Optical fiber current transformer error compensation method based on neural network

Abstract

The invention discloses an error compensation method for an optical fiber current transformer based on a neural network. The error compensation method comprises the following steps: analyzing to obtain a state quantity of an all-optical fiber current transformer; fitting to obtain relation functions of different state quantities and transformation ratios; removing the historical state quantity data and the historical transformation ratio data of which the deviation is greater than a preset deviation threshold value; an error compensation model is constructed based on a neural network, input of an error compensation data set is a vector function composed of power grid environment parameters and state quantity data, and output of the error compensation data set is transformation ratio data; training the error compensation model by adopting the compensation data set; importing power grid environment parameters and state quantity data which are acquired in real time into the error compensation model to obtain predicted transformation ratio data; and performing error compensation on the actually measured transformation ratio data by adopting the predicted transformation ratio data. According to the invention, error compensation can be carried out on the output transformation ratio of the optical fiber current transformer, and the accuracy and reliability of monitoring data are improved.

Description

Technical Field

[0001] The invention relates to the technical field of optical fiber current transformers, and in particular to an optical fiber current transformer error compensation method based on a neural network. Background Art

[0002] The rapid development of modern industry has put forward higher requirements for the transmission and detection of power grids. The traditional means of measuring high voltage and high current will face severe tests. With the development of optical fiber technology and material science, the optical fiber current sensing system has been widely valued by people because of its good insulation and anti-interference ability, high measurement accuracy, easy miniaturization, no potential explosion hazard and other advantages.

[0003] The increase in urban electricity consumption has caused power supply equipment to be often in an overloaded pre-installed state, and the challenges faced by power supply equipment are becoming greater and greater. 60% of the failures of electronic equipment come from the power supply. With the increasing severity of power supply problems, power supply technology has gradually been valued by manufacturers. Power supply technology with sensor detection, sensor sampling, and sensor protection has gradually become a trend. Equipment to protect power supply has also been born, and sensors for detecting current or voltage have come into being. Current sensors refer to sensors that can sense the measured current and convert it into a usable output signal. They are widely used at home and abroad.

[0004] Compared with traditional electromagnetic current transformers, fiber optical current transformers (FOCT) have a series of advantages such as good insulation properties, wide bandwidth, large dynamic range, high detection accuracy, small size and light weight. They are more compatible with modern digital control and protection systems and have become an important research direction for electronic current transformers. During field operation, fiber optical current transformers (FOCT) will be affected by various external factors (such as ambient temperature, vibration, impact, electromagnetic interference, etc.) and internal factors (aging of optoelectronic device performance, increased optical path loss, etc.). These factors will cause changes in the internal state of the fiber optical current transformer, thereby reducing the accuracy and reliability of the product.

[0005] Therefore, there is an urgent need for a method for effectively compensating errors of optical fiber current transformers. Summary of the invention

[0006] The purpose of the present invention is to provide a fiber optic current transformer error compensation method based on a neural network, which can compensate the output ratio of the fiber optic current transformer according to the internal state quantity and environmental parameters of the full fiber optic current transformer, thereby improving the accuracy and reliability of monitoring data.

[0007] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0008] A method for compensating an optical fiber current transformer error based on a neural network, the method comprising the following steps:

[0009] S1, obtaining factory parameters and characteristic parameters of the all-optical current transformer, and analyzing and obtaining the state quantity of the all-optical current transformer;

[0010] S2, obtaining power grid environment parameters and historical state quantity data and historical transformation ratio data of the all-optical current transformer under different power grid environment parameters;

[0011] S3, performing correlation analysis on historical state quantity data and historical transformation ratio data, and fitting relationship functions between different state quantities and transformation ratios; using the relationship function between the state quantity and the transformation ratio, removing historical state quantity data and historical transformation ratio data whose deviation is greater than a preset deviation threshold;

[0012] S4, using the power grid environment parameters, state quantity data and transformation ratio data as a group of samples, constructing a compensation data set;

[0013] S5, constructing an error compensation model based on a neural network, wherein the input of the error compensation data set is a vector function composed of power grid environment parameters and state quantity data, and the output is transformation ratio data; the error compensation model is trained using the compensation data set;

[0014] S6, importing the real-time collected power grid environment parameters and state quantity data into the error compensation model to obtain predicted transformation ratio data; using the predicted transformation ratio data to perform error compensation on the measured transformation ratio data.

[0015] Furthermore, the state quantities of the all-optical current transformer include temperature data, power data, current and voltage data, and compensation data; wherein the temperature data includes the temperature data of each component of the all-optical current transformer and the environmental data of the all-optical current transformer; the power data includes the transmission power, the return power, and the return optical power; the current and voltage data include the modulation voltage applied by the modulator, the modulation phase, and the driving current of the light source; the compensation data includes the sensor ring compensation and the optical module compensation.

[0016] Furthermore, the compensation data set is divided into a validation set and a training set in a ratio of 1:9.

[0017] Furthermore, the difference between the predicted transformation ratio data and the measured transformation ratio data is calculated. If the absolute value of the difference is greater than a first preset difference threshold, the predicted transformation ratio data is used to perform error compensation on the measured transformation ratio data. Otherwise, the measured transformation ratio data is directly output.

[0018] Furthermore, the process of using the predicted transformation ratio data to perform error compensation on the measured transformation ratio data includes the following steps:

[0019] Obtain the grid environment parameters and state quantity data of the previous N acquisition cycles, import them into the error compensation model in turn, and obtain the predicted transformation ratio data of the previous N acquisition cycles;

[0020] Calculate the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles;

[0021] Analyze the difference:

[0022] If the absolute values ​​of the differences are all greater than the second preset difference threshold, the predicted transformation ratio data is directly output;

[0023] If the absolute value of the difference shows an increasing trend, the average value of the absolute value of the difference in the last M acquisition cycles is selected as the error value, the measured transformation ratio data is compensated to make it approach the predicted transformation ratio data, and the compensated measured transformation ratio data is output;

[0024] Otherwise, the measured transformation ratio data is compensated by using the average value of the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles, and the compensated measured transformation ratio data is output.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] The optical fiber current transformer error compensation method based on neural network of the present invention can perform error compensation on the output ratio of the optical fiber current transformer according to the internal state quantity and environmental parameters of the full optical fiber current transformer, thereby improving the accuracy and reliability of monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a flowchart of the optical fiber current transformer error compensation method based on neural network. DETAILED DESCRIPTION

[0028] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0029] See also Figure 1 The present invention discloses a method for compensating an optical fiber current transformer error based on a neural network, and the method for compensating an optical fiber current transformer error comprises the following steps:

[0030] S1, obtaining factory parameters and characteristic parameters of the all-optical current transformer, and analyzing and obtaining the state quantity of the all-optical current transformer.

[0031] In the present invention, the state quantity of the all-optical current transformer includes temperature data, power data, current and voltage data, and compensation data; wherein the temperature data includes the temperature data of each component of the all-optical current transformer and the environmental data of the all-optical current transformer; the power data includes the transmission power, the return power, and the return optical power; the current and voltage data include the modulation voltage applied by the modulator, the modulation phase, and the driving current of the light source; the compensation data includes the sensor ring compensation and the optical module compensation.

[0032] S2, obtaining power grid environment parameters and historical state quantity data and historical transformation ratio data of the all-optical current transformer under different power grid environment parameters.

[0033] S3, performing correlation analysis on historical state quantity data and historical transformation ratio data, and fitting relationship functions between different state quantities and transformation ratios; using the relationship function between the state quantity and the transformation ratio, removing historical state quantity data and historical transformation ratio data whose deviation is greater than a preset deviation threshold.

[0034] S4, using the power grid environment parameters, state quantity data and transformation ratio data as a group of samples, a compensation data set is constructed.

[0035] S5, constructing an error compensation model based on a neural network, wherein the input of the error compensation data set is a vector function composed of power grid environment parameters and state quantity data, and the output is transformation ratio data; the error compensation model is trained using the compensation data set; preferably, the compensation data set is divided into a verification set and a training set in a ratio of 1:9, the error compensation model is first trained using the training set, and then the accuracy of the error compensation model is verified using the verification set until the accuracy of the error compensation model meets the requirements.

[0036] S6, importing the real-time collected power grid environment parameters and state quantity data into the error compensation model to obtain predicted transformation ratio data; using the predicted transformation ratio data to perform error compensation on the measured transformation ratio data.

[0037] For example, the difference between the predicted ratio data and the measured ratio data is calculated. If the absolute value of the difference is greater than a first preset difference threshold, the predicted ratio data is used to perform error compensation on the measured ratio data. Otherwise, the measured ratio data is directly output.

[0038] In the present invention, the process of using the predicted transformation ratio data to perform error compensation on the measured transformation ratio data comprises the following steps:

[0039] Obtain the grid environment parameters and state quantity data of the previous N acquisition cycles, import them into the error compensation model in turn, and obtain the predicted transformation ratio data of the previous N acquisition cycles;

[0040] Calculate the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles;

[0041] Analyze the difference:

[0042] If the absolute values ​​of the differences are all greater than the second preset difference threshold, the predicted transformation ratio data is directly output;

[0043] If the absolute value of the difference shows an increasing trend, the average value of the absolute value of the difference in the last M acquisition cycles is selected as the error value, the measured transformation ratio data is compensated to make it approach the predicted transformation ratio data, and the compensated measured transformation ratio data is output;

[0044] Otherwise, the measured transformation ratio data is compensated by using the average value of the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles, and the compensated measured transformation ratio data is output.

[0045] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0046] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0047] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0048] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0049] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0050] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for compensating optical fiber current transformer error based on neural network, characterized in that: The optical fiber current transformer error compensation method comprises the following steps: S1, obtaining factory parameters and characteristic parameters of the all-optical current transformer, and analyzing and obtaining the state quantity of the all-optical current transformer; S2, obtaining power grid environment parameters and historical state quantity data and historical transformation ratio data of the all-optical current transformer under different power grid environment parameters; S3, performing correlation analysis on historical state quantity data and historical transformation ratio data, and fitting relationship functions between different state quantities and transformation ratios; using the relationship function between the state quantity and the transformation ratio, removing historical state quantity data and historical transformation ratio data whose deviation is greater than a preset deviation threshold; S4, using the power grid environment parameters, state quantity data and transformation ratio data as a group of samples, constructing a compensation data set; S5, constructing an error compensation model based on a neural network, wherein the input of the error compensation data set is a vector function composed of power grid environment parameters and state quantity data, and the output is transformation ratio data; the error compensation model is trained using the compensation data set; S6, importing the real-time collected power grid environment parameters and state quantity data into the error compensation model to obtain predicted transformation ratio data; using the predicted transformation ratio data to perform error compensation on the measured transformation ratio data.

2. The optical fiber current transformer error compensation method based on neural network according to claim 1, characterized in that: The state quantity of the all-optical current transformer includes temperature data, power data, current and voltage data and compensation data; wherein the temperature data includes the temperature data of each component of the all-optical current transformer and the environmental data of the all-optical current transformer; the power data includes the transmission power, the return power and the return optical power; the current and voltage data includes the modulation voltage applied by the modulator, the modulation phase and the driving current of the light source; The compensation data includes sensor ring compensation and optical module compensation.

3. The optical fiber current transformer error compensation method based on neural network according to claim 1, characterized in that: The compensation dataset is divided into a validation set and a training set in a ratio of 1:

9.

4. The optical fiber current transformer error compensation method based on neural network according to claim 1, characterized in that: The difference between the predicted ratio data and the measured ratio data is calculated. If the absolute value of the difference is greater than a first preset difference threshold, the predicted ratio data is used to perform error compensation on the measured ratio data. Otherwise, the measured ratio data is directly output.

5. The method for optical fiber current transformer error compensation based on neural network according to claim 1, characterized in that: The process of using the predicted ratio data to perform error compensation on the measured ratio data includes the following steps: Obtain the grid environment parameters and state quantity data of the previous N acquisition cycles, import them into the error compensation model in turn, and obtain the predicted transformation ratio data of the previous N acquisition cycles; Calculate the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles; Analyze the difference: If the absolute values ​​of the differences are all greater than the second preset difference threshold, the predicted transformation ratio data is directly output; If the absolute value of the difference shows an increasing trend, the average value of the absolute value of the difference in the last M acquisition cycles is selected as the error value, the measured transformation ratio data is compensated to make it approach the predicted transformation ratio data, and the compensated measured transformation ratio data is output; Otherwise, the measured transformation ratio data is compensated by using the average value of the difference between the predicted transformation ratio data and the measured transformation ratio data in the first N acquisition cycles, and the compensated measured transformation ratio data is output.

Images