Optical fiber current transformer phase demodulation method and system, terminal and storage medium

Abstract

The invention discloses an optical fiber current transformer phase demodulation method and system, a terminal and a storage medium, and the method comprises the steps: obtaining a light intensity signal outputted by an all-optical fiber current transformer, a frequency doubling carrier signal and a frequency doubling carrier signal, converting the light intensity signal into an output signal, and carrying out the signal internal modulation, obtaining a first tangent signal and a second tangent signal; performing ellipse fitting on discrete sampling points of the first tangent signal and the second tangent signal to obtain a first initial ellipse parameter set; inputting the first tangent signal and the second tangent signal into the trained neural network for elliptic surface fitting to obtain a second initial elliptic parameter set; and comparing the first initial ellipse parameter with the second initial ellipse parameter to obtain an optimal ellipse parameter, and calculating and processing by combining the optimal ellipse parameter to obtain a correct demodulation signal. According to the invention, the nonlinear distortion of the signal can be effectively suppressed, the demodulation linearity is good, the accuracy is high, and the sensing precision of the all-fiber current transformer is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of all-optical fiber current transformer signal demodulation, and more specifically, relates to an optical fiber current transformer phase demodulation method and system. Background Art

[0002] With the progress of social intelligence and digitalization, higher requirements are put forward for the power transmission industry with growing electricity consumption. Therefore, it is very important to monitor the power system in a timely and effective manner. The most important of these is current measurement, which is a key factor in maintaining the normal operation of the power system. Correspondingly, pulse current method, radio frequency method, fiber optic sensing method, etc. have emerged. The current transformer needs to meet the following harsh conditions: no chemical reaction with the insulating medium (oil or gas) under specific conditions inside the power equipment, high insulation performance, long-term stable operation, and very small size. Among them, sensors based on fiber optic sensing can meet the above application scenarios. The all-fiber current transformer based on the Faraday effect has the advantages of high sensitivity, small size, and strong anti-electromagnetic interference ability, and has good application scenarios in the field of power monitoring.

[0003] Designing a signal demodulation algorithm with good performance is an important part of the design of all-fiber current transformers. Common demodulation methods are mainly divided into three types according to their modulation principles: intensity demodulation, wavelength demodulation, and phase demodulation. Among them, the phase demodulation method has received widespread attention. According to the way of adding carriers, it is divided into internal modulation and external modulation. The external modulation method uses piezoelectric ceramics, electro-optical modulators and other devices to phase modulate the output light, but the external modulation method increases the system volume and reduces the system's ability to resist electromagnetic interference. Internal modulation achieves wavelength (frequency) modulation by adjusting the laser current. However, internal modulation will cause a ripple to be superimposed on the originally stable optical power, introducing the problem of associated amplitude modulation. At the same time, factors such as electromagnetic interference, optical path transmission, and circuit conversion will cause carrier phase delay. In addition, there is a nonlinear factor, the carrier phase modulation depth C, which is related to the EFPI cavity length and laser modulation characteristics. The above nonlinear factors will reduce the accuracy of the demodulation algorithm or even cause demodulation failure, and the signal required for current detection cannot be output from the sensor.

[0004] In the prior art, the phase modulation depth is extracted based on the carrier phase delay and the ellipse fitting parameters, and the phase modulation depth is fed back to the light source modulation module for closed-loop feedback control of the phase modulation depth, thereby realizing signal demodulation.

[0005] The prior art still has at least the following technical problems:

[0006] 1. The ellipse fitting parameters are easily affected by the fluctuation of signal values, resulting in a large deviation from the ideal value, and the system has poor robustness;

[0007] 2. In the prior art, an additional sinusoidal frequency-doubled signal is introduced to calculate the phase modulation depth value for feedback adjustment. The calculation process is easily affected by signal fluctuations, resulting in large fluctuations, which leads to errors in the calculated phase modulation depth value. Feedback adjustment using the wrong phase modulation depth value leads to errors in demodulation. Summary of the invention

[0008] In order to solve the deficiencies in the prior art, the present invention provides a phase demodulation method for an optical fiber current transformer, which can solve the technical problem in the prior art that the phase demodulation signal is inaccurate due to nonlinear factors.

[0009] The present invention adopts the following technical solution.

[0010] A method for phase demodulation of an optical fiber current transformer comprises the following steps:

[0011] The optical intensity signal output by the all-optical current transformer is obtained, as well as the single-frequency carrier signal and the double-frequency carrier signal generated by the carrier generator for carrier modulation of the output signal of the all-optical current transformer, and the optical intensity signal is converted into an output signal I d ;

[0012] For output signal I d , a one-time carrier signal and a two-time carrier signal are used for signal intra-modulation to obtain a first tangent signal P(t) and a second tangent signal Q(t);

[0013] Performing ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set;

[0014] Constructing a neural network and training it, inputting the first tangent signal P(t) and the second tangent signal Q(t) into the trained neural network to perform elliptical surface fitting, and obtaining a second initial elliptical parameter set;

[0015] The first initial ellipse parameter set and the second initial ellipse parameter set are compared to obtain the optimal ellipse parameters, and the correct demodulation signal is obtained by calculating and processing in combination with the optimal ellipse parameters.

[0016] Preferably, obtaining the light intensity signal output by the all-optical current transformer specifically includes:

[0017] The light intensity signal output by the all-fiber current transformer is the light intensity result after the interference of two coherent light beams I1 and I2:

[0018]

[0019] In the formula, A and B are constant terms, A=I1+I2, C is the carrier phase modulation depth, ω0 is the carrier frequency, It is the phase caused by the external influence on the sensor;

[0020] The light intensity signal is converted into an electrical signal by a photodetector as the output signal I d .

[0021] Preferably, the output signal I d , a one-time carrier signal and a two-time carrier signal are modulated intra-signally to obtain a first tangent signal P(t) and a second tangent signal Q(t), specifically including:

[0022] The output signal Id, the one-time carrier signal and the two-time carrier signal are subjected to signal intra-modulation by a carrier generator to obtain a first tangent signal P(t) and a second tangent signal Q(t);

[0023] The signal internal modulation includes: multiplying the output signal Id, the one-time carrier signal and the two-time carrier signal and filtering them through a low-pass filter.

[0024] Preferably, performing ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set specifically includes:

[0025] The direct least square method based on the block matrix is ​​used to perform ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), wherein the ellipse implicit equation used in the ellipse fitting satisfies:

[0026] K(e,x i )=ex 2 +Exy+Fy 2 +Lx+My+N=0

[0027] In the formula, K(e, x i ) represents the point (x i ,y i ) to the quadratic curve K(e, x i ), the algebraic distance of the point (x i ,y i ) are the coordinates of the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively;

[0028] The ellipse constraint conditions are set to obtain the best fitting state of the ellipse, and the parameters E, F, L, M, and N of the ellipse under the best fitting state are obtained as the first initial ellipse parameter set.

[0029] Preferably, constructing a neural network and training it specifically includes:

[0030] Construct a BP neural network, which includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to fit any quadratic surface.

[0031] Obtain a training data set, specifically, use a data acquisition system to collect the output signal of the all-optical current transformer and the first-harmonic frequency and second-harmonic frequency carrier signals to calculate the first tangent signal P(t) and the second tangent signal Q(t), and obtain the corresponding actual phase, with the first tangent signal P(t) and the second tangent signal Q(t) as input, and the ellipse fitting parameters E, F, L, M, and N as output;

[0032] The training data set is sent to the BP neural network for training, and the actual phase is combined to determine whether the training is completed.

[0033] Preferably, comparing the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameters specifically includes:

[0034] Conduct a calibration test under actual working conditions, test and record the minimum sum of squared algebraic distances obtained by fitting the first initial ellipse, calculate the phase signal parameters using the first initial ellipse parameters, compare the calculated phase signal parameters with the actual phase values, and obtain the distortion of the phase signal parameters under different minimum sums of squared algebraic distances.

[0035] The threshold of the minimum sum of squared algebraic distances is determined according to the distortion requirements of the system's working conditions. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set, otherwise the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

[0036] Preferably, the correct demodulation signal is obtained by calculating and processing the optimal ellipse parameters, specifically including:

[0037] According to the first tangent signal P(t), the second tangent signal Q(t) and the parameters E, F, and L in the optimal ellipse parameter set, a demodulated signal value under the influence of nonlinear factors is obtained by an inverse tangent operation;

[0038] The demodulated signal value affected by nonlinear factors is subjected to high-pass filtering to obtain the correct demodulated signal.

[0039] Preferably, the demodulated signal value under the influence of nonlinear factors The calculation formula is as follows:

[0040]

[0041] Where P(t) and Q(t) represent the first tangent signal and the second tangent signal respectively, E, F, and L are the parameters of the elliptic implicit equation in the optimal elliptic parameter set, represents the correct demodulated signal, and θ2 represents the component affected by nonlinear factors.

[0042] A fiber optic current transformer phase demodulation system includes: a signal acquisition module, a signal processing module, an ellipse fitting module, a neural network module, a parameter comparison module and a signal calculation module;

[0043] The signal acquisition module is used to obtain the signal data of the all-optical current transformer;

[0044] The signal processing module is used to perform internal modulation on the light intensity signal through the carrier generator to generate a single-frequency carrier signal and a double-frequency carrier signal, and to generate a single-frequency carrier signal and a double-frequency carrier signal. d , a one-time carrier signal and a two-time carrier signal are subjected to signal intra-modulation and filtering to obtain a first tangent signal and a second tangent signal;

[0045] The ellipse fitting module is used to perform ellipse fitting on discrete sampling points of the first tangent signal and the second tangent signal to obtain a first initial ellipse parameter set;

[0046] The neural network module is used to obtain a second initial ellipse parameter set according to the first tangent signal, the second tangent signal and the trained BP neural network;

[0047] The parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set;

[0048] The signal calculation module calculates and processes the optimal ellipse parameter set to obtain the correct demodulation signal.

[0049] Preferably, the signal acquisition module acquires the light intensity signal output by the all-optical current transformer, specifically including:

[0050] The light intensity signal output by the all-fiber current transformer is the light intensity result after the interference of two coherent light beams I1 and I2:

[0051]

[0052] In the formula, A and B are constant terms, A=I1+I2, C is the carrier phase modulation depth, ω0 is the carrier frequency, It is the phase caused by the external influence on the sensor;

[0053] The light intensity signal is converted into an electrical signal by a photodetector as the output signal I d .

[0054] Preferably, the signal processing module processes the output signal I d , a one-time carrier signal and a two-time carrier signal are modulated intra-signally to obtain a first tangent signal P(t) and a second tangent signal Q(t), specifically including:

[0055] The output signal I d , a one-time carrier signal and a two-time carrier signal, and perform signal intra-modulation through a carrier generator to obtain a first tangent signal P(t) and a second tangent signal Q(t);

[0056] The signal internal modulation includes: d , the first-time carrier signal and the second-time carrier signal are multiplied and then filtered by a low-pass filter.

[0057] Preferably, the ellipse fitting module performs ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set, specifically including:

[0058] The direct least square method based on the block matrix is ​​used to perform ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), wherein the ellipse implicit equation used in the ellipse fitting satisfies:

[0059] K(e,x i )=ex 2 +Exy+Fy 2 +Lx+My+N=0

[0060] In the formula, K(e, x i ) represents the point (x i ,y i ) to the quadratic curve K(e, x i ), the algebraic distance of the point (x i ,y i ) are the coordinates of the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively;

[0061] The ellipse constraint conditions are set to obtain the best fitting state of the ellipse, and the parameters E, F, L, M, and N of the ellipse under the best fitting state are obtained as the first initial ellipse parameter set.

[0062] Preferably, the neural network module is used to construct a neural network and perform training, specifically including:

[0063] Construct a BP neural network, which includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to fit any quadratic surface.

[0064] Obtain a training data set, specifically, use a data acquisition system to collect the output signal of the all-optical current transformer and the first-harmonic frequency and second-harmonic frequency carrier signals to calculate the first tangent signal P(t) and the second tangent signal Q(t), and obtain the corresponding actual phase, with the first tangent signal P(t) and the second tangent signal Q(t) as input, and the ellipse fitting parameters E, F, L, M, and N as output;

[0065] The training data set is sent to the BP neural network for training, and the actual phase is combined to determine whether the training is completed.

[0066] Preferably, the parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set, specifically including:

[0067] Conduct a calibration test under actual working conditions, test and record the minimum sum of squared algebraic distances obtained by fitting the first initial ellipse, calculate the phase signal parameters using the first initial ellipse parameters, compare the calculated phase signal parameters with the actual phase values, and obtain the distortion of the phase signal parameters under different minimum sums of squared algebraic distances.

[0068] The threshold of the minimum sum of squared algebraic distances is determined according to the distortion requirements of the system's working conditions. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set, otherwise the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

[0069] Preferably, the signal calculation module is used to calculate and process the optimal ellipse parameters to obtain a correct demodulation signal, specifically including:

[0070] According to the first tangent signal P(t), the second tangent signal Q(t) and the parameters E, F, and L in the optimal ellipse parameter set, a demodulated signal value under the influence of nonlinear factors is obtained by an inverse tangent operation;

[0071] The demodulated signal value affected by nonlinear factors is subjected to high-pass filtering to obtain the correct demodulated signal.

[0072] Preferably, the signal calculation module obtains the demodulated signal value under the influence of nonlinear factors through inverse tangent operation, specifically including: the demodulated signal value under the influence of nonlinear factors The calculation formula is as follows:

[0073]

[0074] Where P(t) and Q(t) represent the first tangent signal and the second tangent signal respectively, E, F, and L are the parameters of the elliptic implicit equation in the optimal elliptic parameter set, represents the correct demodulated signal, and θ2 represents the component affected by nonlinear factors.

[0075] The present invention also provides a terminal, comprising a processor and a storage medium;

[0076] The storage medium is used to store instructions;

[0077] The processor is used to operate according to the instructions to execute the steps of the optical fiber current transformer phase demodulation method.

[0078] The present invention also proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the optical fiber current transformer phase demodulation method are implemented.

[0079] The beneficial effect of the present invention is that, compared with the prior art, the present invention proposes an internal modulation phase demodulation method for an all-optical current transformer, which generates a carrier and constructs a pair of strictly tangent tangent signals using a direct least squares method based on a block matrix and a neural network, and demodulates the required phase signal after inverse tangent, which can effectively suppress the influence of nonlinear factors in the phase demodulation signal and effectively improve the accuracy of the demodulation algorithm; the present invention introduces a neural network to perform parameter fitting for signals with large fluctuations using the advantage of the high robustness of the neural network, thereby reducing the influence of signal value fluctuations and thus reducing demodulation errors; when the phase modulation depth value deviates greatly from the ideal value, a neural network fitting method is used to demodulate the signal, thereby avoiding demodulation errors caused by phase modulation depth errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 It is a flow chart of the optical fiber current transformer phase demodulation method in the present invention;

[0081] Figure 2 It is a principle diagram of the optical fiber current transformer phase demodulation method in the present invention;

[0082] Figure 3 It is a schematic diagram of the neural network structure used in the present invention;

[0083] Figure 4 It is a structural diagram of the optical fiber current transformer phase demodulation system in the present invention;

[0084] Figure 5 It is a schematic diagram of simulation results under an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only embodiments of a part of the present invention, rather than all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the protection scope of the present invention.

[0086] like Figure 1 , 2 As shown, the present invention proposes a fiber optic current transformer phase demodulation method, the method comprising the following steps:

[0087] S1, obtain the light intensity signal I output by the all-optical current transformer, and the single-frequency carrier signal and double-frequency carrier signal generated by the carrier generator for carrier modulating the output signal of the all-optical current transformer, and convert the light intensity signal into the output signal I d ;

[0088] The optical intensity signal I output by the all-fiber current transformer, the single-frequency carrier signal and the double-frequency carrier signal generated by the carrier generator are collected. The optical intensity signal I output by the all-fiber current transformer is obtained by the interference of two coherent light beams, and then the electrical signal obtained after the conversion by the photoelectric detector is used as the output signal I d , specifically:

[0089] The light intensity signal output by the all-fiber current transformer is the light intensity result after the two coherent light beams 11 and 12 interfere with each other:

[0090]

[0091] Where A and B are the first constant and the second constant respectively, C is the carrier phase modulation depth, ω0 is the carrier frequency, is the phase generated by the external influence on the sensor, and the first constant A and the second constant B satisfy:

[0092] A=I1+I2,

[0093] Furthermore, the carrier frequency is ω0, and the one-time carrier signal and the two-time carrier signal generated by the carrier generator are Gcosω0t and Hcos2ω0t, respectively, where G and H are the amplitudes of the one-time carrier frequency and the two-time carrier frequency, respectively.

[0094] S2, for output signal I d , a one-time carrier signal and a two-time carrier signal are modulated and filtered to obtain a pair of theoretically tangent signals P(t) and Q(t);

[0095] Specifically, according to the output signal I obtained by S1 d , a one-time carrier signal and a two-time carrier signal, and perform signal internal modulation through a carrier generator to obtain a pair of theoretically tangent signals: a first tangent signal P(t) and a second tangent signal Q(t);

[0096] The signal internal modulation includes: the output signal I collected in S1 d , the one-time carrier signal and the two-time carrier signal are multiplied and then passed through a low-pass filter to obtain a pair of theoretically tangent signals P(t) and Q(t), which specifically include:

[0097] In the present invention, the first tangent signal P(t) and the second tangent signal Q(t) are respectively the output signals I obtained after the photoelectric conversion of the light intensity signal I after interference without the influence of nonlinear factors. d The first tangent signal P(t) and the second tangent signal Q(t) obtained by multiplying the carrier frequency signal by one times and the carrier frequency signal by two times and passing through a low-pass filter should ideally satisfy:

[0098]

[0099] In the formula, G and H are the values ​​of the carrier frequency multiplied by one and the carrier frequency multiplied by two, and J n (C) is the Bessel function term related to C, J n (C) is the Bessel function term related to the carrier phase modulation depth C value. To facilitate the inverse tangent operation, the C value in the PGC-Atan algorithm takes the optimal value of 2.63 rad, at which time J1(C)=J2(C) is satisfied, and the change rate of J1(C) / J2(C) is the smallest.

[0100] The above expressions of P(t) and Q(t) are expressions that are not affected by the environment, and the two parameters are orthogonal.

[0101] However, due to the modulation in the light source, if the current is unstable, the optical frequency under the current modulation will also cause the optical power to change, so that an alternating ripple is superimposed on the originally stable optical power, generating an associated amplitude modulation term (1+mcosω0t), where m is the amplitude of the first harmonic of the associated amplitude modulation. In addition, factors such as optical path propagation, circuit transmission, and digital-to-analog conversion will also cause carrier phase delay θ.

[0102] Therefore, after being affected by the associated amplitude modulation and carrier phase delay, the first tangent signal P(t) and the second tangent signal Q(t) are respectively affected by the following two equations:

[0103]

[0104] Obviously, the first tangent signal P(t) and the second tangent signal Q(t) are no longer orthogonal due to nonlinear factors such as current modulation and phase delay. Therefore, a pair of orthogonal signals containing information of the signal to be measured constructed by the present invention are demodulated, and the constructed first tangent signal P(t) and the second tangent signal Q(t) satisfy:

[0105]

[0106] And: the first physical parameter a, the second physical parameter b, the third physical parameter c, the first error influence angle θ1, and the second error influence angle θ2 respectively satisfy:

[0107]

[0108]

[0109] Thus, a pair of orthogonal signals P(t) and Q(t) containing information of the signal to be measured under the influence of nonlinear factors can be obtained.

[0110] S3, performing ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) based on a direct least squares method of a block matrix to obtain a first initial ellipse parameter set;

[0111] S3 specifically includes:

[0112] The initial fitting uses the direct least squares method based on the block matrix to fit the ellipse. This algorithm has low time complexity, strong noise resistance, good stability, and the fitted ellipse is guaranteed to have a unique optimal solution. This unique optimal solution is used as the first initial ellipse parameter.

[0113] Among them, the quadratic curve of the elliptic implicit equation used in ellipse fitting satisfies:

[0114]

[0115] In the formula, K(e,x i ) represents the algebraic distance from point (xi,yi) to the quadratic curve K(e,xi), point (xi,yi) is the discrete sampling point of the first tangent signal P(t) and the second tangent signal Q(t), xi and yi represent the horizontal and vertical coordinates of the discrete sampling points, respectively. are the first elliptic vector and the second elliptic vector formed by the coefficients of the elliptic implicit equation, e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively;

[0116] Among them, the first ellipse vector and the second ellipse vector Respectively meet:

[0117]

[0118] The relationship between E, F, L, M, N and parameters a, b, c, θ1, θ2 is:

[0119]

[0120] L=-2a

[0121]

[0122] N=a 2 -b 2 cos 2 (θ1+θ2)

[0123] Furthermore, the best fitting state of the ellipse is obtained, and the ellipse parameters obtained in this state are used as the first initial ellipse parameter set. The best fitting state of the ellipse must satisfy:

[0124] (1) Algebraic distance constraints:

[0125] Based on the direct least squares method of the block matrix, the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) are fitted with an ellipse, so that the sum of the squares of the algebraic distances from each discrete sampling point to the quadratic curve K(e,xi) is minimized, and the minimum sum of the squares of the algebraic distances is It can be expressed as:

[0126]

[0127] In the formula, is an n×6 matrix,

[0128] (2) Matrix constraints:

[0129] To avoid trivial solutions, such as the zero vector, and to solve for the first elliptic vector The same quadratic curve can be represented by any multiple of , requiring some restrictions on the parameters of the ellipse. The parameters can be scaled arbitrarily, so let the constraint be 4eF-E 2 =1.

[0130] Construct a 6×6 matrix To express this constraint, the matrix constraint must satisfy The details are as follows:

[0131]

[0132] In summary, the two constraints for ellipse fitting are: and

[0133] right By differentiating and introducing the Lagrange factor λ, we can obtain the following simultaneous equations:

[0134]

[0135] structure Then there is Then we can get:

[0136]

[0137] By solving the first ellipse vector And the eigenvalues, and then the ellipse fitting parameters E, F, L, M, N can be obtained, which are used as the first initial ellipse parameter set.

[0138] S4, constructing a neural network and training it, inputting the first tangent signal P(t) and the second tangent signal Q(t) into the trained neural network to perform elliptical surface fitting, and obtaining a second initial elliptical parameter set;

[0139] The neural network in this embodiment adopts BP neural network, such as Figure 3 As shown, Figure 3 This is a typical structural diagram of a BP neural network. The BP neural network includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to achieve the fitting of any quadratic surface.

[0140] Furthermore, the training process of the BP neural network is:

[0141] Prepare training data set: Use data acquisition system to collect the output signal of all-fiber current transformer and the first and second frequency carrier signals, and the phase difference with the actual known The above formulas of S1 and S2 are used for reverse deduction, and together with the corresponding collected signals, they constitute the training data set of the neural network.

[0142] Build the network: Use tools such as Matlab and Python to design and build the corresponding neural network model, and design the input layer, hidden layer, output layer network structure and hyperparameters.

[0143] Training network: Under the hyperparameters set in the previous step, the prepared data set is sent to the designed neural network for training, and the one with the best demodulation effect is selected as the final elliptical surface fitting model.

[0144] The input of the neural network is: a first tangent signal P(t) and a second tangent signal Q(t);

[0145] The output of the neural network is: ellipse fitting parameters E, F, L, M, N;

[0146] Furthermore, the phase is calculated based on the obtained ellipse fitting parameters By comparing the actual phase with the training data set, the parameters of the neural network are adjusted to achieve its training.

[0147] Furthermore, the first tangent signal P(t) and the second tangent signal Q(t) calculated by S2 are input into the trained neural network, and the ellipse fitting parameters E, F, L, M, and N output by the neural network are used as the second initial ellipse parameter set.

[0148] S5, compare the first initial ellipse parameter set with the second initial ellipse parameter set to obtain the optimal ellipse parameter set, and calculate and filter the correct phase signal parameters by combining the optimal ellipse parameter set with the inverse tangent formula.

[0149] Specifically, comparing the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameters specifically includes:

[0150] Carry out calibration test under actual working conditions, test and record the sum of the squares of the minimum algebraic distances obtained by fitting the first initial ellipse The phase signal parameters are calculated using the first initial ellipse parameters The calculated phase signal parameters are compared with the actual phase values ​​to obtain the sum of the squares of the minimum algebraic distances at different Phase signal parameters under the condition Distortion

[0151] According to the distortion requirements of the system working conditions, a threshold of the minimum sum of squared algebraic distances is determined to judge the fitting quality. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is If the value is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set; otherwise, the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

[0152] The above steps ensure that optimized ellipse parameters are used in subsequent steps to improve the accuracy of phase demodulation.

[0153] Furthermore, the parameters a, b, c, θ1, and θ2 required for constructing a strict tangent signal are calculated based on the obtained optimal ellipse parameter set.

[0154] Phase signal parameters The acquisition specifically includes:

[0155] According to the formula, the constructed result can be demodulated to get the corresponding phase information by performing an inverse tangent operation:

[0156] Wherein, the first tangent signal P(t) and the second tangent signal Q(t) respectively satisfy:

[0157]

[0158] Wherein, a, b, c, θ1, θ2 represent the first intermediate parameter, the second intermediate parameter, the third intermediate parameter, the fourth intermediate parameter and the nonlinear factor influencing component respectively;

[0159] The relationship between E, F, L, M, N and parameters a, b, c, θ1, θ2 in the optimal ellipse parameter set is:

[0160]

[0161] L=-2a

[0162]

[0163] N=a 2 -b 2 cos 2 (θ1+θ2)

[0164] A pair of strictly tangent signals are calculated based on the first tangent signal P(t), the second tangent signal Q(t), and the parameters E, F, and L in the optimal ellipse parameter set. and

[0165]

[0166]

[0167] Use the parameters fitted above to construct a strictly tangent signal and get Specifically expressed as:

[0168]

[0169] The numerator and denominator in the above formula are a pair of strictly orthogonal signals. The demodulated signal is calculated according to the above formula: In order to avoid the influence of nonlinear factors, the demodulated signal obtained after the arc tangent operation is Then high-pass filtering is performed to remove the nonlinear factor affecting the component θ2 to obtain the correct demodulated signal. And output.

[0170] like Figure 4As shown, the present invention also proposes a fiber optic current transformer phase demodulation system for implementing the above fiber optic current transformer phase demodulation method, the system comprising: a signal acquisition module, a signal processing module, an ellipse fitting module, a neural network module, a parameter comparison module and a signal calculation module;

[0171] The signal acquisition module is used to obtain the signal data of the all-optical current transformer;

[0172] The signal processing module is used to perform internal modulation on the light intensity signal through the carrier generator to generate a single-frequency carrier signal and a double-frequency carrier signal, and to generate a single-frequency carrier signal and a double-frequency carrier signal. d , a one-time carrier signal and a two-time carrier signal are subjected to signal intra-modulation and filtering to obtain a first tangent signal and a second tangent signal;

[0173] The ellipse fitting module is used to perform ellipse fitting on discrete sampling points of the first tangent signal and the second tangent signal to obtain a first initial ellipse parameter set;

[0174] The neural network module is used to obtain a second initial ellipse parameter set according to the first tangent signal, the second tangent signal and the trained BP neural network;

[0175] The parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set;

[0176] The signal calculation module calculates and processes the optimal ellipse parameter set to obtain the correct demodulation signal.

[0177] Among them, the signal acquisition module obtains the light intensity signal output by the all-optical current transformer, specifically including:

[0178] The light intensity signal I output by the all-fiber current transformer is the light intensity result after the interference of two coherent light beams I1 and I2:

[0179]

[0180] In the formula, A and B are constant terms, A=I1+I2, C is the carrier phase modulation depth, ω0 is the carrier frequency, It is the phase caused by the external influence on the sensor;

[0181] The light intensity signal is converted into an electrical signal by a photodetector as the output signal I d .

[0182] The signal processing module outputs the signal I d , a one-time carrier signal and a two-time carrier signal are modulated intra-signally to obtain a first tangent signal P(t) and a second tangent signal Q(t), specifically including:

[0183] The output signal I d , a one-time carrier signal and a two-time carrier signal, and perform signal intra-modulation through a carrier generator to obtain a first tangent signal P(t) and a second tangent signal Q(t);

[0184] The signal internal modulation includes: d , the first-time carrier signal and the second-time carrier signal are multiplied and then filtered by a low-pass filter.

[0185] The ellipse fitting module performs ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set, which specifically includes:

[0186] The direct least square method based on the block matrix is ​​used to perform ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), wherein the ellipse implicit equation used in the ellipse fitting satisfies:

[0187] K(e,x i )=ex 2 +Exy+Fy 2 +Lx+My+N=0

[0188] In the formula, K(e, x i ) represents the point (x i ,y i ) to the quadratic curve K(e, x i ), the algebraic distance of the point (x i ,y i ) are the coordinates of the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively;

[0189] The ellipse constraint conditions are set to obtain the best fitting state of the ellipse, and the parameters E, F, L, M, and N of the ellipse under the best fitting state are obtained as the first initial ellipse parameter set.

[0190] The neural network module is used to build and train neural networks, including:

[0191] Construct a BP neural network, which includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to fit any quadratic surface.

[0192] Obtain a training data set, specifically, use a data acquisition system to collect the output signal of the all-optical current transformer and the first-harmonic frequency and second-harmonic frequency carrier signals to calculate the first tangent signal P(t) and the second tangent signal Q(t), and obtain the corresponding actual phase, with the first tangent signal P(t) and the second tangent signal Q(t) as input, and the ellipse fitting parameters E, F, L, M, and N as output;

[0193] The training data set is sent to the BP neural network for training, and the actual phase is combined to determine whether the training is completed.

[0194] The parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set, specifically including:

[0195] Conduct a calibration test under actual working conditions, test and record the minimum sum of squared algebraic distances obtained by fitting the first initial ellipse, calculate the phase signal parameters using the first initial ellipse parameters, compare the calculated phase signal parameters with the actual phase values, and obtain the distortion of the phase signal parameters under different minimum sums of squared algebraic distances.

[0196] The threshold of the minimum sum of squared algebraic distances is determined according to the distortion requirements of the system's working conditions. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set, otherwise the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

[0197] The signal calculation module is used to calculate and process the optimal ellipse parameters to obtain the correct demodulation signal, including:

[0198] According to the first tangent signal P(t), the second tangent signal Q(t) and the parameters E, F, and L in the optimal ellipse parameter set, a demodulated signal value under the influence of nonlinear factors is obtained by an inverse tangent operation;

[0199] The demodulated signal value affected by nonlinear factors is subjected to high-pass filtering to obtain the correct demodulated signal.

[0200] The signal calculation module obtains the demodulated signal value under the influence of nonlinear factors through inverse tangent operation, specifically including: the demodulated signal value under the influence of nonlinear factors The calculation formula is as follows:

[0201]

[0202] Where, P9t) and Q(t) represent the first tangent signal and the second tangent signal respectively, E, F, and L are the parameters of the elliptic implicit equation in the optimal elliptic parameter set, represents the correct demodulated signal, and θ2 represents the component affected by nonlinear factors.

[0203] In order to verify the beneficial effects of the present invention, the following simulation examples are used for verification and explanation:

[0204] The signal acquisition module includes a photoelectric detector, an acquisition card, and a computer; the 1.6kHz signal generated by the FP interferometer is converted into an electrical signal by the photoelectric detector, and the signal can be transmitted to the computer by an acquisition card model PICO4225A at a sampling rate of 1Ms / s.

[0205] The carrier signal can be generated by a signal generator. In a specific implementation, the carrier frequency ω0 can be 50kHz. The collected output signal is multiplied by a 1-fold signal of 50kHz and a 2-fold signal of 100kHz, and then sent to a low-pass filter. The low-pass filter is an elliptical filter, the passband cutoff frequency can be 6kHz, the stopband cutoff frequency is 8kHz, the passband attenuation coefficient is 0.1dB, the stopband attenuation coefficient is 85dB, and the filter order is 9. At the same time, the setting parameter of the phase modulation depth C is >1.2rad. The simulation is performed in Labview, and the simulation results are as follows: Figure 5 As shown. When the carrier phase delay is set to 40° and the current modulation degree is m, that is, the influence of the associated amplitude modulation and phase modulation depth deviation, the impact on the demodulation signal-to-noise ratio and the demodulated signal. From the simulation results, it is known that the PGC-BP-Atan algorithm can effectively eliminate the influence of three nonlinear factors. When C≥1rad, m takes any value in the range of 0 to 1, and the demodulation amplitude and SINAD of the algorithm are basically stable. There is an error of about 0.04rad between the demodulation amplitude and the measured value of 1rad, which may be related to the accuracy of filtering and elliptical fitting. What is filtered out by the low-pass filter are high-order harmonics, carriers and their sideband signals.

[0206] like Figure 5 As shown, Figure 5 This is a comparison chart of the anti-nonlinear simulation results of the PGC-BP-Atan algorithm. Figure 5 (a) is the relationship between the current modulation and the demodulation result. Figure 5 (b) is the relationship between the current modulation index and the signal-to-noise ratio. In the simulation example, the carrier phase delay is set to 40° and the current modulation index is m, that is, the influence of the associated amplitude modulation and phase modulation depth deviation on the demodulation signal-to-noise ratio and the demodulated signal.

[0207] From the simulation results, it can be seen that the PGC-BP-Atan algorithm can effectively eliminate the influence of three nonlinear factors. When C≥1rad, m takes any value in the range of 0 to 1, and the demodulation amplitude and SINAD of the algorithm are basically stable. There is an error of about 0.04rad between the demodulation amplitude and the measured value of 1rad, which is much smaller than the error between the demodulation amplitude and the measured value obtained under the prior art. It can be seen that the present invention can improve the accuracy of demodulation.

[0208] The beneficial effect of the present invention is that, compared with the prior art, the present invention introduces a neural network to perform parameter fitting for signals with large fluctuations, utilizing the advantage of the high robustness of the neural network, thereby reducing the impact of signal value fluctuations and thus reducing demodulation errors; when the phase modulation depth value deviates greatly from the ideal value, a neural network fitting method is used to demodulate the signal, thereby avoiding demodulation errors caused by phase modulation depth errors.

[0209] The present invention also provides a terminal, including a processor and a storage medium;

[0210] Wherein, the storage medium is used to store instructions;

[0211] The processor is used to operate according to the instructions to execute the steps of the above-mentioned optical fiber current transformer phase demodulation method.

[0212] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0213] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0214] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0215] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for phase demodulation of an optical fiber current transformer, characterized in that: The following steps are involved: The optical intensity signal output by the all-optical current transformer is obtained, as well as the single-frequency carrier signal and the double-frequency carrier signal generated by the carrier generator for carrier modulation of the output signal of the all-optical current transformer, and the optical intensity signal is converted into an output signal I d ; For output signal I d , a one-time carrier signal and a two-time carrier signal are used for signal intra-modulation to obtain a first tangent signal P(t) and a second tangent signal Q(t); Performing ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set; Constructing a neural network and training it, inputting the first tangent signal P(t) and the second tangent signal Q(t) into the trained neural network to perform elliptical surface fitting, and obtaining a second initial elliptical parameter set; The first initial ellipse parameter set and the second initial ellipse parameter set are compared to obtain the optimal ellipse parameters, and the correct demodulation signal is obtained by calculating and processing in combination with the optimal ellipse parameters.

2. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: Obtain the light intensity signal output by the all-fiber current transformer, including: The light intensity signal I output by the all-fiber current transformer is the light intensity result after the interference of two coherent light beams I1 and I2: In the formula, A and B are constant terms, A=i1+i2, C is the carrier phase modulation depth, ω0 is the carrier frequency, It is the phase caused by the external influence on the sensor; The light intensity signal I is converted by the photodetector to obtain an electrical signal as the output signal I d .

3. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: For output signal I d , a one-time carrier signal and a two-time carrier signal are modulated intra-signally to obtain a first tangent signal P(t) and a second tangent signal Q(t), specifically including: The output signal I d , a one-time carrier signal and a two-time carrier signal, and perform signal intra-modulation through a carrier generator to obtain a first tangent signal P(t) and a second tangent signal Q(t); The signal internal modulation includes: d , the first-time carrier signal and the second-time carrier signal are multiplied and then filtered by a low-pass filter.

4. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: Performing ellipse fitting on discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set specifically includes: The direct least square method based on the block matrix is ​​used to perform ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), wherein the ellipse implicit equation used in the ellipse fitting satisfies: K(e,x i )=ex 2 +Exy+Fy 2 +Lx+My+N=0 In the formula, K(e,x i ) represents the point (x i ,y i ) to the quadratic curve K(e,x i ), the algebraic distance of the point (x i ,y i ) are the coordinates of the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively; The ellipse constraint conditions are set to obtain the best fitting state of the ellipse, and the parameters E, F, L, M, and N of the ellipse under the best fitting state are obtained as the first initial ellipse parameter set.

5. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: Build a neural network and train it, including: Construct a BP neural network, which includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to fit any quadratic surface. Obtain a training data set, specifically, use a data acquisition system to collect the output signal of the all-optical current transformer and the first-harmonic frequency and second-harmonic frequency carrier signals to calculate the first tangent signal P(t) and the second tangent signal Q(t), and obtain the corresponding actual phase, with the first tangent signal P(t) and the second tangent signal Q(t) as input, and the ellipse fitting parameters E, F, L, M, and N as output; The training data set is sent to the BP neural network for training, and the actual phase is combined to determine whether the training is completed.

6. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: Comparing the first initial ellipse parameter set with the second initial ellipse parameter set, the optimal ellipse parameters are obtained, specifically including: Conduct a calibration test under actual working conditions, test and record the minimum sum of squared algebraic distances obtained by fitting the first initial ellipse, calculate the phase signal parameters using the first initial ellipse parameters, compare the calculated phase signal parameters with the actual phase values, and obtain the distortion of the phase signal parameters under different minimum sums of squared algebraic distances. The threshold of the minimum sum of squared algebraic distances is determined according to the distortion requirements of the system's working conditions. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set, otherwise the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

7. The optical fiber current transformer phase demodulation method according to claim 1, characterized in that: Combined with the optimal ellipse parameters, the correct demodulation signal is calculated and processed, including: According to the first tangent signal P(t), the second tangent signal Q(t) and the parameters E, F, and L in the optimal ellipse parameter set, a demodulated signal value under the influence of nonlinear factors is obtained by an inverse tangent operation; The demodulated signal value affected by nonlinear factors is subjected to high-pass filtering to obtain the correct demodulated signal.

8. The optical fiber current transformer phase demodulation method according to claim 7, characterized in that: Demodulated signal value affected by nonlinear factors The calculation formula is as follows: Where P(t) and Q(t) represent the first tangent signal and the second tangent signal respectively, E, F, and L are the parameters of the elliptic implicit equation in the optimal elliptic parameter set, represents the correct demodulated signal, and θ2 represents the component affected by nonlinear factors.

9. A fiber optic current transformer phase demodulation system using the fiber optic current transformer phase demodulation method according to any one of claims 1 to 8, characterized in that: include: Signal acquisition module, signal processing module, ellipse fitting module, neural network module, parameter comparison module and signal calculation module; The signal acquisition module is used to obtain the signal data of the all-optical current transformer; The signal processing module is used to perform internal modulation on the light intensity signal through the carrier generator to generate a single-frequency carrier signal and a double-frequency carrier signal, and to generate a single-frequency carrier signal and a double-frequency carrier signal. d , a one-time carrier signal and a two-time carrier signal are subjected to signal intra-modulation and filtering to obtain a first tangent signal and a second tangent signal; The ellipse fitting module is used to perform ellipse fitting on discrete sampling points of the first tangent signal and the second tangent signal to obtain a first initial ellipse parameter set; The neural network module is used to obtain a second initial ellipse parameter set according to the first tangent signal, the second tangent signal and the trained BP neural network; The parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set; The signal calculation module calculates and processes the optimal ellipse parameter set to obtain the correct demodulation signal.

10. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The signal acquisition module obtains the light intensity signal output by the all-fiber current transformer, which specifically includes: The light intensity signal I output by the all-fiber current transformer is the light intensity result after the interference of two coherent light beams I1 and I2: In the formula, A and B are constant terms, A=I1+I2, C is the carrier phase modulation depth, ω0 is the carrier frequency, It is the phase caused by the external influence on the sensor; The light intensity signal I is converted by the photodetector to obtain an electrical signal as the output signal I d .

11. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The signal processing module processes the output signal I d , a one-time carrier signal and a two-time carrier signal are modulated intra-signally to obtain a first tangent signal P(t) and a second tangent signal Q(t), specifically including: The output signal I d , a one-time carrier signal and a two-time carrier signal, and perform signal intra-modulation through a carrier generator to obtain a first tangent signal P(t) and a second tangent signal Q(t); The signal internal modulation includes: d , the first-time carrier signal and the second-time carrier signal are multiplied and then filtered by a low-pass filter.

12. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The ellipse fitting module performs ellipse fitting on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t) to obtain a first initial ellipse parameter set, which specifically includes: Using the direct least square method based on the block matrix, ellipse fitting is performed on the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), wherein the ellipse implicit equation used in the ellipse fitting satisfies: K(e,x i )=ex 2 +Exy+Fy 2 +Lx+My+N=0 In the formula, K(e,x i ) represents the point (x i ,y i ) to the quadratic curve K(e,x i ), the algebraic distance of the point (x i ,y i ) are the coordinates of the discrete sampling points of the first tangent signal P(t) and the second tangent signal Q(t), e, E, F, L, M, N represent the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, the fifth coefficient and the sixth coefficient of the elliptic quadratic curve respectively; The ellipse constraint conditions are set to obtain the best fitting state of the ellipse, and the parameters E, F, L, M, and N of the ellipse under the best fitting state are obtained as the first initial ellipse parameter set.

13. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The neural network module is used to build and train a neural network, including: Construct a BP neural network, which includes an input layer, a hidden layer, and an output layer. The input features are processed through nonlinear operations in the hidden layer to fit any quadratic surface. Obtain a training data set, specifically, use a data acquisition system to collect the output signal of the all-optical current transformer and the first-harmonic frequency and second-harmonic frequency carrier signals to calculate the first tangent signal P(t) and the second tangent signal Q(t), and obtain the corresponding actual phase, with the first tangent signal P(t) and the second tangent signal Q(t) as input, and the ellipse fitting parameters E, F, L, M, and N as output; The training data set is sent to the BP neural network for training, and the actual phase is combined to determine whether the training is completed.

14. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The parameter comparison module is used to compare the first initial ellipse parameter set and the second initial ellipse parameter set to obtain the optimal ellipse parameter set, specifically including: Conduct a calibration test under actual working conditions, test and record the minimum sum of squared algebraic distances obtained by fitting the first initial ellipse, calculate the phase signal parameters using the first initial ellipse parameters, compare the calculated phase signal parameters with the actual phase values, and obtain the distortion of the phase signal parameters under different minimum sums of squared algebraic distances. The threshold of the minimum sum of squared algebraic distances is determined according to the distortion requirements of the system's working conditions. If the minimum sum of squared algebraic distances of the first initial ellipse fitting result is less than the threshold, the first initial ellipse parameter set is selected as the optimal ellipse parameter set, otherwise the second initial ellipse parameter set is selected as the optimal ellipse parameter set.

15. The optical fiber current transformer phase demodulation system according to claim 9, characterized in that: The signal calculation module is used to calculate and process the optimal ellipse parameters to obtain the correct demodulation signal, including: According to the first tangent signal P(t), the second tangent signal Q(t) and the parameters E, F, and L in the optimal ellipse parameter set, a demodulated signal value under the influence of nonlinear factors is obtained by an inverse tangent operation; The demodulated signal value affected by nonlinear factors is subjected to high-pass filtering to obtain the correct demodulated signal.

16. The optical fiber current transformer phase demodulation system according to claim 15, characterized in that: The signal calculation module obtains the demodulated signal value under the influence of nonlinear factors through inverse tangent operation, specifically including: the demodulated signal value under the influence of nonlinear factors The calculation formula is as follows: Where P(t) and Q(t) represent the first tangent signal and the second tangent signal respectively, E, F, and L are the parameters of the elliptic implicit equation in the optimal elliptic parameter set, represents the correct demodulated signal, and θ2 represents the component affected by nonlinear factors.

17. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instruction to execute the steps of the optical fiber current transformer phase demodulation method according to any one of claims 1-8.

18. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the optical fiber current transformer phase demodulation method described in any one of claims 1 to 8 are implemented.

Images