CN108196616B - Time synchronization method for detecting data of variable-diameter inertial navigation subsystem in pipeline - Google Patents

Abstract

The invention discloses a synchronous operation flow of PIG variable diameter and inertial navigation subsystem output data based on mileage data time. The process first graphs the odometer data column in the variable diameter and inertial navigation output data matrix, and takes the number of sampling points as the horizontal axis; then uses the starting point search method to mark the sampling of the odometer data from stationary to the moment of movement in the two graphs respectively Then, the time corresponding to the mileage data in the respective system clocks from stationary to the moment of movement is obtained; finally, the difference between the two clocks is obtained, and the clocks of the variable diameter and the inertial navigation subsystem are corrected to the same time to realize the data synchronization. The synchronization method of the present invention can effectively solve the synchronization failure problem caused by the failure of the odometer during the PIG movement or the correction of the mileage data in the later stage, as long as the integrity of the mileage data in the stage of the ball is guaranteed, All output data of the two subsystems can be synchronized.

Description

Time synchronization method for detecting variable diameter inertial navigation subsystem data in pipeline

technical field

The invention relates to a method for synchronizing output data from an internal subsystem of a detection device in a pipeline, in particular to two subsystems of variable diameter and inertial navigation in a pipeline detection device, and the combined internal detection of variable diameter and inertial navigation of various pipeline diameters The synchronization method can be used by any system.

Background technique

1. In-pipe inspection

In-pipe inspection device (PIG) can measure pipeline defects or deformation during the movement of pipeline. PIG can usually carry multiple detection subsystems according to the current mission needs. Since each detection subsystem needs to be flexibly combined according to different tasks, each subsystem usually has independent data acquisition and storage capabilities. Since the reducing device occupies a relatively small space in the PIG, it is often installed in the cavity of a PIG with the inertial navigation device, which can be regarded as one in space and can share a mileage data.

2. Synchronization between internal subsystems of PIG

When multiple detection subsystems need to be used together, the synchronization relationship between data is very important. For example, when the two subsystems of variable diameter detection and inertial navigation positioning detection are mounted on the same PIG platform and perform online detection at the same time, theoretically the downloaded data should be highly synchronized, that is, when the detection work is completed, according to the detected data, the customary The objects of guide positioning and variable diameter detection must maintain the same scale in time and space. Specifically, if the variable diameter detector detects a certain pipe wall defect, the inertial navigation positioning device should give the corresponding geographical coordinates of the defect at the same time.

However, due to the different detection principles of the sensors, the electrical structures of the corresponding detection devices are also different, and the sampling rate, signal format, system clock, and even the format of the output data are completely different. Therefore, it is difficult to determine the one-to-one correspondence between the output data of the inertial navigation and variable diameter subsystems in time.

This requires that the two batches of output data be synchronized after the two subsystems complete the data output. The essence is to ensure that the two batches of output data use the same time scale on their respective time axes.

3. Limitations of Existing Synchronization Methods

The previous solution to this problem was to directly use the mileage data as a synchronization reference. The same odometer signal is introduced into different detection subsystems as a column of the output data matrix. The so-called output data matrix means that each row of data includes multiple data items (mileage is one of them), which is called a sampling point; all data at the sampling point comes from the sampling of the detection subsystem at the same time; by introducing the sampling of the subsystem Parameters such as frequency, signal format, etc., can calculate the sampling time of the sampling point of this line under the clock of this subsystem. Since each sampling point contains a mileage data item, the mileage data can be used as the horizontal axis of the coordinates when outputting data. When different detection subsystems use the common mileage data as the horizontal axis of the coordinate system, the synchronization of the output data of different subsystems is realized.

The disadvantages of the above-mentioned direct use of mileage data as the synchronization reference signal are as follows:

1) It is easy to fail. The odometer runs outside the protective housing of the PIG, and has a higher probability of data failure than other detection subsystems. Inside the oil and gas pipeline, the working environment is very harsh, and the mechanical structure of the odometer has a high failure rate. Even if the odometer maintains its mechanical structure without deformation and damage in a harsh environment of about 100 hours, high-viscosity crude oil, highly corrosive natural gas, and defects on the inner wall of the pipeline may cause the odometer to slip, get stuck, and fail to electrical components, etc. Serious Problem. Once the mileage data is partially or completely invalidated, all synchronization operations cannot be performed.

2) The mileage correction causes the related synchronization data to be corrected again. After the measurement, the accuracy of the odometer data also needs to be corrected, such as the compensation calculation for the slippage of the odometer. However, the mileage data is used as the synchronization reference data. Once the correction is performed, the detection data of all the related subsystems should also be corrected separately. That is to say, before the synchronization operation of the two subsystems, in order to ensure that the output data inside the two subsystems and the current mileage output data are "bound" together, the "synchronization" operation must be performed within the output data of the subsystems, and This "synchronization" operation within the subsystem may also be performed multiple times. This obviously increases the complexity of system data processing and reduces system reliability.

SUMMARY OF THE INVENTION

Purpose of invention

The purpose of the present invention is to solve the above problem of the mileage synchronization method of the variable diameter and inertial navigation subsystems, so that the mileage data can still be used after the data of the moving part after leaving the ball cylinder is invalid, or after the mileage data is corrected locally or globally. The start point time is calculated and used as the synchronization reference to ensure relatively high synchronization accuracy.

Technical solutions

The time synchronization method for detecting the data of the variable diameter inertial navigation subsystem in the pipeline, an odometer signal of the PIG system is simultaneously introduced into the two subsystems of the variable diameter and inertial navigation installed in a cabin; in the stage of the ball, the odometer data is complete; During the power-on operation, the respective system clocks of the two subsystems are at the same precision level; it is characterized in that the method steps are as follows:

Step 1: The odometer raw data in the variable diameter and inertial navigation output data matrix are graphed respectively, and are defined as the odometer raw data map of the inertial navigation subsystem and the odometer raw data map of the variable diameter subsystem;

The format of the output data of the inertial navigation subsystem and the variable diameter subsystem is similar. They are organized into a two-dimensional table form. Each row of data includes multiple data items, and the original odometer signal is also one of them. Each row of data is called a sampling point. ; Starting from the normal collection of data after power-on, self-test, and startup, the positive integer sequence of sampling points arranged according to time is called the number of sampling points, which is defined as m, where m is a positive integer;

The first sampling point obtained after power-on is placed on the origin of the coordinate axis, that is, point 0, and the number of subsequent sampling points m corresponds to the positive integer of the horizontal axis, that is, the number of sampling points from 1 to N; the number of sampling points corresponding to each The original signal value of the odometer is used as the vertical axis;

Step 2, define the moment when the PIG transitions from the static state to the moving state in the ball cylinder as the starting point, and the physical time of the starting point of the two subsystems of the variable diameter and inertial navigation is the same; Point search algorithm, the number of sampling points that mark the starting point in the odometry original data map of the inertial navigation subsystem and the odometer original data map of the variable diameter subsystem, respectively, is defined as m1 and m2;

The starting point search algorithm is as follows:

1) Determine the accuracy target search_time of the search start point algorithm, which must match the accuracy target of the entire synchronization algorithm, that is, the accuracy target of the search algorithm is an order of magnitude higher than the accuracy target of the entire synchronization algorithm, that is, the error time search_time of the search algorithm is 1/N of the synchronization algorithm error time goal_time;

2) Determine the search step size step: Define the sampling frequency of a subsystem (the number of sampling points in one second) as sample_times, then the search step size is the accuracy target of the search start point algorithm multiplied by the sampling frequency

step=search_time×sample_times

During the search process, when the number of sampling points in the i-th search is m _i , the number of sampling points in the i+1-th search m _i+1 has

m _i+1 =m _i +step

3) Define the search space S _s :

S _s interval is a collection of sampling points, which is a continuous interval on the abscissa of the original mileage data, defined by the starting point m_start and the ending point m_end of the search;

The principle of defining S _s is to ensure that the search target, that is, the starting point, is between m_start and m_end;

The method of setting S _s is as follows: observe the graph of the original mileage data, the starting point must be in the transition area from the flat area to the oscillating area, and set the left boundary of this transition area as m_start;

m_start is the number of sampling points to start the search, and m_end is the number of sampling points to end the search. For subsequent search calculations, it can be determined that the maximum number of searches N _s is

N _s =(m_end-m_start)/step

Let i be the number of current searches to define the search space

S _s ={m _i ∣m_start≤m _i ≤m_end,i=1,2,...,N _s }

where m _i is the number of sampling points in the ith search;

4) Determine the static interval S of the PIG in the serving cylinder; the static interval, that is, marked in the original signal diagram of the subsystem odometer, the PIG is in a static state, and the collection interval of the sampling points without slight movement, define the static interval The number of boundary sampling points is m _s0 and m _s1 respectively, then it can be defined

S={(m,S _V )∣m _s0 ≤m≤m _s1 }

where m is the number of sampling points in the static interval, and S _V is the original odometer signal corresponding to m;

The method of setting S is as follows: Observe on the original mileage data map, starting from the origin, the image will quickly become stable and enter a flat area, which will last for many sampling points from the abscissa. In the flat area, the image only has high-frequency noise signals, and there will be no large fluctuations in the ordinate. The left boundary of this area is marked as the number of sampling points m _s0 , and the right boundary is marked as the number of sampling points m _s1 ;

On the premise of ensuring that the PIG is in a static state, the S interval should be as large as possible. In a statistical sense, the observed value of S _V at rest can be more accurately obtained, that is, the mean value of all S _V in the static interval S, which is defined as E(S _V );

5) Set the parameters used in the search:

Define the number of target sampling points for the i-th search as m _i , where m _i ∈ S _s ;

Define the two sets that need to be used in the ith search: the left set (S _l ) and the right set (S _r ), define the number of sampling points contained in the left set as m _l , and define the number of sampling points contained in the right set as m _r ; Two sets S _l and S _r , they are respectively the two sampling point sets located on the left and right sides of m _i centered on the number of sampling points currently searched for m _i , and the number of elements in these two sets is set to 1 The number of all sampling points in seconds, that is, sample_times, there are

S _l ={(m _l ,S _V )∣m _i -sample_times≤m _l ≤m _i ,m _i ∈S _s }

S _r ={(m _r ,S _V )∣m _i ≤m _r ≤m _i +sample_times,m _i ∈S _s }

_The mean values of all _SVs of the two sets are defined as _El (SV) and _Er (SV) respectively, that is, _El ( _{SV )} is the observed value of all _SVs in _Sl , and _Er ( _SV ) is the observed value of all S _V in S _r , when _mi is the starting point, then _El (S _V ) must approach E(S _V ), and E _r (S _V ) must be greater than E(S _V ) ; When mi is not the starting point, then E _l (S _V ) and _Er (S _V ) must also approach E(S _V ) or be greater than _{E(S V )} ;

Define the two threshold values ε _l and ε _r used for the determination of this search. The principle of initial setting is that ε _l > 0, and it is as small as possible, and ε _r is an order of magnitude larger than ε _l , that is, ε _r is larger than ε _l M times larger, M is an integer from 2 to 10; if there are more than 1 starting point obtained by the initial search, decrease ε _l and increase ε _r , and search again;

6) In the range of S _s , starting from m_start and ending at m_end, searching for the starting point, a total of N _s search calculations are performed; for the i-th search, i=1,2,...,N _s , when there are

∣E _l (S _V )－E(S _V )∣<ε _l and E _r (S _V )－E(S _V )>ε _r

If established, accept m _i as the starting point; when

∣E _l (m _i )－E(S _V )∣≥ε _l or E _r (m _i )－E(S _V )≤ε _r

If it is established, it is necessary to continue the i+1th search for the number of m _i+1 sampling points, where m _i+1 =m _i +step;

7) If the above search is completed and only one starting point is obtained, the starting point search algorithm of the subsystem ends. If the search is performed according to the current ε _l and ε _r in the entire S _s range, multiple starting points can be obtained. , then you need to adjust ε _l and ε _r , and search again until there is one left; the adjustment method is to decrease ε _l or increase ε _r ; the ε _l and ε _r must be in the original odometer data of the two subsystems are used at the same time, and the starting points m1 and m2 are obtained respectively;

Step 3: According to the respective sampling frequencies sample_times of the variable diameter and inertial navigation subsystems, obtain the corresponding time t of each sampling point m under the respective system clocks in the odometer original signal diagrams of the variable diameter and inertial navigation subsystems respectively. , including the corresponding times t1 and t2 of m1 and m2 in their respective system clocks,

t=m/sample_times

Convert the horizontal axis of the original number of sampling points m to the horizontal axis of the current subsystem clock t;

Step 4: Calculate the difference Δt between t1 and t2: Δt=t2-t1, Δt is used as the difference between the two subsystem clocks;

Step 5: Use Δt to correct the clocks of the two subsystems to the same time to realize the synchronization of the data of the variable diameter and inertial navigation subsystems;

When the variable diameter subsystem clock needs to be used as the standard time for subsequent processing, the corresponding time of all sampling points of the inertial navigation subsystem needs to be reduced by Δt; when the inertial navigation subsystem time is required to be the standard time, all the time parameters of the variable diameter subsystem Need to add Δt.

Advantages and Effects

Compared with the prior art, the beneficial effects of the present invention are:

First of all, the problem that the synchronization operation cannot be performed due to the failure of the odometer is significantly solved. The main part of the odometer is a mechanical structure, and the failure almost always occurs during the movement after leaving the ball barrel, and there is usually no problem in the ball barrel stage. However, the present invention only requires that the data at the stage of the ball is complete, that is, m1 and m2 in step 2 can be found, and the whole algorithm is not affected subsequently. The effect of odometer failure on subsequent synchronization operations is virtually eliminated.

Secondly, the problem of the failure of the previous synchronization operation results caused by the correction of the odometer data is completely solved. The correction of the mileage data is mainly aimed at the problems of slippage and data consistency in the operation of the PIG, and will not affect the data in the static stage of the ball cylinder, so it will not affect the values of m1 and m2 in step 2, and will not affect the subsequent synchronous operation.

Description of drawings

Figure 1 is the original data of the odometer of the inertial navigation subsystem;

Figure 2 is the original data of the variable diameter subsystem odometer;

Figure 3 is the odometer output data of the inertial navigation subsystem;

Figure 4 is the output data of the variable diameter subsystem odometer;

Figure 5 shows the mileage data of variable diameter and inertial navigation output under the same time scale;

Figure 6 is a partial magnification near the starting point of the odometry raw data of the inertial navigation subsystem.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings:

Figure 1 is the original data of the typical inertial navigation detection subsystem mileage, the horizontal axis is the number of sampling points, the vertical axis is the mileage signal containing a lot of noise, the signal is essentially a triangular wave level, the m1 on the horizontal axis is the inertia of the ball The number of sample points for the start point of the derivative system.

Figure 2 is the original data of the typical variable diameter detection subsystem mileage, the horizontal axis is the number of sampling points, the vertical axis is the mileage signal containing a lot of noise, the signal is essentially a triangular wave level, m2 is the starting point of the variable diameter subsystem in the ball cylinder the number of sampling points.

Figure 3 shows the output data of a typical inertial navigation subsystem. The horizontal axis is time, and the vertical axis is mileage, in meters. t1 is the time when the inertial navigation subsystem passes the starting point in the ball barrel calculated by m1 under the clock standard of the inertial navigation subsystem.

Figure 4 shows the output data of a typical variable diameter subsystem. The horizontal axis is time, and the vertical axis is mileage, in meters. t2 is the time that the variable diameter subsystem passes the starting point in the ball cylinder under the standard of the variable diameter subsystem clock, calculated by m2.

Figure 5 puts the mileage data output by the typical variable diameter detection subsystem and the inertial navigation detection subsystem in a coordinate system, and the horizontal axis is time. t1 is the time when the inertial navigation subsystem passes the starting point in the ball barrel calculated by m1 under the clock standard of the inertial navigation subsystem. t2 is the time when the variable diameter subsystem passes the starting point in the ball cylinder calculated by m2 under the standard of the variable diameter subsystem clock; Δt is the synchronization error of the two subsystem clocks of the variable diameter and inertial navigation.

Figure 6 shows the local amplification near the starting point of the odometry raw data of the inertial navigation subsystem, which is the signal near the starting point. At this time, due to the existence of noise and other factors, the artificial method cannot accurately locate the m1 point, so the starting point search algorithm needs to be used.

In-pipe inspection systems (PIGs) are usually equipped with multiple inspection devices, each of which has independent data acquisition and storage capabilities. Due to the size limitation of the ball launcher and the ball receiver, the inertial navigation positioning device (the core of which is the IMU system) and the variable diameter detection device are often used as two subsystems to carry PIG into the pipeline at the same time. When multiple detection devices need to be used together, the synchronization relationship between data is very important. However, due to the different detection principles of the sensors of the variable diameter system and the inertial navigation system, the corresponding electrical structures are also different, and the sampling rate and signal format are also completely different, which may cause the dislocation of different types of sensor data at the same time on the unified time axis. , so a synchronous operation is necessary. This operation mainly uses the calculated time data as the synchronization basis. The premise of its execution is: the odometer signal of the PIG system is simultaneously introduced into the variable diameter and inertial navigation subsystems; the respective system times of the two subsystems are at the same level of accuracy.

The time synchronization method for detecting the data of the variable diameter inertial navigation subsystem in the pipeline, an odometer signal of the PIG system is simultaneously introduced into the two subsystems of the variable diameter and inertial navigation installed in a cabin; in the stage of the ball, the odometer data is complete; During the power-on operation, the respective system clocks of the two subsystems are at the same precision level; it is characterized in that the method steps are as follows:

Step 1: The odometer raw data in the variable diameter and inertial navigation output data matrix are graphed respectively, and defined as the odometer raw data map of the inertial navigation subsystem (as shown in Figure 1) and the variable diameter subsystem odometer raw data map ( as shown in picture 2);

The format of the output data of the inertial navigation subsystem and the variable diameter subsystem is similar, and they are organized into a two-dimensional table form. Each row of data includes multiple data items. The original odometer signal (voltage value, unit V, defined as S _V in the following formula ) is also one of them, and each line of data is called a sampling point; starting from the normal collection of data after power-on, self-test, and startup, the positive integer sequence of sampling points arranged in time is called the number of sampling points, which is defined as m, where m is positive integer;

The first sampling point obtained after power-on is placed on the origin of the coordinate axis, that is, point 0, and the number of subsequent sampling points m corresponds to the positive integer of the horizontal axis, that is, the number of sampling points from 1 to N; the number of sampling points corresponding to each The original signal value of the odometer is used as the vertical axis;

The original signal of the odometer is essentially a triangular wave. When participating in the subsequent data processing, it needs to be converted into the mileage in meters according to the method provided in the odometer product manual, as shown in the ordinates in Figure 3, Figure 4 and Figure 5;

Step 2, define the moment when the PIG transitions from the static state to the moving state in the ball cylinder as the starting point, and the physical time of the starting point of the variable diameter and inertial navigation subsystems is the same. The starting point search algorithm is used for the odometer raw signals of the two subsystems, and the number of sampling points that mark the starting point in the odometry raw data map of the inertial navigation subsystem and the odometer raw data map of the variable diameter subsystem are defined as m1 and m2;

The starting point search algorithm can solve the following problems: due to the pressure of the PIG in the tee, before entering a perceptible and observable motion state, there is actually a slow acceleration motion process, which can last up to tens of seconds. , the number of sampling points involved is related to the sampling frequency of the two subsystems, which can reach tens of thousands at most (as shown in Figure 1 and Figure 6). When the starting point is directly obtained by the method of manual marking, the number of sampling points of the starting point is judged successively in the odometer original data map of the inertial navigation and variable diameter subsystems (as shown in Figure 1 and Figure 2), which will generate 0.1 seconds to 10 Errors ranging from seconds. The two possible starting points A and B shown in Figure 6, when the sampling frequency of the inertial navigation subsystem is 500 Hz, can be converted into time, and the difference can reach more than 5 seconds. The goal of the synchronous operation of the two subsystems is to hope that the system clocks of the two subsystems face the same moment, and the difference of the output time, that is, the synchronization time precision goal_time, is as small as possible. Through the inversion of the target measurement accuracy of the entire PIG detection project, the time accuracy goal_time of the synchronous operation of the two subsystems must be lower than 0.1 seconds, so the error of manual operation cannot be ignored. It is necessary to select a state tracking method executed by a machine to judge the data inflection points of the two graphs on the same scale, that is, to start the point search algorithm. Using this search algorithm, the exact number of sampling points at the start point of the two subsystems of PIG can be obtained at the same scale. This data is crucial to the two subsystems and substantially determines the final time-based synchronization processing accuracy.

The starting point search algorithm must use the original data of the odometers of the two subsystems of inertial navigation and variable diameter (as shown in Figure 1 and Figure 2), which is necessary. Otherwise, some information will be lost after de-noising, granulation, softening, data conversion, etc. of the original data of inertial navigation and variable-path odometer, and the loss of information in different subsystems is also different, and it is difficult to evaluate , so the processed data cannot guarantee a more accurate starting point at the same scale, and the original data must be used.

The starting point search algorithm is as follows:

1) Determine the accuracy target search_time of the search start point algorithm, which must match the accuracy target of the entire synchronization algorithm, that is, the accuracy target of the search algorithm is an order of magnitude higher than the accuracy target of the entire synchronization algorithm, that is, the error time search_time of the search algorithm is 1/N of the synchronization algorithm error time goal_time;

2) Determine the search step size step: Define the sampling frequency of a subsystem (the number of sampling points in one second) as sample_times, then the search step size is the accuracy target of the search start point algorithm multiplied by the sampling frequency

step=search_time×sample_times

During the search process, when the number of sampling points in the i-th search is m _i , the number of sampling points in the i+1-th search m _i+1 has

m _i+1 =m _i +step

3) Fig. 6 defines the search space S _s as an example:

S _s interval is a collection of sampling points, which is a continuous interval on the abscissa of the original mileage data, defined by the starting point m_start and the ending point m_end of the search;

The principle of defining S _s is to ensure that the search target, that is, the starting point, is between m_start and m_end.

The method of setting S _s is as follows: observe the graph of the original mileage data, the starting point must be in the transition area from the flat area to the oscillating area, and the left boundary of this transition area is set to m_start, for example, m_start=195000, and the right boundary is set is m_end, for example m_end=204000;

m_start is the number of sampling points to start the search, and m_end is the number of sampling points to end the search. For subsequent search calculations, it can be determined that the maximum number of searches N _s is

N _s =(m_end-m_start)/step

Let i be the number of current searches to define the search space

S _s ={m _i ∣m_start≤m _i ≤m_end,i=1,2,...,N _s }

where m _i is the number of sampling points in the ith search;

4) Determine the static interval S of the PIG in the serving cylinder; the static interval, that is, marked in the original signal diagram of the subsystem odometer, the PIG is in a static state, and the collection interval of the sampling points without slight movement, define the static interval The number of boundary sampling points is m _s0 and m _s1 respectively, then it can be defined

S={(m,S _V )∣m _s0 ≤m≤m _s1 }

where m is the number of sampling points in the static interval, and S _V is the original odometer signal corresponding to m;

As shown in Figure 1 and Figure 6, the method of setting S is as follows: Observe on the original mileage data map, starting from the origin, the image will quickly become stable and enter a flat area, which is viewed from the abscissa, It will last for a lot of sampling points. In this flat area, the image only has high-frequency noise signals, and there will be no large fluctuations in the ordinate. The left boundary of this area is marked as the number of sampling points m _s0 , and the right boundary is marked as The number of sampling points m _s1 , such as m _s0 =1000, m _s1 =190000, the S area defined in this way ensures that the PIG is placed in the ball cylinder in a static state during this time period, and is not disturbed by other factors;

On the premise of ensuring that the PIG is in a static state, the S interval should be as large as possible. In a statistical sense, the observed value of S _V at rest can be more accurately obtained, that is, the mean value of all S _V in the static interval S, which is defined as E(S _V );

5) Set the parameters used in the search:

Define the number of target sampling points for the i-th search as m _i , where m _i ∈ S _s ;

Define the two sets that need to be used in the ith search: the left set (S _l ) and the right set (S _r ), define the number of sampling points contained in the left set as m _l , and define the number of sampling points contained in the right set as m _r ; Two sets S _l and S _r , they are respectively the two sampling point sets located on the left and right sides of m _i centered on the number of sampling points currently searched for m _i , and the number of elements in these two sets is set to 1 The number of all sampling points in seconds, that is, sample_times, there are

S _l ={(m _l ,S _V )∣m _i -sample_times≤m _l ≤m _i ,m _i ∈S _s }

S _r ={(m _r ,S _V )∣m _i ≤m _r ≤m _i +sample_times,m _i ∈S _s }

_The mean values of all _SVs of the two sets are defined as _El (SV) and _Er (SV) respectively, that is, _El ( _{SV )} is the observed value of all _SVs in _Sl , and _Er ( _SV ) is the observed value of all S _V in S _r , when _mi is the starting point, then _El (S _V ) must approach E(S _V ), and E _r (S _V ) must be greater than E(S _V ) ; When mi is not the starting point, then E _l (S _V ) and _Er (S _V ) must also approach E(S _V ) or be greater than _{E(S V )} ;

When the number of sampling points m _i = 195005 in a certain search, the range of S _l is expressed as follows: S _l ={(m _l ,S _V )∣194505≤m _l ≤195005}; the range of S _r is expressed as follows: S _r ={(m _r ,S _V )∣195005≤m _r ≤195505};

Define the two threshold values ε _l and ε _r used for the determination of this search. The principle of initial setting is that ε _l > 0, and it is as small as possible, and ε _r is an order of magnitude larger than ε _l , that is, ε _r is larger than ε _l M times larger, M is an integer from 2 to 10; if there are more than 1 starting point obtained by the initial search, decrease ε _l and increase ε _r , and search again;

6) Within the range of S _s , start from m_start and end at m_end, search for the starting point, and perform a total of N _s search calculations. For the ith search, i=1,2,...,N _s , when there is

∣E _l (S _V )－E(S _V )∣<ε _l and E _r (S _V )－E(S _V )>ε _r

If established, accept m _i as the starting point; when

∣E _l (m _i )－E(S _V )∣≥ε _l or E _r (m _i )－E(S _V )≤ε _r

If it is established, it is necessary to continue the i+1th search for the number of m _i+1 sampling points, where m _i+1 =m _i +step;

7) If the above search is completed and only one starting point is obtained, the starting point search algorithm of the subsystem ends. If the search is performed according to the current ε _l and ε _r in the entire S _s range, multiple starting points can be obtained. , then you need to adjust ε _l and ε _r , and search again until there is one left; the adjustment method is to decrease ε _l or increase ε _r ; the ε _l and ε _r must be in the original odometer data of the two subsystems are used at the same time, and the starting points m1 and m2 are obtained respectively;

Step 3: According to the respective sampling frequencies sample_times of the variable diameter and inertial navigation subsystems, obtain the corresponding time t of each sampling point m under the respective system clocks in the odometer original signal diagrams of the variable diameter and inertial navigation subsystems respectively. , including the corresponding times t1 and t2 of m1 and m2 in their respective system clocks, as shown in the abscissas of Figure 3 and Figure 4;

t=m/sample_times

Convert the horizontal axis of the original number of sampling points m to the horizontal axis of the current subsystem clock t;

Step 4, find the difference Δt between t1 and t2: Δt=t2-t1, Δt is used as the difference between the two subsystem clocks, as shown in Figure 5;

Step 5: Use Δt to correct the clocks of the two subsystems to the same time to realize the synchronization of the data of the variable diameter and inertial navigation subsystems;

When the variable diameter subsystem clock needs to be used as the standard time for subsequent processing, the corresponding time of all sampling points of the inertial navigation subsystem needs to be reduced by Δt; when the inertial navigation subsystem time is required to be the standard time, all the time parameters of the variable diameter subsystem Need to add Δt.

Practical example:

step one:

Figure 1 is obtained from step 1, which is the original data of the mileage of a typical inertial navigation detection subsystem. The horizontal axis is the number of sampling points, and the vertical axis is the mileage signal containing a lot of noise. The signal is essentially a triangular wave level. m1 is the number of sampling points of the starting point of the inertial navigation subsystem in the ball cylinder, and the S area near the starting point is enlarged in Figure 6;

Figure 2 is the original data of the typical variable diameter detection subsystem mileage obtained from step 1. The horizontal axis is the number of sampling points, and the vertical axis is the mileage signal containing a lot of noise. The signal is essentially a triangular wave level, and m2 is the change in the ball. The number of sample points for the start point of the radius subsystem. Obviously, the waveforms of the data in Figure 2 and Figure 1 are exactly the same, and they are homologous data, but need to be synchronized.

Step 2:

The necessity of starting the point search algorithm:

Figure 6 shows the signal in the S area near the start point. Both A (sampling point number 199246) and B (sampling point number 202419) may be the start point. The sampling frequency is 500Hz. At this time, the time difference between A and B is 6.346 seconds. Therefore, due to noise and other factors, it is difficult to define the m1 point accurately by manual methods.

Start point search algorithm 1)

When the synchronization precision goal_time of the two subsystems is required to reach 0.1 seconds, the precision search_time of the search start point algorithm should be between 0.05 and 0.01 seconds;

Start point search algorithm 2)

Sampling frequency 500Hz, search_time=0.01, then step is 5,

Start point search algorithm 3)

As shown in FIG. 6 , it can be set: m_start=195000, m_end=204000, and N _s =1800.

Start point search algorithm 4)

As shown in Figure 1, it can be determined that m _s0 =1000, m _s1 =190000, E(S _V )=1.772130

Start point search algorithm 5)

Taking the data in Figure 6 as an example, the initial setting value of ε _l is 0.001, and the initial setting value of ε _r is 0.01.

Start point search algorithm 6)

When the first search, that is, i=1, the number of search sampling points m ₁ =195000, and the range of S _l is expressed as follows: S _l ={(m _l ,S _V )∣194500≤m _l ≤195000}, and E _l ( S _V )=1.772128; the S _r range is expressed as follows: S _r ={(m _r ,S _V )∣195000≤m _r ≤195500}, resulting in _Er (S _V )=1.772234. E _r (S _V )-E(S _V )>ε _r is not satisfied, so m ₁ =195000 is not accepted as the starting point.

When the second search, i.e. i=2, the number of search sampling points m ₂ =195005, the range of S _l is expressed as follows: S _l ={(m _l ,S _V )∣194505≤m _l ≤195005}, and E _l ( S _V )=1.772127; the S _r range is expressed as follows: S _r ={(m _r ,S _V )∣195005≤m _r ≤195505}, resulting in _Er (S _V )=1.772151. E _r (S _V )-E(S _V )>ε _r is not satisfied, so m ₂ =195005 is not accepted as the starting point.

When the 1640th search, i.e. i=1640, the number of search sampling points m ₁₆₄₀ =203195, the range of S _l is expressed as follows: S _l ={(m _l ,S _V )∣202695≤m _l ≤203195}, and E _l ( S _V )=1.772517; the S _r range is expressed as follows: S _r ={(m _r , S _V )∣203195≤m _r ≤203695}, resulting in E _r (S _V )=1.783518. If the judgment condition is met, m ₁₆₄₀ =203195 is accepted as the starting point.

Start point search algorithm 7)

After completing 1800 searches, it is found that the searched starting points reach 22, which are not unique. Modify the value of ε _r to 0.03, search again, and obtain m ₁₇₂₄ =203615. The range of S _l is expressed as follows: S _l ={(m _l ,S _V )∣203115≤m _l ≤203615}, and E _l (S _V )=1.772802 is obtained; the range of S _r is expressed as follows: S _r ={(m _r , S _V )∣203615≤m _r ≤204115}, E _r (S _V )=1.803518 is obtained. If the judgment conditions are met, m ₁₇₂₄ = 203615 is accepted as the only starting point of the inertial navigation subsystem, that is, m1 = 203615.

In a similar way, m2=106710 can be obtained.

Step 3:

Figure 3 is obtained from step 3, which is the output data of a typical inertial navigation subsystem. The horizontal axis is time, and the vertical axis is mileage, in meters. t1 is the time 407.23 seconds for the inertial navigation subsystem to pass the starting point in the ball cylinder under the clock standard of the inertial navigation subsystem calculated by m1.

Figure 4 is obtained from step 3, which is the output data of a typical variable diameter subsystem. The horizontal axis is time, and the vertical axis is mileage, in meters. t2 is 213.42 seconds when the variable diameter subsystem passes the starting point in the ball cylinder under the clock standard of the variable diameter subsystem calculated by m2.

As can be seen from Figures 3 and 4, many mileage data in the low-speed range "disappeared" due to the filtering and other processing of the original data. If the starting point is found at this time, the starting point will "move", and the distance moved does not equal Inconsistent, so the starting point must be searched using the raw data.

Step 4:

Figure 5 is obtained from step 4. The mileage data output by the typical variable diameter detection subsystem and the inertial navigation detection subsystem are placed in a coordinate system, Δt is the synchronization error of the two subsystem clocks of the variable diameter and inertial navigation, and Δt=210.08 second.

Step 5:

Taking the variable diameter clock as the standard, the time corresponding to all inertial navigation output data (including but not limited to mileage data) must be subtracted by 210.08 seconds.

Claims (1)

1. A time synchronization method for detecting data of a variable diameter inertial navigation subsystem in a pipeline is characterized in that an odometer signal of a PIG system is simultaneously introduced into the variable diameter inertial navigation subsystem and the inertial navigation subsystem which are arranged in a cabin; in the stage of the serve barrel, mileage data is complete; in the process of power-on operation, respective system clocks of the two subsystems are at the same precision level; the method is characterized in that: the method comprises the following steps:

step one, original odometer data in a variable-diameter output data matrix and an inertial navigation output data matrix are respectively imaged and defined as an original odometer data graph of an inertial navigation subsystem and an original odometer data graph of a variable-diameter subsystem;

the formats of the output data of the inertial navigation subsystem and the variable diameter subsystem are similar, the output data are organized into a two-dimensional table form, each line of data comprises a plurality of data items, the original signal of the odometer is also one of the data items, and each line of data is called a sampling point; starting from normal data acquisition after power-on, self-inspection and starting, the sampling points are called as sampling point numbers according to a positive integer sequence arranged by time, and are defined as m, wherein m is a positive integer;

placing a first sampling point obtained after electrification on an original point of a coordinate axis, namely 0 point, and mapping a subsequent sampling point number m to a positive integer of a transverse axis, namely 1 to N sampling points; the original signal value of the odometer corresponding to each sampling point is taken as a longitudinal axis;

step two, defining the moment when the PIG is converted from a static state to a motion state in the ball serving barrel as a starting point, wherein the physical time of the starting points of the reducing subsystem and the inertial navigation subsystem is the same; starting point search algorithm is adopted for original odometer signals of the two subsystems, and sampling points of starting points are marked in an original odometer data graph of the inertial navigation subsystem and an original odometer data graph of the reducing subsystem respectively and are defined as m1 and m 2;

the starting point search algorithm is as follows:

1) determining a precision target search _ time of a search starting point algorithm, wherein the precision must be matched with a precision target of the whole synchronization algorithm, namely the precision target of the search algorithm is higher than the precision target of the whole synchronization algorithm by one order of magnitude, namely the error time search _ time of the search algorithm is one N times of the error time goal _ time of the synchronization algorithm;

2) determining a search step size step: defining the sampling frequency (the number of sampling points in one second) of a certain subsystem as sample _ times, and then the step length of searching is the precision target of the algorithm of the search starting point multiplied by the sampling frequency

step＝search_time×sample_times

In the searching process, when the sampling point number of the ith searching is m_iThe number m of sampling points of the i +1 th search_i+1Is provided with

m_i+1＝m_i+step

3) Defining a search space S_s：

S_sThe interval is a set formed by sampling points, is a section of continuous interval on the abscissa of the mileage original data, and is defined by a starting point m _ start and an end point m _ end of the search;

definition of S_sThe principle of (1) is as follows: ensuring that the target of searching, namely the starting point, is between m _ start and m _ end;

setting S_sThe method comprises the following steps: observing the graph of the original mileage data, wherein the starting point is necessarily in a transition region from a straight region to an oscillation region, and the left boundary of the transition region is set as m _ start;

m _ start is the number of sampling points for starting search, m _ end is the number of sampling points for finishing search, and subsequent search calculation is carried out, wherein the maximum search frequency N can be determined at the moment_sIs composed of

N_s＝(m_end－m_start)/step

Let i be the current search times, define the search space

S_s＝{m_i∣m_start≤m_i≤m_end,i＝1,2,…,N_s}

Wherein m is_iThe sampling point number of the ith search is;

4) determining a standing interval S of the PIG in the pitching barrel; a standing interval, namely an aggregation interval of sampling points marked in an original signal diagram of the subsystem odometer, wherein the PIG is in a standing state and has no micro movement, and the sampling points on the boundary of the standing interval are respectively defined as m_s0And m_s1Then can define

S＝{(m,S_V)∣m_s0≤m≤m_s1}

Wherein m is the number of sampling points in the standing interval, S_VOriginal signals of the odometer corresponding to m;

the method for setting S is as follows: observing on a mileage original data graph, from an origin, an image quickly tends to be stable and enters a flat area, the area is continuously provided with a plurality of sampling points from the abscissa, in the flat area, the image only has a high-frequency noise signal, and large fluctuation of the ordinate does not occur, and the left landmark of the area is marked as the sampling point m_s0And the right landmark is marked as the number m of sampling points_s1；

On the premise of ensuring that the PIG is in a static state, the S interval is as large as possible, and S in the standing state can be calculated more accurately in a statistical sense_VI.e. all S in the rest interval S_VIs defined as E (S)_V)；

5) Setting parameters used in the search:

defining the number of target sampling points of the ith search as m_i，m_i∈S_s；

Two sets are defined that need to be used in the ith search: left set (S)_l) And right set (S)_r) Defining the number of sampling points contained in the left set as m_lDefining the number of sampling points contained in the right set as m_r(ii) a Two sets S_lAnd S_rThey are respectively the number m of sampling points of the current search_iIs at the center of m_iTwo sets of sampling points on the left and right sides, the number of elements in the two sets is set as the number of all sampling points in 1 second, namely sample _ times, which includes

S_l＝{(m_l,S_V)∣m_i－sample_times≤m_l≤m_i,m_i∈S_s}

S_r＝{(m_r,S_V)∣m_i≤m_r≤m_i+sample_times,m_i∈S_s}

Two sets all S_VAre respectively defined as E_l(S_V) And E_r(S_V) I.e. E_l(S_V) Is S_lAll S in_VObserved value of (E)_r(S_V) Is S_rAll S in_VWhen m is an observed value of_iIs a starting point, E_l(S_V) Inevitably approaching to E (S)_V) And E is_r(S_V) Must be larger than E (S)_V) (ii) a When m is_iNot the starting point, E_l(S_V) And E_r(S_V) Must also approach E (S)_V) Or likewise greater than E (S)_V)；

Two thresholds epsilon defining the present search for decision_lAnd ε_rThe principle of initial setting is epsilon_lIs greater than 0 and as small as possible, and_rthan epsilon_lOne order of magnitude larger, i.e. epsilon_rThan epsilon_lM is larger than M times, and M is an integer from 2 to 10; if the initial setting search results in more than 1 starting point, then epsilon is reduced_lAnd increase ε_rSearching again;

6) at S_sIn the range from m _ start to m _ end, searching for the starting point, and performing N_sCalculating secondary search; for the ith search, i is 1,2, …, N_sWhen there is

∣E_l(S_V)－E(S_V)∣＜ε_lAnd E_r(S_V)－E(S_V)＞ε_r

If true, then m is accepted_iIs a starting point; when in use

∣E_l(m_i)－E(S_V)∣≥ε_lOr E_r(m_i)－E(S_V)≤ε_r

If it is true, then m needs to be matched_i+1SamplingThe point number continues to search for the (i + 1) th time, where m_i+1＝m_i+step；

7) If the above search is completed and only one starting point is obtained, the starting point search algorithm of the subsystem is ended if the whole S_sWithin the range according to the current epsilon_lAnd ε_rSearching for multiple starting points, the epsilon needs to be adjusted_lAnd ε_rSearching again until one is left; the adjustment is made by reducing epsilon_lOr increase ε_r(ii) a The epsilon_lAnd ε_rThe method must be simultaneously used in the original data of the odometers of the two subsystems to respectively obtain starting points m1 and m 2;

step three, according to the sampling frequency sample _ times of the variable diameter subsystem and the inertial navigation subsystem, respectively solving the corresponding time t of each sampling point number m under the system clock of each subsystem in the original signal diagram of the odometer of the variable diameter subsystem and the inertial navigation subsystem, wherein the corresponding time t comprises the corresponding time t1 and t2 of m1 and m2 in the system clock of each subsystem;

t＝m/sample_times

converting a horizontal axis of the original sampling point number m into a horizontal axis of the current subsystem clock t;

step four, solving the difference value delta t between t1 and t 2: t2-t1, wherein the Δ t is used as the difference value of two subsystem clocks;

fifthly, correcting the clocks of the two subsystems to be at the same time by utilizing delta t, and realizing the synchronization of the data of the two subsystems of diameter changing and inertial navigation;

when the variable-diameter subsystem clock is required to be used as standard time for subsequent processing, the corresponding time of all sampling points of the inertial navigation subsystem is required to be reduced by delta t; when the time of the inertial navigation subsystem is required to be standard time, delta t is required to be added to all time parameters of the variable-diameter subsystem.

Images