CN-121997214-A - Method and system for constructing digital twin model of camera based on real-time data fitting

CN121997214ACN 121997214 ACN121997214 ACN 121997214ACN-121997214-A

Abstract

The invention discloses a method and a system for constructing a digital twin model of a camera based on real-time data fitting, wherein the method comprises the steps of obtaining real-time operation data of the camera; based on vibration data in operation data of a camera, extracting vibration spectrum density characteristic quantity by wavelet transformation, using reactive power in the operation data as an input variable, using the vibration spectrum density characteristic quantity as an output variable, performing function fitting by using a least square method to construct a digital twin model between the reactive power and the vibration spectrum density, inputting the reactive power in the operation data acquired in real time into the digital twin model to generate a predicted value of the vibration spectrum density, calculating a standard deviation between the real-time vibration spectrum density and the predicted value, and judging that the camera is in an abnormal operation state and triggering an alarm mechanism when the standard deviation exceeds a preset threshold value. The invention has the advantages of improving the monitoring efficiency of the running state of the camera, reducing the running and maintenance cost of the camera and the like.

Inventors

ZHANG DONGQING
LIAO SIZHUO
JU YUNTAO
ZHANG CHAOFENG
ZHAO WENQIANG
ZHANG GUOHUA
LV JIWEI
HAN SHUAI
WANG ZHENGWEI
YANG JIANMING
ZHOU JUN

Assignees

国家电网有限公司直流技术中心
国网湖南省电力有限公司超高压变电公司
国网青海省电力公司电力科学研究院
国网青海省电力公司
国网湖南省电力有限公司
国家电网有限公司
清华海峡研究院（厦门）

Dates

Publication Date: 20260508
Application Date: 20251201

Claims (17)

1. A method for constructing a digital twin model of a camera based on real-time data fitting is characterized by comprising the following steps: Acquiring real-time operation data of the camera, wherein the operation data comprises vibration data and reactive power; based on vibration data in operation data of a camera, extracting a vibration spectrum density characteristic quantity by wavelet transformation, taking reactive power in the operation data as an input variable, taking the vibration spectrum density characteristic quantity as an output variable, and performing function fitting by a least square method to construct a digital twin model between the reactive power and the vibration spectrum density; reactive power in the operation data acquired in real time is input into the digital twin model, and a predicted value of vibration spectrum density is generated; And when the standard deviation exceeds a preset threshold value, judging that the camera is in an abnormal running state and triggering an alarm mechanism.
2. The method for constructing the digital twin model of the camera based on real-time data fitting according to claim 1, wherein the specific process of extracting the characteristic quantity of the vibration spectrum density by adopting wavelet transformation processing is as follows: carrying out direct current removal and trend removal pretreatment on the collected original vibration signals; Performing continuous wavelet transformation on the preprocessed signals by using a complex-valued wavelet basis function to obtain a time-frequency distribution diagram; averaging in the time dimension of the time-frequency distribution diagram to obtain a frequency-energy spectrum; And identifying energy peaks corresponding to the fundamental frequency, the double fundamental frequency and the triple fundamental frequency from the frequency-energy spectrum, and completing extraction of the vibration spectrum density characteristic quantity.
3. The method for constructing the digital twin model of the camera based on real-time data fitting according to claim 2, wherein the function fitting is performed by adopting a least square method, and the specific process of constructing the digital twin model between reactive power and vibration spectral density is as follows: drawing the acquired vibration spectrum density data into a scatter diagram; and performing curve fitting on the scattered point data by adopting a least square method to obtain a digital twin model between reactive power and vibration spectrum density so as to generate a prediction curve of the vibration spectrum density.
4. The method for constructing digital twin model of camera for a dispatching machine based on real-time data fitting according to claim 1, 2 or 3, wherein the standard deviation The calculation formula of (2) is as follows: Wherein, the Is the first Data; mean of all data; is the number of overall data; the sum of squares of all data and mean differences.
5. A method of constructing a digital twin model of a camera for a motor vehicle based on real-time data fitting as defined in claim 1, 2 or 3, further comprising periodically checking the digital twin model, including revising the fitting parameters and recording the parameter change, and triggering an alarm when the rate of change of the parameter exceeds a threshold.
6. A method of constructing a digital twin model of a camera for a camera in a webcam based on real-time data fitting according to claim 1 or 2 or 3, wherein the conditions for triggering an alarm include at least one of: The standard deviation of the digital twin model predicted value and the real-time acquired data exceeds the limit; The real-time acquisition value exceeds the limit; overrun of the digital twin model predictive value; The coefficient variation of the digital twin model is overrun.
7. The method for constructing the digital twin model of the camera based on real-time data fitting according to claim 1, 2 or 3, wherein the operation data further comprises electrical parameters, mechanical parameters and thermal parameters, the electrical parameters comprise stator current, voltage, power factor, active power, exciting current and voltage, the mechanical parameters comprise rotor rotation speed and bearing displacement, and the thermal parameters comprise stator winding temperature, rotor winding temperature, stator cooling water outlet temperature, rotor cooling water outlet temperature, camera housing temperature and lubricating oil temperature.
8. A computer program product comprising a computer program which, when run by a processor, performs the steps of the method according to any one of claims 1-7.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, performs the steps of the method according to any of claims 1-7.
10. A computer system comprising a memory and a processor connected to each other, the memory having stored thereon a computer program which, when run by the processor, performs the steps of the method according to any of claims 1-7.
11. A system for constructing a digital twin model of a camera based on real-time data fitting, comprising: the data acquisition module is used for acquiring real-time operation data of the camera, wherein the operation data comprise vibration data and reactive power; the model construction module is used for extracting vibration spectrum density characteristic quantity based on vibration data in operation data of the camera by adopting wavelet transformation processing, taking reactive power in the operation data as an input variable, taking the vibration spectrum density characteristic quantity as an output variable, and adopting a least square method to perform function fitting to construct a digital twin model between the reactive power and the vibration spectrum density; The prediction module is used for inputting reactive power in the operation data acquired in real time into the digital twin model to generate a predicted value of vibration spectrum density; And when the standard deviation exceeds a preset threshold value, judging that the camera is in an abnormal running state and triggering an alarm mechanism.
12. The system for constructing a digital twin model of a camera based on real-time data fitting according to claim 11, wherein in the model construction module, the specific process of extracting the characteristic quantity of the vibration spectrum density by wavelet transformation is as follows: carrying out direct current removal and trend removal pretreatment on the collected original vibration signals; Performing continuous wavelet transformation on the preprocessed signals by using a complex-valued wavelet basis function to obtain a time-frequency distribution diagram; averaging in the time dimension of the time-frequency distribution diagram to obtain a frequency-energy spectrum; And identifying energy peaks corresponding to the fundamental frequency, the double fundamental frequency and the triple fundamental frequency from the frequency-energy spectrum, and completing extraction of the vibration spectrum density characteristic quantity.
13. The system for constructing a digital twin model of a camera based on real-time data fitting according to claim 12, wherein in the model construction module, the function fitting is performed by adopting a least square method, and the specific process of constructing the digital twin model between reactive power and vibration spectral density is as follows: drawing the acquired vibration spectrum density data into a scatter diagram; and performing curve fitting on the scattered point data by adopting a least square method to obtain a digital twin model between reactive power and vibration spectrum density so as to generate a prediction curve of the vibration spectrum density.
14. The system for constructing digital twin model of camera based on real-time data fitting according to claim 11, 12 or 13, wherein in the early warning module, standard deviation The calculation formula of (2) is as follows: Wherein, the Is the first Data; mean of all data; is the number of overall data; the sum of squares of all data and mean differences.
15. The system for constructing a digital twin model of a camera based on real-time data fitting according to claim 11, 12 or 13, further comprising a checking module for checking the digital twin model periodically, including revising the fitting parameters and recording parameter changes, and triggering an alarm when the parameter change rate exceeds a threshold value.
16. The digital twin model construction system for a camera based on real-time data fitting according to claim 11, 12 or 13, wherein in the early warning module, the condition for triggering the alarm comprises at least one of: The standard deviation of the digital twin model predicted value and the real-time acquired data exceeds the limit; The real-time acquisition value exceeds the limit; overrun of the digital twin model predictive value; The coefficient variation of the digital twin model is overrun.
17. The digital twin model construction system for a camera of a utility vehicle of claim 11, 12 or 13, wherein in the data acquisition module, the operational data further comprises electrical parameters including stator current, voltage, power factor, active power, excitation current and voltage, mechanical parameters including rotor speed, bearing displacement, and thermal parameters including stator winding temperature, rotor winding temperature, stator cooling water outlet temperature, rotor cooling water outlet temperature, camera housing temperature, and lube oil temperature.

Description

Method and system for constructing digital twin model of camera based on real-time data fitting Technical Field The invention mainly relates to the technical field of camera detection, in particular to a method and a system for constructing a digital twin model of a camera based on real-time data fitting. Background The phase regulator is essentially a synchronous motor in a special running state, runs in a motor state and is not provided with a mechanical load, and the phase regulator is specially used for providing or absorbing reactive power for an electric power system so as to improve the power factor of a power grid and maintain the voltage level of the power grid. In an electrical power system, an asynchronous motor and a transformer act as main loads, drawing a lot of reactive power from the grid for self-excitation, which causes the grid to take on a lot of inductive reactive current, which in turn leads to a reduction of the grid power factor. The power factor is reduced, so that the efficiency of the generator and the power transmission and distribution equipment cannot be fully exerted, the line loss and the voltage loss can be increased, the transmission quality is affected, and even the transmission stability is threatened. With the high-proportion access of the novel energy into the power grid, the camera is used as an important reactive power supporting device of the power grid, plays an important role in improving the stable operation, transregional power transmission and new energy consumption capacity of the power grid, and is required to cope with complex working conditions such as high-frequency voltage fluctuation, long-period full-load operation and the like, and the difficulty of operation state monitoring, fault early warning and operation and maintenance optimization is remarkably improved. The traditional physical modeling method based on the mechanism is too complex in model, large in quantity of parameters to be debugged, difficult to accurately reflect the dynamic characteristics and the multi-physical field coupling effect of equipment, and the pure data driving model lacks of deep association on the inherent attribute of the equipment, so that the management requirement of a camera in a novel power system cannot be met. Disclosure of Invention Aiming at the technical problems existing in the prior art, the invention provides a method and a system for constructing a digital twin model of a camera based on real-time data fitting, which are used for improving the monitoring efficiency of the running state of the camera and reducing the running and maintenance cost of the camera. In order to solve the technical problems, the technical scheme provided by the invention is as follows: a method for constructing a digital twin model of a camera based on real-time data fitting comprises the following steps: Acquiring real-time operation data of the camera, wherein the operation data comprises vibration data and reactive power; based on vibration data in operation data of a camera, extracting a vibration spectrum density characteristic quantity by wavelet transformation, taking reactive power in the operation data as an input variable, taking the vibration spectrum density characteristic quantity as an output variable, and performing function fitting by a least square method to construct a digital twin model between the reactive power and the vibration spectrum density; reactive power in the operation data acquired in real time is input into the digital twin model, and a predicted value of vibration spectrum density is generated; And when the standard deviation exceeds a preset threshold value, judging that the camera is in an abnormal running state and triggering an alarm mechanism. Preferably, the specific process of extracting the vibration spectrum density characteristic quantity by adopting wavelet transformation processing is as follows: carrying out direct current removal and trend removal pretreatment on the collected original vibration signals; Performing continuous wavelet transformation on the preprocessed signals by using a complex-valued wavelet basis function to obtain a time-frequency distribution diagram; averaging in the time dimension of the time-frequency distribution diagram to obtain a frequency-energy spectrum; And identifying energy peaks corresponding to the fundamental frequency, the double fundamental frequency and the triple fundamental frequency from the frequency-energy spectrum, and completing extraction of the vibration spectrum density characteristic quantity. Preferably, the function fitting is carried out by adopting a least square method, and the specific process of constructing the digital twin model between reactive power and vibration spectral density is as follows: drawing the acquired vibration spectrum density data into a scatter diagram; and performing curve fitting on the scattered point data by adopting a least square method to obtain a digita