CN115951173A - Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor - Google Patents
Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor Download PDFInfo
- Publication number
- CN115951173A CN115951173A CN202310097646.5A CN202310097646A CN115951173A CN 115951173 A CN115951173 A CN 115951173A CN 202310097646 A CN202310097646 A CN 202310097646A CN 115951173 A CN115951173 A CN 115951173A
- Authority
- CN
- China
- Prior art keywords
- state
- integrated capacitor
- capacitor
- monitoring
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/30—Reactive power compensation
Abstract
The invention belongs to the technical field of electrical equipment monitoring, and discloses an on-line state evaluation method and device for an integrated capacitor based on a multilayer perceptron. Firstly, acquiring and acquiring key parameter information of an integrated capacitor through an online monitoring unit, and transmitting the key parameter information to a monitoring and analyzing unit; then the monitoring and analyzing unit calculates and obtains index information according to the key parameter information; and then the monitoring and analyzing unit inputs the index information into a multilayer sensor model, and the multilayer sensor model outputs a state index capable of reflecting the running state of the integrated capacitor. The invention can intuitively, reliably and real-timely evaluate the running state of the integrated capacitor.
Description
Technical Field
The invention belongs to the technical field of electrical equipment monitoring, and particularly relates to an on-line state evaluation method and device for an integrated capacitor based on a multilayer sensor.
Background
The set formula condenser is arranged interior unit set in the shell, and inside adopts insulating measure to reduce the spacing distance between the unit, for frame-type condenser, has greatly reduced the condenser size, and capacitor element keeps apart with the external world simultaneously, has and receives environmental impact little, the simple advantage of wiring.
However, the conventional integrated capacitor has the problems of higher failure rate and long repair cycle compared with a single capacitor. The integrated capacitor is technically improved, so that the reliability of the integrated capacitor is improved, and the integrated capacitor is free from damage and maintenance in the whole life cycle; on the other hand, the intelligent improvement is carried out on the integrated capacitor, state data in the operation process of the integrated capacitor is recorded, the existing hidden danger is early warned, the reason for the hidden danger is analyzed on line or afterwards, and a basis is provided for improving the reliability of the integrated capacitor. These measures are a prerequisite and basis for the development of integrated capacitors.
At present, the on-line monitoring device and the technical research on the integrated capacitor are not much, the existing means mainly analyzes the change of the total capacity value of the integrated capacitor through monitoring the phase current and the phase voltage, the sensitivity is poor, and the monitoring data is not comprehensive enough. It is difficult to intuitively, reliably, and real-time evaluate the operating state of the collective capacitor.
Disclosure of Invention
The invention provides an on-line state evaluation method and device of an integrated capacitor based on a multilayer perceptron, and solves the problem that the running state of the integrated capacitor is difficult to evaluate reliably in the prior art.
The invention provides an on-line state evaluation method of an integrated capacitor based on a multilayer perceptron, which comprises the following steps:
step 1, acquiring and acquiring key parameter information of an integrated capacitor through an online monitoring unit, and transmitting the key parameter information to a monitoring and analyzing unit;
step 2, the monitoring analysis unit calculates index information according to the key parameter information;
and 3, inputting the index information into a multilayer sensor model by the monitoring and analyzing unit, and outputting a state index capable of reflecting the running state of the integrated capacitor by the multilayer sensor model.
Preferably, the method for estimating the online state of the integrated capacitor based on the multi-layer perceptron further comprises: and 4, the monitoring and analyzing unit carries out early warning or alarm according to the state index.
Preferably, the key parameter information includes phase current, phase voltage, bridge difference current and operating temperature;
the index information comprises a dielectric loss variation linear regression parameter a l And k l And compensate forInitial unbalanced bridge difference current I d And overvoltage multiple k at capacitor switching-on moment v Over-current multiple k at the moment of opening capacitor c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e And an average harmonic distortion factor F.
Preferably, the multi-layered perceptron model comprises an input layer, a hidden layer and an output layer; the input layer is an 8-dimensional vector and respectively and correspondingly compensates the bridge difference current I of initial unbalance d Dielectric loss variation linear regression parameter a l And k l Overvoltage multiple k v Over-current multiple k c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e Average harmonic distortion factor F; the hidden layer has two layers, namely a 12-dimensional vector and a 6-dimensional vector, and is used for extracting self characteristics and mutual correlation characteristics of each parameter in the input vector; the output layer is a 1-dimensional scalar and corresponds to the state index.
Preferably, index information acquired in the long-time operation process of the plurality of integrated capacitors is taken as a sample, and the corresponding state index is taken as a sample label to construct a sample set; and training and verifying the multilayer perceptron model based on the sample set.
Preferably, during operation of the collective capacitor, a time interval is set at t 0 Representing a time starting point, and storing index information at fixed time; setting a state index by analyzing index information when constructing the sample set;
setting the state index to compensate for the initially unbalanced bridge differential current I d Determining the specific time point t of the abnormal occurrence of the collective capacitor by combining other indexes for a main line 1 Specific point in time t of occurrence of a fault 2 And a process of change of the operating state;
for t 0 Data of the starting time, the running state of the integrated capacitor is considered to be healthy at the moment, and the state index H =0 is set;
for t 0 If the operation state of the data after the moment is not changed, keeping H =0;
for t 1 The state index H = v is set for the time data 1 ；
For t 0 After, t 1 Previous data, if there is a change in the operating state, set the values of H at 0 and v 1 To the change start time to t 1 Linear interpolation is carried out between the two;
for t 2 The state index H = v is set for the time data 2 ；
For t 1 After, t 2 Previous data, if the operation state has a change process, setting the value of H at v 1 And v 2 To the change start time to t 2 Linear interpolation is carried out between the two;
for t 2 If the operation state of the subsequent data is not further changed, setting the state index H = v 2 。
Preferably, a dielectric loss factor tan δ is calculated based on the acquired phase current, the acquired phase voltage and the acquired operating temperature; performing linear regression analysis on the measured values of all dielectric loss factors in a preset time period to obtain tan delta = a l +k l t, fitting the past measurement value of the dielectric loss factor to obtain a dielectric loss change linear regression parameter a l And k l 。
Preferably, said bridge difference current I compensated for initial imbalance d ＝I b -I 0 Wherein, I b Indicating the monitored unbalance current, I 0 Representing the initial unbalance current.
Preferably, when the integrated capacitor is abnormal, switching is performed, or a specified time interval is reached, the monitoring and analyzing unit stores all relevant information to the database.
In another aspect, the present invention provides an integrated capacitor online status evaluation device based on a multi-layer perceptron, including: the online monitoring unit is communicated with the monitoring analysis unit; the online monitoring unit is integrated with the capacitor body to form an integrated capacitor; the monitoring and analyzing unit stores a multilayer perceptron model; the online state evaluation device of the integrated capacitor based on the multilayer perceptron is used for realizing the steps in the online state evaluation method of the integrated capacitor based on the multilayer perceptron.
One or more technical schemes provided by the invention at least have the following technical effects or advantages:
firstly, acquiring and acquiring key parameter information of an integrated capacitor through an online monitoring unit, and transmitting the key parameter information to a monitoring and analyzing unit; then the monitoring and analyzing unit calculates and obtains index information according to the key parameter information; and then the monitoring and analyzing unit inputs the index information into a multilayer sensor model, and the multilayer sensor model outputs a state index capable of reflecting the running state of the integrated capacitor. The method combines a multilayer perceptron model, comprehensively utilizes various indexes in the operation process of the integrated capacitor to evaluate the operation state of the capacitor, thereby achieving a more stable result than the result judged according to a single parameter, and obtaining a state index which is visual and clear. In conclusion, the invention can intuitively, reliably and real-timely evaluate the running state of the integrated capacitor.
Drawings
FIG. 1 is a schematic diagram of the internal structure of one phase of the integrated capacitor and bridge differential current sampling;
FIG. 2 is a flowchart of an online state evaluation method for an integrated capacitor based on a multi-layer perceptron according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a multilayer perceptron model adopted in an online state evaluation method for an integrated capacitor based on a multilayer perceptron according to an embodiment of the present invention.
Detailed Description
In order to better understand the technical scheme, the technical scheme is described in detail in the following with reference to the attached drawings of the specification and specific embodiments.
The integrated capacitor integrates all the inner units of single phase or three phases in the shell, and the inner units of each phase are connected together in a series-parallel combination mode. Referring to fig. 1, the inside of a capacitor of a certain phase is simplified as C 1 ～C 4 Four internal unitsThe internal units are composed of series-parallel structures of internal capacitor elements, and the capacitance values of the four internal units are equal. The bridge difference current signal is obtained by the CT (current transformer) and subsequent acquisition circuitry in fig. 1. The value of the acquired bridge difference current is 0 in the normal state of the capacitor, when an internal element of a certain internal unit is damaged, the balance relation is broken, and the bridge difference current can be sampled to obtain unbalanced current. Meanwhile, phase voltage of the phase capacitor is collected through a PT (potential transformer), phase current of the phase capacitor is collected through a CT (current transformer), temperature (namely running temperature) of the hottest point in the capacitor is collected through a temperature sensor, and the collected data are used as key parameter information of online monitoring.
Referring to fig. 2, in the online state evaluation method for the integrated capacitor based on the multilayer perceptron, the online monitoring unit (including the current transformer CT, the voltage transformer PT, the temperature sensor, etc.) collects and acquires key parameters such as phase current, phase voltage, bridge difference current, temperature, etc. of the integrated capacitor, and then converts the acquired information into digital signals and transmits the digital signals to the monitoring and analyzing unit (for example, the background state evaluation and calculation server) through the optical cable or the cable. The server receives and stores the signals. And then, the background analysis software running on the server carries out post-processing on the stored data, and the indexes and data such as dielectric loss, bridge difference current, initial unbalanced current, harmonic waves, overcurrent and overvoltage in the transient process are obtained through calculation. And then, calculating the state index of the integrated capacitor from the indexes through a multilayer perceptron (MLP) model, and performing early warning and alarming based on a threshold value. The invention can analyze and evaluate the operation condition of the integrated capacitor, and can feed back the operation condition to a user through a human-computer interface.
The present invention will be specifically explained below.
The embodiment of the invention provides an on-line state evaluation method of an integrated capacitor based on a multilayer perceptron, which is shown in figure 2 and comprises the following steps:
step 1, acquiring and acquiring key parameter information of the integrated capacitor through an online monitoring unit, and transmitting the key parameter information to a monitoring and analyzing unit.
Wherein the key parameter information comprises phase current, phase voltage, bridge difference current and operating temperature.
And 2, calculating index information by the monitoring and analyzing unit according to the key parameter information.
Wherein the index information comprises a dielectric loss variation linear regression parameter a l And k l Compensating the bridge differential current I of the initial unbalance d And the over-voltage multiple k at the moment of closing the capacitor v Over-current multiple k at the moment of opening capacitor c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e And an average harmonic distortion factor F.
Specifically, the dielectric loss factor tan δ can be calculated by calculating the phase current and the phase difference of the phase voltage and supplementing according to the operating temperature. Considering that the measurement of the dielectric loss factor is influenced by temperature, working voltage and environment, the measurement error is large, and the precision requirement of evaluating the performance of the capacitor through the measurement error cannot be met. Therefore, the invention adopts the steps of storing all measured dielectric loss tangent values in the last long time period T and carrying out linear regression analysis on the dielectric loss tangent values, namely carrying out linear regression analysis on all measured values of dielectric loss factors in a preset time period to obtain the dielectric loss tangent value with tan = a l +k l t, fitting the past measurement value of the dielectric loss factor to obtain a dielectric loss change linear regression parameter a l And k l . Dielectric loss variation linear regression parameter a l And k l Reflecting the level and the trend of dielectric loss. Although the real situation cannot be reflected by the accuracy problem of the instant dielectric loss measurement value, the variation trend of the instant dielectric loss measurement value in a period can reflect the variation of the capacitor operation condition. The invention adopts a fitting parameter a l And k l As the evaluation basis of the state of the integrated capacitor, the problem of direct insufficient precision can be avoided, and the reliable evaluation result of the online state of the integrated capacitor can be obtained.
Said bridge differential current I compensated for initial imbalance d ＝I b -I 0 Wherein, I b Indicating the monitored unbalance current, I 0 Representing the initial unbalance current. Specifically, as shown in FIG. 1, in general, C is the factor 1 ＝C 2 ＝C 3 ＝C 4 Thus bridge difference current I b Is balanced and is substantially 0. When C is 1 ～C 4 When a capacitor element of one internal unit fails, a fuse of the element is blown out to cut the element out of a line, the capacitance value of the branch circuit changes correspondingly, and bridge difference unbalanced current is generated, wherein I b Is not 0, and I b Increases as the number of damaged elements increases, thereby increasing I b A step-like change is present. Thus monitoring the bridge differential current I b Can determine the operating condition of the capacitor. Theoretically, the initial bridge differential current is balanced, essentially 0. In practice, however, the initial value of the bridge differential current is not absolutely zero but a small value I due to manufacturing process, operating environment, etc 0 . The system measures and compensates for the initial unbalance current by observing the unbalance current I b Relative to (I) 0 Change value of (I) d ＝I b -I 0 The operation condition of the capacitor is judged, so that the initial unbalanced current is further eliminated, and the monitoring is more sensitive.
The harmonic levels of the phase current phase voltages running on the capacitors affect the safe operation of the capacitor bank, and when the harmonic components are outside the allowable range, the equipment is prone to cause damage and tripping during operation. But at the same time, the harmonic level is influenced by the environment, can show certain fluctuation along with time, and is not suitable for judging the running condition of the current capacitor in real time. Therefore, the present invention calculates the average value of the timing sampling values of the harmonic distortion rate in a longer time period, such as the average value F of 84 harmonic distortion rate indicators sampled at each half-small time interval in seven days, as the estimation of the harmonic level of the capacitor bank operation.
During the operation of the capacitor, defects of partial elements cannot be reflected on unbalanced current when the defects are not developed to be damaged, but the defects may be reflected in the transient process during the switching of the capacitor bank. Therefore, the invention monitors the switching process of the capacitor and obtains the overvoltage of the capacitor at the moment of switching onMultiple k v And over-current multiple k at the moment of opening c And the stable time t of the transient switching-on and switching-off process s 、t e From these parameters, the mini-risk condition of the collective capacitor bank is evaluated.
And 3, inputting the index information into a multilayer sensor model by the monitoring and analyzing unit, and outputting a state index capable of reflecting the running state of the integrated capacitor by the multilayer sensor model.
The monitoring indexes obtained by calculation in the above steps include: dielectric loss variation index a l And k l Compensating the bridge differential current I of the initial unbalance d Overvoltage multiple k v And over-current multiple k c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e Once the capacitor has hidden trouble in operation, the indexes change correspondingly. The mapping relation between the index and the occurrence of the hidden danger or the fault is difficult to calculate through an analytical expression. The method utilizes the multilayer perceptron model shown in figure 3 to learn the intrinsic relation based on sample data and represents the intrinsic relation in the parameters of the perceptron model.
Specifically, referring to fig. 3, the multi-layered sensor model includes an input layer, a hidden layer, and an output layer; the input layer is an 8-dimensional vector and respectively and correspondingly compensates the bridge difference current I of initial unbalance d Dielectric loss variation linear regression parameter a l And k l Overvoltage multiple k v Over-current multiple k c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e Average harmonic distortion factor F; the hidden layer has two layers, namely a 12-dimensional vector and a 6-dimensional vector, and is used for extracting self characteristics and mutual correlation characteristics of each parameter in the input vector; the output layer is a 1-dimensional scalar and corresponds to the state index.
Index information acquired in the long-time operation process of the plurality of integrated capacitors is taken as a sample, and the corresponding state index is taken as a sample label to construct a sample set; and training and verifying the multilayer perceptron model based on the sample set so as to enable the multilayer perceptron model to learn the internal correlation of indexes and state indexes.
The samples required by the training of the multilayer perceptron model can be acquired by the timed acquisition of a system and the combination of expert diagnosis. For example, during operation of the collective capacitor (or capacitor evaluation system), a time interval is set at t 0 Representing a time starting point, and storing index information at fixed time; after the system is operated for a long time, all saved indexes and data can be analyzed by related experts. Namely, when the sample set is constructed, the state index is set through analyzing the index information. Setting the state index to compensate for the initially unbalanced bridge differential current I d Determining the specific time point t of the abnormality of the integrated capacitor by combining other indexes as a main line 1 Specific time t at which a failure occurs 2 And a process of change of the operation state. For example, the bridge difference current I compensated with the initial imbalance is analyzed d As a main line, judging the approximate time point of the capacitor when the capacitor is abnormal and has faults, and determining the specific time point t of the capacitor when the capacitor is abnormal by combining dielectric loss, harmonic wave and transient process 1 Specific point in time t of occurrence of a fault 2 And a variation process. For t 0 Data of the starting time, the running state of the integrated capacitor is considered to be healthy at the moment, and the state index H =0 is set; for t 0 If the operation state of the data after the moment is not changed, keeping H =0; for t 1 The state index H = v is set for the data of the time 1 (ii) a Can be based on the unbalanced current I d The number of damaged elements estimated by combining other indexes is used as an index key value v 1 E.g. v 1 =3; for t 0 After, t 1 Previous data, if there is a change in the operating state, set the values of H at 0 and v 1 To the change start time to t 1 Linear interpolation is carried out between the two; for t 2 The state index H = v is set for the data of the time 2 (ii) a The index key value v is also specified with the previously described criteria 2 Specific values of (e.g. v) 2 =5; for t 1 After, t 2 Previous data, if the operation state has a change process, setting the value of H at v 1 And v 2 In the above-mentioned manner,for the change starting time to t 2 Linear interpolation is carried out between the two; for t 2 If the operation state of the subsequent data is not further changed, setting the state index H = v 2 。
And 4, the monitoring and analyzing unit carries out early warning or alarm according to the state index.
For example, with the exponential key value v described above 1 、v 2 And (3) early warning and alarming the capacitor for a threshold value: when H is present>v 1 In time, early warning is carried out to indicate that the capacitor is abnormal, and hidden danger reasons need to be concerned and prevented from continuing to develop; when H is present>v 2 And in time, alarming is carried out to indicate that the capacitor has a fault and needs to be processed in time to remove the fault.
In addition, when the integrated capacitor is abnormal, switched or reaches a specified time interval, the monitoring and analyzing unit stores all relevant information to the database. The method can be used as a sample for training a multilayer perceptron model and for experts to analyze abnormal or fault reasons afterwards.
Corresponding to the method, the invention also provides an integrated capacitor online state evaluation device based on the multilayer perceptron, which comprises: the online monitoring unit is communicated with the monitoring analysis unit; the online monitoring unit is integrated with the capacitor body to form an integrated capacitor; the monitoring and analyzing unit stores a multilayer perceptron model; the online state evaluation device of the integrated capacitor based on the multilayer perceptron is used for realizing the steps in the online state evaluation method of the integrated capacitor based on the multilayer perceptron.
The method and the device for evaluating the online state of the assembled capacitor based on the multilayer perceptron at least have the following technical effects:
1. according to the method, the operation state of the capacitor is evaluated by establishing a multilayer perceptron model and comprehensively utilizing index parameters in various aspects in the operation process of the integrated capacitor, so that a more stable result is obtained compared with a result judged according to a single parameter, and a reliable analysis and evaluation result can be obtained. In addition, through sample learning, the system can monitor the slow development process of the abnormal condition before the jump of the evaluation state, and a user can intervene and remove the fault before the abnormality or the fault occurs.
2. The dielectric loss of the capacitor is difficult to monitor with high precision by the conventional means, the problem of insufficient precision of direct dielectric loss monitoring is avoided, the variation trend of the dielectric loss of the integrated capacitor is monitored, slow and slight variation of the dielectric loss of the capacitor in a long-time process can be effectively extracted, and more accurate analysis and evaluation results can be obtained.
3. The invention monitors bridge difference current and compensates the initial unbalanced current, thereby improving the sensitivity of capacitor monitoring and being beneficial to obtaining more accurate analysis and evaluation results.
4. Bridge difference current monitoring is carried out on an internal unit of one phase of the integrated capacitor, and an online monitoring unit (comprising a sensor, an acquisition device and the like) is integrated with the capacitor body, so that the monitoring device is free of secondary installation and convenient to use; and the performance and the influence on the operation of the body can be evaluated and tested while leaving the factory, so that the working process is very stable and reliable.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to examples, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.
Claims (10)
1. An on-line state evaluation method of an integrated capacitor based on a multilayer perceptron is characterized by comprising the following steps:
step 1, acquiring and acquiring key parameter information of an integrated capacitor through an online monitoring unit, and transmitting the key parameter information to a monitoring and analyzing unit;
step 2, the monitoring and analyzing unit calculates index information according to the key parameter information;
and 3, inputting the index information into a multilayer sensor model by the monitoring and analyzing unit, and outputting a state index capable of reflecting the running state of the integrated capacitor by the multilayer sensor model.
2. The multi-layer perceptron-based integrated capacitor online state assessment method according to claim 1, further comprising:
and 4, the monitoring and analyzing unit carries out early warning or alarm according to the state index.
3. The multi-layer perceptron-based integrated capacitor online status assessment method according to claim 1, characterized in that the key parameter information includes phase current, phase voltage, bridge difference current and operating temperature;
the index information comprises dielectric loss variation linear regression parameter a l And k l Compensating the bridge differential current I of the initial unbalance d And overvoltage multiple k at capacitor switching-on moment v Over-current multiple k at the moment of opening capacitor c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e And an average harmonic distortion rate F.
4. The multilayer perceptron-based collective capacitor online status evaluation method according to claim 3, characterized in that the multilayer perceptron model comprises an input layer, a hidden layer and an output layer; the input layer is an 8-dimensional vector and respectively and correspondingly compensates bridge difference current I of initial unbalance d Dielectric loss variation linear regression parameter a l And k l Overvoltage multiple k v Over-current multiple k c And the stable time t of the switching-on transient process s The stable time t of the switching-off transient process e Average harmonic distortion factor F; the hidden layer has two layers, namely a 12-dimensional vector and a 6-dimensional vector, and is used for extracting self characteristics and mutual correlation characteristics of each parameter in the input vector; the output layer is a 1-dimensional scalar and corresponds to the state index.
5. The method for evaluating the on-line state of the assembled capacitor based on the multilayer perceptron as claimed in claim 4, wherein index information obtained in the long-time running process of a plurality of assembled capacitors is taken as a sample, and a sample set is constructed by taking the corresponding state index as a sample label; and training and verifying the multilayer perceptron model based on the sample set.
6. The method as claimed in claim 5, wherein a time interval is set to t during the operation of the integrated capacitor 0 Representing a time starting point, and storing index information at fixed time; setting a state index by analyzing index information when constructing the sample set;
setting the state index to compensate for the bridge difference current I of the initial imbalance d Determining the specific time point t of the abnormality of the integrated capacitor by combining other indexes as a main line 1 Specific point in time t of occurrence of a fault 2 And a process of change of the operating state;
for t 0 Data of the starting time, the running state of the integrated capacitor is considered to be healthy at the moment, and the state index H =0 is set;
for t 0 If the operation state of the data after the moment is not changed, keeping H =0;
for t 1 The state index H = v is set for the time data 1 ；
For t 0 After, t 1 Previous data, if there is a change in the operating state, set the values of H at 0 and v 1 To the change start time to t 1 Linear interpolation is carried out between the two;
for t 2 The state index H = v is set for the time data 2 ；
For t 1 After t, t 2 Previous data, if the operation state has a change process, setting the value of H at v 1 And v 2 To the change start time to t 2 Linear interpolation is carried out between the two;
for t 2 If the operation state of the subsequent data is not further changed, setting the state index H = v 2 。
7. The multilayer perceptron-based integrated capacitor online state assessment method according to claim 3, characterized in that a dielectric loss factor tan δ is calculated based on the collected phase current, the phase voltage and the operating temperature; performing linear regression analysis on the measured values of all dielectric loss factors in a preset time period to obtain tan delta = a l +k l t, fitting the historical measured value of the dielectric loss factor to obtain a dielectric loss change linear regression parameter a l And k l 。
8. The multilayer perceptron-based integrated capacitor online state assessment method according to claim 3, characterized in that the bridge difference current I that compensates for initial imbalance d ＝I b -I 0 Wherein, I b Indicating the monitored unbalance current, I 0 Representing the initial unbalance current.
9. The method for evaluating the online state of the integrated capacitor based on the multi-layer perceptron as claimed in claim 1, wherein when the integrated capacitor is abnormal, switched or reaches a specified time interval, the monitoring and analyzing unit stores all the related information to the database.
10. An on-line state evaluation device of an integrated capacitor based on a multilayer perceptron is characterized by comprising: the online monitoring unit is communicated with the monitoring analysis unit; the online monitoring unit is integrated with the capacitor body to form an integrated capacitor; a multilayer perceptron model is stored in the monitoring and analyzing unit; the multi-layer perceptron-based integrated capacitor online state evaluation device is used for realizing the steps in the multi-layer perceptron-based integrated capacitor online state evaluation method according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310097646.5A CN115951173A (en) | 2023-01-28 | 2023-01-28 | Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310097646.5A CN115951173A (en) | 2023-01-28 | 2023-01-28 | Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115951173A true CN115951173A (en) | 2023-04-11 |
Family
ID=87291299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310097646.5A Pending CN115951173A (en) | 2023-01-28 | 2023-01-28 | Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115951173A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116316667A (en) * | 2023-05-09 | 2023-06-23 | 深圳华强电子网集团股份有限公司 | Intelligent integrated power capacitor compensation device |
-
2023
- 2023-01-28 CN CN202310097646.5A patent/CN115951173A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116316667A (en) * | 2023-05-09 | 2023-06-23 | 深圳华强电子网集团股份有限公司 | Intelligent integrated power capacitor compensation device |
| CN116316667B (en) * | 2023-05-09 | 2023-08-08 | 深圳华强电子网集团股份有限公司 | Intelligent integrated power capacitor compensation device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105513252B (en) | A kind of electrical fire monitoring method for early warning and equipment based on integral algorithm | |
| CN116879662B (en) | Transformer fault detection method based on data analysis | |
| EP2992340B1 (en) | System and method for ups battery monitoring and data analysis | |
| EP2989705B1 (en) | Method and apparatus for defect pre-warning of power device | |
| CN112598298A (en) | Power transformer health management system and management method | |
| GB2476246A (en) | Diagnosing an operation mode of a machine | |
| KR100931992B1 (en) | Metal enclosed switchgear system and method for diagnosis by oneself insulation aging and abnormal temperature | |
| CN115951173A (en) | Multilayer perceptron-based on-line state evaluation method and device for integrated capacitor | |
| CN115238831B (en) | Fault prediction method, device, computer equipment and storage medium | |
| CN108418304B (en) | Transformer substation secondary circuit state monitoring method, device and system | |
| CN115877205A (en) | Intelligent fault diagnosis system and method for servo motor | |
| CN116609651A (en) | Detection method of multifunctional circuit breaker for photovoltaic | |
| CN117318310B (en) | Remote unmanned intelligent monitoring system for explosion-proof power distribution cabinet | |
| CN113985284A (en) | System and method for monitoring additional diesel generator of nuclear power station | |
| CN106802384B (en) | Electrical equipment fault detection method based on infrared and ultraviolet photoelectric sensing fusion | |
| CN117284151A (en) | Method and system for monitoring battery power of self-adaptive new energy vehicle | |
| JP2020187955A (en) | Method and device for diagnosing switchgear | |
| CN114895163A (en) | Cable inspection positioning device and method based on cable insulation performance | |
| CN116087718A (en) | Method and device for on-line monitoring of collective capacitor based on parameter fusion | |
| KR102604708B1 (en) | Switchboard diagnosis system based on artificial intelligence and switchboard diagnosis method based on artificial intelligence | |
| CN116342110B (en) | Intelligent fault diagnosis and fault tolerance measurement method for multiple temperature measurement loops of train | |
| CN115358353B (en) | Multi-source fusion transformer fault diagnosis method | |
| Li et al. | Transformer fault detection method based on multi-source sensing data correlation analysis | |
| RU2814857C1 (en) | System for monitoring and diagnosing state of turbine generator | |
| CN117849691B (en) | Multi-dimensional collaborative operation monitoring and early warning system and method for capacitive voltage transformer |
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 |