CN117390566B - Intelligent power plant abnormality detection method based on convolutional neural network algorithm - Google Patents
Intelligent power plant abnormality detection method based on convolutional neural network algorithm Download PDFInfo
- Publication number
- CN117390566B CN117390566B CN202311587541.4A CN202311587541A CN117390566B CN 117390566 B CN117390566 B CN 117390566B CN 202311587541 A CN202311587541 A CN 202311587541A CN 117390566 B CN117390566 B CN 117390566B
- Authority
- CN
- China
- Prior art keywords
- data
- abnormal data
- power plant
- abnormal
- equipment
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 64
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 15
- 230000005856 abnormality Effects 0.000 title claims abstract description 14
- 230000002159 abnormal effect Effects 0.000 claims abstract description 156
- 238000000034 method Methods 0.000 claims abstract description 38
- 238000007689 inspection Methods 0.000 claims abstract description 6
- 238000012544 monitoring process Methods 0.000 claims abstract description 6
- 238000003491 array Methods 0.000 claims description 9
- 238000010248 power generation Methods 0.000 claims description 6
- 239000003550 marker Substances 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 5
- 238000012795 verification Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to the technical field of equipment data detection, in particular to an intelligent power plant abnormality detection method based on a convolutional neural network algorithm, which comprises the following steps: s10: marking each device in the power plant by using a relevance marking method, and endowing each device in the power plant with a corresponding relevance marking system; s20: configuring a corresponding detection terminal for each device in the power plant, and realizing intermittent detection of the operation data of the device; s30: intermittently acquiring operation images of each device in the power plant by using the set video monitoring unit; s40: comparing the operation data detected by the detection terminal to each device with the standard data range of the normal operation of the device; s50: the abnormal data is checked by a checking method. The equipment to be detected is marked by the association marking method, so that the abnormal data is efficiently checked by matching with the checking method, the authenticity of data acquisition is ensured, and the efficient inspection of the operation of the power plant equipment is improved.
Description
Technical Field
The invention relates to the technical field of equipment data detection, in particular to an intelligent power plant abnormality detection method based on a convolutional neural network algorithm.
Background
A power plant refers to a power plant that converts some form of primary energy into electrical energy for stationary facilities or transportation. In order to ensure the power generation continuity, an auxiliary system is generally adopted to detect the power generation condition of the whole power plant in real time, and the acquired data is subjected to actual analysis so as to timely process abnormal equipment.
The existing equipment detection applied to a power plant generally adopts a mode that a plurality of detection equipment are used for respectively and independently detecting the plurality of equipment to obtain corresponding data, and a patrol inspector is commissioned to check and recheck if the data are abnormal, but the authenticity of the data of the single equipment obtained by the mode is completely dependent on the performance of the detection equipment, and the condition of data detection errors can exist.
Therefore, in the existing power plant operation data detection link, how to ensure the authenticity of the detection data is particularly important, and in view of the fact, we propose an intelligent power plant abnormality detection method based on a convolutional neural network algorithm.
Disclosure of Invention
Aiming at the defects existing in the prior art, the invention provides an intelligent power plant abnormality detection method based on a convolutional neural network algorithm, which can effectively solve the problem that the authenticity of detection data aiming at equipment is not checked effectively when the existing power plant operates, so that the cost of checking the data is high.
In order to achieve the above purpose, the invention is realized by the following technical scheme:
the invention provides an intelligent power plant abnormality detection method based on a convolutional neural network algorithm, which comprises the following steps:
s10: marking each device in the power plant by using a relevance marking method, and endowing each device in the power plant with a corresponding relevance mark which reflects the logical relation of the running states of the devices;
s20: configuring a corresponding detection terminal for each device in the power plant, realizing intermittent detection of the operation data of the device, and marking the operation data of the device detected by the detection terminal by utilizing a corresponding association mark of the device as association data;
s30: the method comprises the steps that a set video monitoring unit is utilized to intermittently acquire operation images of each device in a power plant, the operation images are compared with standard images of the devices in normal operation by using a convolutional neural network algorithm, and if difference exists in comparison results, a corresponding device attention mark is given;
s40: comparing the operation data detected by the detection terminal to the standard data range of the normal operation of each device, marking the operation data of the corresponding device as abnormal data if the two are different, and endowing the device with an abnormal mark;
s50: and checking the abnormal data by using a checking method, if the abnormal data is checked without errors, assigning inspection personnel to inspect the equipment for detecting the abnormal data, and if the abnormal data is checked with errors, waiting for further judgment of the next data detection.
Further, the association degree marking method comprises the following steps:
s11: grouping a plurality of devices according to the specific functions of the devices in the power plant as classification standards, and endowing the devices with the same function with the same level association degree mark;
s12: sequentially marking each device according to the power generation sequence of the power plant, and respectively endowing corresponding devices with corresponding logic association degree marks according to the functional logic relation between adjacent devices on the basis of the sequential marking;
the peer association degree marks and the logic association degree marks belong to parallel relations, and can reflect the association degree between a single device and other devices at the same time;
in the equipment with the same-level relevance marks, equipment operation data detected by the corresponding detection terminals belong to the same type of data, and can be directly compared; and the equipment operation data detected by the corresponding detection terminal are different types of data among the equipment with the logic association degree marks, and the two types of data can be compared by adopting a set proportion or an increase relation.
Further, the method of screening in step 40 includes the steps of:
s41: judging whether equipment corresponding to a plurality of normal data has a attention mark or not in the plurality of normal data within the standard data range;
s42: if the attention mark is provided, the normal data is also defined as abnormal data. .
Further, step 40 further includes a sorting method for sorting the checking sequence of the devices detecting the abnormal data, including the following steps:
s43: defining a data difference value between the abnormal data and the standard data range as a difference value, judging the percentage between the difference value and the standard data range, and arranging a plurality of abnormal data in the sequence from big to small according to the percentage to form an initial abnormal data arrangement array;
s44: if the arranged initial abnormal data arrangement arrays have the marks with the same level association degree at adjacent positions, rejecting the abnormal data with small percentage in the initial abnormal data arrangement arrays, retaining the abnormal data with large percentage to obtain the abnormal data arrangement arrays, finding out corresponding equipment, and realizing the checking sequence arrangement of the abnormal marking equipment to obtain a checking sequence.
Further, step S44 further includes:
s441: a device having a attention marker is found in the collation sequence;
s442: one or more devices with attention markers are arranged at the head end of the collation sequence.
Further, the collation method includes the steps of:
s51: firstly judging whether the abnormal data have the same level association degree mark or not, if so, comparing a plurality of abnormal data with the same level association degree mark, and if the plurality of abnormal data are higher than or lower than the standard data range, judging that the abnormal data with the same level association degree mark in the plurality of abnormal data are first reasonable abnormal data;
s52: judging whether the abnormal data have the same logic association degree marks or not, if so, comparing the abnormal data with the same logic association degree marks, and if the abnormal data are matched with the functional logic relationship of the corresponding equipment, judging that the abnormal data with the same logic association degree marks are reorganized into second reasonable abnormal data;
s53: if the first reasonable abnormal data and the second reasonable abnormal data are overlapped, the plurality of abnormal data overlapped by default are third reasonable abnormal data, and if the first reasonable abnormal data and the second reasonable abnormal data are not overlapped, the non-overlapped abnormal data are defined as unreasonable abnormal data, and the detection of the next data needs to be waited for further judgment.
Advantageous effects
Compared with the known public technology, the technical scheme provided by the invention has the following beneficial effects:
the method marks the equipment to be detected by the association marking method, realizes efficient checking of abnormal data by matching with the checking method, ensures the authenticity of data acquisition and improves efficient inspection of the operation of power plant equipment.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It is evident that the drawings in the following description are only some embodiments of the present invention and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art.
FIG. 1 is a schematic flow chart of a detection method of the present invention;
FIG. 2 is a flow chart of the sorting method according to the present invention;
FIG. 3 is a schematic flow chart of the checking method of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention is further described below with reference to examples.
Examples:
a convolution neural network algorithm-based intelligent power plant abnormality detection method comprises the following steps:
s10: marking each device in the power plant by using a relevance marking method, and endowing each device in the power plant with a corresponding relevance mark which reflects the logical relation of the running states of the devices;
specifically, the association degree marking method comprises the following steps:
s11: grouping a plurality of devices according to the specific functions of the devices in the power plant as classification standards, and endowing the devices with the same function with the same level association degree mark;
s12: sequentially marking each device according to the power generation sequence of the power plant, and respectively endowing corresponding devices with corresponding logic association degree marks according to the functional logic relation between adjacent devices on the basis of the sequential marking;
the peer association degree marks and the logic association degree marks belong to parallel relations, and can reflect the association degree between a single device and other devices at the same time;
in the equipment with the same-level relevance marks, equipment operation data detected by the corresponding detection terminals belong to the same type of data, and can be directly compared; and the equipment operation data detected by the corresponding detection terminal are different types of data among the equipment with the logic association degree marks, and the two types of data can be compared by adopting a set proportion or an increase relation.
In this case, the flag for the degree of association between the devices is to be used for uniformly comparing the data having a certain relationship when detecting the operation data of the corresponding devices, where the certain relationship may be understood as the same kind of relationship or may be a logical relationship, and is specifically:
for example, in the actual operation process of the power plant, the generated power is required to be transformed step by step, and at least one transformer is adopted to realize transformation, so that in the process, a plurality of transformers can be understood to be in a same type of relation, and correspondingly in the detection method, when the detection terminal detects the transformers, the operation data of the plurality of transformers obtained through detection are endowed with the same relevance mark, specifically the same-level relevance mark; further, the transformer may have a grade division due to the difference of basic parameters, for example, the voltage generated by the generator of the power plant is 10kv at the beginning (of course, the voltage actually generated by each power plant is related to the generator set), which is not specifically described in this case, when the voltage generated by the power plant is 10kv, the plurality of primary transformers will reduce the 10kv to 5kv, and the plurality of secondary transformers will reduce the voltage of 5kv to 1kv until the voltage can be reduced to the input of the commercial power; then in the above device, multiple primary transformers are assigned the same level association label, while primary and secondary transformers are assigned the same logical association label;
aiming at the primary transformer and the secondary transformer, certain relation exists in logic, specifically, for example, the data type detected by the detection terminal is the equipment temperature, then in the actual operation process, certain difference exists between the operation temperatures of the primary transformer and the secondary transformer, and under the normal condition, the difference can approximately judge the accuracy of the detected operation data by using a growing relation;
or in the actual thermal power generation process, the corresponding detection terminal of the boiler equipment can reflect the running state of the boiler in a mode of detecting the temperature of the equipment; in the subsequent electrical processing link, in order to determine whether the transformer is normally operated, the detection terminal can detect the current in the transformer to determine whether the transformer is normally operated, then the temperature data and the voltage data belong to different types of data and cannot be directly compared, so that when the situation is met, the two data can be judged to be reasonable or not by adopting a preset proportion.
S20: configuring a corresponding detection terminal for each device in the power plant, realizing intermittent detection of the operation data of the device, and marking the operation data of the device detected by the detection terminal by utilizing a corresponding association mark of the device as association data;
specifically, in the case, the abnormality detection method adopts intermittent detection of operation data of each device, which is understood to be that collection of operation data of all devices is realized at the same time, and then collection of operation data of all devices is performed again after a period of time; and the operation data of the equipment detected by the detection terminal is correspondingly marked according to the corresponding association degree mark on the corresponding equipment, namely the association degree data, so that the subsequent comparison and verification are facilitated.
S30: the method comprises the steps that a set video monitoring unit is utilized to intermittently acquire operation images of each device in a power plant, the operation images are compared with standard images of the devices in normal operation by using a convolutional neural network algorithm, and if difference exists in comparison results, a corresponding device attention mark is given;
in the scheme, a video monitoring unit adopts a camera mode and is consistent with the time point of detecting equipment operation data acquisition by a detection terminal, so that the image acquisition during the equipment operation is completed, the standard image is compared, and the operation image is compared with the standard image, wherein the comparison direction is mainly two characteristics of the appearance size and the color of the equipment;
specifically, regarding the external dimensions, with respect to the cable used for transmitting power, when the voltage or current transmitted by the cable exceeds the rated parameter thereof, the insulating layer outside the cable is deformed, and the deformed position is always present at the joint, when the cable is deformed, the running image acquired by the video monitoring unit is different from the standard image, and the corresponding device with which the cable is connected is given attention mark; regarding color comparison, when the temperature of the cable is abnormal in the transmission process, an insulating layer outside the cable is blackened, so that the difference between the insulating layer and the original color of the cable in the standard image occurs, and at the moment, corresponding equipment which is communicated with the cable and is kept connected with the cable is endowed with attention marks;
it should be noted that, in this case, the relevant content of extracting features from the acquired image by using the convolutional neural network algorithm is already a common application, and will not be described in detail in this case.
S40: comparing the operation data detected by the detection terminal to the standard data range of the normal operation of each device, marking the operation data of the corresponding device as abnormal data if the two are different, and endowing the device with an abnormal mark;
a screening method further included in step 40 includes the steps of:
s41: judging whether equipment corresponding to a plurality of normal data has a attention mark or not in the plurality of normal data within the standard data range;
s42: if the attention mark is provided, the normal data is also defined as abnormal data.
Step S40, comparing the collected operation data of each equipment with the standard data range to judge whether the operation of the current equipment is normal, and if the detected operation data is not in the standard data range, marking the corresponding operation data as abnormal data;
however, it should be noted that, for the operation data within the standard data range, the corresponding device may also have an abnormality, specifically, for example, as mentioned above, when the cable is subjected to a certain high temperature, it is deformed, but the deformation is within a certain range, and then the cable can still be normally used, so in this case, the operation data of the corresponding detected device is accurate, but the continuous use is risky, so in this case, the attention of the device is marked for checking whether there is: if the abnormal condition exists in the equipment but the detected operation data is normal, the abnormal condition can still be marked as abnormal data even if the operation data of the equipment currently detected belongs to the standard data range, so that the high-precision operation state detection of the equipment in the power plant is realized.
After the above-mentioned complete data acquisition has been clarified, statistics of all abnormal data is clear, then to the multiple abnormal data that gathers, it needs to carry out subsequent check one by one, but to some equipment, if do not make the adjustment in time in a certain time, it can appear the condition of damaging and lead to the electricity generation suspension of power plant, therefore when checking a plurality of abnormal data, in this scheme, still can carry out the order of checking according to the importance degree of a plurality of abnormal data, check to the abnormal data of important equipment preferentially, and the priority of minor equipment can be arranged backward, guarantee the scientific and reasonable that abnormal data checked.
Specifically, step 40 further includes a sorting method for sorting the checking sequence of the devices detecting the abnormal data, including the following steps:
s43: defining a data difference value between the abnormal data and the standard data range as a difference value, judging the percentage between the difference value and the standard data range, and arranging a plurality of abnormal data in the sequence from big to small according to the percentage to form an initial abnormal data arrangement array;
s44: if the arranged initial abnormal data arrangement arrays have the marks with the same level association degree at adjacent positions, rejecting the abnormal data with small percentage in the initial abnormal data arrangement arrays, retaining the abnormal data with large percentage to obtain the abnormal data arrangement arrays, finding out corresponding equipment, and realizing the checking sequence arrangement of the abnormal marking equipment to obtain a checking sequence.
The synchronization step S43-S44 can reasonably sort the checking sequence of the plurality of abnormal data, and firstly, judge the degree of abnormality of all the abnormal data;
for example, the abnormal data is abnormal data of a temperature type, and the abnormal data corresponding to the a device, the B device and the C device are respectively: 55 ℃, 60 ℃ and 70 ℃; the standard temperature range is 40+/-2 ℃; the corresponding difference data are respectively: 15 ℃, 20 ℃ and 30 ℃; the corresponding percentages are: 37.5%, 50% and 75%, respectively, the ordering of the three devices is: c equipment, B equipment and A equipment; when checking, the equipment C is checked preferentially, then the equipment B is checked, and finally the equipment A is checked;
specific percentages are obtained for different data types, and the data types are arranged in sequence from large to small; however, for abnormal data with the same level of association, the percentages can have slight differences, but the abnormal data usually appear at the same position, in which case, only the abnormal data with the maximum percentages need to be reserved, and the corresponding abnormal marking equipment can also be found;
specifically, if the device a, the device B, and the device C are a plurality of transformers at the same stage, respectively, and the three have the same level association degree mark, the calculated corresponding percentages are as follows: 16%, 15.5% and 15%, and in the actual initial abnormal data arrangement array, the above three percentages are located at adjacent positions, and only 16% of the abnormal data with the same level association degree mark is reserved for subsequent verification by adopting the mode of S44, which is understood that only one most abnormal data needs to be reserved, when the abnormal data is verified to be abnormal, the subsequent data does not need to be verified, and the probability is also problematic.
The step S44 further includes:
s441: a device having a attention marker is found in the collation sequence;
s442: one or more devices with attention markers are arranged at the head end of the collation sequence.
The above-mentioned already obtained check sequence, on this basis, also can find out the corresponding apparatus in the check sequence, judge this apparatus has attention degree mark, there is attention degree mark, indicate this apparatus has already appeared the problem in the physical level, for example the expansion of the cable, take place deformation, in this case, will further raise the check priority of this some apparatuses, check as soon as possible, if there is abnormality, need to arrange the personnel of patrolling and examining in time to deal with, reduce losses.
S50: and checking the abnormal data by using a checking method, if the abnormal data is checked without errors, assigning inspection personnel to inspect the equipment for detecting the abnormal data, and if the abnormal data is checked with errors, waiting for further judgment of the next data detection.
The checking method comprises the following steps:
s51: firstly judging whether the abnormal data have the same level association degree mark or not, if so, comparing a plurality of abnormal data with the same level association degree mark, and if the plurality of abnormal data are higher than or lower than the standard data range, judging that the abnormal data with the same level association degree mark in the plurality of abnormal data are first reasonable abnormal data;
s52: judging whether the abnormal data have the same logic association degree marks or not, if so, comparing the abnormal data with the same logic association degree marks, and if the abnormal data are matched with the functional logic relationship of the corresponding equipment, judging that the abnormal data with the same logic association degree marks are reorganized into second reasonable abnormal data;
s53: if the first reasonable abnormal data and the second reasonable abnormal data are overlapped, the plurality of abnormal data overlapped by default are third reasonable abnormal data, and if the first reasonable abnormal data and the second reasonable abnormal data are not overlapped, the non-overlapped abnormal data are defined as unreasonable abnormal data, and the detection of the next data needs to be waited for further judgment.
Step S40-S442, the checking sequence of the abnormal data can be reasonably and continuously arranged, a plurality of abnormal data needs to be checked later, and when the checking finds that the abnormal data is accurate, the equipment has a problem and needs to be overhauled;
specifically, in the checking method, firstly, statistics is uniformly performed on a plurality of abnormal data, specifically, step S51 may be understood that, for example, the transformers a, B and C in the same stage have the same level association degree marks, in the checking sequence, the transformers a, B and C are not adjacent to each other (if the transformers are adjacent, the transformers are simplified by step S44), the types of operation data detection are all currents, and the corresponding currents detected by the transformers a, B and C are all greater than the standard voltage range, at this time, the corresponding three sets of current abnormal data are considered to be the first reasonable abnormal data, but if the current data corresponding to the transformer B is lower than the standard voltage range, but if the current data corresponding to the transformer a and the transformer C is higher than the standard voltage range, the current data corresponding to the transformer B is not the first reasonable abnormal data, and further judgment is required to wait for the detection of the next data;
the corresponding S52 corresponds to the case that the collected data is at least three groups of data with different types, and the three groups of data have the same logical association degree mark, and the three groups of data are: group a is abnormal data representing the temperature of the boiler; group b is abnormal data representing the rotational speed of the rotating blades; group c is abnormal data representing the amount of emissions; the embodiment on the abnormal data is as follows: the data of the group a and the data of the group b are both lower than the standard data range, but the data of the group c is higher than the standard data range, and the specific conditions are that the temperature of a boiler is lower, the rotating speed of a rotating blade is lower, but the generating capacity is higher than the standard data range, so that the judgment can be carried out, the data of the group c is unreasonable as a whole, and the data of the group c is not the second reasonable abnormal data, and the detection of the next data is required to be further judged;
through steps S51-S52, when the remaining data have coincidence, the default coincidence of multiple abnormal data is third reasonable abnormal data, the part of data is truly abnormal, the inspection personnel needs to be commissioned to examine, and for the abnormal data without coincidence, we consider that the data may have data acquisition errors, and need to wait for further judgment of the next detection of the data.
If the feedback is wrong for multiple data acquisition of a certain device, whether the corresponding detection terminal has a problem or not needs to be considered, so that the detection terminal can be replaced in time, and the data acquisition is ensured to be reasonable.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; these modifications or substitutions do not depart from the essence of the corresponding technical solutions from the protection scope of the technical solutions of the embodiments of the present invention.
Claims (4)
1. The intelligent power plant abnormality detection method based on the convolutional neural network algorithm is characterized by comprising the following steps of:
s10: marking each device in the power plant by using a relevance marking method, and endowing each device in the power plant with a corresponding relevance mark which reflects the logical relation of the running states of the devices;
s20: configuring a corresponding detection terminal for each device in the power plant, realizing intermittent detection of operation data of each device in the power plant, and marking the operation data of the devices detected by the detection terminal by using the corresponding association mark of the devices as association data;
s30: the method comprises the steps that a set video monitoring unit is utilized to intermittently acquire operation images of each device in a power plant, the operation images are compared with standard images of the devices in normal operation by using a convolutional neural network algorithm, and if difference exists in comparison results, a corresponding device attention mark is given;
s40: comparing the operation data detected by the detection terminal to the standard data range of the normal operation of each device, marking the operation data of the corresponding device as abnormal data if the two are different, and endowing the device with an abnormal mark;
s50: checking the abnormal data by using a checking method, if the abnormal data is checked without errors, assigning inspection personnel to inspect the equipment for detecting the abnormal data, and if the abnormal data is checked with errors, waiting for further judgment of the detection of the next data;
the association degree marking method comprises the following steps:
s11: grouping a plurality of devices according to the specific functions of the devices in the power plant as classification standards, and endowing the devices with the same function with the same level association degree mark;
s12: sequentially marking each device according to the power generation sequence of the power plant, and respectively endowing corresponding devices with corresponding logic association degree marks according to the functional logic relation between adjacent devices on the basis of the sequential marking;
the peer association degree marks and the logic association degree marks belong to parallel relations, and can reflect the association degree between a single device and other devices at the same time;
in the equipment with the same-level relevance marks, equipment operation data detected by the corresponding detection terminals belong to the same type of data, and can be directly compared; the equipment operation data detected by the corresponding detection terminal are different types of data, and the two types of data can be compared by adopting a set proportion or an increase relation;
the checking method comprises the following steps:
s51: firstly judging whether the abnormal data have the same level association degree mark or not, if so, comparing a plurality of abnormal data with the same level association degree mark, and if the plurality of abnormal data are higher than or lower than the standard data range, judging that the abnormal data with the same level association degree mark in the plurality of abnormal data are first reasonable abnormal data;
s52: judging whether the abnormal data have the same logic association degree marks or not, if so, comparing the abnormal data with the same logic association degree marks, and if the abnormal data are matched with the functional logic relationship of the corresponding equipment, judging that the abnormal data with the same logic association degree marks are reorganized into second reasonable abnormal data;
s53: if the first reasonable abnormal data and the second reasonable abnormal data are overlapped, the plurality of abnormal data overlapped by default are third reasonable abnormal data, and if the first reasonable abnormal data and the second reasonable abnormal data are not overlapped, the non-overlapped abnormal data are defined as unreasonable abnormal data, and the detection of the next data needs to be waited for further judgment.
2. The method for detecting abnormal conditions in a smart power plant based on a convolutional neural network algorithm according to claim 1, wherein the step 40 further comprises a screening method comprising the steps of:
s41: judging whether equipment corresponding to a plurality of normal data has a attention mark or not in the plurality of normal data within the standard data range;
s42: if the attention mark is provided, the normal data is also defined as abnormal data.
3. The method for detecting abnormal conditions in a smart power plant based on a convolutional neural network algorithm according to claim 2, wherein the step 40 further comprises a sorting method for sorting the checking sequence of the devices detecting abnormal data, comprising the steps of:
s43: defining a data difference value between the abnormal data and the standard data range as a difference value, judging the percentage between the difference value and the standard data range, and arranging a plurality of abnormal data according to the sequence from large to small by using the percentage between the difference value and the standard data range to form an initial abnormal data arrangement array;
s44: if the arranged initial abnormal data arrangement arrays have the marks with the same level association degree at adjacent positions, rejecting the abnormal data with small percentage in the initial abnormal data arrangement arrays, retaining the abnormal data with large percentage to obtain the abnormal data arrangement arrays, finding out corresponding equipment, and realizing the checking sequence arrangement of the abnormal marking equipment to obtain a checking sequence.
4. The intelligent power plant abnormality detection method based on the convolutional neural network algorithm according to claim 3, wherein after step S44, further comprises:
s441: a device having a attention marker is found in the collation sequence;
s442: one or more devices with attention markers are arranged at the head end of the collation sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311587541.4A CN117390566B (en) | 2023-11-27 | 2023-11-27 | Intelligent power plant abnormality detection method based on convolutional neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311587541.4A CN117390566B (en) | 2023-11-27 | 2023-11-27 | Intelligent power plant abnormality detection method based on convolutional neural network algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117390566A CN117390566A (en) | 2024-01-12 |
| CN117390566B true CN117390566B (en) | 2024-03-29 |
Family
ID=89470225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311587541.4A Active CN117390566B (en) | 2023-11-27 | 2023-11-27 | Intelligent power plant abnormality detection method based on convolutional neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117390566B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120071A (en) * | 2018-10-29 | 2019-01-01 | 国家电网有限公司 | It draws water electric power storage station equipment automatic tour inspection system and automatic detecting method |
| CN110135273A (en) * | 2019-04-19 | 2019-08-16 | 中铁第一勘察设计院集团有限公司 | Contact net video image cloud intellectual monitoring and fault recognition method |
| WO2021208875A1 (en) * | 2020-04-17 | 2021-10-21 | 华为技术有限公司 | Visual detection method and visual detection apparatus |
| CN115293508A (en) * | 2022-07-05 | 2022-11-04 | 国网江苏省电力有限公司南通市通州区供电分公司 | Visual optical cable running state monitoring method and system |
| CN115393118A (en) * | 2022-07-15 | 2022-11-25 | 国网四川电力送变电建设有限公司 | Secondary system fault positioning method and system based on monitoring data analysis |
| CN116205226A (en) * | 2023-03-09 | 2023-06-02 | 国网陕西省电力有限公司电力科学研究院 | Power equipment state evaluation method, device, equipment and readable storage medium |
| CN116300724A (en) * | 2023-02-17 | 2023-06-23 | 南京鑫起点智能制造技术有限公司 | Digital management system for intelligent factory production |
| CN116631087A (en) * | 2023-07-20 | 2023-08-22 | 厦门闽投科技服务有限公司 | Unmanned aerial vehicle-based electric power inspection system |
| CN116699329A (en) * | 2023-06-05 | 2023-09-05 | 国网江苏省电力有限公司南通供电分公司 | Substation space voiceprint visual imaging method |
| CN117037059A (en) * | 2023-08-07 | 2023-11-10 | 上海控创信息技术股份有限公司 | Equipment management method and device based on inspection monitoring and electronic equipment |
| CN117115937A (en) * | 2023-10-20 | 2023-11-24 | 湖南半岛医疗科技有限公司 | Equipment running state monitoring method and device, cloud equipment and storage medium |
-
2023
- 2023-11-27 CN CN202311587541.4A patent/CN117390566B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109120071A (en) * | 2018-10-29 | 2019-01-01 | 国家电网有限公司 | It draws water electric power storage station equipment automatic tour inspection system and automatic detecting method |
| CN110135273A (en) * | 2019-04-19 | 2019-08-16 | 中铁第一勘察设计院集团有限公司 | Contact net video image cloud intellectual monitoring and fault recognition method |
| WO2021208875A1 (en) * | 2020-04-17 | 2021-10-21 | 华为技术有限公司 | Visual detection method and visual detection apparatus |
| CN115293508A (en) * | 2022-07-05 | 2022-11-04 | 国网江苏省电力有限公司南通市通州区供电分公司 | Visual optical cable running state monitoring method and system |
| CN115393118A (en) * | 2022-07-15 | 2022-11-25 | 国网四川电力送变电建设有限公司 | Secondary system fault positioning method and system based on monitoring data analysis |
| CN116300724A (en) * | 2023-02-17 | 2023-06-23 | 南京鑫起点智能制造技术有限公司 | Digital management system for intelligent factory production |
| CN116205226A (en) * | 2023-03-09 | 2023-06-02 | 国网陕西省电力有限公司电力科学研究院 | Power equipment state evaluation method, device, equipment and readable storage medium |
| CN116699329A (en) * | 2023-06-05 | 2023-09-05 | 国网江苏省电力有限公司南通供电分公司 | Substation space voiceprint visual imaging method |
| CN116631087A (en) * | 2023-07-20 | 2023-08-22 | 厦门闽投科技服务有限公司 | Unmanned aerial vehicle-based electric power inspection system |
| CN117037059A (en) * | 2023-08-07 | 2023-11-10 | 上海控创信息技术股份有限公司 | Equipment management method and device based on inspection monitoring and electronic equipment |
| CN117115937A (en) * | 2023-10-20 | 2023-11-24 | 湖南半岛医疗科技有限公司 | Equipment running state monitoring method and device, cloud equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 一种基于多源事件流事件关联的设备异常检测方法;曹云梦;《全国优秀硕士学位论文全文数据库》;20190715(第第7期期);C042-376 * |
| 基于深度学习的电厂设备运行参数异常检测;张鹏锋;《全国优秀硕士学位论文全文数据库》;20220515;C042-692 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117390566A (en) | 2024-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017148336A1 (en) | Battery pack string fault identification method in photoelectric system, apparatus and device | |
| CN103616579B (en) | Fault diagnosis method used for intelligent substation secondary system | |
| CN107395121B (en) | Based on Grubbs test method and outlier detection photovoltaic array fault detection method | |
| CN108879956B (en) | Method for actively judging and repairing system fault based on equipment running state | |
| CN103278244A (en) | Monitoring method and monitoring system for overheat fault of transformer | |
| WO2022057555A1 (en) | Fault detection method and apparatus, and photovoltaic power generation system | |
| CN112269812A (en) | Intelligent power distribution network safety monitoring management system based on big data | |
| CN113485432A (en) | Photovoltaic power station electroluminescence intelligent diagnosis system and method based on unmanned aerial vehicle | |
| CN117390566B (en) | Intelligent power plant abnormality detection method based on convolutional neural network algorithm | |
| CN110932406B (en) | Remote signaling information checking method and device for intelligent substation scheduling master station | |
| CN116743079A (en) | Photovoltaic string fault processing method and device, photovoltaic management system and medium | |
| CN113054906B (en) | Fault determination method and device for photovoltaic power station, processor and photovoltaic system | |
| CN113884787A (en) | Intelligent oscillograph tripping and closing verification function test method and system | |
| CN116975766A (en) | Power distribution network fault self-healing method and system based on machine learning | |
| CN116089790A (en) | Calculation method and device for generating capacity loss of photovoltaic module and electronic equipment | |
| CN206292339U (en) | A kind of fuse switch monitoring device based on image recognition | |
| CN113743626A (en) | Primary equipment fault processing method and equipment | |
| LU504189B1 (en) | A system for monitoring the surface temperature of photovoltaic modules | |
| CN116707445B (en) | Photovoltaic module fault positioning method and system | |
| CN217279619U (en) | Main network equipment defect identification system | |
| CN114915022B (en) | Wireless communication method and system for intelligent power distribution network | |
| CN211236173U (en) | Detection equipment for fan power panel of frequency converter of wind turbine generator | |
| CN220543085U (en) | Transformer winding deformation detection system based on Internet of things | |
| CN115877275A (en) | Self-adaptive disconnection-free relay protection outlet matrix testing device and method | |
| Xu et al. | Diagnosis of Substation Equipment Operating Status Based on Target Detection and Information Extraction Technique in Infrared Image |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |