CN118392255A - Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant - Google Patents
Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant Download PDFInfo
- Publication number
- CN118392255A CN118392255A CN202410864625.6A CN202410864625A CN118392255A CN 118392255 A CN118392255 A CN 118392255A CN 202410864625 A CN202410864625 A CN 202410864625A CN 118392255 A CN118392255 A CN 118392255A
- Authority
- CN
- China
- Prior art keywords
- model
- anomaly
- abnormal
- recognition
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 76
- 238000012544 monitoring process Methods 0.000 title claims abstract description 73
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 46
- 230000002159 abnormal effect Effects 0.000 claims description 133
- 230000005856 abnormality Effects 0.000 claims description 76
- 238000012549 training Methods 0.000 claims description 53
- 230000008569 process Effects 0.000 claims description 43
- 238000010276 construction Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 5
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 238000005273 aeration Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000010865 sewage Substances 0.000 description 1
- 238000012795 verification 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a monitoring method, a system and equipment for the operation condition of electromechanical equipment in a water plant, which belong to the technical field of the operation monitoring of the electromechanical equipment.
Description
Technical Field
The invention belongs to the technical field of operation monitoring of electromechanical equipment, and particularly relates to a method, a system and equipment for monitoring operation conditions of electromechanical equipment of a water plant.
Background
In an industrial environment, for example, monitoring the operation of electromechanical devices (mainly water pumps and aeration fans) in water plants has an important role for the safe operation of the devices. The current mode for monitoring the operation condition of the electromechanical equipment mainly comprises the following steps: (1) The method comprises the steps of collecting relevant operation data by using a sensor, and then judging a threshold value based on the collected operation data, so as to judge whether the equipment has an abnormality or not; (2) Based on a large amount of historical data, the abnormal equipment is identified by combining an artificial intelligence algorithm, for example, a neural network or other machine learning algorithms are used, and the working condition parameters of the electromechanical equipment comprise current, voltage, speed, temperature and other historical data to learn, judge and identify whether the equipment is in a normal running state or not.
However, the above prior art has the following drawbacks: the abnormal condition is simply identified by utilizing the numerical parameters, namely the data transmitted by the sensor is used for monitoring and judging the abnormal condition, and the mode is easy to misjudge, because the current, the voltage and the like are not decisive factors for judging the abnormality for electromechanical equipment such as water pumps or fans in water plants, and the abnormal identification accuracy is not high due to misjudgment caused by the abnormal data and the like caused by the sensor.
Disclosure of Invention
In order to solve the problem of low abnormal recognition accuracy in the prior art, the invention provides a monitoring method, a system and equipment for the operation condition of electromechanical equipment of a water plant.
The invention is realized by the following technical scheme:
A method for monitoring the operating condition of electromechanical equipment in a water mill, the method comprising:
respectively establishing respective single-source abnormal recognition models aiming at each monitoring parameter of the electromechanical equipment; wherein the monitoring parameters comprise numerical monitoring parameters and non-numerical monitoring parameters;
Based on the identification results of the single-source abnormal identification models and the corresponding weight coefficients, constructing and initializing a comprehensive abnormal identification model, thereby obtaining an equipment abnormal identification model consisting of the single-source abnormal identification model and the comprehensive abnormal identification model;
Training the equipment anomaly recognition model by using a training set, and iteratively adjusting each weight coefficient in the equipment anomaly recognition model according to the actual anomaly condition of the electromechanical equipment until the recognition accuracy requirement is met; the training set is constructed by historical measurement data of all monitoring parameters of the electromechanical equipment and marks thereof, the marks are determined by checking actual abnormal conditions of the electromechanical equipment, and the marks comprise abnormal conditions and normal conditions;
and carrying out electromechanical equipment abnormality recognition by using the trained equipment abnormality recognition model.
In some embodiments, the iterative adjustment process of the weight coefficient specifically includes:
when the recognition result of the comprehensive anomaly recognition model is anomaly, if the recognition result is anomaly, increasing the weight of the anomaly item in the comprehensive anomaly recognition model, and if the recognition result is normal, decreasing the weight of the anomaly item in the comprehensive anomaly recognition model;
When the recognition result of the comprehensive anomaly recognition model is normal, if the recognition result is marked as anomaly, increasing the weight of the item with correct anomaly recognition in the comprehensive anomaly recognition model, and simultaneously reducing the weight of the item with incorrect anomaly recognition in the comprehensive anomaly recognition model, and if the recognition result is marked as normal, keeping the current weight unchanged.
In some embodiments, for numerical class monitoring parameters, the single source anomaly identification model construction process specifically includes:
comparing the measured value of the numerical value monitoring parameter with a corresponding threshold range, and outputting an identification result according to the comparison result;
the threshold range is determined according to the actual operation requirement and the working parameter of the electromechanical equipment;
For non-numerical class monitoring parameters, the construction process of the single-source anomaly identification model specifically comprises the following steps:
Acquiring waveforms of a plurality of electromechanical devices in normal operation according to a preset period, respectively converting the waveforms into a data matrix, marking the data matrix as a normal model, and constructing an initial model library;
Acquiring waveforms of the electromechanical equipment in operation according to a preset period as waveforms to be identified;
and converting the waveform to be identified into a data matrix to be identified, carrying out similarity calculation on the data matrix to be identified and normal models in the model library one by one, and outputting an identification result according to a similarity calculation result.
In some implementations, for the numeric class parameter, the model training process further includes:
Checking the identification result of the single-source abnormal identification model according to the actual abnormal condition of the electromechanical equipment, and updating the threshold range of the single-source abnormal identification model if the identification result does not accord with the actual abnormal condition of the electromechanical equipment;
for non-numeric class parameters, the model training process further includes:
and checking the identification result of the single-source abnormal identification model according to the actual abnormal condition of the electromechanical equipment, and updating the model library of the single-source abnormal identification model if the identification result does not accord with the actual abnormal condition of the electromechanical equipment.
In some embodiments, the comprehensive anomaly identification model building process specifically includes:
establishing an exception function, wherein the exception function is expressed as: the anomaly value is equal to the sum of products of the single-source anomaly identification results and the corresponding weight coefficients divided by the sum of the weight coefficients corresponding to the single-source anomaly identification results;
Outputting a final recognition result according to the comparison result of the abnormal value and the threshold value;
The process for initializing the comprehensive anomaly identification model specifically comprises the following steps:
All weight coefficients are initialized to be the same.
In some embodiments, the method further comprises:
and dynamically adjusting each weight coefficient in the equipment abnormality recognition model in the electromechanical equipment abnormality recognition process by using the equipment abnormality recognition model.
In some embodiments, the method further comprises:
And carrying out electromechanical equipment abnormality early warning by adopting a trained equipment abnormality early warning model.
In some embodiments, the device anomaly early warning model training process is:
constructing a time sequence model and a training sample, wherein the training sample is constructed by a time sequence formed by each single-source abnormal recognition model and a historical recognition result output by comprehensive abnormal recognition and an actual abnormal mark of electromechanical equipment of the time sequence;
and training the time sequence model by using the training sample to obtain an equipment abnormality early warning model.
In a second aspect, the present invention provides a system for monitoring the operation condition of electromechanical equipment in a water plant, the system comprising:
The system comprises a single-source abnormal model construction module, a single-source abnormal model analysis module and a single-source abnormal model analysis module, wherein the single-source abnormal model construction module respectively establishes respective single-source abnormal recognition models for all monitoring parameters of the electromechanical equipment; wherein the monitoring parameters comprise numerical monitoring parameters and non-numerical monitoring parameters;
The comprehensive anomaly model construction module is used for constructing and initializing a comprehensive anomaly identification model based on the identification result of each single-source anomaly identification model and the corresponding weight coefficient, so as to obtain an equipment anomaly identification model consisting of a unit anomaly identification model and the comprehensive anomaly identification model;
The model training module is used for training the equipment anomaly identification model by utilizing a training set, and iteratively adjusting each weight coefficient in the equipment anomaly identification model according to the actual anomaly condition of the electromechanical equipment until the identification accuracy requirement is met; the training set is constructed by historical measurement data of all monitoring parameters of the electromechanical equipment and marks thereof, the marks are determined by checking actual abnormal conditions of the electromechanical equipment, and the marks comprise abnormal conditions and normal conditions;
And the equipment abnormality recognition module is used for carrying out electromechanical equipment abnormality recognition by using the trained equipment abnormality recognition model.
In a third aspect, the invention proposes a computer device comprising a memory storing a computer program and a processor implementing the steps of the method of the invention when the processor executes the computer program.
According to the monitoring method, the system and the equipment for the operation condition of the electromechanical equipment of the water plant, the single-source abnormal recognition models are respectively built for different monitoring parameters, then the comprehensive abnormal recognition model is built and initialized based on the recognition results of the single-source abnormal recognition models and the weight coefficients corresponding to the single-source abnormal recognition models, all weight coefficients are the same during initialization, the model is trained by using historical measurement data, and all weight coefficients in the model are iteratively adjusted in the process, so that the abnormal recognition precision and reliability are improved; in addition, in the model training process, not only the data set is used for training the model, but also the single-source abnormal recognition model is subjected to iterative correction in combination with the abnormal verification, so that the abnormal recognition precision is further improved.
According to the monitoring method, the system and the equipment for the operation condition of the electromechanical equipment of the water plant, which are provided by the invention, in the process of real-time monitoring by using the equipment anomaly identification model, each weight coefficient in the model can be further dynamically adjusted, so that the identification precision is further improved;
according to the monitoring method, the system and the equipment for the operation condition of the electromechanical equipment of the water plant, provided by the invention, based on the historical abnormal recognition result, the electromechanical equipment can be accurately predicted by combining the time sequence model, so that early warning is realized, and a basis is provided for the monitoring and operation and maintenance decision of the electromechanical equipment.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:
FIG. 1 is a flow chart of a method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an equipment anomaly identification model architecture constructed according to an embodiment of the present invention;
FIG. 3 is a flowchart of iterative adjustment of weight coefficients according to an embodiment of the present invention;
Fig. 4 is a system schematic block diagram of an embodiment of the present invention.
Detailed Description
For the purpose of making apparent the objects, technical solutions and advantages of the present invention, the present invention will be further described in detail with reference to the following examples and the accompanying drawings, wherein the exemplary embodiments of the present invention and the descriptions thereof are for illustrating the present invention only and are not to be construed as limiting the present invention.
Examples: in the prior art, abnormal conditions are mainly monitored and judged based on numerical value data transmitted by a sensor, and false recognition is easy to generate, because numerical value parameters such as current and voltage are not decisive factors for judging the abnormal conditions in many times, and meanwhile, the numerical value data such as current and voltage are abnormal due to the fault of the sensor, so that the problem of low recognition accuracy is caused. In view of this, this embodiment provides a monitoring method for the operation condition of the electromechanical device of the water plant, and the method provided in this embodiment firstly constructs respective single-source abnormal recognition models for different parameters, then constructs a comprehensive prediction model based on the abnormal recognition results of a plurality of single-source abnormal recognition models, and iteratively updates the weights of the abnormal factors in the comprehensive prediction model, thereby improving the recognition accuracy.
As shown in fig. 1, the method provided in this embodiment specifically includes the following steps:
And 100, respectively establishing respective single-source anomaly identification models for all monitoring parameters of the electromechanical equipment. For numerical parameters such as temperature, humidity, current, voltage and the like, constructing a corresponding single-source abnormality identification model by utilizing threshold judgment logic; for non-numerical parameters such as vibration or noise waveforms, a corresponding single-source anomaly identification model is constructed by using a similarity comparison mode.
Optionally, for the numerical class parameter, the single-source anomaly identification model construction process specifically includes: and comparing the numerical value class data acquired by the sensor with a corresponding threshold range, and outputting an identification result according to the comparison result. Specifically, if the threshold range is exceeded, the device is identified as abnormal, otherwise, the device is identified as normal. The threshold range may be selected based on actual operating requirements and operating parameters of the electromechanical device, etc.
For non-numerical parameters such as waveforms of equipment vibration, noise and the like, the construction process of the single-source anomaly identification model specifically comprises the following steps:
Step 101, acquiring waveforms of a plurality of electromechanical devices in normal operation according to a preset period, respectively converting the waveforms into a data matrix and marking the data matrix as a normal model to construct an initial model library; specifically, the preset period may be set to 30 seconds, that is, a section of waveform may be recorded every 30 seconds, and data is collected in advance, for example, the waveform is recorded every 30 seconds within 15 days, so that a plurality of normal models may be obtained;
step 102, acquiring waveforms of the electromechanical equipment in operation as waveforms to be identified according to a preset period;
Step 103, converting the waveform to be identified into a data matrix to be identified, carrying out similarity calculation on the data matrix to be identified and normal models in a model library one by one, and outputting an identification result according to a similarity calculation result. Specifically, if the matching degree is greater than the upper threshold, the matching degree is identified as normal, if the matching degree is less than the lower threshold, the matching degree is identified as abnormal, otherwise, the matching degree is identified as suspected abnormal; specifically, in this embodiment, a similarity calculation method may be used to perform matching, that is, calculate the similarity between the data matrix of the waveform to be identified and the normal model in the model library, and identify the waveform as normal if the similarity is greater than 95% (which may be set according to the actual situation), identify the waveform as abnormal if the similarity is less than 80% (which may be set according to the actual situation), and identify the waveform as suspected abnormal if the similarity is between 90% and 80%.
Step 200, based on the identification results of the single-source anomaly identification models and the corresponding weight coefficients, constructing and initializing a comprehensive anomaly identification model, thereby obtaining an equipment anomaly identification model composed of the single-source anomaly identification model and the comprehensive anomaly identification model.
Optionally, the process of constructing and initializing the comprehensive anomaly identification model in this embodiment specifically includes:
Step 201, constructing a comprehensive anomaly identification model based on the identification result of each single-source anomaly identification model and the corresponding weight coefficient thereof; specifically, the single-source anomaly identification results corresponding to each monitoring parameter are P 1、P2、…、Pn respectively, and the corresponding weight set to each monitoring parameter is w 1、w2、…、wn, and then the constructed comprehensive anomaly identification model is as follows:
wherein, P is an abnormal value, if P is greater than a threshold value, the comprehensive abnormal recognition model recognition result is abnormal (marked as 1), and if P is less than or equal to the threshold value, the comprehensive abnormal recognition model recognition result is normal (marked as 0); if the monitoring parameter according to the i-th item predicts that the monitoring parameter is abnormal, the corresponding single-source abnormal prediction result P i (i=1, 2, …, n) is 1, otherwise, the single-source abnormal prediction result P i is 0; n is the number of monitored parameter items of the electromechanical device.
Step 202, all weights in the comprehensive anomaly identification model are initially the same, and an initialized comprehensive anomaly identification model is obtained. That is, the weight of each recognition result (i.e., each monitoring parameter) affects the final abnormality recognition to the same extent at the time of initialization. Based on this, the obtained device abnormality recognition model is shown in fig. 2.
And 300, training the equipment anomaly recognition model by using a training set, and iteratively adjusting each weight coefficient in the anomaly recognition model according to the actual anomaly condition of the electromechanical equipment until the recognition accuracy requirement is met. The training set is constructed by historical data collected by the sensor and marks thereof, and the marks can be realized by checking the actual abnormal condition of the electromechanical equipment and are divided into abnormal (marked as 1) and normal (marked as 0).
As shown in fig. 3, the iterative adjustment process of the weight coefficient may specifically include:
When the recognition result of the comprehensive anomaly recognition model is anomaly, if the recognition result is anomaly, increasing the weight of an anomaly item in the comprehensive anomaly recognition model; if the mark is normal, reducing the weight of the abnormal item in the comprehensive abnormal identification model; through the dynamic weight adjustment mode, when the abnormality occurs next time, the probability of correctly and comprehensively identifying the abnormality is increased, and the probability of incorrectly and comprehensively identifying the abnormality is reduced, so that the misjudgment rate is reduced, and the identification precision is improved;
When the recognition result of the comprehensive anomaly recognition model is normal, if the recognition result is marked as anomaly, increasing the weight of the item with correct anomaly recognition in the comprehensive anomaly recognition model, and simultaneously reducing the weight of the item with incorrect anomaly recognition in the comprehensive anomaly recognition model; if the mark is normal, keeping the current weight unchanged; the comprehensive recognition is normal, but the electromechanical equipment is actually abnormal, namely the abnormal recognition does not respond at all, which means that the weight setting in the comprehensive abnormal recognition model is unreasonable, and the adjustment strength of the weight is required to be larger, so that the weight of the item which is correctly judged by the abnormality can be increased at the moment, and the weight of the item which is incorrectly judged by the abnormality can be reduced, thereby quickly adjusting the influence degree of each item on the abnormality prediction, realizing timely abnormality response and improving the abnormality recognition precision; and under the condition that the comprehensive identification is normal and the electromechanical equipment is also normal, the comprehensive abnormal identification model is correctly operated, and the weight adjustment is not required at the moment. By adopting the weight dynamic real-time adjustment technology, after a period of iteration, various weights in the comprehensive anomaly identification model can change, and aiming at electromechanical equipment, when anomaly occurs, the weight of the item with higher anomaly influence degree is increased, and the weight of the item with smaller anomaly influence degree is reduced, so that the model with higher anomaly identification accuracy is obtained, the accuracy and reliability of anomaly identification of the electromechanical equipment are improved, and the operation safety and reliability of the electromechanical equipment are ensured.
In this embodiment, the water pump is taken as an example to describe the comprehensive anomaly monitoring process in detail, and each monitoring parameter involved in anomaly identification includes: temperature (P 1), humidity (P 2), smoke (P 3), current (P 4), voltage (P 5), water pressure (P 6), vibration (P 7) and noise intensity (P 8) are eight in total, and these eight monitoring parameters return to 1 if it is predicted that an abnormality is present, otherwise return to 0. The initial weights are all set to 1, and the comprehensive anomaly identification model is as follows:
And (3) monitoring the operation of the water pump by using the comprehensive abnormality recognition model, and when monitoring If the number is greater than 0.5, it is considered that abnormality occurs, and if the number is less than or equal to 0.5, it is considered that abnormality (i.e., normal) does not occur.
(1) When the comprehensive recognition is abnormal and the actual abnormality of the water pump is checked, the recognition is correct, the abnormality prediction of the temperature, the current, the voltage and the vibration is 1 in the early warning process, namely, the abnormality is judged to be 0 in the rest of the abnormalities, namely, the abnormality is not judged, the weight of the abnormal item in the early warning process is improved by 3%, namely, the adjusted weight is 1.03 of the temperature, 1.03 of the current, 1.03 of the voltage and 1.03 of the vibration, and the adjusted comprehensive abnormality recognition model is as follows:
in this case, the anomaly identification value does not have a negative value, and the item weight of the occasional successful participation in the early warning has a small variation, and the item weight of the continuous successful participation many times is rapidly increased.
(2) When the comprehensive recognition is abnormal, but the water pump is checked to be actually abnormal, the recognition error is indicated, in the early warning process, the abnormality prediction of the temperature, the current, the voltage and the vibration is 1, namely the abnormality, the rest items are judged to be 0, namely the abnormality is not, the weight of the item judged to be abnormal in the early warning process is adjusted by 3%, namely the adjusted weight is 0.97, the current is 0.97, the voltage is 0.97, the vibration is 0.97, and the adjusted comprehensive abnormality recognition model is as follows:
(3) When the comprehensive recognition is normal, but the water pump is checked to be actually abnormal, the early warning is not responded, the weight setting is unreasonable, so that the change strength of the weight is increased, as described above, in the early warning process, the abnormal prediction of temperature, current, voltage and vibration is 1, namely, the abnormality is judged that the rest items are 0, namely, the abnormality is not abnormal, the weight of the items which are judged to be correct in the early warning process is increased by 10%, the weight of the items which are judged to be incorrect in the early warning process is reduced by 5%, namely, the adjusted weight is 1.1 of the temperature, 1.1 of the current, 1.1 of the voltage, 1.1 of the vibration, 0.95 of the humidity, 0.95 of the water pressure, 0.95 of the noise strength and 0.95 of the smoke, and thus, the weight of the items which are not responded to the abnormality can be changed rapidly, and the adjusted comprehensive abnormality recognition model is as follows:
(4) When the comprehensive identification is normal and the water pump is checked to be actually normal, the correct early warning is indicated, and the weight adjustment can not be carried out at the moment.
Optionally, for the numerical class parameter, the model training process may specifically further include: according to the actual abnormal feedback condition of the electromechanical equipment, checking the identification result of the single-source abnormal identification model, and if the identification result does not accord with the actual abnormal condition of the electromechanical equipment, updating the threshold range, thereby further improving the identification accuracy.
Optionally, for the non-numeric class parameter, the model training process may specifically further include:
Further checking the identification result according to the actual abnormal feedback condition of the electromechanical equipment; the checking is to avoid abnormal misjudgment, for example, if the equipment is not abnormal according to feedback of abnormal conditions, but is identified as abnormal, the equipment is checked as normal; and marking the checked waveforms to be identified according to the corresponding prediction results, and updating a model library, namely marking the normal running state as a normal model logging model library, and marking the abnormal running state as an abnormal model logging model library. Optionally, the data in the model library can also be used as training data to assist in establishing an abnormality early warning model so as to realize more accurate abnormality identification early warning.
And 400, performing electromechanical equipment abnormality recognition by using the trained equipment abnormality recognition model. Optionally, the step 400 further includes that the equipment anomaly identification model can dynamically adjust various weight coefficients in the model in real time in the actual monitoring process, so as to further improve the identification accuracy and reliability. It should be noted that, the specific adjustment manner of the weights in step 400 is described in step 300, and will not be repeated here.
Compared with the existing technology which only takes the traditional data set as the training set training model for carrying out the anomaly identification, the method provided by the embodiment firstly establishes respective single-source anomaly identification models for different monitoring parameters respectively; and then constructing and initializing a comprehensive anomaly identification model based on the identification result of each single-source anomaly identification model and the corresponding weight coefficient, wherein each weight coefficient is the same during initialization, namely each monitoring parameter has the same influence degree on the anomaly identification result, training the model by using a training set, and in the process, iteratively adjusting each weight coefficient to improve the anomaly identification precision and reliability. In addition, in the method provided by the embodiment, in the model training process, not only the data set training model is utilized, but also the anomaly checking is combined to carry out iterative correction on the single-source anomaly identification model, so that the anomaly identification precision is further improved. In the method provided by the embodiment, in the process of utilizing the equipment anomaly identification model to monitor in real time, various weight coefficients in the model can be further dynamically adjusted, and the identification precision is further improved.
Optionally, the embodiment predicts whether the electromechanical equipment in the water service scene such as the water pump, the fan and the like is abnormal based on the historical abnormal recognition result by combining the time sequence model, so that early warning is realized. Based on this, the method proposed in this embodiment further includes:
and 500, carrying out electromechanical equipment abnormality early warning by adopting a trained equipment abnormality early warning model.
Specifically, the training process of the equipment abnormality early warning model specifically comprises the following steps:
Step 501, a time series model and training samples are constructed. Wherein, the time sequence model can adopt ARIMA, LSTM and the like. The training sample is constructed by a time sequence of historical anomaly identification results (comprising a single-source anomaly identification result and a comprehensive anomaly identification result) and actual anomaly marks of the equipment.
Step 502, training the time sequence model by using a training sample, thereby obtaining an equipment abnormality early warning model.
And inputting the abnormality recognition result time sequence obtained in real time into an equipment abnormality early warning model to obtain a prediction result (abnormal or normal), thereby providing a basis for electromechanical equipment monitoring and operation and maintenance strategy formulation.
The training and application process of the equipment abnormality early warning model is further described by the water pump related data in a certain water plant.
(1) When the abnormality occurs, namely after the monitoring is completed through the equipment abnormality identification model, the water outlet pressure condition of 24 hours before the abnormality occurs, the flow condition and the pressure data of each pressure monitoring point in the pipe network hydraulic distribution are recorded. An abnormal data group is formed, and the data is a characteristic value of data whose tag is abnormal. For example, if the collection frequency is once an hour, data of one abnormality is formed. Assuming that the number of anomalies is 20, there are 20 sets of 24X anomaly data sets, as shown in table 1 below.
TABLE 1
(2) The water pump works continuously for more than two weeks, no abnormality occurs, and the water pump is regarded as a normal condition, and continuous data quantity equal to the occurrence quantity of the abnormality is extracted to form a normal data group. For example, if 20 anomalies occur in total in the history and 20 days of anomaly data before the anomalies occur are summed up, a corresponding database is formed by selecting 20 days of normal data in the normal days, and the characteristic values of the corresponding normal data are formed according to the anomaly table (that is, the data of the normal day in which no anomalies occur in a normal period of time, in general, a partial report of anomalies occurs, but the situation that the whole is gradually and automatically returned to 0 is reflected, and the anomaly values of partial points are negligible), and then 20 sets of 24×x normal data sets exist, as shown in table 2 below.
TABLE 2
(3) And analyzing characteristic values in the normal data set and the abnormal data set, namely a single-source abnormal recognition model and a comprehensive abnormal recognition model recognition result, and training by adopting a time sequence model to form an equipment abnormal early warning model, namely a data model in 24 hours before the water pump is abnormal.
(4) And inputting the real-time data into the equipment abnormality early warning model for prediction, namely, inputting the data into the corresponding data 24h before the current moment to obtain a prediction result at the current moment as normal or abnormal.
The training and application process of the equipment abnormality early warning model is further described by fan related data in a certain water plant.
First, some relevant water quality parameters in the water treatment process, such as cod, water flow, etc., are combined, as shown in table 3.
TABLE 3 Table 3
And similarly, establishing and predicting a subsequent process according to the water pump model to obtain an abnormal prediction result of the aeration fan in the sewage plant environment.
Based on the same technical concept, the embodiment also provides a monitoring system for the operation condition of the electromechanical equipment of the water plant, as shown in fig. 4, and the system provided by the embodiment specifically comprises:
The system comprises a single-source abnormal model construction module, a single-source abnormal model analysis module and a single-source abnormal model analysis module, wherein the single-source abnormal model construction module respectively establishes respective single-source abnormal recognition models for all monitoring parameters of the electromechanical equipment, and the corresponding single-source abnormal recognition models are constructed for numerical parameters such as temperature, humidity, current, voltage and the like by utilizing threshold judgment logic; for non-numerical parameters such as vibration or noise waveforms, constructing a corresponding single-source abnormal recognition model by using a similarity comparison mode;
The comprehensive anomaly model construction module is used for constructing and initializing a comprehensive anomaly identification model based on the identification result of each single-source anomaly identification model and the corresponding weight coefficient thereof, so as to obtain an equipment anomaly identification model consisting of the single-source anomaly identification model and the comprehensive anomaly identification model;
And the model training module is used for training the equipment anomaly identification model by utilizing the training set, and iteratively adjusting each weight in the comprehensive anomaly identification model according to the anomaly feedback condition of the electromechanical equipment until the identification accuracy is met.
And a device anomaly identification module that performs electromechanical device anomaly identification using the trained device anomaly identification model.
Optionally, the system provided in this embodiment further includes:
And the equipment abnormality early warning module adopts a trained equipment abnormality early warning model to carry out electromechanical equipment abnormality early warning.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A method for monitoring the operation condition of electromechanical equipment of a water plant, which is characterized by comprising the following steps:
respectively establishing respective single-source abnormal recognition models aiming at each monitoring parameter of the electromechanical equipment; wherein the monitoring parameters comprise numerical monitoring parameters and non-numerical monitoring parameters;
Based on the identification results of the single-source abnormal identification models and the corresponding weight coefficients, constructing and initializing a comprehensive abnormal identification model, thereby obtaining an equipment abnormal identification model consisting of the single-source abnormal identification model and the comprehensive abnormal identification model;
Training the equipment anomaly recognition model by using a training set, and iteratively adjusting each weight coefficient in the equipment anomaly recognition model according to the actual anomaly condition of the electromechanical equipment until the recognition accuracy requirement is met; the training set is constructed by historical measurement data of all monitoring parameters of the electromechanical equipment and marks thereof, the marks are determined by checking actual abnormal conditions of the electromechanical equipment, and the marks comprise abnormal conditions and normal conditions;
and carrying out electromechanical equipment abnormality recognition by using the trained equipment abnormality recognition model.
2. The method for monitoring the operation condition of the electromechanical equipment of the water mill according to claim 1, wherein the iterative adjustment process of the weight coefficient specifically comprises the following steps:
when the recognition result of the comprehensive anomaly recognition model is anomaly, if the recognition result is anomaly, increasing the weight of the anomaly item in the comprehensive anomaly recognition model, and if the recognition result is normal, decreasing the weight of the anomaly item in the comprehensive anomaly recognition model;
When the recognition result of the comprehensive anomaly recognition model is normal, if the recognition result is marked as anomaly, increasing the weight of the item with correct anomaly recognition in the comprehensive anomaly recognition model, and simultaneously reducing the weight of the item with incorrect anomaly recognition in the comprehensive anomaly recognition model, and if the recognition result is marked as normal, keeping the current weight unchanged.
3. The method for monitoring the operation condition of the electromechanical equipment of the water plant according to claim 1 or 2, wherein for the numerical monitoring parameters, the single-source anomaly identification model construction process specifically comprises the following steps:
comparing the measured value of the numerical value monitoring parameter with a corresponding threshold range, and outputting an identification result according to the comparison result;
the threshold range is determined according to the actual operation requirement and the working parameter of the electromechanical equipment;
For non-numerical class monitoring parameters, the construction process of the single-source anomaly identification model specifically comprises the following steps:
Acquiring waveforms of a plurality of electromechanical devices in normal operation according to a preset period, respectively converting the waveforms into a data matrix, marking the data matrix as a normal model, and constructing an initial model library;
Acquiring waveforms of the electromechanical equipment in operation according to a preset period as waveforms to be identified;
and converting the waveform to be identified into a data matrix to be identified, carrying out similarity calculation on the data matrix to be identified and normal models in the model library one by one, and outputting an identification result according to a similarity calculation result.
4. A method of monitoring the operation of an electromechanical device in a water mill according to claim 3, wherein for numerical class parameters, the model training process further comprises:
Checking the identification result of the single-source abnormal identification model according to the actual abnormal condition of the electromechanical equipment, and updating the threshold range of the single-source abnormal identification model if the identification result does not accord with the actual abnormal condition of the electromechanical equipment;
for non-numeric class parameters, the model training process further includes:
and checking the identification result of the single-source abnormal identification model according to the actual abnormal condition of the electromechanical equipment, and updating the model library of the single-source abnormal identification model if the identification result does not accord with the actual abnormal condition of the electromechanical equipment.
5. The method for monitoring the operation condition of the electromechanical equipment of the water plant according to claim 1 or 2, wherein the comprehensive anomaly identification model construction process specifically comprises the following steps:
establishing an exception function, wherein the exception function is expressed as: the anomaly value is equal to the sum of products of the single-source anomaly identification results and the corresponding weight coefficients divided by the sum of the weight coefficients corresponding to the single-source anomaly identification results;
Outputting a final recognition result according to the comparison result of the abnormal value and the threshold value;
The process for initializing the comprehensive anomaly identification model specifically comprises the following steps:
All weight coefficients are initialized to be the same.
6. A method of monitoring the operation of an electromechanical device in a water mill according to claim 1 or 2, further comprising:
and dynamically adjusting each weight coefficient in the equipment abnormality recognition model in the electromechanical equipment abnormality recognition process by using the equipment abnormality recognition model.
7. A method of monitoring the operation of an electromechanical device in a water mill according to claim 1 or 2, further comprising:
And carrying out electromechanical equipment abnormality early warning by adopting a trained equipment abnormality early warning model.
8. The method for monitoring the operation condition of the electromechanical equipment of the water plant according to claim 7, wherein the training process of the equipment abnormality early warning model is as follows:
constructing a time sequence model and a training sample, wherein the training sample is constructed by a time sequence formed by historical recognition results output by each single-source abnormal model and the comprehensive abnormal recognition model and an actual abnormal mark of electromechanical equipment of the time sequence;
and training the time sequence model by using the training sample to obtain an equipment abnormality early warning model.
9. A system for monitoring the operating condition of an electromechanical device of a water mill, the system comprising:
The system comprises a single-source abnormal model construction module, a single-source abnormal model analysis module and a single-source abnormal model analysis module, wherein the single-source abnormal model construction module respectively establishes respective single-source abnormal recognition models for all monitoring parameters of the electromechanical equipment; wherein the monitoring parameters comprise numerical monitoring parameters and non-numerical monitoring parameters;
The comprehensive anomaly model construction module is used for constructing and initializing a comprehensive anomaly identification model based on the identification result of each single-source anomaly identification model and the corresponding weight coefficient, so as to obtain an equipment anomaly identification model consisting of a unit anomaly identification model and the comprehensive anomaly identification model;
The model training module is used for training the equipment anomaly identification model by utilizing a training set, and iteratively adjusting each weight coefficient in the equipment anomaly identification model according to the actual anomaly condition of the electromechanical equipment until the identification accuracy requirement is met; the training set is constructed by historical measurement data of all monitoring parameters of the electromechanical equipment and marks thereof, the marks are determined by checking actual abnormal conditions of the electromechanical equipment, and the marks comprise abnormal conditions and normal conditions;
And the equipment abnormality recognition module is used for carrying out electromechanical equipment abnormality recognition by using the trained equipment abnormality recognition model.
10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1-8 when the computer program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410864625.6A CN118392255B (en) | 2024-07-01 | 2024-07-01 | Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410864625.6A CN118392255B (en) | 2024-07-01 | 2024-07-01 | Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118392255A true CN118392255A (en) | 2024-07-26 |
| CN118392255B CN118392255B (en) | 2024-09-27 |
Family
ID=91998267
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410864625.6A Active CN118392255B (en) | 2024-07-01 | 2024-07-01 | Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118392255B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542262A (en) * | 2012-01-04 | 2012-07-04 | 东南大学 | Waveform identification method based on operating-characteristic working condition waveform library of high-speed rail |
| US8306931B1 (en) * | 2009-08-06 | 2012-11-06 | Data Fusion & Neural Networks, LLC | Detecting, classifying, and tracking abnormal data in a data stream |
| CN109524139A (en) * | 2018-10-23 | 2019-03-26 | 中核核电运行管理有限公司 | A kind of real-time device performance monitoring method based on equipment working condition variation |
| US20200097810A1 (en) * | 2018-09-25 | 2020-03-26 | Oracle International Corporation | Automated window based feature generation for time-series forecasting and anomaly detection |
| CN112258620A (en) * | 2018-12-28 | 2021-01-22 | 英特尔公司 | Apparatus and method for ray tracing instruction processing and execution |
| CN113673600A (en) * | 2021-08-23 | 2021-11-19 | 中海石油气电集团有限责任公司 | Industrial signal abnormity early warning method, system, storage medium and computing equipment |
| CN114548236A (en) * | 2022-01-28 | 2022-05-27 | 郑州信大捷安信息技术股份有限公司 | Production condition monitoring method and system |
| US20220187140A1 (en) * | 2020-12-14 | 2022-06-16 | Yokogawa Electric Corporation | Abnormal temperature detection device, abnormal temperature detection method, and non-transitory computer readable storage medium |
| CN115438747A (en) * | 2022-09-29 | 2022-12-06 | 中国农业银行股份有限公司 | Abnormal account recognition model training method, device, equipment and medium |
| JP2023018316A (en) * | 2021-07-27 | 2023-02-08 | 京セラドキュメントソリューションズ株式会社 | Image recognition method, image recognition device, and image recognition program |
| CN115902646A (en) * | 2023-01-06 | 2023-04-04 | 中国电力科学研究院有限公司 | Energy storage battery fault identification method and system |
| CN116434341A (en) * | 2023-04-23 | 2023-07-14 | 昆明理工大学 | Student classroom abnormal behavior identification method and system |
| CN117762913A (en) * | 2023-12-14 | 2024-03-26 | 广西电网有限责任公司 | Automatic identification method and system for abnormal power data of offshore wind power |
-
2024
- 2024-07-01 CN CN202410864625.6A patent/CN118392255B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8306931B1 (en) * | 2009-08-06 | 2012-11-06 | Data Fusion & Neural Networks, LLC | Detecting, classifying, and tracking abnormal data in a data stream |
| CN102542262A (en) * | 2012-01-04 | 2012-07-04 | 东南大学 | Waveform identification method based on operating-characteristic working condition waveform library of high-speed rail |
| US20200097810A1 (en) * | 2018-09-25 | 2020-03-26 | Oracle International Corporation | Automated window based feature generation for time-series forecasting and anomaly detection |
| CN109524139A (en) * | 2018-10-23 | 2019-03-26 | 中核核电运行管理有限公司 | A kind of real-time device performance monitoring method based on equipment working condition variation |
| CN112258620A (en) * | 2018-12-28 | 2021-01-22 | 英特尔公司 | Apparatus and method for ray tracing instruction processing and execution |
| US20220187140A1 (en) * | 2020-12-14 | 2022-06-16 | Yokogawa Electric Corporation | Abnormal temperature detection device, abnormal temperature detection method, and non-transitory computer readable storage medium |
| JP2023018316A (en) * | 2021-07-27 | 2023-02-08 | 京セラドキュメントソリューションズ株式会社 | Image recognition method, image recognition device, and image recognition program |
| CN113673600A (en) * | 2021-08-23 | 2021-11-19 | 中海石油气电集团有限责任公司 | Industrial signal abnormity early warning method, system, storage medium and computing equipment |
| CN114548236A (en) * | 2022-01-28 | 2022-05-27 | 郑州信大捷安信息技术股份有限公司 | Production condition monitoring method and system |
| CN115438747A (en) * | 2022-09-29 | 2022-12-06 | 中国农业银行股份有限公司 | Abnormal account recognition model training method, device, equipment and medium |
| CN115902646A (en) * | 2023-01-06 | 2023-04-04 | 中国电力科学研究院有限公司 | Energy storage battery fault identification method and system |
| CN116434341A (en) * | 2023-04-23 | 2023-07-14 | 昆明理工大学 | Student classroom abnormal behavior identification method and system |
| CN117762913A (en) * | 2023-12-14 | 2024-03-26 | 广西电网有限责任公司 | Automatic identification method and system for abnormal power data of offshore wind power |
Non-Patent Citations (1)
| Title |
|---|
| 柳青秀;马红占;褚学宁;马斌彬;王峥;: "基于长短时记忆―自编码神经网络的风电机组性能评估及异常检测", 计算机集成制造系统, no. 12, 15 December 2019 (2019-12-15), pages 233 - 243 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118392255B (en) | 2024-09-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116757534B (en) | Intelligent refrigerator reliability analysis method based on neural training network | |
| JP2021064370A (en) | Method and system for semi-supervised deep abnormality detection for large-scale industrial monitoring system based on time-series data utilizing digital twin simulation data | |
| CN111460728B (en) | Method and device for predicting residual life of industrial equipment, storage medium and equipment | |
| CN108204341B (en) | method and device for identifying operating state of wind power plant | |
| CN106663086A (en) | Apparatus and method for ensembles of kernel regression models | |
| CN114692507B (en) | Soft measurement modeling method for count data based on stacked poisson self-encoder network | |
| CN112214006A (en) | Intermittent process fault detection method and system considering two-dimensional dynamic characteristics | |
| CN111930601A (en) | Deep learning-based database state comprehensive scoring method and system | |
| CN118386024A (en) | Machine tool based on artificial intelligence and fault detection method thereof | |
| CN118351970A (en) | Prediction method for PM2.5 concentration in ecological environment | |
| CN103279030B (en) | Dynamic soft measuring modeling method and device based on Bayesian frame | |
| CN117235624B (en) | Emission data falsification detection method, device and system and storage medium | |
| CN118392255B (en) | Method, system and equipment for monitoring operation condition of electromechanical equipment of water plant | |
| CN114970709B (en) | Improved GA-based data-driven AHU multi-fault diagnosis feature selection method | |
| CN117032080A (en) | Implementation method and application of algorithm for improving state monitoring capability of numerical control machine tool | |
| CN111524023A (en) | Greenhouse adjusting method and system | |
| Liu et al. | Batch process monitoring with two-dimensional LSTM-variational auto-encoder | |
| CN116127831A (en) | Soft measurement method for difficult-to-measure parameters of heavy gas turbine | |
| CN115759438A (en) | Atmospheric pollution analysis and prediction method and system based on machine learning | |
| CN117933093B (en) | Fault mode and influence analysis method based on water treatment exceeding accident dynamic simulation | |
| CN111443686A (en) | Industrial alarm design method based on multi-objective optimization and evidence iterative update | |
| CN113822580B (en) | Equipment working condition assessment method and related equipment | |
| CN117237165B (en) | Method for detecting fake data | |
| CN118863609A (en) | Quality control system in mutton course of working | |
| CN115421467A (en) | Large-scale industrial process fault diagnosis method based on multi-block probability kernel partial least squares |
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 |