module can fully display real-time monitoring data so as to serve as a production domain management basis.
Preferably, the index monitoring module receives the prediction data of the data processing module, and performs simulation deletion based on the prediction data to obtain and record a prediction ratio of the data processing module when the prediction result is correct. The method can timely feed back various index completion degrees under the condition that a prediction result is correct based on prediction data and simulation deletion, and can also carry out defect occurrence prediction again after hidden danger occurrence of defects is avoided based on the prediction data, so that various index completion degrees under the condition that the prediction result is correct after hidden danger is obtained.
Preferably, the statistical data stored by the defect management module further includes types and models of the production equipment operating in the region, and numbers, types and models of the retired production equipment, the scrapped production equipment and the pre-commissioning production equipment. The method is beneficial to recording the equipment which does not work in the production domain, and can make reference for the retirement and maintenance time of other production equipment.
Preferably, the central control client marks real-time defect production equipment, defect grades, predicted defect production equipment and predicted defect grades in the form of pictures, characters, symbols and three-dimensional models.
Preferably, the system further comprises a value prediction module, wherein the value prediction module is used for receiving the data processed by the defect management module and the data processing module and then carrying out regional value evaluation according to the data stored in the database, including a damaged equipment price table, a production equipment connecting line table, the power supply quantity of a production equipment connecting line and the building distribution passed by the production equipment connecting line; and the value prediction module is combined with the prediction model to carry out prediction and evaluation on the economic loss generated by predicting the occurrence of the defects. Based on the real-time defect production equipment data, the economic loss caused by the real-time defect production equipment can be evaluated, and the follow-up more accurate loss calculation is facilitated. And the economic loss evaluation under the correct state of the prediction result based on the prediction result can provide another reference factor for a specific deletion strategy, and the economic loss evaluation is taken as a consideration factor of deletion planning by combining with specific index completion rate and the like.
Compared with the prior art, the invention has the beneficial effects that: the production equipment in the production domain system is monitored in real time, so that defect information is fed back in time, defects are eliminated in time, and loss caused by defects is reduced. More importantly, the historical records from defect occurrence to defect elimination completion in the production domain are effectively utilized based on machine learning, so that the defect occurrence in the production domain is predicted, defect elimination preparation can be made in advance, prediction of defect objects can be focused in time, defect hidden dangers can be found in time, the purpose of avoiding is achieved, and the operation stability of the whole power grid system is improved. And reference data can be provided for all aspects of the whole power grid production domain operation based on the prediction model, and the power grid production domain operation can be subjected to defect elimination simulation based on at least the reference data and combined with index completion and value evaluation to gradually improve a defect elimination strategy, so that an optimal defect elimination strategy is provided after actual defects occur.
Drawings
FIG. 1 is a schematic structural diagram of the present invention.
Fig. 2 is a visualization panel (one) of the central control client according to the present invention.
Fig. 3 is a second visualization panel of the central control client according to the present invention.
Fig. 4 is a visualization panel (three) of the central control client of the present invention.
Fig. 5 is a visualization panel (iv) of the central control client according to the present invention.
Fig. 6 is a visualization panel (v) of the central control client of the present invention.
Fig. 7 is the visualization panel (six) of the central control client of the present invention.
Detailed Description
The drawings are only for purposes of illustration and are not to be construed as limiting the invention. For a better understanding of the following embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
Examples
As shown in fig. 1, the present embodiment discloses a power grid production domain operation monitoring system, which includes:
the data acquisition module is used for acquiring basic data reflecting the state of the production equipment, distribution data of the production equipment and line data of the production equipment in real time, and judging whether the production equipment is in a defect state, and the defect type and the defect grade under the defect state based on the basic data; the basic data are data reflecting the state of the production equipment in the operation and maintenance stage of the production equipment, and comprise defects of a production equipment body, changes of the distance to the ground of a production equipment connecting circuit, load changes of the production equipment connecting circuit, position changes of the production equipment relative to the ground and the like, the distribution data of the production equipment comprise geographical distribution of the production equipment, the terrain where the production equipment is located, the quantity of the production equipment in the whole production domain and the like, and the data of the production equipment located lines comprise line distribution of the production equipment. The defect types comprise the types of defects of specific equipment in an overhead power transmission line and a power transmission cable line, such as inclination of a tower or damage of a cable support and the like; the defect grades comprise an emergency defect, a major defect and a general defect, namely, the defect grades are commonly used in the prior art for distinguishing. In the embodiment, data acquisition is performed on the power grid production equipment and the production equipment connecting line by deploying a data acquisition module;
the defect management module is used for receiving the data acquired by the data acquisition module, counting the defect number and defect distribution of the production equipment in the management area within specific time so as to provide a defect elimination basis for workers, and counting the defect elimination number and the time from the beginning of defect elimination to the completion of defect elimination of each defect production equipment; the defect management module also stores the statistical data to form historical data for subsequent monitoring. As shown in fig. 2, the defect management statistics panel at least includes statistics of defect frequency, defect elimination time rate of production equipment, defect elimination time, expired defects, and occurrence status of each defect level. The time unit for counting the data such as the erasure rate is year, and the statistics can be carried out according to the monthly and the quarterly according to the needs. In addition, as shown in fig. 4, at least the specific defect device, defect appearance, and defect distribution are also visualized;
the data processing module is used for performing machine learning based on historical record data of the defect management module and establishing a prediction model to predict the occurrence of defects of production equipment in the region;
the central control client end is used for visualizing the original data and the processed data of the data acquisition module, the defect management module and the data processing module and sending warning information based on the defect management module and prediction warning information based on the data processing module.
In this embodiment, the central control client acquires and visualizes data of all modules, which is helpful for personnel at the central control client to perform manual query and management. And the personnel defect elimination arrangement is carried out through the real-time defect generation warning information based on the defect management module, and the important attention and the preparation for eliminating defects in advance can be realized through the prediction warning information based on the data processing module. In this embodiment, the defect management module, the data processing module, and the central control client are all computer devices and are connected at least through a remote wireless network. As shown in fig. 3, the method at least includes panels and marks that are needed to be visualized by the power grid in practical applications, such as defect monitoring of production equipment, line monitoring, defect factor decomposition, analysis histogram based on geographical distribution of production equipment, and the like.
In this embodiment, the statistical data stored in the defect management module further includes types and models of the in-region operating production devices, and numbers, types and models of the retired production devices, the scrapped production devices and the pre-commissioning production devices. The method is beneficial to recording the equipment which does not work in the production domain, and can make reference for the retirement and maintenance time of other production equipment.
In this embodiment, the central control client marks the real-time defect production equipment, the defect level, the predicted defect production equipment, and the predicted defect level in a form including a picture, a character, a symbol, and a three-dimensional model.
In this embodiment, the data processing module performs machine learning based on an artificial intelligence learning system tensrflow by using a random forest algorithm and a convolutional neural network algorithm.
In this embodiment, the data acquisition module includes current transformer, voltage transformer, positioning sensor, humidity transducer, temperature sensor, image sensor, the data transmission part that sets up on production facility and production facility interconnecting link. The current transformer and the voltage transformer are common detection equipment in a power grid system, the positioning sensor can provide a distance geographical position for distribution of production equipment, the humidity sensor and the temperature sensor can detect changes of surrounding environment of the production equipment and help to judge defect types, the image sensor monitors the production equipment and the surrounding environment, automatic fault analysis is preferably performed through an image recognition technology, and the data transmission component is used for transmitting collected data.
In this embodiment, still include a plurality of basic data, production facility distribution data, the removal end of the line data that production facility is located that are used for artifical entering reaction production facility state, the removal end is equipped with locating part, and is equipped with in the removal end and sets up the characteristic information who removes end personnel. When the data acquisition module is invalid or needs manual data acquisition, the mobile terminal personnel are arranged for acquisition and recording, so that the situation that data cannot be acquired is avoided. In addition, the mobile terminal is also used as communication equipment for receiving the information of the central control client terminal, so that personnel provided with the mobile terminal are guided to eliminate defects and maintain; the mobile terminal personnel characteristic information comprises the personal basic data, the technical post, the defect eliminating history and the maintenance range of the mobile terminal personnel.
In this embodiment, the scheduling module is further configured to plan missing people, acquire defect production equipment information and defect production equipment distribution information through the central client, acquire mobile terminal position information and feature information of personnel who are allocated to the mobile terminal through the mobile terminal, and send scheduling information to the mobile terminal where missing people are scheduled to be allocated according to a specific scheduling policy. The information of the defect production equipment at least comprises information of defect types, defect production equipment grades, defect production equipment models and the like, and the scheduling module schedules personnel according to all the received information and a specific scheduling strategy, so that the fastest real-time defect elimination is planned. Or after receiving the predicted warning information, arranging personnel and time to realize planning for removing hidden troubles caused by defects or planning for eliminating defects in advance. As shown in fig. 5, the entire process from defect discovery to defect elimination is scheduled and monitored.
In this embodiment, the data processing module receives all information acquired by the scheduling module, establishes a vanishing binary network according to existing production equipment distribution and mobile terminal location information, and simultaneously establishes a partial order relationship for a defect equipment set according to the defect production equipment distribution, and models historical record data of a defect management module and feature information of personnel provided with a mobile terminal to obtain a deep neural network model for planning and scheduling, which is not the same model as the prediction model; and (4) strengthening or weakening deep learning according to the effect generated after the mobile terminal personnel are arranged in the historical vacancy elimination record, so that the optimal specific scheduling strategy is determined and sent to the scheduling module. In the process of establishing the deep neural network model, bipartite graphs are used for matching production equipment and personnel at a mobile terminal, the personnel and the production equipment are used as nodes and connected, and the weights on the connecting sides are continuously changed along with the change of missing data in a historical record. Meanwhile, a partial order relation is established for the defective production equipment according to the geographical distribution of the defective production equipment, and a matching strategy is optimized. In addition, a partial order relation can be established for all the production equipment according to the geographical distribution of all the production equipment, and then a distribution optimization matching strategy of the defective production equipment is combined.
In this embodiment, after the scheduling module obtains the specific scheduling policy, route planning is performed according to the existing map, information of the defective production devices, distribution of the defective production devices, location information of the mobile terminal, and feature information of personnel provided with the mobile terminal, and the specific devices required for eliminating the defect are prompted. The number of the allocated personnel and the required technical posts are obtained through the specific information of the defect production equipment, then the personnel with the corresponding technical posts and capable of reaching the defect eliminating position as fast as possible are scheduled according to the position of the mobile terminal and the pre-planned traveling route, and preferably, the required equipment for defect elimination can be prompted according to the defect type, so that the personnel can prepare the equipment for defect elimination.
In this embodiment, the data processing module further simulates and predicts an optimal specific scheduling policy in a defect occurrence state by combining a prediction model and a deep neural network for planning and scheduling. And storing the predicted defect occurrence as simulated historical data, so that an optimal specific scheduling strategy is formulated when a predicted result occurs by using a deep neural network for planning and scheduling, and the simulated historical data is deleted after the optimal specific scheduling strategy is obtained, thereby reducing the load.
In this embodiment, the defect management system further includes an index monitoring module, where the index monitoring module is configured to count a ratio of defect occurrence times, defect elimination rate, and timely defect elimination rate included in a management area of the defect management module to a preset threshold of the index monitoring module. The index monitoring module can fully display real-time monitoring data so as to serve as a production domain management basis.
In this embodiment, the index monitoring module receives the prediction data of the data processing module, and performs simulation deletion based on the prediction data to obtain and record a prediction ratio of the data processing module when the prediction result is correct. The method can timely feed back various index completion degrees under the condition that a prediction result is correct based on prediction data and simulation deletion, and can also carry out defect occurrence prediction again after hidden danger occurrence of defects is avoided based on the prediction data, so that various index completion degrees under the condition that the prediction result is correct after hidden danger is obtained.
In this embodiment, the system further comprises a value prediction module, wherein the value prediction module is configured to perform regional value evaluation according to a data stored in a database, the data including a damaged equipment price table, a production equipment connection line table, a power supply amount of a production equipment connection line, and a building distribution through which the production equipment connection line passes, after receiving data processed by the defect management module and the data processing module; and the value prediction module is combined with the prediction model to carry out prediction and evaluation on the economic loss generated by predicting the occurrence of the defects. Based on the real-time defect production equipment data, the economic loss caused by the real-time defect production equipment can be evaluated, and the follow-up more accurate loss calculation is facilitated. And the economic loss evaluation under the correct state of the prediction result based on the prediction result can provide another reference factor for a specific deletion strategy, and the economic loss evaluation is taken as a consideration factor of deletion planning by combining with specific index completion rate and the like. As shown in fig. 6 and 7, at least value, asset analysis and visualization distributed in various regions are included.
It should be understood that the above-mentioned embodiments of the present invention are only examples for clearly illustrating the technical solutions of the present invention, and are not intended to limit the specific embodiments of the present invention. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention claims should be included in the protection scope of the present invention claims.