Reusable liquid rocket engine on-line fault detection method

Abstract

The application provides a reusable liquid rocket engine online fault detection method, aiming at the large range variable thrust of the reusable engine, following the engine working condition change, realizing real-time online adjustment of the fault threshold, forming a red line shutdown strategy based on the online dynamic threshold, giving early warning and judgment to the faults occurring in the engine working process, so as to take effective measures in time, and protect the safety of the carrier rocket and the load.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of spaceflight reliability, and particularly relates to an online fault detection method for a reusable liquid rocket engine. BACKGROUND

[0002] Reusable launch vehicles are an important development direction for future spaceflight, and liquid rocket engines, as the core products of launch vehicles, directly determine the success or failure of the flight mission of the launch vehicle whether the engine works normally.

[0003] With the development of reusable launch technology, reusable launch vehicles will have a certain power redundancy capability, that is, when one or more engines fail, the faulty engine can be shut down for processing, and the thrust and swing angle of other normal working engines can be adjusted to ensure the success of the flight mission. For example, the first stage of the Starship of the SpaceX company is composed of 33 Raptor engines. According to the carrying capacity and attitude adjustment capacity of the Starship, the flight mission can still be successful even if five engines fail. Through power redundancy, although the load transportation capacity is sacrificed, the requirement for engine mission reliability is reduced, and the mission reliability level of the launch vehicle is greatly improved.

[0004] Reusable liquid rocket engines will have power redundancy capability, and timely and effective measures can be taken after engine failure to protect the safety of the launch vehicle and the load. The engine online fault diagnosis is of great significance, and the occurrence and development of engine failure are extremely rapid and destructive. If the failure is not handled in time after the failure occurs, an explosion may occur, affecting the safety of the launch vehicle and the load. Therefore, the fault diagnosis of the engine needs to be timely.

[0005] The fault diagnosis of a liquid rocket engine usually uses the size of signals such as pressure, flow, rotation speed, vibration, temperature and the like measured by sensors at key positions of the engine to determine whether the engine is abnormal. In the normal working condition of the engine, the parameters are in a relatively stable interval, and once a fault occurs, some signals may increase or decrease greatly, and once exceeding a certain range, the engine is determined to be faulty. For a one-time use engine, the regulation range of the thrust during flight is very small, generally within 5%, so a fixed threshold red line shutdown strategy is usually adopted, that is, in the rated thrust, the key parameter signals such as pressure, flow, rotation speed, vibration, temperature and the like at the positions of the engine are in a relatively stable range, and within this range, a certain margin, generally 10%-30%, is added upward and downward respectively as the criterion for online fault diagnosis of the engine. For a reusable engine, large range variable thrust is required during working, and the regulation range will exceed 40%. The large regulation of the engine thrust will cause the key parameters at the positions of the engine to change greatly, and during the change process, the parameters are in a non-stable state, so the fixed threshold determination method cannot meet the needs of online fault diagnosis of the engine. SUMMARY

[0006] Therefore, the purpose of the present application is to provide an online fault detection method for a reusable liquid rocket engine, which can realize online fault detection of the engine.

[0007] An online fault detection method for a reusable liquid rocket engine, comprising a steady state working phase online fault detection method, specifically comprising:

[0008] For the collected fault characteristic parameter sequence, an autoregressive moving average model ARMA is used for fitting, and the future fault characteristic parameter values are predicted:

[0009]

[0010] In the formula, x t represents the predicted value of a certain fault characteristic parameter at time t; x t-1 , x t-2 , …, x t-n represent the measured values of a certain fault characteristic parameter at times t-1, t-2, …, t-n; a t-1 , a t-2 , …, a t-m respectively represent the difference between the measured value at time t-1 and the measured value at time t-2, the difference between the measured value at time t-2 and the measured value at time t-3, and the difference between the measured value at time t-m+1 and the measured value at time t-m of the fault characteristic parameter; represents an autoregressive parameter; N represents the autoregressive order; θ j represents a moving average parameter; M represents the moving average order; represents white noise interference;

[0011] The value of the fault characteristic parameter at each moment is multiplied by a set coefficient to obtain a red line shutdown threshold, and the measured value of the fault characteristic parameter is compared with the red line shutdown threshold to determine whether a fault occurs.

[0012] Further, the method further comprises an online fault detection method in a variable working condition stage, and specifically comprises:

[0013] 1) a function relationship between a steady working condition and an engine fault characteristic parameter is constructed:

[0014] K=a·F+b (2)

[0015] In the formula, K represents the fault characteristic parameter of the engine; F represents the working condition of the engine before adjustment;

[0016] 2) for a variable working condition delay response and a non-steady state change process, the change law of each parameter in the working condition change process of the engine is simulated, parameters t d , t, F0, F1 are taken as inputs, and the fault characteristic parameter K is taken as output, a machine learning method is used for training, a change law of the engine in the working condition change process is formed, that is, a mapping relationship between the fault characteristic parameter K and the parameters t d , t, F0, F1 is obtained:

[0017] K=f(t d ,t,F0,F1) (3)

[0018] In the formula, t d represents the working condition adjustment delay time; t represents the starting moment of the working condition adjustment of the engine; F0 represents the working condition before adjustment of the engine; F1 represents the working condition after adjustment of the engine;

[0019] 3) according to the working condition adjustment value before and after the working condition change, the fault characteristic parameter value K0 before working condition adjustment and the fault characteristic parameter value K1 after working condition adjustment are calculated according to formula (2), that is, the upper and lower limit values of the fault characteristic parameters before and after working condition adjustment are obtained; then the curve of the fault characteristic parameter K is obtained according to formula (3);

[0020] 4) based on the curve of the fault characteristic parameter K obtained above, multiply a set coefficient to obtain the red line shutdown threshold at each moment under the variable working condition, compare the measured value of the fault characteristic parameter at each moment with the red line shutdown threshold, and determine whether there is a fault.

[0021] Further, in the steady state working phase online fault detection method, before predicting the future fault characteristic parameter value, the traditional fixed red line method is used to judge whether the extracted fault characteristic parameter data is within the red line range, if it is outside the red line, an abnormal signal is sent, and it is judged as a fault.

[0022] Further, in the steady state working phase online fault detection method, before predicting the future fault characteristic parameter value, it is judged whether the data is stable, if the data is not stable, difference calculation and other methods are used to smooth the data, if the data is stable, ARMA model is used for fitting, and the future fault characteristic parameter value is predicted.

[0023] The present application has the following beneficial effects:

[0024] The purpose of the present application is to provide an online fault detection method for reusable liquid rocket engine, which can adjust the fault threshold value in real time online, form a red line shutdown strategy based on online dynamic threshold, and give early warning and judgment for the faults occurred in the engine working process, so as to take effective measures in time and protect the safety of the launch vehicle and the load. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 It is a certain type of engine working condition adjustment curve in the embodiment of the present application;

[0026] Figure 2 It is a certain key part pressure measured value in the steady state working process of a certain type of engine in the embodiment of the present application;

[0027] Figure 3 It is the implementation flow chart of the adaptive threshold and alarm strategy of the present application;

[0028] Figure 4 It is a comparison chart of actual measurement value and prediction result of a certain parameter of a certain engine in the embodiment of the present application;

[0029] Figure 5 It is the adaptive threshold calculation result of a certain parameter of a certain engine in the steady state working condition in the embodiment of the present application;

[0030] Figure 6 It is a key parameter change curve chart of the engine in the working condition change process in the embodiment of the present application;

[0031] Figure 7 It is the function relationship between the working condition and the key parameter of a certain type of engine in the embodiment of the present application;

[0032] Figure 8A comparison chart of actual curve and calculation result of key parameter change of engine in working condition change process in the embodiment of the present application is shown in the following figure.

[0033] Figure 9 An online dynamic threshold alarm area of a certain type of engine in the whole working process in the embodiment of the present application is shown in the following figure. DETAILED DESCRIPTION

[0034] The present application is described in detail below with reference to the accompanying drawings and embodiments.

[0035] A liquid rocket engine needs to be adjusted in working condition according to the needs of a flight mission in the flight mission, and the engine mainly adjusts the engine thrust to achieve the purpose of working condition adjustment. The working condition adjustment and the change curve of each key parameter of a certain type of engine are shown in the following figure. Figure 1 The sampling frequency is 100 Hz, the abscissa is the sampling sequence, the nth sampling point in actual work is the working condition value at 10*n milliseconds; the ordinate is the percentage of rated thrust, and 100% working condition is the rated thrust.

[0036] The engine mainly monitors the working state through pressure, flow, speed, temperature, vibration and other sensor signals. The threshold of each sensor is set by the red line method, and whether the parameters collected by the sensor exceed the red line threshold is used as the basic basis for engine fault diagnosis. In order to minimize the possibility of false shutdown, a multi-sensor joint decision strategy is used to determine whether the engine needs to execute the shutdown program. As the smallest detection unit for fault diagnosis, the determination of each sensor threshold affects the timeliness and accuracy of fault diagnosis. The engine mainly uses the fixed red line method for fault detection in one use, and if the actual measurement of the key parameters exceeds the threshold, the engine will be shut down urgently. The engine needs to be adjusted in working condition according to the needs of the launch vehicle during flight, in order to improve the fault detection accuracy and make the fault detection more timely, a red line shutdown method based on dynamic threshold is studied. As can be seen from the thrust adjustment curve in the above figure, the engine mainly includes a steady state working stage and a variable working condition stage during the working process, and an online fault detection method is given based on the adaptive threshold adjustment scheme for the two cases.

[0037] Embodiment 1: Online fault detection method in steady state working stage

[0038] Due to the harsh working conditions of the engine, there are intense combustion and high-speed rotating parts, and even in the steady state working stage, the key parameters of the engine still have certain volatility, as shown in the following figure. According to the working condition adjustment curve, Figure 1 in the 0s-4s time period, the engine does not adjust the working condition, and the actual measured value of the pressure of a certain key part of the engine is shown in the following figure. Figure 2

[0039] ​Under stable working condition, the engine does not send working condition adjustment signal, the working condition does not change, the engine pressure, flow, speed and other fault characteristic parameters are in a relatively stable state, and the data acquisition system collects relevant data according to a certain sampling frequency, generally 100 Hz. For the fault characteristic parameters under stable working condition, it is a time series signal with stationary characteristics, that is:

[0040] If any subset of the time series satisfies: the expectation μ t =E(y t ) = μ y is a finite constant, that is, it does not change with time t; the covariance γ s,t =Cov(y s ,y t ) exists and only depends on the interval size |s-t| of s and t; then the sequence is called weakly stationary (covariance stationary, wide stationary) sequence, which is usually referred to as stationary sequence.

[0041] For stationary time series, autoregressive moving average model (ARMA model) is used for fitting and prediction of future time series. ARMA model is a combination of autoregressive model and moving average model, which describes the complex sequence dependence in a very simple model form. The basic form of ARMA model is shown in the following formula:

[0042]

[0043] In the formula: x t represents the predicted value of a certain fault characteristic parameter at time t; x t-1 , x t-2 , …, x t-n represent the measured values of a certain fault characteristic parameter at time t-1, t-2, …, t-n; a t-1 , a t-2 , …, a t-m respectively represent the difference between the measured value at time t-1 and the measured value at time t-2, the difference between the measured value at time t-2 and the measured value at time t-3, and the difference between the measured value at time t-m+1 and the measured value at time t-m; represents the autoregressive parameter; N represents the autoregressive order; θ j represents the moving average parameter; M represents the moving average order; represents white noise interference.

[0044] According to the characteristics of the engine pressure, speed, flow and other fault characteristic parameter data collection, using the collected test data, the ARMA model is used to fit the data of the steady state condition, and then the fault characteristic parameters are predicted. The predicted value is compared with the actual value. If it is greater than the red line shutdown threshold, an alarm signal of engine abnormal working is given. The adaptive threshold of steady state condition and the implementation process of alarm strategy are shown in Figure 3

[0045] After the engine starting process or a period of time after the variable working condition signal is sent, the engine working condition is in non-steady state. The threshold algorithm for variable working condition process will be studied in the next section. After the engine working condition is stable, the following process is used for judgment:

[0046] 1) Extract the fault characteristic parameter data of the t0-t1 time period (set t s = t1-t0);

[0047] 2) In this period of time, use the fixed red line method to judge whether the fault characteristic parameters in the t0-t1 time period are within the red line range. If it is outside the red line, an abnormal signal is sent;

[0048] 3) If the fault characteristic parameter data of the t0-t1 time period is within the red line range, use the Augmented Dickey-Fuller method to test whether the data is stable;

[0049] 4) If the data is not stable, use difference calculation and other methods to smooth the data;

[0050] 5) If the data has stationarity, use the data of the t0-t1 time period to determine the order of the ARMA model. Through the AIC criterion, the order n of the AR model and the order m of the MA model are determined respectively to ensure the fitting accuracy of the model;

[0051] 6) After the model order is determined, use the data of the t0-t1 time period to train the ARMA model and determine the model parameters, including the autoregressive parameters and the moving average parameters;

[0052] 7) Use the trained ARMA model to predict the future time fault characteristic parameter value P f ;

[0053] 8) The actual measured value P m is compared with the predicted value P f . If the difference exceeds the set threshold range, an abnormal signal is sent;

[0054] ​9) If the threshold range is not exceeded, the latest ts time period data is used to predict the next time fault characteristic parameters, until the variable working condition instruction is received, then the steady state phase interpretation is ended;

[0055] 10) After the variable working condition, the next steady state phase is entered, and the above process is repeated to calculate the online adaptive threshold and determine whether the parameters are abnormal, so as to realize online fault detection.

[0056] The online dynamic threshold analysis method based on ARMA time series mainly establishes the mapping relationship between the future signal size and the current signal size by fitting the time series with stationarity, predicts the signal fluctuation, and thus predicts the future value based on the current measured value (stationary signal establishment). When the predicted future value deviates from the actual measured value to a certain extent, it is determined that the system has failed, such as Figure 4 As shown in the figure, it is the actual measurement value and prediction result comparison diagram of a certain parameter of an engine in the embodiment of the application.

[0057] According to the given upper and lower threshold proportion, such as ±10%, the adaptive threshold in the steady state working condition can be obtained, such as Figure 5 As shown in the figure. If the actual measurement value exceeds the upper and lower threshold range, the engine will alarm.

[0058] Example 2:

[0059] Variable working condition online fault detection method:

[0060] The variable working condition of the engine is generally achieved by adjusting the opening of the valve and changing the flow of the propellant to adjust the thrust of the engine. Through working condition adjustment, the engine can realize thrust regulation within a certain range to meet the flight task requirements. During the engine working condition adjustment process, the key parameters such as thrust, flow, pressure, and speed related to the shutdown red line will also change, and are positively correlated with the working condition adjustment range. Therefore, during the working condition adjustment process, the shutdown red line of the key parameters should also be adjusted accordingly to realize adaptive adjustment of the threshold value.

[0061] The working condition change instruction is issued by the control computer. After the control computer issues the working condition adjustment signal, the valve control motor starts to respond to drive the valve opening adjustment, thereby realizing the adjustment of the engine working condition. Since the motor response and the response of the engine after the valve opening adjustment both have a certain delay effect, during the variable working condition process, the engine parameters change nonlinearly. Taking the speed data of a certain engine as an example, as shown in the figure, Figure 6 The working condition adjustment instruction is issued at 66s, and lasts for nearly 1s, and the steady state work of the new working condition is completed at 67s.

[0062] For the variable working condition stage, the adaptive threshold is determined according to the following steps:

[0063] 1) Fitting the mapping relationship between engine steady state condition and key parameters

[0064] Through the engine historical test data or theoretical calculation model, the least square method is used to obtain the mapping relationship between engine thrust and key parameters. The sensor parameters of a certain engine and the working condition are approximately linear, and the function relationship between the steady state condition and the engine fault characteristic parameters is constructed by fitting the least square method:

[0065] K=a·F+b (2)

[0066] In the formula, K represents the engine fault characteristic parameter; F represents the engine condition before adjustment; the fitting effect is shown in Figure 6 .

[0067] 2) Construct the function change relationship under non-steady state process:

[0068] For variable working condition delay response and non-steady state change process, the change law of each parameter in the engine working condition change process is simulated, and the parameters t d , t, F0, F1 are taken as input, and the fault characteristic parameter K is taken as output. Machine learning methods such as neural network are used for training to form the engine change law of working condition change process, that is, the mapping relationship between fault characteristic parameter K and parameters t d , t, F0, F1 is obtained:

[0069] K=f(t d ,t,F0,F1) (3)

[0070] In the formula, t d represents the working condition adjustment delay time; t represents the engine working condition adjustment starting time; F0 represents the engine condition before adjustment; F1 represents the engine condition after adjustment.

[0071] 3) Description of key parameters of variable working condition process engine

[0072] According to the working condition adjustment value before and after the working condition change, the fault characteristic parameter value K0 before working condition adjustment and the fault characteristic parameter value K1 after working condition adjustment are calculated according to formula (2), that is, the upper and lower limit values of the fault characteristic parameters before and after working condition adjustment are obtained; then according to formula (3), the variable working condition process is fitted to form the description of the key parameters of the engine from the steady state condition to the variable working condition process, and then to the next steady state condition, that is, the curve of the fault characteristic parameter K is obtained.

[0073] 4) Fault detection

[0074] Based on the curve of the fault characteristic parameter K obtained above, the red line threshold is obtained by multiplying a set coefficient, and the measurement value of the fault characteristic parameter at each time is compared with the red line threshold to determine whether it is abnormal.

[0075] In the engine working condition adjustment process, the speed parameter change and the fitting result are as shown in Figure 8 The actual parameter measurement value of the engine will have certain fluctuations, the green fluctuation broken line is the true data, and the blue dotted line is the calculation data, the calculation value can well represent the actual change process, and the determination coefficient reaches 0.9982.

[0076] The "steady state working stage adaptive red line shutdown threshold determination method" and the "variable working condition stage adaptive red line shutdown threshold determination method" are used to respectively describe the steady state working condition and the variable working condition process, in order to prevent false shutdown, the fixed multiple of the actual working condition is determined as the threshold of the engine red line key, and the online dynamic threshold warning red line of a certain type of engine in the whole working process is as shown in Figure 9 The yellow area is the safe area, and once the signal actually collected by the sensor exceeds the yellow area, it is determined to be abnormal.

[0077] In summary, the above is only a preferred embodiment of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for online fault detection of a reusable liquid rocket engine, characterized in that, This includes online fault detection methods during the steady-state operation phase, specifically: The collected fault characteristic parameter sequences were fitted using an autoregressive moving average (ARMA) model, and future fault characteristic parameter values ​​were predicted. ；（1） In the formula: This represents the predicted value of a certain fault characteristic parameter at time t. This represents the measured value of a certain fault characteristic parameter at times t-1, t-2, ..., tn; These represent the differences between the measured values ​​of the fault characteristic parameters at time t-1 and t-2, the differences between the measured values ​​at time t-2 and t-3, and the differences between the measured values ​​at time t-m+1 and tm, respectively. Indicates the autoregressive parameters; N represents the autoregressive order; represents the moving average parameter; M represents the moving average order. Indicates white noise interference; The predicted fault characteristic parameter value at each moment is multiplied by a set coefficient to obtain the red line shutdown threshold. The measured value of the fault characteristic parameter is compared with the red line shutdown threshold to determine whether a fault has occurred. It also includes online fault detection methods for variable operating conditions, specifically: 1) Construct a functional relationship between steady-state operating conditions and engine fault characteristic parameters: ； （2） In the formula, K represents the engine's fault characteristic parameters; F represents the engine's operating condition before adjustment; 2) For the delayed response and unsteady-state change process under varying operating conditions, simulate the variation law of each parameter during the engine operating condition change process, and set the parameter t d Using t, F0, and F1 as inputs and fault characteristic parameter K as output, machine learning methods are used for training to form the engine variation law during the operating condition change process, that is, to obtain the fault characteristic parameter K and parameter t. d Mapping relationship between t, F0, and F1: ； （3） In the formula, t d This indicates the delay time for adjusting the engine operating conditions; t indicates the start time of the engine operating condition adjustment; F0 indicates the engine operating condition before adjustment; F1 indicates the engine operating condition after adjustment. 3) Based on the operating condition adjustment values ​​before and after the change, calculate the fault characteristic parameter values ​​before the operating condition adjustment according to formula (2). Compared with the fault characteristic parameter values ​​after operating condition adjustment That is, the upper and lower limits of the fault characteristic parameters before and after the working condition adjustment are obtained; then the curve of the fault characteristic parameter K is obtained according to formula (3); 4) Based on the curve of the obtained fault characteristic parameter K, multiply it by the set coefficient to obtain the red line shutdown threshold at each time under the variable operating condition. Compare the measured value of the fault characteristic parameter at each time with the red line shutdown threshold to determine whether there is a fault.

2. The online fault detection method for a reusable liquid rocket engine as described in claim 1, characterized in that, In the online fault detection method during the steady-state operation phase, before predicting future fault characteristic parameter values, the traditional fixed red line method is used to determine whether the fault characteristic parameters are within the red line range. If they are outside the red line, an abnormal signal is issued, indicating a fault.

3. The online fault detection method for a reusable liquid rocket engine as described in claim 1, characterized in that, In the online fault detection method during the steady-state operation phase, before predicting future fault characteristic parameter values, it is determined whether the data is stable. If the data is unstable, methods such as difference calculation are used to stabilize the data. If the data is stable, an ARMA model is used for fitting, and future fault characteristic parameter values ​​are predicted.

Images