CN117218495A - Risk detection method and system for electric meter box - Google Patents
Risk detection method and system for electric meter box Download PDFInfo
- Publication number
- CN117218495A CN117218495A CN202311286205.6A CN202311286205A CN117218495A CN 117218495 A CN117218495 A CN 117218495A CN 202311286205 A CN202311286205 A CN 202311286205A CN 117218495 A CN117218495 A CN 117218495A
- Authority
- CN
- China
- Prior art keywords
- risk
- parameter
- information
- judgment
- electric meter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 68
- 238000012545 processing Methods 0.000 claims abstract description 74
- 230000007613 environmental effect Effects 0.000 claims description 29
- 238000012549 training Methods 0.000 claims description 25
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 238000000034 method Methods 0.000 claims description 11
- 125000004122 cyclic group Chemical group 0.000 claims description 7
- 238000003915 air pollution Methods 0.000 claims description 5
- 239000000428 dust Substances 0.000 claims description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 238000000605 extraction Methods 0.000 description 10
- 239000000284 extract Substances 0.000 description 6
- 230000000630 rising effect Effects 0.000 description 4
- 238000012502 risk assessment Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000013480 data collection Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a risk detection method and a risk detection system for an electric meter box, which relate to the field of maintenance of electric facilities and comprise the steps of acquiring various parameter information influencing normal operation of the electric meter box, respectively inputting the various parameter information into a first joint relation judgment model and a second joint relation judgment model and an electric meter box risk primary judgment model, outputting first-class and second-class association factor change state judgment parameter object information and primary prediction risk parameter judgment object information, inputting the first-class and second-class association factor change state judgment parameter object information into a third joint influence index judgment model, outputting third-class influence index judgment parameter object information, inputting the information and the operation parameter information under the acquisition normal operation state into an electric meter box risk final judgment model, outputting final-stage prediction risk parameter judgment object information, judging the risk condition of the electric meter box and matching a prediction risk processing scheme. The risk detection method and system for the ammeter box provided by the invention are used for extracting, predicting and early warning the risk.
Description
Technical Field
The invention relates to the field of electric power facility maintenance, in particular to a risk detection method and system for an ammeter box.
Background
The electric meter box is an infrastructure for intensively installing electric meters, switches, wires and other devices, and plays an important role in the transmission, metering and centralized control management of electric power meters. The internal volume of the single ammeter box is smaller, and the environment is more complex, so that potential safety hazards possibly existing in the ammeter box are predicted in time.
In the traditional technology, mainly adopted is the mode of manual inspection detects the environment that the ammeter case was located, however, ammeter incasement environment is comparatively complicated, and traditional manual inspection mode is difficult to adapt to the inside dynamic environment of ammeter case for detect work efficiency is lower, and has certain hysteresis quality, and consequently, design a reasonable scheme that is used for ammeter case risk to detect is very necessary.
Disclosure of Invention
In order to overcome the defects in the prior art, the application provides a risk detection method and a risk detection system for an ammeter box.
In a first aspect, the present application provides a risk detection method for an electric meter box, where the method includes:
acquiring various parameter information of the electric meter box affecting the normal operation of the electric meter box in each cycle detection interval period or each cycle detection interval temperature period, respectively inputting the various parameter information into a first class-class joint relation judgment model and an electric meter box risk primary judgment model, and outputting first class-class correlation factor change state judgment parameter object information and primary prediction risk parameter judgment object information;
Inputting the first type association factor change state judgment result and the second type association factor change state judgment result into a third type joint influence index judgment model, and outputting third type influence index judgment parameter object information;
collecting operation parameter information of the ammeter box in a normal operation state, inputting the operation parameter information and third-class influence index judgment parameter object information in the normal operation state into an ammeter box risk final level judgment model, and outputting final level prediction risk parameter judgment object information;
judging the risk condition of the ammeter box based on the primary and final predicted risk parameter judging object information, and outputting an ammeter box risk prediction judging early warning result;
and based on the risk prediction judging and early warning result of the electric meter box, matching corresponding prediction risk processing schemes, and outputting scheme group comparison and selection implementation processing information.
By adopting the technical scheme, various parameter information influencing the operation of the electric meter box in each set interval section is input into a preset first-class and second-class joint relation judging model and an electric meter box risk primary judging model respectively, so that three different parameter object information for judging the risk condition of the electric meter box are obtained, the first-class and second-class joint factor change state judging parameter object information obtained through training output is input into a third-class joint influence index judging model, the associated judging parameter object information under the mutual influence between the two parameters is trained and output, the obtained third-class influence index judging parameter object information and the operating parameter information under the normal operating state are input into an electric meter box risk final judging model at the same time, the risk condition of the electric meter box is obtained through judging the primary and final-class predicted risk parameter judging object information to a certain extent, corresponding processing scheme information is timely made and matched, the early warning and risk analysis and the risk analysis are carried out on the early warning and the risk analysis of the risk condition of the electric meter box by acquiring the parameter information of various types and the correlation influence between the parameter information and the parameter information, and the early warning risk analysis of the early warning information is timely expanded.
Preferably, the change rate of the environmental temperature of the ammeter box is obtained, a preset rate determination value is set, whether the increase rate of the environmental temperature is larger than the preset rate determination value is judged, if so, a cycle detection interval temperature stage is output, and if not, a cycle detection interval period is output;
based on the cycle detection interval temperature stage and the cycle detection interval period, acquiring environmental parameter information of the electric meter box in each cycle detection interval period or each cycle detection interval temperature stage, which influences the normal operation of the electric meter box, wherein the environmental parameter information comprises environmental temperature, environmental humidity and air pollution dust index parameter information;
performing data filtering processing and training data packaging processing according to time logic on the environmental parameter information, and outputting a trainable sample data set;
and inputting the trainable sample data set into a first type joint relation judging model, wherein the first type joint relation judging model is obtained by a cyclic neural network, and outputting first type association factor change state judging parameter object information.
By adopting the technical scheme, the analysis and judgment of a certain rising rate are carried out on the environmental temperature information of the ammeter box, so that under the condition of high-temperature rapid rising, various intermittent parameter information can be acquired according to the rising rate, if the rate is normal, the parameter information of each time period can be acquired in the intermittent time period, the timeliness of information acquisition is improved, the environmental parameter information affecting the normal operation of the ammeter box in a certain type of interval is acquired, the parameter information is specially input into a model, the prediction judgment parameter object information affecting the ammeter box risk is trained according to the parameter information, the training treatment of the mutual influence relevance is carried out on small-class information in a certain large-class information, and the accuracy of the judgment result of the data information is improved through the diversity treatment of the influence degree analysis between various data.
Preferably, image information of various types of damages of the ammeter box, which influence the normal operation of the ammeter box, of the ammeter box in each cycle detection interval period or each cycle detection interval temperature period is obtained, wherein the image information comprises images of damages and color changes in each circuit of the ammeter box, and damage of outer layers of components and parts and damage of a shell of the ammeter box;
and dividing training sample image information from the image information, inputting the training sample image information into a second type joint relation judging model, wherein the second type joint relation judging model is obtained by a convolutional neural network, and outputting second type correlation factor change state judging parameter object information.
Through adopting above-mentioned technical scheme, through gathering the image information that the ammeter case of influencing ammeter case normal operating in the interval section of above-mentioned certain type is multi-class damaged, through the image information input of multi-class damage to a certain model in, carry out the processing of certain image information to analyze the relevance between multi-class image information and to the relevance of ammeter case risk influence degree, and train out prediction judgement parameter object information according to the relevance, through the diversity processing to influence degree analysis between the multiple data, in order to improve the accuracy of data information judgement result.
Preferably, operation parameter information of the ammeter box affecting the normal operation of the ammeter box in each cycle detection interval period is obtained;
and screening the operation parameter information of the trainable sample, inputting the screened operation parameter information of the trainable sample into an ammeter box risk primary judgment model, wherein the ammeter box risk primary judgment model is obtained by a cyclic neural network, and outputting primary predicted risk parameter judgment object information.
By adopting the technical scheme, the operation parameter information affecting the normal operation of the electric meter box in the certain type of interval section is collected, namely the operation parameter information of an internal circuit of the electric meter box is input into the primary risk judging model of the electric meter box, the predicted judgment parameter object information is obtained according to the training of the operation parameter information by the model, and the information is taken as judgment object information for finally judging whether the electric meter box has risks and the types of the risks, and is one judgment object information for judging the risks of the electric meter box relatively easily.
Preferably, extracting the trainable parameter information set from the first type of association factor change state judgment parameter object information and the second type of association factor change state judgment parameter object information;
And inputting the parameter information set into a third type of combined influence index judgment model, wherein the third type of combined influence index judgment model is obtained by a long-short-term memory neural network, and outputting third type of influence index judgment parameter object information.
Preferably, the trainable sample operation parameter information in the normal operation state is collected and extracted;
and inputting the extracted trainable sample operation parameter information and the third type influence index judgment parameter object information under the normal operation state into an ammeter box risk final stage judgment model, wherein the ammeter box risk final stage judgment model is obtained by a long-short-period memory neural network, and outputting final stage prediction risk parameter judgment object information.
By adopting the technical scheme, the trainable parameter information set in the obtained first-type and second-type association factor change state judgment parameter object information is input into the third-type joint influence index judgment model, and the trainable sample operation parameter information in the normal operation state and the obtained third-type influence index judgment parameter object information are input into the ammeter box risk final judgment model, so that the judgment parameter object information is predicted according to the influence association degree between the two-type parameter object information, the processing of various data is expanded, the analysis and processing of ductility relations among various information are improved, and the reliability of ammeter box risk judgment and the accuracy of analysis and judgment based on the data information are improved.
Preferably, based on the primary and final prediction risk parameter judgment object information, historical risk operation parameter information corresponding to the primary and final prediction risk parameter judgment object information is screened out from the historical risk operation parameter information of the ammeter box;
carrying out amplitude difference processing on the historical risk operation parameter information and primary and final-stage prediction risk parameter judgment object information, and outputting amplitude difference pre-judgment information;
and setting a preset amplitude difference judging section value, judging whether the amplitude difference preset judging information falls in the preset amplitude difference judging section value, if so, outputting a normal operation judging result of the electric meter box, and if not, outputting a risk prediction judging early warning result of the electric meter box.
By adopting the technical scheme, the historical risk operation parameter information corresponding to the category attribute in the primary and final prediction risk parameter judgment object information is screened out from the historical risk operation parameter information of the electric meter box, so that comparison judgment based on the screened historical risk operation parameter information is conducted on the two kinds of prediction judgment object information, namely whether the electric meter box has risk can be judged by judging the amplitude difference value obtained through comparison, and corresponding early warning information is made according to the judged result, so that timely notification is conducted on the risk prediction result and corresponding data information of the electric meter box.
Preferably, based on the risk prediction judgment result of the ammeter box, a history processing scheme information set corresponding to the screened history risk operation parameter information is obtained;
splitting and recombining the scheme steps with high similarity of the historical processing scheme information set, and outputting a novel processing scheme information set;
and the result of the history processing scheme information group and the new processing scheme information group is displayed, and the processing information is implemented by comparing and selecting the output scheme group
By adopting the technical scheme, the corresponding historical treatment scheme information sets are extracted from the historical risk treatment scheme information sets in a concentrated mode according to the screened historical risk operation parameter information, and the corresponding novel treatment scheme information sets are researched by splitting and recombining part of scheme treatment steps with high degree of identity and high treatment success rate on the historical treatment scheme information sets, so that the predicted small-difference change possibly occurring in the risk of the electric meter box is improved, the risk treatment efficiency of the electric meter box is improved, and the novel treatment scheme information sets and the historical treatment scheme information sets are displayed in a contrasting mode so as to facilitate the intuitiveness and the research ductility of information difference analysis.
In a second aspect, a risk detection system for an electric meter box includes:
the data acquisition unit is used for acquiring various parameter information of the electric meter box in each preset interval section, which influences the normal operation of the electric meter box, and acquiring operation parameter information of the electric meter box in a normal operation state and historical risk operation parameter information of the electric meter box;
the data processing unit is used for respectively carrying out simple processing on the trainable parameter information on the various parameter information, respectively inputting the processed parameter information into the first and second type joint relation judging model and the ammeter box risk primary judging model, and outputting first and second type correlation factor change state judging parameter object information and primary prediction risk parameter judging object information.
Preferably, the data processing unit includes a first parameter input unit, a second parameter input unit and a risk judging and early warning unit, where the first parameter input unit is configured to input the first and second type association factor change state judging results to a third type joint impact index judging model to obtain third type impact index judging parameter object information;
the second parameter input unit is used for inputting the operation parameter information and the third type influence index judgment parameter object information in the normal operation state to the ammeter box risk final level judgment model so as to obtain final level prediction risk parameter judgment object information;
And the risk judging and early warning unit is used for judging the parameter amplitude difference value between the primary predicted risk parameter judging object information and the final predicted risk parameter judging object information and the corresponding historical risk operation parameter information so as to confirm the risk condition of the electric meter box and early warn in advance.
Compared with the prior art, the invention has the following characteristics and beneficial effects:
1. the method comprises the steps of obtaining various parameter information of an ammeter box affecting normal operation of the ammeter box in each cycle detection interval period or each cycle detection interval temperature period, respectively inputting the various parameter information into a first type and a second type joint relation judgment model and an ammeter box risk primary judgment model, and continuously inputting the obtained judgment parameter object information into a third type joint influence index judgment model and an ammeter box risk final judgment model, so that different model training is conducted on the parameter information of various types of attributes, namely, different prediction judgment parameter object information is trained according to the correlation of the mutual influence among the analyzed various parameter information, analysis of diversity information is enlarged, and accuracy of subsequent ammeter box risk judgment and reliability of information analysis judgment are enhanced.
2. The primary and final prediction risk parameter judgment object information obtained through training is respectively judged by amplitude difference values with the historical risk operation parameter information corresponding to various types of attributes of the primary and final prediction risk parameter judgment object information in the historical risk operation parameter information of the electric meter box, so that whether the risk of the electric meter box exists or not is judged, early warning notification is made according to a judgment result, and therefore timeliness of predicting the risk of the electric meter box, efficiency of implementing corresponding processing scheme information timeliness of the risk of the electric meter box and ductility research of a novel processing scheme are improved.
Drawings
Fig. 1 is a block diagram of steps of a risk detection method and system for an electric meter box according to the present embodiment.
Fig. 2 is a block diagram of the steps of the S1 substep mainly embodied in the present embodiment.
Fig. 3 is a block diagram of the steps of the sub-step S1004 mainly embodied in the present embodiment.
Fig. 4 is a block diagram of the steps of the sub-step S1006 mainly embodied in the present embodiment.
Fig. 5 is a block diagram of the steps of outputting the first and second type association factor change state determination parameter object information, which is mainly embodied in the present embodiment.
Fig. 6 is a block diagram of the steps of the A2 substep mainly embodied in the present embodiment.
Fig. 7 is a block diagram of the steps of the a2002 sub-step, which is mainly embodied in the present embodiment.
Fig. 8 is a block diagram of the steps of the a2005 sub-step mainly embodied in the present embodiment.
Fig. 9 is a block diagram of a risk detection system of an electric meter box mainly embodied in this embodiment.
Fig. 10 is a block diagram of steps of a data processing unit mainly embodied in the present embodiment.
Reference numerals illustrate: 1. a data acquisition unit; 2. a data processing unit; 21. a first parameter input unit; 22. a second parameter input unit; 23. and a risk judging and early warning unit.
Detailed Description
The invention is described in further detail below in connection with the following examples.
Referring to fig. 1, a risk detection method and system for an electric meter box includes the following steps:
s1, acquiring various parameter information of an ammeter box in each cycle detection interval period or each cycle detection interval temperature period, which influences the normal operation of the ammeter box, respectively inputting the various parameter information into a first class-class joint relation judgment model and an ammeter box risk primary judgment model, and outputting first class-class correlation factor change state judgment parameter object information and primary prediction risk parameter judgment object information.
S2, inputting the first type association factor change state judgment result and the second type association factor change state judgment result into a third type joint influence index judgment model, and outputting third type influence index judgment parameter object information.
S3, collecting operation parameter information of the electric meter box in a normal operation state, inputting the operation parameter information and third-class influence index judgment parameter object information in the normal operation state into an electric meter box risk final stage judgment model, and outputting final stage prediction risk parameter judgment object information.
S4, judging the risk condition of the ammeter box based on the primary and final prediction risk parameter judgment object information, and outputting an ammeter box risk prediction judgment early warning result.
S5, based on the risk prediction judgment early warning result of the electric meter box, matching corresponding prediction risk processing schemes, and outputting scheme group comparison and selection implementation processing information.
Specifically, various parameter information which can influence the normal operation of the ammeter box in the set circulation interval section is collected through the data collection unit, and the various parameter information comprises: detecting that damage information of an outer layer of an ammeter box or a component belongs to one type of parameter information, detecting that external environment condition information of the ammeter box belongs to one type of parameter information and monitoring that operation condition information in an internal circuit of the ammeter box belongs to one type of parameter information, transmitting the various types of parameter information to a data processing unit, respectively inputting the different types of parameter information into a preset first type and second type joint relation judging model and an ammeter box risk primary judging model through the data processing unit, training the various models based on different neural networks, simultaneously feeding back and inputting first and second correlation factor change state judging parameter object information output by the first model and the second model to a first parameter input unit through the data processing unit, judging a model through a third type joint influence index preset in the first parameter input unit, the model is obtained by the same step of building the models and the models, the trained and output third-class influence index judgment parameter object information is fed back and transmitted to a second parameter input unit, the parameter object information is fed back to a data acquisition unit, a trend data acquisition unit outputs data acquisition instruction information and starts to acquire operation parameter information in a normal operation state of the ammeter box, the operation parameter information refers to operation data information in a circuit in the ammeter box, the operation parameter information is fed back and transmitted to the second parameter input unit, the input operation parameter information and the third-class influence index judgment parameter object information are trained to a certain degree through the second parameter input unit to obtain final-class prediction risk parameter judgment object information, and the operation parameter information is the same as the steps, and the risk judgment and early warning unit performs certain comparison between the primary predicted risk parameter judgment object information which is reserved in advance and the corresponding historical risk parameter information of the electric meter box by combining the final predicted risk parameter judgment object information, judges whether the current electric meter box has potential risks, if so, performs timely early warning notification, and makes a corresponding processing scheme according to the predicted risk category, and transmits the analyzed parameter information and the processing scheme information to the management terminal for information display.
Referring to fig. 2, a specific step S1 includes the following sub-steps:
s1001, acquiring the change rate of the environmental temperature of the ammeter box, setting a preset rate determination value, judging whether the increase rate of the environmental temperature is greater than the preset rate determination value, if so, outputting a cycle detection interval temperature stage, and if not, outputting a cycle detection interval period.
S1002, acquiring environmental parameter information of the electric meter box in each cycle detection interval period or each cycle detection interval temperature period, which influences the normal operation of the electric meter box, based on the cycle detection interval temperature period and the cycle detection interval period, wherein the environmental parameter information comprises environmental temperature, environmental humidity and air pollution dust index parameter information.
S1003, carrying out data filtering processing and training data packaging processing according to time logic on the environmental parameter information, and outputting a trainable sample data set.
S1004, inputting the trainable sample data set into a first type joint relation judging model, wherein the first type joint relation judging model is obtained by a cyclic neural network, and outputting first type association factor change state judging parameter object information.
Specifically, before collecting various data information of the ammeter box, certain analysis and judgment are required to be performed for a preset period to be collected, for example: if the cycle time interval of the collected data information is directly set to be one month or two months, the risk detection of the electric meter box may have certain hysteresis, so that the data collection of the cycle time interval of the two forms is set to have certain timeliness, and certain analysis and judgment can be performed according to the speed of the electric meter box affected by the external temperature, for example: if the rising rate of the environmental temperature data information is kept in a certain value which has risk influence on the electric meter box within one month or within two weeks, starting to acquire and analyze the parameter information in each interval phase based on the various parameter information of the electric meter box in the set cycle detection interval temperature phase, wherein the set cycle detection interval temperature phase and the cycle detection interval period are respectively preset through a period acquisition timing module arranged in a data acquisition unit, and one period detection mode of the cycle detection interval temperature phase and the cycle detection interval period can be selected according to actual conditions so as to improve the timeliness of information detection, the environmental parameter information which is acquired in the certain acquisition mode and influences the normal operation of the electric meter box is received through a first parameter input unit in a data processing unit, and the environmental parameter information comprises the environmental temperature, the environmental humidity and the index parameter information of air pollution dust, and the three information have non-negligible relevance, such as: the influence of ambient temperature and humidity on the circuit of ammeter case is affected damp and risk conditions such as circuit generates heat is comparatively obvious, and the influence degree between the two can overlap, air pollution dust index can carry out certain chemical reaction with the steam in the air because of ambient temperature and ambient temperature's change, with lead to the corruption to ammeter incasement circuit, cause the circuit damage, initiate multiple ammeter case risk, but through the information extraction module that transmits the environmental parameter information who gathers to setting up in data processing unit carries out the extraction of trainable sample dataset, and input the dataset that extracts to the first type joint relation judgement model that obtains based on cyclic neural network, with output first type association factor change state judgement parameter object information, and continue to input it to training in the next model.
Referring to fig. 3, the specific step S1004 includes the following sub-steps:
s1005, acquiring image information of various damages of the ammeter box affecting the normal operation of the ammeter box in each cycle detection interval period or each cycle detection interval temperature period, wherein the image information comprises images of damages and color changes in each circuit of the ammeter box, and damage of outer layers of components and parts and damage of a shell of the ammeter box.
S1006, dividing training sample image information from the image information, inputting the training sample image information into a second type of joint relation judging model, wherein the second type of joint relation judging model is obtained by a convolutional neural network, and outputting second type of correlation factor change state judging parameter object information.
Specifically, the data processing unit receives the first type of association factor change state judgment parameter object information, and before training different models of the first type of association factor change state judgment parameter object information, the data acquisition unit also receives the first type of association factor change state judgment parameter object information at the same time, the data acquisition unit outputs a data acquisition instruction, and starts to continuously acquire image information of multiple types of damages of the electric meter box, which affects the normal operation of the electric meter box, wherein the image information comprises images of damage and color change in each circuit of the electric meter box, damage of outer layers of components and damage of a shell of the electric meter box, such as: the damage of the line insulating layer, damage information such as damage degree, damage of the component shell, damage of the ammeter box shell and the like, damage degree information of ammeter box at different positions and under different conditions has influence of degree superposition and influence of relativity among information, the obtained various image information is subjected to image impurity removal and other processes by an image optimization processing module arranged in a data acquisition unit, the processed image information is transmitted to an information extraction module to extract trainable sample image information, the sample image information is input to a data processing unit, a second type joint relation judgment model obtained based on a convolutional neural network in the data processing unit is used for training to obtain second type joint factor change state judgment parameter object information, and parameter training is carried out continuously on next model transmission.
Referring to fig. 4, a specific step S1006 includes the following sub-steps:
s1007, acquiring operation parameter information of the ammeter box affecting the normal operation of the ammeter box in each cycle detection interval period.
S1008, screening the operation parameter information of the trainable sample operation parameter information, inputting the screened trainable sample operation parameter information into an ammeter box risk primary judgment model, wherein the ammeter box risk primary judgment model is obtained by a cyclic neural network, and outputting primary predicted risk parameter judgment object information.
Specifically, the data processing unit receives the second-type association factor change state judgment parameter object information, and before training different models, the data acquisition unit also receives the second-type association factor change state judgment parameter object information at the same time, the data acquisition unit outputs a data acquisition instruction, starts to acquire operation parameter information of the ammeter box affecting the normal operation of the ammeter box in each cycle detection period, the operation parameter information refers to operation parameter information such as resistance, voltage and current in a circuit, the operation parameter information is extracted by the information extraction module from trainable sample operation parameter information in the same steps, the extracted parameter information is input into an ammeter box risk primary judgment model obtained based on a cycle neural network, initial predicted risk parameter judgment object information is obtained, and parameter training is carried out continuously for next model transmission.
Referring to fig. 5, the step of outputting the first and second class association factor change state decision parameter object information includes the sub-steps of:
A1. and extracting the trainable parameter information set by using the first type of association factor change state judgment parameter object information and the second type of association factor change state judgment parameter object information.
A2. And inputting the parameter information set into a third type of combined influence index judgment model, wherein the third type of combined influence index judgment model is obtained by a long-short-term memory neural network, and outputting third type of influence index judgment parameter object information.
Specifically, the first parameter input unit receives the first and second type association factor change state judgment parameter object information transmitted by the data processing unit, and the information extraction module firstly extracts certain trainable parameter information sets of the two parameter object information so as to improve the accuracy of data information analysis, and the same steps as the step of performing parameter training are carried out, namely the extracted trainable parameter information sets are input into a third type joint influence index judgment model obtained by a long-short-period memory neural network in the first parameter input unit so as to obtain third type influence index judgment parameter object information, parameter training is continuously carried out on the next model, and the relevance among various parameter information is analyzed to a certain extent by parameter training of different types of models combined among the various parameter information, so that the reliability of risk judgment of a subsequent degree ammeter box and the ductility of analysis and research on the various data are enhanced.
Referring to fig. 6, a specific step A2 includes the following sub-steps:
A2001. and acquiring and extracting the operation parameter information of the trainable sample in the normal operation state.
A2002. And inputting the extracted trainable sample operation parameter information and third type influence index judgment parameter object information in the normal operation state into an ammeter box risk final level judgment model, wherein the ammeter box risk final level judgment model is obtained by a long-period and short-period memory neural network, and outputting final level prediction risk parameter judgment object information.
Specifically, when receiving the second type of related factor change state judgment parameter object information transmitted by the first parameter input unit in a feedback manner, the data acquisition unit outputs a data acquisition instruction, starts to acquire operation parameter information of the ammeter box in a normal operation state, extracts trainable sample operation parameter information in the normal operation state, inputs the extracted trainable sample operation parameter information into an ammeter box risk final-stage judgment model obtained based on a long-short-period memory neural network, and simultaneously, the second parameter input unit receives third type of influence index judgment parameter object information transmitted by the first parameter input unit, and the ammeter box risk final-stage judgment model carries out model training on the two types of parameter information, namely, the influence degree between the trainable sample operation parameter information and the third type of influence index judgment parameter object information is analyzed, the final-stage prediction risk parameter judgment object information of the trainable sample operation parameter information under the influence condition of the third type of influence index judgment parameter object information is trained, and the data comparison analysis and judgment of final judgment ammeter box predictability are carried out on the judgment object information.
And the third type of influence index judgment parameter object information is input into the ammeter box risk final stage judgment model.
Referring to fig. 7, a specific step a2002 includes the following sub-steps:
A2003. and screening historical risk operation parameter information corresponding to the primary and final prediction risk parameter judgment object information from the historical risk operation parameter information of the ammeter box based on the primary and final prediction risk parameter judgment object information.
A2004. And carrying out amplitude difference processing on the historical risk operation parameter information and primary and final-stage prediction risk parameter judgment object information, and outputting amplitude difference pre-judgment information.
A2005. Setting a preset amplitude difference judging section value, judging whether amplitude difference preset judging information falls in the preset amplitude difference judging section value, if so, outputting an ammeter box normal operation judging result, and if not, outputting an ammeter box risk prediction judging early warning result.
Specifically, when the data acquisition unit receives the final-level predicted risk parameter judgment object information output by the second parameter input unit through feedback, the information extraction module starts to acquire historical risk operation parameter information of the ammeter box from the management terminal, and receives the primary predicted risk parameter judgment object information and the final-level predicted risk parameter judgment object information, so as to acquire category attribute information corresponding to the two judgment object information, and extracts historical risk operation parameter information matched with the category attribute information corresponding to the judgment object information from the historical risk operation parameter information through the information extraction module, wherein the historical risk operation parameter information is used as standard comparison reference information for judging whether the two judgment object information are in a certain standard comparison reference information, namely, a preset amplitude difference judgment interval value is internally set through an information judgment module arranged in a risk judgment early warning unit, the extracted historical risk operation parameter information is respectively processed with the primary and final-level predicted risk parameter judgment object information through a data budget module so as to obtain amplitude difference pre-judgment information, and the information is transmitted to the judgment module so as to obtain amplitude difference value judgment information, when the preset amplitude difference value is in the ammeter box, the current risk judgment information cannot be displayed in a preset amplitude pre-stage, and if the current risk judgment information cannot be displayed in a preset amplitude pre-stage, the current risk judgment information is not displayed in a preset risk judgment interval, and the current risk judgment information can be obtained when the current risk judgment information is in a preset risk judgment interval, and the current risk judgment information is not displayed in a preset stage, and the current risk judgment information has a preset risk judgment stage, and the risk judgment information is displayed in a preset stage, so as to extract the early warning notice.
Referring to fig. 8, a specific step a2005 includes the following sub-steps:
A2006. acquiring a history processing scheme information group corresponding to the screened history risk operation parameter information based on an ammeter box risk prediction judgment result;
A2007. splitting and recombining the scheme steps with high similarity of the historical processing scheme information set, and outputting a novel processing scheme information set;
A2008. and the historical processing scheme information set and the new processing scheme information set are displayed as a result, and the processing information is selected and implemented by comparing the output scheme sets.
Specifically, when the data acquisition unit receives the judged risk prediction judgment result of the electric meter box through feedback, the data acquisition unit outputs a data acquisition instruction, starts to acquire a historical processing scheme information group corresponding to the historical risk operation parameter information extracted by the electric meter box, extracts steps of marking one or more implementation processes in the scheme information group with high similarity from the historical processing scheme information group through an information extraction module, carries out splitting according to the steps of marking one or more implementation processes extracted by an information splitting module arranged in the data processing unit, randomly combines the split scheme steps, extracts the type attribute of the corresponding processing risk after combination from the scheme information group after random combination through the information extraction module, carries out extraction of the type attribute corresponding to the risk category attribute of the judged risk parameter information, so as to obtain a novel processing scheme information group, receives the novel processing scheme information group and the obtained and extracted historical processing scheme information group through an information display module arranged in the risk judgment early warning unit, carries out information result comparison display, so as to obtain novel scheme information group comparison, select implementation processing information, and timely detect the difference between management information and management information of management terminals, and timely detect the risk information, and timely notice the difference is made between management information and the prediction information.
9-10, acquiring various parameter information of the ammeter box in each preset interval section, which influences the normal operation of the ammeter box, and acquiring operation parameter information of the ammeter box in a normal operation state and historical risk operation parameter information of the ammeter box through a data acquisition unit 1, and reserving the acquired parameter information so as to be transmitted to a next unit for inputting and processing the parameter information; the data processing unit 2 is used for respectively carrying out simple processing on the trainable parameter information on the various parameter information, respectively inputting the processed parameter information into a first class-class joint relation judgment model and a second class joint relation judgment model and an ammeter box risk primary judgment model, and outputting first class-class correlation factor change state judgment parameter object information and primary prediction risk parameter judgment object information; the data processing unit 2 comprises a first parameter input unit 21, a second parameter input unit 22 and a risk judging and early warning unit 23, and inputs the first and second type association factor change state judging results into a third type joint influence index judging model through the first parameter input unit 21 so as to obtain third type influence index judging parameter object information; the operation parameter information and the third type influence index judgment parameter object information in the normal operation state are input into the ammeter box risk final level judgment model through the second parameter input unit 22 so as to obtain final level prediction risk parameter judgment object information; and judging the parameter amplitude difference value between the primary predicted risk parameter judgment object information and the final predicted risk parameter judgment object information and the corresponding historical risk operation parameter information through a risk judgment early warning unit 23 so as to confirm the risk condition of the electric meter box and perform early warning in advance.
The above embodiments are not intended to limit the scope of the present application, so: all equivalent changes in structure, shape and principle of the application should be covered in the scope of protection of the application.
Claims (10)
1. The risk detection method of the ammeter box is characterized by comprising the following steps of:
acquiring various parameter information of the electric meter box affecting the normal operation of the electric meter box in each cycle detection interval period or each cycle detection interval temperature period, respectively inputting the various parameter information into a first class-class joint relation judgment model and an electric meter box risk primary judgment model, and outputting first class-class correlation factor change state judgment parameter object information and primary prediction risk parameter judgment object information;
inputting the first type association factor change state judgment parameter object information and the second type association factor change state judgment parameter object information into a third type joint influence index judgment model, and outputting third type influence index judgment parameter object information;
collecting operation parameter information of the ammeter box in a normal operation state, inputting the operation parameter information and third type influence index judgment object information in the normal operation state into an ammeter box risk final level judgment model, and outputting final level prediction risk parameter judgment object information;
Judging the risk condition of the ammeter box based on the primary and final predicted risk parameter judging object information, and outputting an ammeter box risk prediction judging early warning result;
and based on the risk prediction judging and early warning result of the electric meter box, matching corresponding prediction risk processing schemes, and outputting scheme group comparison and selection implementation processing information.
2. The method for detecting risk of an electric meter box according to claim 1, wherein the step of obtaining various parameter information of the electric meter box affecting normal operation of the electric meter box in each cycle detection interval period or each cycle detection interval temperature period comprises the steps of:
acquiring the change rate of the environmental temperature of the ammeter box, setting a preset rate determination value, judging whether the increase rate of the environmental temperature is greater than the preset rate determination value, if so, outputting a cycle detection interval temperature stage, and if not, outputting a cycle detection interval period;
based on the cycle detection interval temperature stage and the cycle detection interval period, acquiring environmental parameter information of the electric meter box in each cycle detection interval period or each cycle detection interval temperature stage, which influences the normal operation of the electric meter box, wherein the environmental parameter information comprises environmental temperature, environmental humidity and air pollution dust index parameter information;
Performing data filtering processing and training data packaging processing according to time logic on the environmental parameter information, and outputting a trainable sample data set;
and inputting the trainable sample data set into a first type joint relation judging model, wherein the first type joint relation judging model is obtained by a cyclic neural network, and outputting first type association factor change state judging parameter object information.
3. The method for risk detection of an electric meter box according to claim 2, wherein the step of outputting the first type of association factor change state determination parameter object information includes:
acquiring image information of various types of damages of the electric meter box, which influence the normal operation of the electric meter box, of the electric meter box in each cycle detection interval period or each cycle detection interval temperature period, wherein the image information comprises images of damage and color change in each circuit of the electric meter box, damage of the outer layer of components and damage of the shell of the electric meter box;
and dividing training sample image information from the image information, inputting the training sample image information into a second type joint relation judging model, wherein the second type joint relation judging model is obtained by a convolutional neural network, and outputting second type correlation factor change state judging parameter object information.
4. A risk detection method for an electric meter box according to claim 3, wherein the step of outputting second type of association factor change state determination parameter object information comprises:
acquiring operation parameter information of an ammeter box affecting normal operation of the ammeter box in each cycle detection interval period;
and screening the operation parameter information of the trainable sample, inputting the screened operation parameter information of the trainable sample into an ammeter box risk primary judgment model, wherein the ammeter box risk primary judgment model is obtained by a cyclic neural network, and outputting primary predicted risk parameter judgment object information.
5. The method for detecting risk of an electric meter box according to claim 1, wherein the step of outputting the first and second kinds of association factor change state determination parameter object information includes:
extracting a trainable parameter information set from the first-type association factor change state judgment parameter object information and the second-type association factor change state judgment parameter object information;
and inputting the parameter information set into a third type of combined influence index judgment model, wherein the third type of combined influence index judgment model is obtained by a long-short-term memory neural network, and outputting third type of influence index judgment parameter object information.
6. The method for risk detection of an electric meter box according to claim 5, wherein the step of outputting third-class impact index determination parameter object information includes:
collecting and extracting trainable sample operation parameter information in a normal operation state;
and inputting the extracted trainable sample operation parameter information and the third type influence index judgment parameter object information under the normal operation state into an ammeter box risk final stage judgment model, wherein the ammeter box risk final stage judgment model is obtained by a long-short-period memory neural network, and outputting final stage prediction risk parameter judgment object information.
7. The method for risk detection of an electric meter box according to claim 6, wherein the step of outputting final-stage predicted risk parameter judgment object information includes:
based on the primary and final prediction risk parameter judgment object information, historical risk operation parameter information corresponding to the primary and final prediction risk parameter judgment object information is screened out from the historical risk operation parameter information of the ammeter box;
carrying out amplitude difference processing on the historical risk operation parameter information and primary and final-stage prediction risk parameter judgment object information, and outputting amplitude difference pre-judgment information;
And setting a preset amplitude difference judging section value, judging whether the amplitude difference preset judging information falls in the preset amplitude difference judging section value, if so, outputting a normal operation judging result of the electric meter box, and if not, outputting a risk prediction judging early warning result of the electric meter box.
8. The method for risk detection of an electric meter box according to claim 7, wherein the step of outputting the risk prediction determination and early warning result of the electric meter box comprises:
acquiring a historical processing scheme information set corresponding to the screened historical risk operation parameter information based on the ammeter box risk prediction judgment result;
splitting and recombining the scheme steps with high similarity of the historical processing scheme information set, and outputting a novel processing scheme information set;
and the historical processing scheme information set and the new processing scheme information set are subjected to result display, and the processing information is selected and implemented by comparing the scheme sets.
9. A risk detection system for an electric meter box, adapted to a risk detection method for an electric meter box according to any one of claims 1-8, comprising:
the data acquisition unit (1) is used for acquiring various parameter information of the ammeter box in each preset interval section, which influences the normal operation of the ammeter box, and acquiring operation parameter information of the ammeter box in a normal operation state and historical risk operation parameter information of the ammeter box;
The data processing unit (2) is used for respectively carrying out simple processing on the trainable parameter information on the various parameter information, respectively inputting the processed parameter information into the first and second type joint relation judging model and the ammeter box risk primary judging model, and outputting first and second type joint factor change state judging parameter object information and primary prediction risk parameter judging object information.
10. A risk detection system for an electric meter box, the data processing unit comprising:
a first parameter input unit (21) for inputting the first and second type association factor change state determination results into a third type joint impact index determination model to obtain third type impact index determination parameter object information;
the second parameter input unit (22) is used for inputting the operation parameter information and the third type of influence index judgment parameter object information in the normal operation state to the electric meter box risk final level judgment model so as to obtain final level prediction risk parameter judgment object information;
and the risk judging and early warning unit (23) is used for judging the parameter amplitude difference value between the primary predicted risk parameter judging object information and the final predicted risk parameter judging object information and the corresponding historical risk operation parameter information so as to confirm the risk condition of the electric meter box and perform early warning in advance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311286205.6A CN117218495A (en) | 2023-10-08 | 2023-10-08 | Risk detection method and system for electric meter box |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311286205.6A CN117218495A (en) | 2023-10-08 | 2023-10-08 | Risk detection method and system for electric meter box |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117218495A true CN117218495A (en) | 2023-12-12 |
Family
ID=89049351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311286205.6A Pending CN117218495A (en) | 2023-10-08 | 2023-10-08 | Risk detection method and system for electric meter box |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117218495A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117523808A (en) * | 2024-01-04 | 2024-02-06 | 珠海派诺科技股份有限公司 | Electrical fire early warning system and method capable of being monitored in real time based on Internet of things |
| CN117593310A (en) * | 2024-01-19 | 2024-02-23 | 江苏红相蓝瑞电力科技有限公司 | Image detection method and device for electric energy meter assembly quality detection |
| CN117791869A (en) * | 2023-12-28 | 2024-03-29 | 湖北华中电力科技开发有限责任公司 | Data online monitoring method and system based on intelligent power distribution cabinet |
-
2023
- 2023-10-08 CN CN202311286205.6A patent/CN117218495A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117791869A (en) * | 2023-12-28 | 2024-03-29 | 湖北华中电力科技开发有限责任公司 | Data online monitoring method and system based on intelligent power distribution cabinet |
| CN117523808A (en) * | 2024-01-04 | 2024-02-06 | 珠海派诺科技股份有限公司 | Electrical fire early warning system and method capable of being monitored in real time based on Internet of things |
| CN117523808B (en) * | 2024-01-04 | 2024-04-09 | 珠海派诺科技股份有限公司 | Electrical fire early warning system and method capable of being monitored in real time based on Internet of things |
| CN117593310A (en) * | 2024-01-19 | 2024-02-23 | 江苏红相蓝瑞电力科技有限公司 | Image detection method and device for electric energy meter assembly quality detection |
| CN117593310B (en) * | 2024-01-19 | 2024-04-02 | 江苏红相蓝瑞电力科技有限公司 | Image detection method and device for electric energy meter assembly quality detection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117218495A (en) | Risk detection method and system for electric meter box | |
| CN105467971B (en) | A kind of second power equipment monitoring system and method | |
| CN105425768B (en) | A kind of second power equipment monitoring device and method | |
| Zaki et al. | Deep‐learning–based method for faults classification of PV system | |
| CN111241154A (en) | Storage battery fault early warning method and system based on big data | |
| EP2478423A1 (en) | Supervised fault learning using rule-generated samples for machine condition monitoring | |
| CN109886328B (en) | Electric vehicle charging facility fault prediction method and system | |
| CN111444169A (en) | Transformer substation electrical equipment state monitoring and diagnosis system and method | |
| CN115238785A (en) | Rotary machine fault diagnosis method and system based on image fusion and integrated network | |
| CN117375237B (en) | Substation operation and maintenance method and system based on digital twin technology | |
| CN111738156A (en) | Intelligent inspection management method and system for state of high-voltage switchgear | |
| CN116523506A (en) | Circuit fault prediction analysis system based on digital twin transformer station technology | |
| CN116961215A (en) | Rapid fault response processing method for power system | |
| CN117612345A (en) | Power equipment state monitoring and alarming system and method | |
| CN112305388A (en) | On-line monitoring and diagnosing method for partial discharge fault of generator stator winding insulation | |
| CN116418117A (en) | Equipment detection system for intelligent power grid | |
| CN116029699A (en) | Power information system based on data twinning and operation and maintenance method | |
| CN117110794A (en) | Intelligent diagnosis system and method for cable faults | |
| CN113836816A (en) | Generator carbon brush temperature monitoring system based on infrared image and temperature prediction method | |
| Alsumaidaee et al. | Fault detection for medium voltage switchgear using a deep learning Hybrid 1D-CNN-LSTM model | |
| Firos et al. | Fault detection in power transmission lines using AI model | |
| Catterson et al. | On-line transformer condition monitoring through diagnostics and anomaly detection | |
| Sarma et al. | A long short-term memory based prediction model for transformer fault diagnosis using dissolved gas analysis with digital twin technology | |
| CN106443238A (en) | High-voltage equipment state evaluation method, high-voltage equipment on-line monitoring device evaluation method and apparatuses | |
| CN117648237B (en) | Automatic monitoring method for performance test process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |