• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The method according to the invention comprises: a step of obtaining (E10) a measurement of the monitored parameter carried out by a sensor and corresponding to an operating point of the engine, this operating point being defined by at least one parameter of engine regulation; a step of estimating (E20) a value of the parameter monitored for this operating point, from a regulated value or a filtered setpoint value of said at least one engine regulation parameter defining the point of functioning; a step of comparing (E40) an error between the measurement of the monitored parameter and its estimation with respect to at least one threshold determined from an uncertainty on this evaluated error for the operating point; and - an emission step (E60) of a notification in case of crossing said at least one threshold.
    公开号:FR3019592A1
    申请号:FR1452954
    申请日:2014-04-03
    公开日:2015-10-09
    发明作者:Antoine Romet；Gonidec Serge Le；Dimitri Malikov；Jonathan Gazagnes
    申请人:SNECMA SAS；
    IPC主号:
    专利说明:

    [0001] BACKGROUND OF THE INVENTION The invention relates to the general field of aeronautics. It relates more particularly to the monitoring of a rocket engine.
    [0002] In known manner, the rocket engine systems are made to operate on a wider operating range, and are equipped with means for verifying their proper operation. The verification implemented is performed firstly in real time, ensuring that the engine operating parameters (eg engine pressure and temperature, etc.) remain within predefined acceptable limits, and secondly in deferred time, by comparing measurements of the operating parameters performed during tests on the engine with respect to predetermined rectilinear monitoring templates, an output of a template revealing a measuring defect or of the engine. The parameterization (i.e. determination) limits and monitoring templates is done manually and presents a risk of error especially when this setting is not in line with the point of operation of the engine. In addition, a change in engine profile and test conditions leads to a tedious and risk-free update of the limits and monitoring templates.
    [0003] OBJECT AND SUMMARY OF THE INVENTION The invention makes it possible to overcome these drawbacks by proposing a method of monitoring a parameter of a rocket engine comprising: a step of obtaining a measurement of the monitored parameter carried out by a sensor and corresponding to an operating point of the engine, this operating point being defined by at least one engine control parameter; a step of estimating a value of the parameter monitored for this operating point, from a regulated value or a filtered setpoint value of said at least one engine control parameter defining the operating point; a step of comparing an error between the measurement of the monitored parameter and its estimation with respect to at least one threshold determined from an uncertainty on this evaluated error for the operating point; and a step of transmitting a notification in case of crossing said at least one threshold. Correlatively, the invention also provides a device for monitoring a parameter of a rocket engine comprising: a module for obtaining a measurement of the monitored parameter made by a sensor and corresponding to an operating point of the engine, operating point being defined by at least one motor control parameter; a module for estimating a value of the parameter monitored for this operating point, from a regulated value or a filtered setpoint value of said at least one engine regulation parameter defining the operating point; a module for comparing an error between the measurement of the monitored parameter and its estimation with respect to at least one threshold determined from an uncertainty on this evaluated error for the operating point; and a transmission module of a notification in case of crossing of said at least one threshold. Crossing the threshold by the error means that the error is greater than the threshold if it defines an upper limit not to be exceeded, or that the error is lower than the threshold if the latter defines a low limit not to exceed not exceed. In addition, it should be noted that the step of comparing the error can be implemented in the context of the invention in different ways, obviously for the skilled person. For example, the error can be defined as the absolute value of the difference between the measurement of the monitored parameter and its estimate, or simply as the difference between the measurement of the monitored parameter and its estimate. In addition, during this step, the error can be estimated and then compared to the threshold, or alternatively, the measure of the monitored parameter can be compared to its estimate to which the threshold has been added (or possibly subtracted), etc. This comparison step is equivalent regardless of how it is implemented to a comparison of the measurement of the monitored parameter to a template defined from the estimate of the monitored parameter and the threshold. The invention thus provides a monitoring of the parameters of a rocket engine that dynamically and automatically evolves with the operating point of this engine. The parameters monitored are for example: a motor pressure; and / or a temperature of the engine; and / or a rotational speed of an element of the engine; and / or a flow rate of a fluid flowing in the engine; and / or a vibratory behavior of the engine. The operating point of the motor is defined from one or more parameters used for motor control. Such regulation parameters are for example a pressure of a combustion chamber of the engine, a mixing ratio (oxygen / hydrogen) at the inlet of an engine pump, a speed of rotation of an oxygen turbopump, a speed rotation of a hydrogen turbopump, etc. These regulation parameters are regulated (ie controlled) in a closed loop during operation of the motor, in a manner known per se, by a control system able to act on various variable engine geometries such as, for example, the position of control valves such as only VBPO (ByPass Oxygen Valve) and VBPH (ByPass Hydrogen Valve) valves.
    [0004] The monitoring proposed by the invention advantageously adapts to the dynamic and non-linear character of the rocket engine systems, this character being linked in particular to the variation as a function of time of the engine regulation setpoints, these instructions being limited in values and in gradient by the regulation system. For this purpose, it relies on an estimation of the monitored parameter (s) (eg by simulation or using models) from a regulated and therefore validated value of the regulation parameters, so as to limit the uncertainties related in particular to the valves and to detect anomalies affecting the engine only, or a filtered setpoint of the control parameters so as to detect anomalies affecting the entire chain downstream of the control loop, that is to say ie not only affecting the motor but also the actuators for controlling the variable geometry of the engine. The estimate thus obtained is then compared to a measurement of the monitored parameter taking into account a threshold determined automatically according to the operating point of the engine. The threshold is determined from the particular uncertainties of measurement and estimation of the parameter to be monitored encountered at the point of operation.
    [0005] Thus, the invention does not rely, contrary to the state of the art, on a preset monitoring template or preset limits at which one (I) we compare a measurement of the parameter to monitor. On the contrary, it automatically and dynamically defines the limits and / or the rocket engine monitoring gauge based on an estimation of the parameters to be monitored and a measurement and estimation precision (standard deviation) evaluated. both taking into account the operating point of the engine. In a particular embodiment, the operating point used to evaluate the error uncertainty is determined from the regulated value of said at least one control parameter. This embodiment is based on an accurate estimate of the operating point. It makes it possible to monitor the rocket engine both in steady state and transient conditions. Indeed, it makes it possible to overcome, in transient conditions, sometimes significant delays that may be encountered between the measurements of the control parameters and the setpoint values of these parameters during the control. As a variant, the operating point used to evaluate the error uncertainty can be determined from the filtered setpoint value of the at least one control parameter or the setpoint value of said at least one control parameter (this last being ahead of the actual operating point of the engine). It should be noted that several thresholds may be taken into account during the monitoring (for example, if G designates the error uncertainty according to the operating point of the motor, the high and low thresholds at +/- 3o and at +/- 6o can be considered), each threshold can be associated with a response to the transmitted notification distinct and appropriate depending on the severity of the anomaly detected (eg raising an alert, maintenance action on the engine, stopping the engine, etc.).
    [0006] The monitoring thus provided by the invention thus makes it possible to detect anomalies affecting the rocket engine in real time as well as in deferred time. The monitoring device can therefore be hosted on the same entity as the engine control device (eg on a computer closest to the engine), or on a separate entity, on board the rocket powered by the engine (ex. carried by the flight computer) or on a rocket engine test stand. In a preferred embodiment, the step of estimating the value of the monitored parameter is performed using an artificial neural network having as input (s) the regulated value or the filtered set value of said at least one a motor control parameter.
    [0007] The use of a neural network makes it easier to reproduce the nonlinear behaviors of the rocket engine system, whatever the appearance of these non-linear behaviors and the complexity of the system. In addition, the use of an artificial neural network makes it possible to reach a compromise between estimation accuracy and computational load that allows a real-time application of the invention. The number of calculations necessary to estimate the value of the parameter to be monitored is indeed limited via the use of such a model. In this embodiment, the uncertainty on the error can take into account in particular an estimation uncertainty of the monitored parameter determined as a function of: an uncertainty on the input of the neural network; and / or - an uncertainty on the neural network; and / or - uncertainty on a learning base used to construct the neural network. Alternatively, other estimators than an artificial neural network may be used, such as an offline nonlinear estimator based on nonlinear regression, etc.
    [0008] The inventors have found that, in the presence of abrupt variations of the setpoint values of a regulation parameter, the error between the measurement of the monitored parameter and its estimation can be very close to the threshold determined in accordance with the invention, depending on the from the operating point. To make the monitoring method more robust to this type of phenomenon, it is possible, in a particular embodiment of the invention, to also take into account, during the comparison step, the dynamics of change of the operating point. of the engine, so in particular to define a larger template around the measurement of the monitored parameter in case of abrupt transitions of the operating point of the engine. This template can be defined, for example: by an upper bound obtained by multiplying the estimate of the monitored parameter by the index response of a predefined order 2 filter modeling the sudden changes in the set value or the filtered set value the regulation parameter or parameters, and adding to the result of this operation the previously determined uncertainty; and / or by a lower bound obtained by multiplying the estimate of the parameter monitored by the index response of a predefined order 1 filter modeling the soft variations of the set value or the filtered set point of the parameter or parameters of regulation, and subtracting from the result of this operation the previously determined uncertainty.
    [0009] In other words, in an equivalent way, the error between the measurement of the monitored parameter and its weighted estimate by the response of a predefined second-order filter at a level representative of the value of setpoint or the filtered setpoint of said at least one control parameter; and at a second so-called low threshold, the error between the measurement of the monitored parameter and its estimate weighted by the response of a predefined first-order filter at said level, a crossing of one or the other of the thresholds resulting from the issue of a crossing notification of a threshold. In a particular embodiment, the various steps of the monitoring method are determined by computer program instructions.
    [0010] Consequently, the invention also relates to a computer program on an information medium, this program being able to be implemented in a monitoring device or more generally in a computer or in a computer, this program comprising instructions adapted to the implementation of the steps of a monitoring method as described above.
    [0011] This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape. The invention also relates to a computer-readable information medium, comprising instructions of a computer program as mentioned above.
    [0012] The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a diskette (floppy disc) or a disk hard. On the other hand, the information medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network. Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question. The invention also relates to a rocket engine comprising a monitoring device according to the invention.
    [0013] It may also be envisaged, in other embodiments, that the monitoring method, the monitoring device and the rocket motor according to the invention present in combination all or part of the aforementioned characteristics. In addition, it should be noted that the monitoring method according to the invention can be applied to other controlled systems that a rocket engine, such as for example an aircraft engine. BRIEF DESCRIPTION OF THE DRAWINGS Other features and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate embodiments having no limiting character. In the figures: FIG. 1 schematically represents a rocket engine comprising a monitoring device according to the invention, in a particular embodiment; FIG. 2 diagrammatically represents the hardware architecture of the monitoring device of FIG. 1; FIG. 3 represents, in the form of a flow chart, the main steps of a monitoring method according to the invention, as implemented by the monitoring device of FIG. 1; FIG. 4 represents an artificial neural network used by the monitoring device of FIG. 1; FIG. 5 represents an exemplary table of values that can be used during the monitoring method represented in FIG. 3; FIG. 6 illustrates an example of variations in the measurement and estimation of a PSPO pressure of the system monitored by the monitoring device of FIG. 1; and FIG. 7 illustrates the evolution as a function of time of the thresholds considered by the monitoring device, in a second embodiment of the invention. DETAILED DESCRIPTION OF THE INVENTION FIG. 1 schematically represents, in its environment, a rocket engine 1 which is envisaged to be monitored in accordance with the invention, in a particular embodiment. In a manner known per se, a rocket engine uses liquid hydrogen and oxygen, which are burned during operation of the engine in a combustion chamber. In the example under consideration, the supply of hydrogen and oxygen is controlled by means of control valves which supply turbopumps. The control valves make it possible to control in particular the flow rates of oxygen and hydrogen, the speeds of the turbines, etc. They are controlled by actuators and their positions are controlled by a computer, for example by the calculator or the controller of the rocket engine. For this purpose, the computer relies on various engine control parameters, such as, for example, the pressure in the combustion chamber denoted PGC, the oxygen / hydrogen mixture ratio at the pump inlet denoted by RMEP, the rotation speeds of the oxygen and hydrogen turbopumps, etc.
    [0014] The operating principle of the regulation system 2 thus implemented is illustrated schematically in FIG. 1. It should be noted that in the example envisaged in FIG. 1, it is assumed that the regulation logic is implemented by the calculator. However, this assumption is not limiting, and this control logic can be implemented by another on-board computer on board the rocket such as for example the computer operating the rocket (also known under the name of OBC for "On Board Computer". The regulation system 2 operates in a closed loop. More specifically, a setpoint yC is created by the computer for each engine control parameter in a manner known per se, and supplied to a tracking dynamics filter 3. The filtered setpoint yCf obtained at the output of the filter 3 supplies an error calculation module 4 which is also supplied with the regulated value yR of the regulation parameter. The error e evaluated by the module 4 is provided to a correction module 5 of the rocket engine, which evaluates a command u to correct the error e (i.e. to make it disappear). The command u is a position command of a control valve intended to allow readjustment of the value of the regulation parameter so that it complies with the instruction. The command u is transmitted to a correction module 6 driving the actuator 7 of the control valve according to the command u. A sensor 8 of the position of the control valve thus driven back to the corrector module 6 a measurement u 'of the position of the control valve. The rocket motor 1 is furthermore equipped with a sensor 9 making it possible to measure the "regulated" value of the regulation parameter or to estimate it from the measured position u 'of the valve by the sensor 8. This sensor 9 is for example a pressure sensor, a speed sensor, a position sensor, a flow meter, etc., depending on the control parameter considered. The regulated value, after having been validated by a module 10 for validating the computer (able in particular to eliminate the outliers), is supplied to the error calculation module 4 to adapt the command u to reach the setpoint yC, etc. Since the operation of the regulation system 2 is known to those skilled in the art, it is not described further here. The proper operation of the rocket motor 1 is ensured by means of a monitoring device 11 according to the invention. This monitoring device 11 is configured to monitor one or more members or elements 12 of the rocket engine 1 from measurements m1,..., MP provided by sensors 13, known per se, and able to measure various parameters p1, ..., pP of the engine at various points of operation of the latter, P designating any integer greater than or equal to 1. These parameters monitored by the monitoring device 11 are for example: the output pressure of the engine pumps; the temperature of the nozzle; rotational speeds of turbopumps; the flow rates of fluids (oxygen, hydrogen) at the inlet of the pumps of the engine; a vibratory behavior of the engine; etc. For the sake of simplification in the following description, consider a single parameter noted p monitored by the monitoring device 11 (ie P = 1), for example the output pressure of the oxygen pump of the engine 1 noted PSPO (for Output Pressure Pump Oxygen).
    [0015] However, no limitation is attached to the number P of engine parameters monitored by the monitoring device 11 nor to the nature of these parameters. Each operating point considered of the rocket motor 1 is defined by the value of one or more control parameters used by the control system 2, as mentioned above. In the example envisaged here, each operating point considered of the motor is defined by a pair of control parameters, namely the pair formed by the value of the pressure of the combustion chamber PGC and the value of the oxygen mixing ratio. / Hydrogen inlet pump RMEP. However, this assumption is not limiting in itself and other control parameters, in addition to the above-mentioned control parameters or in replacement thereof, can be envisaged to define the operating point of the rocket motor 1, as per example the speed of rotation of turbopumps with oxygen and hydrogen. It is even possible to consider only a single regulation parameter. In the embodiment described here, the monitoring device 11 is integrated in the rocket engine 1, and more precisely in its computer or controller, thus making it easy to implement real-time monitoring and flight of the rocket engine 1. It is based on the hardware architecture of the computer which, in the embodiment described here, is that of a computer, as illustrated schematically in FIG. 2. Thus, the monitoring device 11 comprises in particular a processor 11A , a read-only memory 11B, a random access memory 11C, a non-volatile memory 11D and communication means 11E with different elements of the computer and the rocket engine 1, such as in particular the sensors 13 and the regulation system 2. These communication means 11E include for example data buses. The read-only memory 11B of the monitoring device 11 constitutes a recording medium readable by the processor 11A and on which is recorded a computer program according to the invention, comprising instructions for executing the steps of a method of monitoring according to the invention, the steps of this method being described later with reference to Figures 3 to 5 in a particular embodiment.
    [0016] This computer program equivalently defines functional modules of the monitoring device 11 (software modules here) such as a module 14 for obtaining measurements of the monitored parameters, a module 15 for estimating the values of these parameters, comparison module 16 of the measurements and estimated values and a transmission module 17 of an activated threshold crossing notification if necessary. The functions of these various modules are described in more detail with reference to the steps of the monitoring method. In another embodiment, the monitoring device 11 is housed in a rocket engine test stand, or in the rocket board computer (in the OBC mentioned above). No limitation is attached to the entity hosting the monitoring device 11. In particular, the control system 2 and the monitoring device 11 can be hosted by separate entities. We will now, with reference to Figure 3, describe the main steps of the monitoring method implemented by the monitoring device 11 in a particular embodiment of the invention.
    [0017] As mentioned above, the monitoring device 11 implements a monitoring of the rocket motor 1 from measurements of the parameter p collected by one of the sensors 13 equipping the engine, for different operating points of the latter. Such a measurement m of the parameter p (or more generally ml, ..., mP of the parameters p1, ..., pP when P parameters are monitored according to the invention) is obtained from the sensor 13 by the obtaining module 14 of the monitoring device 11 via the communication means 11E of the monitoring device, for an operating point PF of the rocket motor 1 (step E10). In the embodiment described here, the obtaining module 14 is able to process the measurements received from the sensor 13, and more specifically to filter them and eliminate outliers where appropriate. Such treatments are conventionally used and are not described in more detail here. They make it possible to overcome, in particular, metrological problems. The measurement m 'of the parameter p resulting from the processing of the measurement m is stored by the obtaining module 14 in the random access memory 11C of the monitoring device 11. Furthermore (in parallel or after the step E10), the module estimate 15 of the monitoring device 11 estimates the value of the monitored parameter p for the operating point PF (step E20). It uses for this purpose an estimator constructed from a learning database and modeling the nonlinear behavior of the rocket engine 11. This estimator takes as input the value of the operating point PF, and outputs an estimate p_ is the parameter p for this operating point value PF. In the embodiment described here, the operating point PF supplied to the input of the estimator is defined by the regulated values yR of the control parameters PGC and RMEP. In another embodiment, the operating point PF supplied at the input of the estimator is defined by the filtered setpoint values yCf of the regulation parameters.
    [0018] The learning database used to construct the estimator is derived here from a consolidated mathematical model with experimental data. Obtaining such a learning base presents no difficulty for those skilled in the art and is not described further here.
    [0019] In the embodiment described herein, the estimator used by the estimation module 15 is an artificial neural network RNA constructed from the training database. This neural network is represented in FIG. 4. More specifically, the considered RNA neural network is a non-feedback PMC (Perceptron Multi Couches) network, comprising a Lin input layer, an Lout output layer, and one or more intermediate layers called hidden Lhid. For the sake of simplification, only one hidden layer is represented in FIG. 4. The input layer Lin comprises M + 1 inputs (or neurons) forming a vector [Vin, 1], Vin denoting a vector of dimension M comprising the M control parameters defining the operating point considered of the rocket engine. M is an integer greater than or equal to 1. In the example considered, the input layer Lin therefore comprises, in addition to the unitary input, M = 2 inputs respectively corresponding to the pressure PGC of the combustion chamber and the RMEP mixing ratio at the inlet of the motor pump. The output layer Lout of the network comprises P outputs (or neurons) forming a vector Vout respectively corresponding to the estimates of P parameters monitored by the monitoring device 11. P is an integer greater than or equal to 1. In the example envisaged here, P = 1. The hidden layer Lhid comprises N neurons, where N denotes an integer greater than or equal to 1. It relies here on a sigmoidal function F (x) with real vector output at N dimensions, defined by: F (x) = 1 + e-2x 1 Alternatively, other functions of the sigmoidal type can be envisaged. Thus, the output vector Vout of the neural network is calculated according to: Vout = W2. [F [Vin] 11 1 where W1 denotes the matrix of synaptic weights between the input layer Lin and the hidden layer Lhid, the last column of this matrix representing the biases of the hidden layer neurons, and W2 denotes the synaptic weight matrix between the hidden layer Lhid and the output layer Lout, the last column of this matrix representing the biases of the neurons of the output layer. The synaptic weights and biases contained in the matrices W1 and W2 are obtained for the rocket engine 1 using an automatic learning algorithm based on the conjugate gradient method and applied to the training database. known to those skilled in the art and not described here. Alternatively, other known learning algorithms may be implemented to construct the RNA neural network from a training database. In the example envisaged in FIG. 4, a single hidden layer Lhid is represented with N neurons. The numbers of hidden layers and neurons per hidden layer can be optimized to reach a network of a given accuracy, this accuracy being measurable for example by means of a square root error criterion. As a variant, other criteria can also be taken into account, such as, for example, the calculation time required to estimate a parameter using the neural network, etc., possibly weighted according to their relative importance. In addition, other neural network architectures can be envisaged such as, for example, architectures with feedback. Alternatively, other non-linear models than an artificial neural network can be used to estimate the value of the monitored parameter p, such as an offline nonlinear model using nonlinear regression defined from existing structural relationships. between the control parameters and the monitored parameters. Such structural relationships are known per se or can be determined by experience and / or from the training database. With reference to FIG. 3, the measurement m 'and the estimate p_est of the parameter p are respectively provided by the obtaining module 14 and by the estimation module 15 to the comparison module 16 of the monitoring device 11. On reception of these values, the comparison module 16 first evaluates the error E between the measurement m 'and the estimate p_est of the parameter p monitored (step E30) according to: E = 1m' - p_esti where lm '- p_est I denotes the absolute value of the difference between the measure m 'and the estimate p_est. Then it compares the error E with one or more thresholds 51, ..., SK, K designating an integer greater than or equal to 1 (step E40). According to the invention, the thresholds S1,..., SK are determined by the comparison module 16 of the monitoring device 11 based on an uncertainty Cl (ie standard deviation) on the error E evaluated for the point. PF operation of considered the rocket motor 1, that is to say for the operating point of the rocket motor 1 which was acquired the measurement m, and for which was estimated p_est value. For example, the monitoring device 11 considers two distinct thresholds S1 = 30 and S2 = 60 (or equivalent, if we consider the error s = m 'p_est, the thresholds S1 = + / - 3o and S2 = +/- 6a).
    [0020] These examples are given for illustrative purposes only, and of course other multiples of the uncertainty a can be considered depending on the anomalies that one wishes to detect. The choice of the thresholds is guided by the concern to detect the drifts of the elements of the rocket engine 1 (including sensors used) sufficiently early while limiting the false alarms, as detailed further later. In the embodiment described here, the uncertainty takes into account two factors namely the uncertainty al on the measure m of the parameter p and the uncertainty o2 on the estimate p_est of the parameter p. More precisely: O = V (61) 2 + (o-2) 2 Alternatively, the uncertainty can be deduced from the uncertainties al and O2 using a function other than a quadratic sum. The uncertainty 01 on the measurement m of the parameter p is known for a given sensor. It can for example be extracted or determined from the specifications provided by the manufacturer of the sensor 13 (and possibly validated by means of test recipes performed on the sensor). It should be noted that this uncertainty (or equivalently the measurement inaccuracy of the sensor) is not necessarily constant over the entire operating range of the rocket motor 1 but may be varied according to the operating point of the sensor. engine. The monitoring device 11 therefore uses here to evaluate the uncertainty ol, a table TAB1 of predetermined values from the specifications of the manufacturer of the sensor 13 in particular, and giving, for different torque values (PGC, RMEP) defining the operating point PF, the value of the resulting ol uncertainty. The uncertainty cy2 on the estimate p_est of the parameter p depends here on several sources of uncertainty, including: uncertainty (or, conversely, the accuracy, on the contrary) on the input data of the estimator used by the estimation module 15, and more specifically here, regulated values of the control parameters PGC and RMEP. Indeed, the accuracy of the closed loop implemented by the control system 2 depends on the measurement of the regulation parameter or parameters by the sensor 9 and the width of the zone in which the error between the filtered setpoint yCf and the measurement yR is forced to zero to avoid nonlinear instabilities (limit cycle) at the level of the actuators 7 (related to the limit of their resolution); the uncertainty (or, conversely, the precision) on the estimator used, namely here the artificial neural network RNA, which is not identically zero but depends on the operating point of the engine. The accuracy of the estimator characterizes its ability to faithfully reconstruct the learning database from which it derives; and uncertainty (or equivalently, conversely, accuracy) on the basis of training data used to construct the estimator, i.e. here the artificial neural network RNA. The estimator may have excellent accuracy and therefore the value of the parameter estimated using this estimator may be far from the measurement of this parameter if the learning base is not reliable, for example when this database is not reliable. learning is itself derived from a mathematical model. A bias can then be applied during the construction of the estimator to take into account the difference between the real engine and its modeling by the learning base in order to limit this uncertainty. In the embodiment described here, the monitoring device 11 uses, to evaluate the uncertainty a2 resulting from the combination of the aforementioned uncertainties, a table TAB2 of determined values, for example, by simulation using a Monte Carlo method. classic known to those skilled in the art. FIG. 5 represents an example of such a table of values TAB2. This example is for illustrative purposes only. The table TAB2 gives for various values of the pair of control parameters (PGC, RMEP), the value of the resulting uncertainty a2 (in the unit of the parameter p monitored, that is to say in the example envisaged here in bar, the monitored parameter being a pressure). In the example illustrated in FIG. 5, the domain of the values of the regulation parameters PGC and RMEP has been arbitrarily divided into 195 subdomains (these domains are not all represented for the sake of simplification) in each of which prints were performed according to the Monte Carlo method. The number of domains considered depends on a compromise between complexity and homogeneity of the value of the uncertainty within the same domain. Alternatively, one can consider the uncertainty 02 can be evaluated from an artificial neural network rather than using a table such as the TAB2 table. The comparison module 16 of the monitoring device therefore extracts tables of values TAB1 and TAB2 above the uncertainties al and a2 corresponding to the operating point PF considered of the rocket engine 1. It uses for this purpose here as values of the control parameters PGC and RMEP defining this operating point PF, the regulated and validated values yR of the control parameters PGC and RMEP. In an alternative embodiment, the comparison module 16 uses as the operating point PF to extract the uncertainties al and a2 from the tables TAB1 and TAB2 respectively, the filtered setpoint values yCf of these control parameters (ie available at the output of the filter 3 tracking dynamics). In yet another variant embodiment, the comparison module 16 uses, as operating point PF, to extract the uncertainties al and a2 from the tables TAB1 and TAB2 respectively, the setpoint values yC fixed by the regulation system 2 for these control parameters. . Then it evaluates from the uncertainties al and a2, the value of the uncertainty 0. It deduces from this value, the thresholds S1 and S2. For example here, S1 = 30 and S2 = 6a. As mentioned previously, the comparison module 16 compares the error E with the thresholds thus determined (steps E40). In a variant, it compares the measurement m 'of the monitored parameter p with p_est +/- 3a and with p_est +/- 6a.
    [0021] If it determines during this comparison that the error E is greater than the threshold S1 (yes response to the test step E50 and crossing the threshold S1), a notification of the crossing (overrun here, the error being defined so as to always be positive) of the threshold S1 is emitted by the notification module 17 of the monitoring device 11 to an entity 18 for managing the alerts and the firing sequence (step E60). Similarly, if the comparison module 16 determines that the error E is greater than the threshold S2 (yes response to the test step E50 and crossing the threshold S2), a notification of the crossing of the threshold S2 is issued by the module of notification 17 of the monitoring device 11 to an entity 18 for managing the alerts and the firing sequence (step E60). In the embodiment described here, the entity 18 for managing the alerts and the firing sequence implements a majority logic to determine the appropriate action to be taken if necessary on the rocket engine 1 in response to the notifications of threshold exceeded received. A modular and gradual response depending on the exceeded speed can be implemented (eg maintenance to be provided, motor stop, adaptation of the set value yC, etc.). If no threshold is crossed (or exceeded here), the monitoring continues in accordance with steps E10 to E60 previously described for a new operating point of the rocket engine 1. It is the same after notification of the entity 18 of management.
    [0022] In the embodiment described here, the thresholds considered by the monitoring device are set only according to the uncertainty a (eg equal to multiples of this uncertainty), which itself depends on the operating point PF considered of the Rocket Engine 1. In a second embodiment of the invention, to obtain a more robust monitoring of the rocket engine 1, the monitoring device 11 determines the thresholds used during the comparison step from the uncertainty. a and also takes into account the dynamics of change of the operating point of the motor and more specifically, of the setpoints (raw or filtered) of the regulation parameters. The inventors have indeed found that when the set values of the regulation parameters determined by the control system 2 vary sharply, the error E approaches very close to the thresholds defined from the uncertainty o, as illustrated in FIG. FIG. 6 for the PSPO pressure at the outlet of the oxygen turbopump with a threshold S1 = 30. The risk associated with this transient behavior is that an alert (threshold crossing) is erroneously raised by the monitoring device 11 to the entity 18 for managing the alerts and the firing sequence. To take account of this phenomenon, during the comparison step, not only the monitoring device 11 uses thresholds determined from the uncertainty a and the operating point PF of the engine, but it also takes account of the dynamics of change of the operating point of the motor, and more precisely of the set values of the control parameter or parameters. This consideration is reflected in the definition of a larger monitoring template around the measurement of the monitored parameter in case of abrupt transitions (ie fast) of the operating point of the motor, and on the contrary narrower during smooth transitions of the operating point (ie no or little change of the latter). In the second embodiment described here, this template is defined by: an upper bound obtained by multiplying the estimate of the parameter monitored by the index response of a predefined order 2 filter modeling the sudden variations of the set value or the filtered reference value of the regulation parameter (s), and adding to the result of this operation the uncertainty previously determined (or a multiple of this uncertainty according to the considered threshold); and by a lower bound obtained by multiplying the estimate of the parameter monitored by the index response of a predefined order 1 filter modeling the soft variations of the set value or the filtered set point of the regulation parameter or parameters, and deducting from the result of this operation the previously determined uncertainty o (or a multiple of this uncertainty according to the threshold considered). The step considered for evaluating the index responses of the order 1 and 2 filters is defined by the raw target value yC of the control parameter or parameters. In a variant, the monitoring device 11 considers the step defined by the filtered set point value yCf of the regulation parameter or parameters. The choice of the cut-off frequencies and the delays of the filters of order 1 and 2 is realized offline for a given monitored parameter and a given rocket engine, for example in an experimental way, by calculations or by tests, starting from the knowledge of the dynamics of the monitored parameter as a function of time and variations of the setpoint values of the control parameters. In a variant, filters of different orders (of higher orders in particular) can be envisaged. Thus, in this second embodiment, the monitoring device 11 compares: at a first threshold said high defined from the uncertainty o (eg Slhaut = + 30), the error between the measurement m 'of the monitored parameter and its estimate p_ is weighted (ie multiplied) by the response of the filter of order 2 to a step representative of the set value yC (or of the filtered set value yCf); and at a second so-called low threshold defined from the uncertainty o (eg S1bas = -3a), the error between the measurement m 'of the monitored parameter and its estimate p_ is weighted by the response of the predefined order 1 filter. step audit. FIG. 7 represents, by way of illustration, the monitoring template elaborated in this second embodiment around the measurement m 'of the monitored parameter p.
    [0023] The dynamic of the measurement of the monitored parameter p in response to a step f0 reflecting the raw reference value applied to the regulation parameters is represented by the curve f1. By dynamics, here we mean the variations of its amplitude as a function of time. The curve f2 represents the estimate of the monitored parameter p_ is weighted by the output of the filter of order 1. Similarly, the curve f3 represents the estimate of the monitored parameter p_ is weighted by the output of the filter of order 2. The terminals Lower Binf and higher Bsup are obtained respectively by subtracting and adding the uncertainty G to the curves f2 and f3. These terminals Binf and Bsup define a mask for monitoring the measurement m 'of the monitored parameter p, the crossing of one of these terminals by the monitored parameter triggering according to the invention the transmission of a notification by the monitoring device. 11 to the entity 18 for the management of the alerts and the firing sequence. It should be noted that these terminals Binf and Bsup are caused to evolve as a function of time, since they vary as a function of the setpoint values of the regulation parameters, in other words of the operating point of the motor. As previously mentioned, in the embodiments described herein, the invention is applied to a rocket engine. However, the invention can be applied to other regulated systems such as for example an aircraft engine.
    权利要求:
    Claims (13)
    [0001]
    REVENDICATIONS1. A method of monitoring a parameter of a rocket engine (1) comprising: a step of obtaining (E10) a measurement (m ') of the monitored parameter performed by a sensor (13) and corresponding to a point of operation of the engine, this operating point being defined by at least one engine control parameter; a step of estimating (E20) a value of the parameter monitored for this operating point, from a regulated value (yR) or a filtered setpoint value (yCf) of said at least one control parameter of the motor defining the operating point; a step of comparing (E40) an error between the measurement of the monitored parameter and its estimate with respect to at least one determined threshold (51,52) from an uncertainty on this evaluated error for the operating point; and an emission step (E60) of a notification in case of crossing said at least one threshold.
    [0002]
    2. Method according to claim 1 wherein the step of estimating (E20) the value of the monitored parameter is performed using an artificial neural network (RNA) having as input the regulated value or the value of filtered setpoint of said at least one engine control parameter.
    [0003]
    3. Method according to claim 1 or 2 wherein, during the comparison step, the operating point used to evaluate the error uncertainty is determined from: - the regulated value (yR) of said minus one regulation parameter; or - the filtered set point value (yCf) of said at least one regulation parameter; or - the set value (yC) of said at least one regulating parameter.
    [0004]
    4. Method according to any one of claims 1 to 3 wherein the operating point is defined by at least one of the following control parameters: - a pressure of a combustion chamber of the engine; a mixing ratio at the inlet of an engine pump.
    [0005]
    5. Method according to any one of claims 1 to 4 wherein during the comparison step, it takes into account a dynamic change of the operating point of the engine.
    [0006]
    6. Method according to claim 5, wherein, during the comparison step, comparing: at a first threshold said high, the error between the measurement of the monitored parameter and its estimate weighted by the response of a filter d command 2 predefined at a step representative of the set value or the filtered set value of said at least one control parameter; and at a second threshold said low, the error between the measurement of the monitored parameter and its estimate weighted by the response of a predefined order 1 filter to said step.
    [0007]
    7. Method according to any one of claims 1 to 6 wherein the parameter monitored is a parameter among: a pressure of the engine; a motor temperature; a rotational speed of an element of the engine, a flow rate of a fluid flowing in the engine; a vibratory behavior of the engine.
    [0008]
    The method according to any one of claims 1 to 7 wherein the error uncertainty considers at least one uncertainty among an estimation uncertainty of the monitored parameter and a measurement uncertainty of the monitored parameter.
    [0009]
    9. The method of claim 2 wherein the uncertainty on the error takes into account an estimation uncertainty of the monitored parameter determined according to: - an uncertainty on the input of the artificial neural network; - an uncertainty on the neural network; - an uncertainty on a learning base used to build the neural network.
    [0010]
    A computer program comprising instructions for executing the steps of the monitoring method according to any one of claims 1 to 9 when said program is executed by a computer.
    [0011]
    A computer-readable recording medium on which is recorded a computer program comprising instructions for performing the steps of the monitoring method according to any one of claims 1 to 9.
    [0012]
    12. Device for monitoring (11) a parameter of a rocket engine comprising: a module for obtaining (14) a measurement of the monitored parameter carried out by a sensor (13) and corresponding to an operating point of the motor, this operating point being defined by at least one engine control parameter; an estimation module (15) of a value of the parameter monitored for this operating point, from a regulated value (yR) or a filtered set value (yCf) of said at least one engine control parameter defining the operating point; a module for comparing (16) an error between the measurement of the monitored parameter and its estimation with respect to at least one threshold determined from an uncertainty on this evaluated error for the operating point; and a transmission module (17) of a notification in case of crossing of said at least one threshold.
    [0013]
    13. Stub engine (1) comprising a device according to claim 12.10
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3126659B1|2019-12-11|Method and device for monitoring a parameter of a rocket engine
    CA2828704C|2019-03-26|Method and device for monitoring a servo-valve actuation system
    EP3123139B1|2018-05-23|Method for estimating the normal or abnormal character of a measured value of a physical parameter of an aircraft motor
    EP3152528B1|2020-01-15|Method and system for evaluation of a fluid flow
    CA2877210C|2021-05-04|Method and device for adjusting a threshold value of a fuel flow rate
    FR2862793A1|2005-05-27|Motor vehicle components e.g. electro-hydraulic circuit pad, monitoring method, involves classifying component as free from error or erroneous based on whether measured temporal rate is within or outside narrow and wide tolerance limits
    FR3027061B1|2019-10-25|METHOD AND DEVICE FOR NOTIFYING A COMPLETE STOP AUTHORIZATION OF AN AIRCRAFT GAS TURBINE ENGINE
    FR3040220A1|2017-02-24|SYSTEM FOR REGULATING A REGULATED PARAMETER
    CA2950347A1|2015-12-03|Method and device for control of a thrust of a turbojet engine
    FR3029239A1|2016-06-03|METHOD FOR MONITORING THE OPERATION OF A TURBOMACHINE
    FR3035232A1|2016-10-21|SYSTEM FOR MONITORING THE HEALTH CONDITION OF AN ENGINE AND ASSOCIATED CONFIGURATION METHOD
    FR3043432A1|2017-05-12|METHOD AND DEVICE FOR ADJUSTING SET VALUES OF A PLURALITY OF VARIABLE GEOMETRIES DETERMINED OF A TURBOMACHINE
    CA2745977C|2017-10-24|Method for estimating a jet temperature in a jet engine
    FR2972528A1|2012-09-14|Device for reliabilization of value of e.g. variable, measured by measuring equipments, has parametric adaptation module readjusting estimate of measured variable during detection of defect of measuring equipment
    FR3030624A1|2016-06-24|METHOD AND DEVICE FOR OBTAINING REFERENCE DIFFERENTIAL PRESSURE OF A FLUID CROSSING A FILTER OF AN AIRCRAFT ENGINE
    FR3024803A1|2016-02-12|METHOD FOR MONITORING SHORT CIRCUITS ON TORQUE MOTORS
    WO2019207244A1|2019-10-31|Method and system for processing a temperature measurement signal delivered by a sensor
    EP3332108B1|2019-05-29|Method for controlling the pressure and a mixture ratio of a rocket engine, and corresponding device
    FR3028054A1|2016-05-06|IMPROVED CONTROLLER OF STATE OF A PILOT SYSTEM BY AN ORDER
    WO2022043629A1|2022-03-03|Method for determining a fault for a set of sensors
    FR3097063A1|2020-12-11|Method for determining a predictive model of a pressure ratio for a double-flow turbomachine
    EP3942168A1|2022-01-26|Method for monitoring the operating state of a hydromechanical unit
    FR2933451A3|2010-01-08|Electronic fuel injection type diesel engine's idle speed control system for motor vehicle, has proportional integral regulator with proportional gain equal to specific equation constituting gain margin of open loop and engine gain
    同族专利:
    公开号 | 公开日
    RU2016143197A3|2018-05-04|
    FR3019592B1|2016-04-22|
    JP2017524851A|2017-08-31|
    EP3126659B1|2019-12-11|
    JP6585077B2|2019-10-02|
    CN106460727A|2017-02-22|
    RU2654310C2|2018-05-17|
    RU2016143197A|2018-05-04|
    WO2015150706A1|2015-10-08|
    CN106460727B|2018-09-21|
    US10267265B2|2019-04-23|
    CA2944120A1|2015-10-08|
    EP3126659A1|2017-02-08|
    US20170175680A1|2017-06-22|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    EP1496414A1|2003-07-07|2005-01-12|Airbus France|Device and method for controlling the validity of at least one parameter calculated by an aircraft anemometer station|
    US20050222747A1|2004-03-30|2005-10-06|Vhora Mohamad H|Model-based detection, diagnosis of turbine engine faults|
    FR2965305A1|2010-09-28|2012-03-30|Snecma|PROPULSIVE SPATIAL LAUNCHER SYSTEM USING A METHOD OF CONTROLLING THE CONSUMPTION OF ERGOLS|
    WO2012156615A2|2011-05-17|2012-11-22|Snecma|Power supply system and method for eliminating the pogo effect|
    EP2592253A1|2011-11-08|2013-05-15|Thales|FADEC for aeroengine|
    JPH04211859A|1990-02-28|1992-08-03|Chiyoda Corp|Abnormality recognizing method|
    JP3624546B2|1996-05-17|2005-03-02|株式会社日立製作所|Preventive maintenance method and apparatus|
    US5857321A|1996-06-11|1999-01-12|General Electric Company|Controller with neural network for estimating gas turbine internal cycle parameters|
    JP4046309B2|1999-03-12|2008-02-13|株式会社東芝|Plant monitoring device|
    US7216071B2|2002-04-23|2007-05-08|United Technologies Corporation|Hybrid gas turbine engine state variable model|
    DE102007017259B4|2007-04-12|2009-04-09|Siemens Ag|Method for computer-aided control and / or regulation of a technical system|
    CN101408752B|2008-10-21|2014-03-26|中国人民解放军海军航空工程学院|Neural network extreme control method and system based on chaos annealing and parameter destabilization|
    JP2011145846A|2010-01-14|2011-07-28|Hitachi Ltd|Anomaly detection method, anomaly detection system and anomaly detection program|
    FR2956701B1|2010-02-24|2012-07-06|Snecma|DETECTION AND TROUBLESHOOTING SYSTEM AND ENGINE COMPRISING SUCH A SYSTEM|
    CN102374038B|2011-09-06|2013-09-11|天津大学|VVT control method capable of combining self-learning feed-forward and active anti-interference feedback|
    CN103632033A|2013-11-05|2014-03-12|奇瑞汽车股份有限公司|MAP calibrating method based on BP neural network|RU2660173C1|2017-04-13|2018-07-05|Общество с ограниченной ответственностью "Компекс-Т"|Device for introduction of frequency signals of flow and number of turnover in control systems for bench test of rocket and space equipment|
    RU2692591C1|2018-06-29|2019-06-25|федеральное государственное автономное образовательное учреждение высшего образования "Самарский национальный исследовательский университет имени академика С.П. Королёва"|Traction measuring device for testing liquid thrust low-thrust engines in stationary mode of operation|
    RU2711813C1|2019-01-22|2020-01-23|федеральное государственное автономное образовательное учреждение высшего образования "Самарский национальный исследовательский университет имени академика С.П. Королёва"|Traction device for low-thrust liquid-propellant rocket engines in pulsed operation modes|
    CN109779791B|2019-03-24|2021-01-01|西安航天动力测控技术研究所|Intelligent diagnosis method for abnormal data in solid rocket engine|
    JP2021124045A|2020-02-04|2021-08-30|国立研究開発法人宇宙航空研究開発機構|Failure diagnostic system for spacecraft liquid propulsion system, and failure diagnostic method for spacecraft liquid propulsion system|
    FR3106862B1|2020-02-04|2022-02-04|Arianegroup Sas|Chilling method using an artificial neural network|
    KR102289048B1|2020-03-03|2021-08-11|국방과학연구소|Sensor measurement device of ground fire test|
    RU2750875C1|2020-07-24|2021-07-05|Акционерное общество "НПО Энергомаш имени академика В.П. Глушко"|Method for localizing malfunction of liquid-propellant rocket engine during fire test|
    RU2758781C1|2020-07-24|2021-11-01|Акционерное общество "НПО Энергомаш имени академика В.П. Глушко"|Method for monitoring the technical condition of a liquid rocket engine in transitional modes of a bench test|
    CN112539119B|2020-11-30|2021-10-01|北京宇航系统工程研究所|Variable threshold-based propellant utilization system adjusting method|
    CN112377333B|2020-12-02|2021-11-16|西安航天动力研究所|Real-time fault monitoring method for turbopump of high-thrust liquid rocket engine|
    法律状态:
    2015-04-14| PLFP| Fee payment|Year of fee payment: 2 |
    2016-04-14| PLFP| Fee payment|Year of fee payment: 3 |
    2017-04-06| PLFP| Fee payment|Year of fee payment: 4 |
    2018-03-22| PLFP| Fee payment|Year of fee payment: 5 |
    2018-06-29| CD| Change of name or company name|Owner name: SAFRAN AIRCRAFT ENGINES, FR Effective date: 20170719 |
    2020-03-19| PLFP| Fee payment|Year of fee payment: 7 |
    2021-03-23| PLFP| Fee payment|Year of fee payment: 8 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1452954A|FR3019592B1|2014-04-03|2014-04-03|METHOD AND DEVICE FOR MONITORING A PARAMETER OF A ROTOR MOTOR|FR1452954A| FR3019592B1|2014-04-03|2014-04-03|METHOD AND DEVICE FOR MONITORING A PARAMETER OF A ROTOR MOTOR|
    RU2016143197A| RU2654310C2|2014-04-03|2015-04-02|Method and device for the rocket engine parameter monitoring|
    PCT/FR2015/050858| WO2015150706A1|2014-04-03|2015-04-02|Method and device for monitoring a parameter of a rocket engine|
    CA2944120A| CA2944120A1|2014-04-03|2015-04-02|Method and device for monitoring a parameter of a rocket engine|
    CN201580025959.3A| CN106460727B|2014-04-03|2015-04-02|Method and apparatus for the parameter for monitoring rocket engine|
    JP2016560577A| JP6585077B2|2014-04-03|2015-04-02|Method and apparatus for monitoring rocket engine parameters|
    EP15718541.4A| EP3126659B1|2014-04-03|2015-04-02|Method and device for monitoring a parameter of a rocket engine|
    US15/301,496| US10267265B2|2014-04-03|2015-04-02|Method and device for monitoring a parameter of a rocket engine|
    [返回顶部]