CN-122001511-A - High-precision clock synchronization method

CN122001511ACN 122001511 ACN122001511 ACN 122001511ACN-122001511-A

Abstract

The invention provides a high-precision clock synchronization method, which relates to the technical field of clock synchronization and comprises the steps of S1, collecting historical data of a constant-temperature crystal oscillator, S2, training a BP neural network by using the constant-temperature crystal oscillator data to obtain a BP neural network frequency prediction model, S3, respectively receiving corresponding satellite clock information by utilizing a BDS receiver and a GPS receiver and resolving to obtain time stamps of the BDS and the GPS, determining a source clock, S4, determining standard time, S5, when the BDS and the GPS are in communication interruption, enabling the constant-temperature crystal oscillator to enter a timekeeping state, performing frequency adjustment according to the obtained prediction frequency, then converting the virtual prediction time stamp into prediction time, taking the prediction time as the standard time, S6, sending the standard time information to a base station in a message mode, S7, sending the standard clock information to local equipment, and performing clock correction synchronization according to the standard time. The method has the beneficial effects that the stability can be improved while the accuracy of synchronization is improved.

Inventors

HU JUNJIE
HE YAOXI
MA LIN
ZHAN YONG
XIANG ZHENWEN

Assignees

中国长江电力股份有限公司
长电新能有限责任公司

Dates

Publication Date: 20260508
Application Date: 20260128

Claims (10)

1. A high-precision clock synchronization method is characterized by comprising the following steps: S1, collecting historical data of a constant-temperature crystal oscillator; s2, training the BP neural network by using the constant-temperature crystal oscillator data to obtain a BP neural network frequency prediction model; s3, respectively receiving corresponding satellite clock information by utilizing a BDS receiver and a GPS receiver, resolving to obtain time stamps of the BDS and the GPS, and determining a source clock; S4, carrying out time service on the constant-temperature crystal oscillator by using a 1PPS signal of the source clock, and determining standard time; S5, when communication interruption occurs in the BDS and the GPS, the constant-temperature crystal oscillator enters a timekeeping state, at the moment, frequency adjustment is carried out according to the prediction frequency obtained by the prediction model, and then the virtual prediction time stamp is converted into prediction time, and the prediction time is taken as standard time; s6, sending the standard time information to the base station in a message form; and S7, the base station sends the standard clock information to the local equipment, and the local equipment performs clock correction synchronization according to the standard time.
2. The method for synchronizing a high-precision clock according to claim 1, wherein the BP neural network frequency prediction model comprises an input layer, a hidden layer and an output layer; the frequency prediction model comprises an aging factor, a current instantaneous temperature, a current EMI peak value, a current voltage fluctuation, a historical frequency mean value, a historical temperature standard deviation and a historical aging accumulated value; The aging factors consist of core aging factors and auxiliary aging factors, wherein the expression of the core aging factors is as follows: (1); Wherein, the As a core age factor, the number of the cells is, For the initial ageing coefficient of the glass, the glass is, In order to accumulate the operating time period, Is a temperature accelerated aging term based on a physical principle; the expression of the auxiliary ageing factor is: (2); Wherein, the In order to start and stop the number of times, The calculation process of the effective value for the fluctuation of the power supply voltage is as follows: (3); Wherein, the For the number of points of the data, Is the first Data point voltage ripple maximum; Aging factor The core aging factor and auxiliary aging factor weighting composition: (4); Wherein, the 、 The weights of the core aging factor and the auxiliary aging factor are respectively; the output of the model is a constant-temperature crystal oscillator frequency predicted value; EMI peaks represent the maximum intensity of electromagnetic interference; the output of the output layer is the frequency predicted value of the constant-temperature crystal oscillator.
3. The method for synchronizing a high-precision clock according to claim 2, wherein the weights of the core aging factor and the auxiliary aging factor are expressed as follows: (5)。
4. the method for synchronizing a high-precision clock according to claim 1, wherein in the step S3, the source clock is determined as follows: S301, determining evaluation indexes including real-time stability indexes and historical reliability indexes; Real-time stability index The expression of (2) is: (6); wherein, the SNR is the signal-to-noise ratio, For the maximum signal-to-noise ratio, p is the phase jitter value, Is a jitter threshold; Historical reliability index The expression of (2) is: (7); Wherein, the Is the average value of the synchronous errors of the standby time service source within nearly 1 hour, Is the maximum allowable error; then overall score The expression of (2) is: (8); Wherein, the 、 Respectively stabilizing weight and historical weight; s302, determining a source clock according to the total score, wherein when only a single standby time service source is available, the standby time service source is directly used as the source clock; When 2 alternate time service sources are available, the source clock is determined using the total score, which is for BDS and GPS Can be respectively marked as And Setting a scoring threshold Judging And When (1) is related to When BDS is used as source clock When GPS is used as the source clock When using a weighted fusion of timestamps As a source clock, the expression is: (9); Wherein, the 、 The time weights of BDS and GPS respectively, 、 The time stamps for BDS and GPS, respectively.
5. The method for synchronizing a high-precision clock as defined in claim 4, wherein in step S4, the time service process of the constant-temperature crystal oscillator is as follows: S401, after receiving the 1PPS signal of the satellite, carrying out frequency division operation on the constant temperature crystal oscillator by taking the signal as an initial signal, dividing the local signal of 10MHz into the local 1PPS signal of 1Hz, and considering the full period receiving error Is used for the control of the (c), The expression of (2) is as follows: (10); Wherein, the 、 Respectively an acquisition error and a noise error; Using Correcting the 1PPS signal of the satellite: (11); Wherein, the 、 The rising edge time of the 1PPS signal before correction and after correction is respectively; S402, measuring rising edge phase difference between 1PPS and local 1PPS signals of the satellite; s403, filtering the rising edge phase difference by using a Kalman filter; S404, determining the voltage required to be adjusted by the constant-temperature crystal oscillator voltage control section according to the filtered phase difference by using a PID controller ; For the voltage to be regulated, the regulation error should be considered Transmission error Synchronization error ; Adjustment error The expression of (2) is: (12); Wherein, the 、 Respectively the theoretical output voltage of the PID controller with the last synchronization period and the actual input voltage of the voltage-controlled section of the constant-temperature crystal oscillator with the last synchronization period, Is a voltage error conversion coefficient; Transmission error Obtained by measuring the communication link delay of the base station and the local device; synchronization error The time sequence delay parameter of the local equipment synchronization module can be calculated; Will adjust for errors Transmission error Synchronization error The sum is recorded as the residual error ; When (when) At the position of When the ratio of the total amount exceeds 50%, the ratio of Correcting; S405, determining and adjusting voltage digital data by using a digital-to-analog converter, converting the digital data into analog data, and inputting the analog data into a constant-temperature crystal oscillator voltage-controlled section to finish time service of the constant-temperature crystal oscillator; When (when) At the position of When the ratio of the time stamp is more than 50%, the time stamp of time service needs to be adjusted: (13); Wherein, the Is the adjusted timestamp; S406, setting a phase difference threshold, and repeating the steps S401-S406 when the phase difference is larger than the phase difference threshold.
6. A high precision clock synchronization method as defined in claim 5, wherein the corrected adjustment voltage The expression of (2) is: (14)。
7. The method of claim 6, wherein the phase difference is an absolute value of a phase difference between a 1PPS signal of the satellite and a local PPS signal.
8. The method of claim 6, wherein the phase difference threshold is a very small positive real number to limit the accuracy of the timing.
9. The method for synchronizing the clock with high precision according to claim 2, wherein the activation function f in (-) of the hidden layer is a hyperbolic tangent function, and the expression is: (15); The output layer activation function f out (·) is a sigmoid function, and the expression is: (16)。
10. the method for synchronizing a high-precision clock as recited in claim 1, wherein in the step S5, the frequency is adjusted according to the following formula: (17); Wherein u is a regulating voltage, u 0 is a standard voltage, K is voltage-controlled sensitivity, f is a predicted frequency, and f 0 is a standard frequency.

Description

High-precision clock synchronization method Technical Field The invention relates to the technical field of clock synchronization, in particular to a high-precision clock synchronization method. Background Along with the development of technology, higher and higher requirements are placed on the precision of clock synchronization, and high-precision clock synchronization technology plays an increasingly important role in the fields of aerospace, power systems, communication and the like. In an electric power system, a collecting line is a multi-architecture multi-type complex system, once faults occur, the detection difficulty is high, a traveling wave method is generally selected to detect fault points at present, in order to ensure the detection precision, the clock of a signal receiver on the line needs to realize high-precision synchronization, and the time synchronization problem is the key of whether the traveling wave method can accurately detect the positions of the fault points or not, and is also a problem to be solved urgently in the current collecting line fault detection. Disclosure of Invention In order to solve the problems in the background art, the invention provides a high-precision clock synchronization method which can reduce the dependence degree of a GPS system, improve the synchronization precision and ensure high-precision synchronization time service when communication is interrupted. The invention provides a high-precision clock synchronization method, which comprises the following steps: S1, collecting historical data of a constant-temperature crystal oscillator; s2, training the BP neural network by using the constant-temperature crystal oscillator data to obtain a BP neural network frequency prediction model; s3, respectively receiving corresponding satellite clock information by utilizing a BDS receiver and a GPS receiver, resolving to obtain time stamps of the BDS and the GPS, and determining a source clock; S4, carrying out time service on the constant-temperature crystal oscillator by using a 1PPS signal of the source clock, and determining standard time; S5, when communication interruption occurs in the BDS and the GPS, the constant-temperature crystal oscillator enters a timekeeping state, at the moment, frequency adjustment is carried out according to the prediction frequency obtained by the prediction model, and then the virtual prediction time stamp is converted into prediction time, and the prediction time is taken as standard time; s6, sending the standard time information to the base station in a message form; and S7, the base station sends the standard clock information to the local equipment, and the local equipment performs clock correction synchronization according to the standard time. Further, the BP neural network frequency prediction model comprises an input layer, a hidden layer and an output layer; the frequency prediction model comprises an aging factor, a current instantaneous temperature, a current EMI peak value, a current voltage fluctuation, a historical frequency mean value, a historical temperature standard deviation and a historical aging accumulated value; The aging factors consist of core aging factors and auxiliary aging factors, wherein the expression of the core aging factors is as follows: (1); Wherein, the As a core age factor, the number of the cells is,For the initial ageing coefficient of the glass, the glass is,In order to accumulate the operating time period,Is a temperature accelerated aging term based on a physical principle; the expression of the auxiliary ageing factor is: (2); Wherein, the In order to start and stop the number of times,The calculation process of the effective value for the fluctuation of the power supply voltage is as follows: (3); Wherein, the For the number of points of the data,Is the firstData point voltage ripple maximum; Aging factor The core aging factor and auxiliary aging factor weighting composition: (4); Wherein, the 、The weights of the core aging factor and the auxiliary aging factor are respectively; the output of the model is a constant-temperature crystal oscillator frequency predicted value; EMI peaks represent the maximum intensity of electromagnetic interference; the output of the output layer is the frequency predicted value of the constant-temperature crystal oscillator. Further, the expression of the weights of the core aging factor and the auxiliary aging factor is: (5)。 Further, in the step S3, the determination process of the source clock is as follows: S301, determining evaluation indexes including real-time stability indexes and historical reliability indexes; Real-time stability index The expression of (2) is: (6); wherein, the SNR is the signal-to-noise ratio, For the maximum signal-to-noise ratio, p is the phase jitter value,Is a jitter threshold; Historical reliability index The expression of (2) is: (7); Wherein, the Is the average value of the synchronous errors of the standby t