Abnormal output current detection method for battery pack during the earth shadow period of BeiDou-3 navigation satellite

Abstract

The present invention provides a method for detecting abnormal output current of a battery pack during the earth's shadow period of a BeiDou-3 navigation satellite, comprising the following steps: a first step: determining the earth's shadow period of the BeiDou-3 satellite; a second step: aligning and summing the output current of the battery; a third step: detecting abnormal output current of the battery pack and outputting the abnormal output current during the abnormal period; the method can effectively overcome the diagnosis difficulties caused by the short latent time and sudden occurrence of the power supply abnormality of the BeiDou-3 navigation satellite, and gain time for early detection and rapid disposal of the power supply abnormality of the BeiDou-3 series satellites during the earth's shadow period.

Description

Technical Field

[0001] The invention belongs to the field of on-orbit spacecraft telemetry anomaly detection and analysis, and specifically relates to a method for detecting anomaly in output current of a battery pack during an earth shadow period of a Beidou-3 navigation satellite. Background Art

[0002] As the main energy storage device of spacecraft, the battery pack mainly needs to meet the load requirements of the spacecraft during the Earth's shadow period. However, after multiple cycles, the capacity and voltage of the battery pack will inevitably decay. The BeiDou-3 navigation satellite has an Earth's shadow period of about 92 days each year, and deep discharge during the Earth's shadow period seriously affects the safety of the spacecraft. Real-time monitoring of the output current of the battery pack during the Earth's shadow period to avoid reaching the safety threshold is an important measure to effectively prevent over-discharge.

[0003] At present, there are many methods for abnormal detection or fault diagnosis of battery over-discharge, but most of them require modeling and simulation or correlation mining of multiple factors that affect the battery pack attenuation process, such as average discharge depth, number of cycles, temperature control, initial discharge capacity, and whether reprocessing is performed on orbit. This is not conducive to cross-model promotion and it is difficult to meet real-time requirements. Summary of the invention

[0004] Aiming at the problem of untimely detection of power system anomalies of BeiDou-3 navigation satellite after it enters the Earth's shadow period, a method for anomaly detection based on multiple historical Earth's shadow cycles and extracting the output current envelope of parameter storage batteries was proposed.

[0005] The invention is implemented as follows:

[0006] A method for detecting power anomaly during the earth shadow period of a BeiDou-3 navigation satellite (hereinafter referred to as "BeiDou-3" satellite) specifically comprises the following steps:

[0007] Step 1: Determination of the Earth shadow period of the BeiDou III satellite;

[0008] Step 2: Alignment and summation of battery output current;

[0009] Step 3: Abnormal detection of power supply and output during abnormal period.

[0010] The first step as described above specifically includes the following steps:

[0011] Step 1.1: First, you need to complete the prediction of the "Bei San" satellite's earth shadow period. The "Bei San" satellite's earth shadow period has two periods, starting around the spring equinox and autumn equinox each year, and each period is about 46 days. Record the earth shadow period start time DT s , Earth's shadow period ends at DT e Parameter Battery working mode identification (denoted as P mode ) is a discrete quantity, P mode =0 means illuminator, Pmode =255 indicates the Earth's shadow period. mode The time when it jumps from 0 to 255 is DT s , P mode The time when the value changes from 255 to 0 is DT e .

[0012] Step 1.2: Verify the prediction result of step 1.1 and output the accurate value of the earth's shadow period. The battery output current is recorded as P I Take a very short period of time, record it as ΔT (ΔT satisfies greater than P I The sampling interval is usually 30 seconds. Record the battery output current at DTs-ΔT and DTs+ΔT (denoted as P I ) are P Is0 and P Is1 , if ||P Is0 -P Is1 ||>P Is0 , (||·|| means taking the absolute value of the variable in the brackets), then the exact value of the beginning time of the Earth's shadow period is DT s ; Otherwise, record DT s =DT s -ΔT, repeat the above process until the conditions are met. Similarly, the exact value of the end time of the Earth's shadow period can be determined DT e .

[0013] The second step as described above specifically includes the following steps:

[0014] Step 2.1: Alignment of battery output current. Earth shadow period S i Internal I The time series is TS i '=[(t i0 , v i0 ), (t i1 , v i1 ),…,(t iNi-1 , v iNi-1 ), (t iNi , v iNi )],t iNi For S i At the Ni moment, v iNi For iNi , P I The value of N i For S i The length of the sequence. The most recent earth shadow period of the Beisan satellite is S j , the shadow period here is also the abnormal detection interval. j There are K earth shadow periods before, and the shadow periods of each place are S0, ..., S i , ..., Sk , each eclipse period P I The sequence lengths are N0, …, N i , …, N k . Take N as the minimum of the above sequence lengths, i.e., N = min(N0, …, N i , …, N k ). Then the time series of P within the new eclipse periods Si is TS I = [(t i , v i0 ), (t i0 , v i1 ), …, (t i1 , v iN-1 ), (t iN-1 , v iN ), (t iN )].

[0015] Step 2.2: Complete the summation calculation of the battery output current. Then, for the first K eclipse periods, the summation result of the time series of P I is TS c = [(t c0 , v c0 ), (t c1 , v c1 ), …, (t cN-1 , v cN-1 ), (t cN , v cN )], where t cN = t iN , v cN = (v 0N + v 1N + … v K-1N + v KN ) / K. At the same time, it is necessary to save the summation result TS c , and perform anomaly detection according to the third step when entering the eclipse period S j .

[0016] The third step as described above specifically includes the following steps:

[0017] Step 3.1: Align the time scales in the anomaly detection interval. Denote the time series of P j within S I as TS j ’ = [(t j0 , v j0 ), (t j1 , v j1 ), …, (t jM-1 , v jM-1 ), (t jM , v jM )]. If M < N, then TS c=[(t c0 , v c0 ), (t c1 , v c1 ),…,(t cM-1 , v cM-1 ), (t cM , v cM )], TS j '=TS j ; On the contrary, if M>N, then TS j =[(t j0 , v j0 ), (t j1 , v j1 ),…,(t jN-1 , v jN-1 ), (t jN , v jN )], that is, only detect P in the [0,N] time interval I Sequence, default (N, M] time interval P I The sequence is normal. Because the length of each earth shadow period of the same navigation satellite is basically the same, the difference between N and M is close to 0.

[0018] Step 3.2: Detect the battery output current exceeding the limit. k , if ||v j -v C ||>0.1*v c , then the abnormal flag AF at time t is marked t True; otherwise, if ||v j -v C ||≤0.1*v c , then the abnormal flag AFt at time t is False, repeat the above steps to mark S j Internal I The abnormal flag of each moment in the sequence. t Add the time t at which True is added to the candidate alarm list L AF .

[0019] Step 3.3: Complete the alarm output during the abnormal detection period. The output process is as follows: If AFt is True at time t, there are two cases: i) AF t-2 and AF t-1 , AF t-1 and AF t+1 , AF t+1 and AF t+2 If any of the three groups does not satisfy the condition True at the same time, the time is removed from the candidate alarm list L AF ii) Otherwise, do not remove the candidate alarm list. AFAll elements in the output final alarm list L AF .

[0020] The beneficial effects of the present invention are:

[0021] 1) A universal method for detecting power anomalies of BeiDou-3 navigation satellites during the Earth's shadow period is designed. The battery pack status is determined by multiple historical period parameters. This method is independent of the navigation satellite platform and the power supply component model, thus avoiding threats to the safety of the on-orbit spacecraft platform caused by battery anomalies during the Earth's shadow period.

[0022] 2) This method can effectively overcome the diagnostic difficulties caused by the short latency and sudden occurrence of power anomalies of BeiDou-3 navigation satellites, and buy time for early detection and rapid disposal of power anomalies of BeiDou-3 series satellites during the Earth shadow period. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of a method for detecting abnormal output current of a battery pack during the Earth shadow period of the "Bei San" satellite;

[0024] Figure 2 It is a schematic diagram of an implementation method of a method for detecting abnormal output current of a battery group during the Earth shadow period of the "Bei San" satellite. DETAILED DESCRIPTION

[0025] The present invention is further described below in conjunction with the accompanying drawings and embodiments. The present invention includes but is not limited to the following embodiments.

[0026] Reference Figure 1 , Figure 2 , which is a schematic diagram of an application case of the present invention, is divided into three main steps:

[0027] 1) Determination of the Earth Shadow Period of the BeiDou Satellite. Select a BeiDou navigation satellite and mode Get the nearest shadow period S i The estimated start and end times of all three previous Earth shadow periods (respectively denoted as DT s and DT e ). DT for the above 6 Earth Shadow Periods s and DT e , compare the P corresponding to the 30 seconds before and the 30 seconds after the moment I Value (respectively denoted as P I0 and P I1 ) The absolute value of the difference. The results are all greater than P I0 . Therefore, the start and end time of the above three earth shadow periods are determined as DT s and DT e .

[0028] 2) Alignment and summation of battery output current. To demonstrate the effect, only the above three earth shadow periods ( Figure 2 The 100 sampling points of the 22nd to 24th days when the discharge depth is the largest (middle dashed line). The above three sequences are summed and averaged, and recorded as TS c , and save the calculation result TS c .

[0029] 3) Battery abnormality detection and abnormal period output. j ( Figure 2 The solid line in the middle) has 45 days, and the same 100 sampling points are taken from the 22nd to the 24th day. According to step 3.1, TS j Sequence and TS c The absolute value of the difference between the 100 corresponding points and 0.1 times TS c The 1st to 18th, 73rd to 75th and 78th to 100th points have abnormal flags AF t True, add these thirty-four points to the candidate alarm list AF t According to step 3.3, the 18th, 73rd to 75th and 78th points meet the requirements of front and rear AF. t Not all are True at the same time, so move these five moments out of AF t , output the final alarm list.

[0030] In summary, this method selects the typical parameter of the energy subsystem, the battery output current, and learns the performance trend characteristics of the BeiDou-3 power supply during the Earth shadow period based on multiple historical Earth shadow cycles.

[0031] This method can effectively overcome the diagnostic difficulties caused by the short latency and sudden occurrence of power anomalies of Beidou-3 navigation satellites, and buy time for early detection and rapid disposal of power anomalies of Beidou-3 series satellites during the Earth shadow period.

Claims

1. A method for detecting abnormal output current of a battery pack during the Earth shadow period of BeiDou-3 navigation satellite, characterized in that: The steps include: Step 1: Determine the Earth shadow period of BeiDou-3 satellite; Step 2: Alignment and summation of battery output current; Step 3: Abnormal detection of storage power supply and output during abnormal period; The first step specifically includes the following steps: Step 1.1: First, you need to complete the prediction of the earth shadow period of the Beidou-3 satellite, record the start time of the earth shadow period DTs, and the end time of the earth shadow period DTe; the parameter battery working mode identifier is recorded as Pmode as a discrete quantity, Pmode = 0 represents the illuminator, Pmode = 255 represents the earth shadow period, record the moment when Pmode jumps from 0 to 255 as DTs, and the moment when Pmode jumps from 255 to 0 as DTe; Step 1.2: Verify the prediction result of step 1.1 and output the accurate value of the earth's shadow period; record the battery output current as PI, take a very small time period, record it as ΔT, where ΔT satisfies the sampling interval greater than PI, generally 30 seconds; record the battery output current at DTs-ΔT and DTs+ΔT, record the values ​​of PI as PIs0 and PIs1 respectively, if ||P Is0 -P Is1 ||>PIs0, ||·|| means taking the absolute value of the variable in the brackets, then the precise value of the beginning time of the earth's shadow period is DTs; otherwise, DTs=DTs-ΔT, repeat the above process until the condition is met, and the precise value of the ending time of the earth's shadow period DTe can be determined in the same way; The second step specifically includes the following steps: Step 2.1: Alignment of battery output current: The time series of PI in the earth shadow period Si is TSi'=[(ti0, vi0), (ti1, vi1), ..., (tiNi-1, viNi-1), (tiNi, viNi)], tiNi is the Ni-th moment in Si, viNi is the value of PI at tiNi, and Ni is the sequence length of Si; the most recent earth shadow period of BeiDou-3 satellite is Sj, which is also the abnormality detection interval, and the time series of BeiDou-3 satellite from From the launch to Sj, there are K earth shadow periods, and the shadow periods are S0, ..., Si, ..., Sk, and the length of the PI sequence of each earth shadow period is N0, ..., Ni, ..., Nk; N is taken as the minimum value of the above sequence length, that is, N = min (N0, ..., Ni, ..., Nk); then the time series of PI in each new shadow period Si is TSi = [(ti0, vi0), (ti1, vi1), ..., (tiN-1, viN-1), (tiN, viN)]; Step 2.2: After the summation calculation of the battery output current is completed, the time series summation result of PI in the first K earth shadow periods is TSc=[(tc0, vc0), (tc1, vc1), ..., (tcN-1, vcN-1), (tcN, vcN)], where tcN=tiN, vcN=(v0N+v1N+...vK-1N+vKN) / K. At the same time, the summation result TSc needs to be saved, and anomaly detection is performed according to the third step when entering the earth shadow period Sj; The third step specifically includes the following steps: Step 3.1: Align the time scales of the anomaly detection intervals; Denote the time series of PI within Sj as TSj’ = [(tj0, vj0), (tj1, vj1), …, (tjM-1, vjM-1), (tjM, vjM)]; If M < N, then TSc = [(tc0, vc0), (tc1, vc1), …, (tcM-1, vcM-1), (tcM, vcM)], and TSj‘ = TSj; Conversely, if M > N, then TSj = [(tj0, vj0), (tj1, vj1), …, (tjN-1, vjN-1), (tjN, vjN)], that is, only the PI sequence within the time interval [0, N] is detected, and the PI sequence in the time interval (N, M] is defaulted to be normal because the length of each eclipse period of the same navigation satellite is basically the same, and the difference between N and M is close to 0; Step 3.2: Detect the battery output current over-limit abnormality; at a certain time tk, if ||v j -v C ||>0.1*vc, then the abnormal flag AFt at time t is True; otherwise, if ||v j -v C ||≤0.1*vc, then mark the abnormal flag AFt at time t as False, repeat the above steps to mark the abnormal flag of each moment of the PI sequence in Sj; add the moment t when AFt is True to the candidate alarm list LAF; Step 3.3: Complete the alarm output for the anomaly detection period, and the output process is as follows: If AFt is True at time t, then there are the following two cases: i) If any one of the three groups of AFt-2 and AFt-1, AFt-1 and AFt+1, AFt+1 and AFt+2 is not both True, then remove the time from the candidate alarm list LAF; ii) Otherwise, do not remove it from the candidate alarm list, traverse all elements in LAF according to the above steps, and output the final alarm list LAF.

Images