A Method and Device for Cleaning Flight Telemetry Data of a Liquid Rocket Engine

Abstract

A method for cleaning flight telemetry data of a liquid rocket engine, comprising: dividing the flight telemetry data into a steady section and an unsteady section according to the working process of the engine; determining a criterion for the sampling stability of the flight telemetry data, and classifying the flight telemetry data into normal sampling points, censored sampling points, and overlapping sampling points according to this criterion; performing data cleaning on the flight telemetry data according to the following situations: if the sampling point is a censored sampling point in the unsteady section, using the constant filling method for data cleaning; if the sampling point is a censored sampling point in the steady section, using the statistical value filling method for data cleaning; if the sampling point is an overlapping sampling point, using the sequential identification and deletion method for data cleaning; if the sampling point is a normal sampling point, no data cleaning is required; rearranging all sampling points in chronological order, and thus completing the cleaning of the flight telemetry data.

Description

Technical Field

[0001] The present invention relates to a method and device for cleaning flight telemetry data of a liquid rocket engine, belonging to the field of telemetry technology. Background Art

[0002] During the flight of a liquid rocket engine, a large amount of operation monitoring data is collected and transmitted to a ground monitoring station through a telemetry transmission method. Due to interference from conditions such as temperature, distance, and noise during the telemetry data transmission process, the transmission signal often has the characteristic of unstable sampling intervals, resulting in poor accuracy of data analysis results. Improving the quality of flight telemetry data of a liquid rocket engine has a significant impact on improving the accuracy of data analysis and is of great significance for engine anomaly detection and fault diagnosis.

[0003] In the prior art, research has been conducted on general-purpose data cleaning methods, such as statistical methods, data mining methods, pattern rule methods, etc. However, the above methods are not applicable to cleaning unstable telemetry data of liquid rocket engines and cannot meet the requirements for improving the quality of different types of flight telemetry data of liquid rocket engines. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and solve the problem of cleaning flight telemetry data of a liquid rocket engine.

[0005] The object of the present invention is achieved through the following technical solutions:

[0006] A method for cleaning flight telemetry data of a liquid rocket engine, comprising:

[0007] Dividing the flight telemetry data into a steady section and an unsteady section according to the working process of the engine;

[0008] Determining a sampling stability criterion for the flight telemetry data, and classifying the flight telemetry data into normal sampling points, censored sampling points, and overlapping sampling points according to this criterion;

[0009] Performing data cleaning on the flight telemetry data according to the following situations:

[0010] If the sampling point is a censored sampling point in the unsteady section, using the constant filling method for data cleaning;

[0011] If the sampling point is a censored sampling point in the steady state section, using the statistical value filling method for data cleaning;

[0012] If the sampling point is an overlapping sampling point, using the sequential identification and deletion method for data cleaning;

[0013] If the sampling point is a normal sampling point, no data cleaning is required;

[0014] Rearranging all the sampling points in chronological order completes the cleaning of flight telemetry data.

[0015] Preferably, the stable section is N seconds after the liquid rocket engine ignition command moment to the shutdown command moment; the unstable section is within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment; among them, the flight telemetry data is unstable within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment, and the flight telemetry data is stable at other times.

[0016] Preferably, the sampling stability criterion for the flight telemetry data is determined according to the sampling interval and the ideal sampling interval of two sampling points in the measured data.

[0017] Preferably, the sampling stability criterion for the flight telemetry data is:

[0018]

[0019] In the formula, δ i = T real,i - T ideal

[0020] Among them, T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval.

[0021] Preferably, the data cleaning by the constant filling method includes:

[0022] Calculating the number of censored points n:

[0023] n = abs(T real,i | T ideal )

[0024] Among them, | represents integer division, and abs represents taking the absolute value; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval;

[0025] Generating the censored point time series t del :

[0026] t del = [t i + T ideal , t i + 2T ideal ,..., t i + nT ideal T

[0027] Among them, ti Represents the time of the i-th sampling point, T ideal is the ideal sampling interval, T is the transpose symbol;

[0028] Generate the censored point amplitude sequence a del,c :

[0029]

[0030] Generate the censored point sequence p del,c :

[0031] p del,c =[t del ,a del,c .

[0032] Preferably, the data cleaning using the statistical value filling method includes:

[0033] Calculate the number of censored points n:

[0034] n = abs(T real,i |T ideal )

[0035] where, | represents integer division, abs represents taking the absolute value; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, T ideal is the ideal sampling interval;

[0036] Generate the censored point time sequence t del :

[0037] t del =[t i +T ideal ,t i +2T ideal ,...,t i +nT ideal T

[0038] where, t i represents the time of the i-th sampling point, T ideal is the ideal sampling interval, T is the transpose symbol;

[0039] Generate the censored point amplitude sequence a del,s :

[0040]

[0041] where, a i represents the amplitude of the i-th sampling point;

[0042] Generate the censored point sequence p del,s : ​

[0043] p del,s = [t del , a del,s .

[0044] Preferably, the data cleaning using the sequential recognition and deletion method includes:

[0045] Calculate the number of overlapping points m:

[0046] m = abs(ceil(T real,i / T ideal ))

[0047] where ceil represents the ceiling operation and abs represents the absolute value; T real,i is the sampling interval between the i-th and (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval;

[0048] Obtain the time series of overlapping points t over :

[0049] t over = [t i-m , t i-m+1 ,..., t i+m+1 T

[0050] where t i represents the time of the i-th sampling point, and T is the transpose symbol;

[0051] Obtain the amplitude sequence of overlapping points a over :

[0052] a over = [a i-m , a i-m+1 ,..., a i+m+1 T

[0053] where a i represents the amplitude of the i-th sampling point;

[0054] Correct the time series of overlapping points t' over :

[0055] t' over = [t i-m , t i-m+1 ,..., t i T

[0056] Correct the amplitude sequence of overlapping points a' over :

[0057] a' over = [a​​​i-m , a i-m+1 ,..., a i T

[0058] Generate the censoring point sequence p over :

[0059] p over = [t' over , a' over .

[0060] A flight telemetry data cleaning device for a liquid rocket engine, comprising:

[0061] A data partitioning module, configured to partition flight telemetry data into a steady segment and a non-steady segment according to the working process of the engine;

[0062] A discrimination module, configured to determine a sampling stability criterion for flight telemetry data and classify the flight telemetry data into normal sampling points, censoring sampling points, and overlapping sampling points according to this criterion;

[0063] A data cleaning module, configured to perform data cleaning on the flight telemetry data according to the following situations:

[0064] If the sampling point is a censoring sampling point in the non-steady segment, use the constant filling method for data cleaning;

[0065] If the sampling point is a censoring sampling point in the steady segment, use the statistical value filling method for data cleaning;

[0066] If the sampling point is an overlapping sampling point, use the sequential recognition and deletion method for data cleaning;

[0067] If the sampling point is a normal sampling point, no data cleaning is required;

[0068] A sorting module, configured to rearrange all sampling points in chronological order, that is, complete the cleaning of flight telemetry data.

[0069] The present invention has the following beneficial effects compared with the prior art:

[0070] (1) The present invention adopts specific data cleaning methods for telemetry data in different working segments of a liquid rocket engine, improves the effectiveness of the cleaning result of the engine flight telemetry data, solves the problem of evaluation errors caused by low-quality data in the process of engine health status evaluation, and provides an important guarantee for the normal completion of space launch missions;

[0071] ​(2) For different error types of the flight telemetry data of liquid rocket engines, the present invention adopts corresponding data cleaning methods, improving the accuracy of the cleaning results of the engine flight telemetry data, enhancing the accuracy of engine fault diagnosis, and providing important support for the treatment of engine weak links and the improvement of reliability;

[0072] (3) The present invention can simultaneously implement the cleaning of flight telemetry data under different working segments and error types of liquid rocket engines. The proposed method can significantly improve the cleaning efficiency of the flight telemetry data of liquid rocket engines and reduce the time cost of data processing;

[0073] (4) The present invention has the feature of adapting to the improvement of the quality of flight telemetry data of various models of liquid rocket engines and is suitable for large-scale applications in the aerospace field. Description of the Drawings

[0074] Figure 1 is the flowchart of the steps of the method of the present invention.

[0075] Figure 2 is the evaluation result of the sampling stability index of the telemetry data in the non-steady section of the engine.

[0076] Figure 3 is the telemetry data in the non-steady section of the engine after cleaning.

[0077] Figure 4 is the evaluation result of the sampling stability index of the telemetry data in the steady section of the engine.

[0078] Figure 5 is the telemetry data in the steady section of the engine after cleaning. Detailed Embodiment

[0079] To make the objectives, technical solutions and advantages of the present invention clearer, the following will further describe in detail the embodiments of the present invention with reference to the drawings.

[0080] A method for cleaning flight telemetry data of a liquid rocket engine includes the following steps:

[0081] (1) Divide the flight telemetry data into a steady section and a non-steady section according to the working process of the engine;

[0082] (2) Construct an evaluation index for the sampling stability of the engine telemetry data and formulate corresponding judgment criteria, and divide all sampling points into normal sampling points, censored sampling points and overlapping sampling points;

[0083] (3) Adopt the constant filling method, the statistical value filling method or the sequential identification and deletion method for data cleaning according to the different segments and abnormal types to which the sampling points belong;

[0084] (4) Rearrange the normal sampling points and the abnormal sampling points after cleaning in chronological order to obtain high-quality flight telemetry data.

[0085] In step (1), the stable section of the engine is the time period between 3 s after the ignition command and the shutdown command, and the non-stable section is the time period of 3 s after the ignition command and the shutdown command respectively.

[0086] In step (2), the sampling stability evaluation index δ of the engine telemetry data is shown in the following formula.

[0087] δ i = T real,i - T ideal (1)

[0088] Where, T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T· is the ideal sampling interval.

[0089] In step (2), the judgment criterion of the engine telemetry data is shown in the following formula.

[0090]

[0091] In the formula, T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval.

[0092] In step (3), if the sampling point is a censored sampling point in the non-steady state section, the constant filling method is used for data cleaning; if the sampling point is a censored sampling point in the steady state section, the statistical value filling method is used for data cleaning; if the sampling point is an overlapping sampling point in the steady state section or the non-steady state section, the sequential recognition and deletion method is used for data cleaning; if the sampling point is a normal sampling point in the steady state section or the non-steady state section, no data cleaning is required.

[0093] In step (3), the constant filling method first calculates the number of censored points n as shown in the following formula.

[0094] n = abs(T real,i | T ideal ) (3)

[0095] Where, | represents integer division, and abs represents taking the absolute value; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval.

[0096] Secondly, generate the censored point time series t del as shown in the following formula.

[0097] t del = [t i+T ideal ,t i +2T ideal ,...,t i +nT ideal T (4)

[0098] where t i represents the time at the i-th sampling point, T ideal is the ideal sampling interval, n is the number of censored points, and T is the transpose symbol.

[0099] Then, generate the censored point amplitude sequence a del,c as shown in the following formula.

[0100]

[0101] where n is the number of censored points.

[0102] Finally, generate the censored point sequence p del,c as shown in the following formula.

[0103] p del,c =[t del ,a del,c (6)

[0104] In the statistical value filling method in step (3), first calculate the number of censored points n as shown in formula (3).

[0105] Secondly, generate the censored point time sequence t del as shown in formula (4).

[0106] Then, generate the censored point amplitude sequence a del,s as shown in the following formula.

[0107]

[0108] where a i represents the amplitude at the i-th sampling point, and n is the number of censored points.

[0109] Finally, generate the censored point sequence p del,s as shown in the following formula.

[0110] p del,s =[t del ,a del,s (8)

[0111] In the sequential recognition and deletion method in step (3), first calculate the number of overlapping points m as shown in the following formula.

[0112] m = abs(ceil(T real,i / T ideal )) (9) ​

[0113] where ceil represents the ceiling operation and abs represents the absolute value operation; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval.

[0114] Secondly, obtain the overlapping point time series t over as shown in the following formula.

[0115] t over = [t i-m , t i-m+1 ,..., t i+m+1 T (10)

[0116] where t i represents the time of the i-th sampling point, and m is the number of overlapping points.

[0117] Then, obtain the overlapping point amplitude series a over as shown in the following formula.

[0118] a over = [a i-m , a i-m+1 ,..., a i+m+1 T (11)

[0119] where a i represents the amplitude of the i-th sampling point, and m is the number of overlapping points.

[0120] Then, correct the overlapping point time series t' over as shown in the following formula.

[0121] t' over = [t i-m , t i-m+1 ,..., t i T (12)

[0122] where t i represents the time of the i-th sampling point, and m is the number of overlapping points.

[0123] Then, correct the overlapping point amplitude series a' over as shown in the following formula.

[0124] a' over = [a i-m , a i-m+1 ,..., a i T (13)

[0125] where a i ​​​​Represents the amplitude of the i-th sampling point, and m is the number of overlapping points.

[0126] Finally, generate the censored point sequence p over See the following formula.

[0127] p over = [t' over , a' over (14)

[0128] Example 1

[0129] (1) Clean the flight telemetry data of a certain type of liquid rocket engine. According to the engine command, select 142.5 - 145.5 s and 322.7 - 325.7 s as the non-steady sections, and select 145.5 - 322.7 s as the steady section.

[0130] (2) Evaluate the original telemetry data using the constructed sampling stability index for engine telemetry data, as Figure 2 shown. It can be found that there are large deviations (up to 0.1 s) in the sampling interval error of the original data near 143.2 s and 143.4 s. Since the ideal sampling interval T ideal = 0.195 ms, according to formula (2), δ ∈ [10T ideal , +∞), so the above two sampling points are censored sampling points, and the rest of the sampling points are normal sampling points.

[0131] (3) Since the censored sampling points are located in the non-steady section, the constant filling method is used for data cleaning. Fill in the censored values according to formulas (3) - (6), and rearrange the normal sampling points and the cleaned sampling points in chronological order to obtain the cleaned flight telemetry data as Figure 3 shown.

[0132] Example 2

[0133] (1) Clean the flight telemetry data of a certain type of liquid rocket engine. According to the engine command, select 142.5 - 145.5 s and 322.7 - 325.7 s as the non-steady sections, and select 145.5 - 322.7 s as the steady section.

[0134] (2) Evaluate the original telemetry data using the constructed sampling stability index for engine telemetry data, as Figure 4 shown. It can be found that there is a large deviation (up to 1.8 s) in the sampling interval error of the original data near 224 s. Since the ideal sampling interval T ideal = 0.195 ms, according to formula (2), δ ∈ [10T ideal , +∞), so the above sampling point is a censored sampling point, and the rest of the sampling points are normal sampling points.

[0135] (3) Since the censored sampling points are located in the steady state section, the statistical value filling method is used for data cleaning. The censored values are filled according to Equations (3), (4), (7), (8) and (9), and the normal sampling points and the sampled points after cleaning are rearranged in chronological order to obtain the flight telemetry data after cleaning as Figure 5 shown below.

[0136] Example 3

[0137] A flight telemetry data cleaning device for a liquid rocket engine, comprising:

[0138] A data division module, configured to divide flight telemetry data into a steady state section and a non-steady state section according to the working process of the engine;

[0139] A discrimination module, configured to determine a sampling stability criterion for flight telemetry data, and divide the flight telemetry data into normal sampling points, censored sampling points and overlapping sampling points according to the criterion;

[0140] A data cleaning module, configured to clean the flight telemetry data according to the following situations:

[0141] If the sampling point is a censored sampling point in the non-steady state section, the constant filling method is used for data cleaning;

[0142] If the sampling point is a censored sampling point in the steady state section, the statistical value filling method is used for data cleaning;

[0143] If the sampling point is an overlapping sampling point, the sequential recognition and deletion method is used for data cleaning;

[0144] If the sampling point is a normal sampling point, no data cleaning is required;

[0145] A sorting module, configured to rearrange all sampling points in chronological order, that is, complete the cleaning of flight telemetry data.

[0146] The steady state section is N seconds after the liquid rocket engine ignition command moment to the shutdown command moment; the non-steady state section is within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment; wherein within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment, the flight telemetry data is unstable, and the flight telemetry data is stable at other times.

[0147] The sampling stability criterion for the flight telemetry data is determined according to the sampling interval and the ideal sampling interval of two sampling points in the measured data.

[0148] The sampling stability criterion for the flight telemetry data is:

[0149]

[0150] In the formula, δ i = T real,i - T ideal

[0151] Wherein, T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval.

[0152] The content not detailed in the description of the present invention belongs to the well-known technology of those skilled in the art.

[0153] Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solution of the present invention by using the methods and technical content disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for cleaning flight telemetry data of a liquid rocket engine, characterized in that, Including: Dividing the flight telemetry data into a stable segment and an unstable segment according to the engine working process; Determining the criterion for the sampling stability of the flight telemetry data, and classifying the flight telemetry data into normal sampling points, censored sampling points, and overlapping sampling points according to this criterion; the criterion for the sampling stability of the flight telemetry data is determined according to the sampling interval and the ideal sampling interval of two sampling points in the measured data; Performing data cleaning on the flight telemetry data according to the following situations: If the sampling point is a censored sampling point in the unstable segment, using the constant filling method for data cleaning; If the sampling point is a censored sampling point in the stable segment, using the statistical value filling method for data cleaning; If the sampling point is an overlapping sampling point, using the sequential identification and deletion method for data cleaning; If the sampling point is a normal sampling point, no data cleaning is required; Rearranging all sampling points in chronological order, that is, completing the cleaning of the flight telemetry data; Performing data cleaning using the constant filling method includes: Calculating the number of censored points n: n = abs(T real,i |T ideal ) where, | represents integer division, and abs represents taking the absolute value; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, and T ideal is the ideal sampling interval; Generate the censored point time series t del : t del = [t i + T ideal , t i + 2T ideal ,..., t i + nT ideal T ​ where, t i represents the time of the i-th sampling point, T ideal is the ideal sampling interval, and T is the transpose symbol; Generate the censored point amplitude sequence a del,c : Generate the censored point sequence p del,c : p del,c = [t del , a del,c ; Performing data cleaning using the statistical value filling method includes: Generate the censored point amplitude sequence a del,s : where a i represents the amplitude of the i-th sampling point; Generate the censoring point sequence p del,s : p del,s = [t del , a del,s ; Performing data cleaning using the sequential identification and deletion method includes: Calculating the number of overlapping points m: m = abs(ceil(T real,i / T ideal )) Where ceil represents the ceiling operation; Obtain the overlapping point time series t over : t over = [t i-m , t i-m+1 ,..., t i+m+1 T ​ Obtain the amplitude sequence a of overlapping points over : a over = [a i-m , a i-m+1 ,..., a i+m+1 T ​ Corrected overlapping point time series t′ over : t' over = [t i-m , t i-m+1 ,..., t i T ​ Corrected overlapping point amplitude sequence a′ over : a' over = [a i-m , a i-m+1 ,..., a i T ​ Generate the censored point sequence p over : p over = [t' over , a' over .

2. The flight telemetry data cleaning method according to claim 1, wherein The stable segment is N seconds after the liquid rocket engine ignition command moment to the shutdown command moment; the unstable segment is within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment; among them, within N seconds after the liquid rocket engine ignition command moment and within M seconds after the liquid rocket engine shutdown command moment, the flight telemetry data is unstable, and the flight telemetry data is stable at other times.

3. The flight telemetry data cleaning method according to claim 1, wherein The criterion for the sampling stability of the flight telemetry data is: where δ i = T real,i - T ideal .

4. A flight telemetry data cleaning device for a liquid rocket engine, characterized in that, Including: A data division module for dividing the flight telemetry data into a stable segment and an unstable segment according to the engine working process; A discrimination module for determining the criterion for the sampling stability of the flight telemetry data and classifying the flight telemetry data into normal sampling points, censored sampling points, and overlapping sampling points according to this criterion; the criterion for the sampling stability of the flight telemetry data is determined according to the sampling interval and the ideal sampling interval of two sampling points in the measured data; A data cleaning module for performing data cleaning on the flight telemetry data according to the following situations: If the sampling point is a censored sampling point in the unstable segment, using the constant filling method for data cleaning; If the sampling point is a censored sampling point in the stable segment, using the statistical value filling method for data cleaning; If the sampling point is an overlapping sampling point, using the sequential identification and deletion method for data cleaning; If the sampling point is a normal sampling point, no data cleaning is required; A sorting module for rearranging all sampling points in chronological order, that is, completing the cleaning of the flight telemetry data; Performing data cleaning using the constant filling method includes: Calculating the number of censored points n: n = abs(T real,i |T ideal ) Among them, | represents integer division, and abs represents absolute value; T real,i is the sampling interval between the i-th and the (i + 1)-th sampling points in the measured data, T ideal is the ideal sampling interval; Generate the censored point time series t del : t del = [t i + T ideal , t i + 2T ideal ,..., t i + nT ideal T ​ where t i represents the time of the i-th sampling point, T ideal is the ideal sampling interval, and T is the transpose symbol; generate the censored point amplitude sequence a del,c : Generate the censored point sequence p del,c : p del,c = [t del , a del,c ; Performing data cleaning using the statistical value filling method includes: Generate the censored point amplitude sequence a del,s : Among them, a i represents the amplitude of the i-th sampling point; Generate the censoring point sequence p del,s : p del,s = [t del , a del,s ; Performing data cleaning using the sequential identification and deletion method includes: Calculating the number of overlapping points m: m = abs(ceil(T real,i / T ideal )) Where ceil represents the ceiling operation; Obtain the overlapping point time series t over : t over = [t i-m , t i-m+1 ,..., t i+m+1 T ​ Obtain the amplitude sequence a of overlapping points over : a over = [a i-m , a i-m+1 ,..., a i+m+1 T ​ Corrected overlapping point time series t′ over : t' over = [t i-m , t i-m+1 ,..., t i T ​ Amplitude sequence a' of corrected overlapping points over : a' over = [a i-m , a i-m+1 ,..., a i T ​ Generate the censoring point sequence p over : p over = [t' over , a' over .

5. The flight telemetry data cleaning device according to claim 4, wherein The steady section is from N seconds after the ignition command moment of the liquid rocket engine to the shutdown command moment; the non-steady section is within N seconds after the ignition command moment of the liquid rocket engine and within M seconds after the shutdown command moment of the liquid rocket engine; among them, the flight telemetry data is unstable within N seconds after the ignition command moment of the liquid rocket engine and within M seconds after the shutdown command moment of the liquid rocket engine, and the flight telemetry data is stable at other times.

Images