METHOD AND A DEVICE FOR GENERATING A CONTROL OF A FLOW OF FUEL INTENDED TO BE INJECTED INTO A COMBUSTION CHAMBER OF A TURBO MACHINE^ Background of the invention The present invention relates to the general field of turbine engines, and it applies in preferred manner to the field of aviation. The invention relates more particularly to regulating the flow rate of fuel for a turbine engine of an aircraft, such as a turbojet, for example, during a stage of starting the aircraft. In known manner, the flow rate of fuel for a turbojet is regulated by generating appropriate fuel flow rate commands for the purpose of ensuring that the mass flow rate of fuel injected into the combustion chamber of the turbojet does not exceed a certain (lower or upper) limit beyond which a malfunction of the turbojet may be encountered, such as for example shutdown or surging of a compressor of the turbojet. Such regulation is conventional.ly performed in an open-loop using fuel flow rate commands generated on the basis of a relationship, or more precisely on the basis of a network of pre-established relationships, giving the flow rate of fuel to be injected into the combustion chamber for various different reduced (normalized) speed values of a compressor of the turbojet (e.g. a highpressure compressor for a two-spool turbojet). In general, two distinct groups of networks of preestablished relationships are taken into consideration: a first group of relationship networks for ensuring ignition in the combustion chambgr, and providing a fuel flow rate command written WMCmd as a function of at least the reduced speed of the compressor which is written XNr, in other words: WFCmd = f (XNr); and I Translation of the title as established ex oflcio. a second group of relationship networks, also known as C/P limits (referring to the ratio of the flow rate C of fuel injected into the combustion chamber divided by the static pressure P measured at the outlet 5 from the combustion chamber), specifying the fuel flow rate for managing the turbojet spin-up stage until it reaches idling speed. In known manner, one such limit may be written in particular in the following form: 10 where WF is the fuel flow rate, PS is the static pressure in the combustion chamber, T is the total temperature at the inlet to the high-pressure compressor, XNr is the reduced speed of the high-pressure spool, and PT is the total pressure at the inlet to the fan. 15 These various relationship networks are drawn up so as to take account of the specific features of the turbojet and also its sensitivity to various parameters, such as, for example: outside temperature, flight domain, etc. 2 0 Presently-designed turbojets present ever increasing performance, and their components (compressor, turbine, etc.) are optimized for operating at high speed, to the detriment of l~ow speeds, and in particular durinrj the starting stage. 2 5 This leads to modern turbojets being very sensitive to external conditions (e.g. thermal state of the turbojet, outside temperature, accuracy with which fuel is metered, type of fuel injected, outside temperature, aging of the jet, etc.), and it also leads to wide 30 dispersion in behavior between turbojets. The operability limits of turbojets that are taken into account during open-loop regulation are thus subject to a large degree of variability from one turbojet to another, which is difficult to predict. Furthermore, the very great sensitivity of such turbojets to numerous parameters makes it laborious, if not impossible, to adjust the above-mentioned command relationships. 5 It should be observed that for a turbojet having a high-pressure compressor with a compression ratio that is high relative to the number of stages in the compressor, this very great sensitivity also leads to the existence of a relatively narrow corridor between the surging limit 10 and the stagnation limit. There therefore exists a need for a mechanism for regulating the flow rate of fuel for a turbine engine that is effective and appropriate for the starting stage, which mechanism takes account of the above-mentioned 15 constraints that are imposed by the turbine engines being designed nowadays. Object and summary of the invention The present invention satisfies this need in 20 particular by proposing a method of generating a command of a fuel flow rate to be injected into a combustion chamber of a turbine engine for propelling an aircraft, the method being for use during a stage of starting the turbine engine, and comprising: 25 an open-loop generating step of generating a fuel flow rate command from at least one pre-established relationship; and a closed-loop monitoring step of monitoring at least one operating parameter of the turbine engine 30 selected from: a rate of acceleration of a cqmpressor of the turbine engine; and a temperature at the outlet from a turbine of the turbine engine; 35 this monitoring step comprising maintaining the operating parameter in a determined range of values by using at least one corrector network associated with the parameter and suitable for delivering a signal for correc-tj.ng the open-loop generated fuel flow rate command soas to maintain the operating parameter in the determined range of values. 5 Correspondingly, the invention also provides a device for generating a command of a fuel flow rate to be injected into a combustion chamber of a turbine engine for propelling an aircraft, the device comprising means that are activated during a stage of starting the turbine 10 engine and that comprise: a generator module for open-loop generation of a command of the fuel flow rate from at least one preestablished relationship; and a monitoring module for closed-loop monitoring of 15 at least one operating parameter of the turbine engine selected from: a rate of acceleration of a compressor of the turbine engine; and a temperature at the outlet from a turbine of 20 the turbine engine; the monitoring module being suitable for maintaining the operating parameter in a determined range of values, and comprising at least one corrector network associated with that parameter and suitable for delivering a 25 correction signal for correcting the open-loop generated fuel flow rate command so as to enable the operating parameter to be maintained in the determined range of values, and correction means that are activated, where appropriate, for correcting the fuel flow rate command 30 generated by the generator module by using the correction signal delivered by the corrector network., The invention thus proposes introducing closed-loop regulation of the fuel flow rate that is to be injected into the combustion chamber of the turbine engine, 35 thereby enabling certain suitably selected operating parameters of the turbine engine to be contained within a determined range of values so as to keep the turbine engine within conditions of operability. Such operating parameters are typically the rate of acceleration of a compressor of the turbine engine and the temperature at the outlet from the -turbine of the turbine engine. The i-nvention thus defines a'control corridor (or in equivalent manner a range of values that are authorized) around the regulation relationship that is conventionally used in an open-loop: so long as the operating parameters of the turbine engine continue to have current values that are contained within the corridor, then the fuel flow rate is regulated by commands generated on the basis of conventional control relationships for open-loop regulation of the fuel flow rate. In contrast, once the current value of any one of the operating parameters leaves or is likely to leave the corridor, a control-loop is implemented in accordance with the invention in order to correct (i.e. adjust) the fuel flow rate as established from such conventional open-loop control relationships, so that, where necessary, these values for the operating parameters are returned to and maintain within the control corridor. In accordance with the invention, the closed-loop regulation of the fuel flow rate that is used is thus not a full authority regulation-loop: it comes into operation only when certain operating parameters of the turbine engine cross or are about to cross pre-established setpoint values that are deduced from the operability limits of the turbine engine. For this purpose, the closed-loop regulation proposed by the invention relies advantageously on corrector networks associated with the operating parameter(s) that are being monitored, and more precisely on correction signals that those corrector networks deliver, when necessary, for the purpose of enabling operating parameters to be maintainedwithin the intended control corridor. The correction signals are applied to the open-loop generated command so that the command as corrected in this way serves to maintain the operating values within the range of values defining the control 5 corridor. Consequently, the invention is particularly original in that for the purpose of regulating the flow rate of fuel injected into the combustion chamber of the turbine engine it proposes relying on a main control that is 10 open-loop generated, and that is adjusted, if necessary, by means of a closed-loop relying on corrector networks suitable for ensuring that the rate of acceleration of the compressor and/or the temperature at the outlet from the turbine are contained within a range of predetermined 15 values so as to guarantee operability of the turbine engine. In other words, the invention is relatively easy to implement. It does not require knowledge of how the operating parameters vary as a function of the injected 20 fuel flow rate, but only requires control templates to be defined for these operating parameters, i.e. ranges of values within which these operating parameters ought to lie, which is particularly simple to undertake. Consequently, the invention can be incorporated very 25 easily in existing control architectures that are based on open-loop regulation of the fuel flow rate. The invention makes it possible to benefit from advantages that result from closed-loop regulation of the fuel flow rate (i.e. effectiveness, better accuracy), 30 while guaranteeing simplicity and ease of implementation. These operating parameters that are t,aken into consideration for closed-loop regulation as proposed by the invention comprise in particular a rate of acceleration of a compressor of the turbine engine (e.g. ' 35 the high-pressure compressor in a two-spool turbine engine), and a temperature at the outlet from the turbine of the turbine engine, also known as the exhaust gas temperature (EGT) . In known manner, such operating parameters are already measured by using sensors of the aircraft or of the turbine engine, or in a variant they are evaluated on the basis of measurements coming from such sensors, and they participate in the monitoring and control of the turbine engine as performed by the full authority digital engine control (FADEC) system of the aircraft. There is therefore no need to include new sensors on board the aircraft or the turbine engine in order to implement the invention. Monitoring the ra-te of acceleration makes it possible advantageously to detect stagnation or surging of the turbine engine. Thus, during the monitoring step, the rate of acceleration of the compressor of the turbine engine is preferably maintained between a minimum acceleration setpoint value (in order to avoid a risk of stagnation) and a maximum acceleration setpoint value (in order to avoid a risk of surging) by using two distinct corrector networks. Monitoring the temperature at the outlet from the turbine serves to detect behavior of the turbine engine that runs the risk of requiring starting to be interrupted. In order to avoid such interruption, during the monitoring step, the temperature at the outlet from the turbine of the turbine engine is preferably maintained below a maximum temperature setpoint value. Naturally, the invention is not limiGed to the above-mentioned operating parameters, namely the rate of acceleration and the temperature at the outlet from the turbine, and it is also possible to envisage monitoring other operating parameters in addition to the abovementioned parameters that have an impact on the behavior of the turbine engine on starting, such as for example the pressure in the combustion chamber. In a particular implementation, both the rate of acceleration of the compressor and the temperature at the outlet from the turbine are monitored, and the monitoring step includes selecting one of the signals from among the correction signals generated by the corrector networks associated with the rate of acceleration of the compressor and with the temperature at the outlet from the turbine, the selected signal being used for correcting the open-loop generated fuel flow rate command. Correspondingly, in a particular embodiment, the monitoring module comprises a plurality of corrector networks and the means for selecting one of the correction signals from among the correction signals delivered by the corrector networks, the selected signal being delivered to the correction means for correcting the fuel flow rate command as generated in an open-loop by the generator module. It should be observed that, at any given instant, the corrector networks do not necessarily all provide a respective correction signal (i.e. the corrector networks do not need to be activated continuously). This depends in particular on the current value of the operating parameter being monitored by each corrector network, which value may lie in a range of values that are acceptable (i.e. "valid" or "authorized") for that parameter and that enable the turbine engine to operate, such that properly speaking there is no need for any correction to the open-loop generated fuel? flow rate command. The selection that is performed during the monitoring step, where appropriate, serves to organize the correction signals delivered by the various corrector networks in hierarchical manner, in particular so as to limit any divergencies that might appear between the correction signals. By way of example, such selection may be performed by a succession of components suitable for selecting the minimum value or the maximum value from among the signals present at their inputs, and suitably arranged between the outputs of the corrector networks. By way of illustration, in certain situations, it can happen that the temperature at the outlet from the turbine and the rate of acceleration of the compressor both depart from their respective control corridors. In particular, it can happen that the rate of acceleration of the compressor approaches a minimum setpoint value representative of abnormal stagnation of the turbine engine, while the temperature at the outlet from the turbine begins to exceed a maximum setpoint value. In such a situation, it is necessary to select the most appropriate correction signal from among the correction signals delivered by the corrector networks. For this purpose, it is preferable to give precedence to high setpoints, i.e. the selected correction signal is the correction signal that is generated by the corrector network that is associated with the temperature at the outlet from the turbine, and that delivers a correction signal enabling the value of the temperature at the outlet from the turbine to be kept below a maximum setpoint value. This ensures that the turbine engine is not damaged irremediably as a result of overheating, which can be fatal. In a particular implementation, each,corrector network is of the proportional integral type (e.g. class 1 proportional integral PI, or proportional double integral PI-I), and is suitable for delivering a correction signal for correcting the fuel flow rate command, which correction signal is evaluated from a difference between a current value of the operating parameter with which it is associated and a determined setpoint value. This implementation is relatively easy to perform by adjusting parameters of each corrector network (e.g. gain, activation of the network, etc.). Thus, the gain of each network may depend in particular on a static pressure in the combustion chamber and on a total pressure at the inlet of a fan of the turbine engine. In a preferred implementation, the regulator means comprise a corrector network for each monitored parameter and for each setpoint value established for that parameter. Thus, by way of illustration, if the operating parameters taken into consideration are the rate of acceleration of a compressor of the turbine engine and the temperature EGT, and if the monitoring module of the device is configured to maintain the rate of acceleration of the compressor between a minimum acceleration setpoint value and a maximum acceleration setpoint value, and to keep the temperature EGT below a maximum temperature setpoint value, then the generator device of the invention may have three corrector networks. When the regulator device has a plurality of corrector networks, the corrector networks may advantageously share a common integrator, preferably a saturated integrator. By way of example, this saturation of the common integrator can be performed as a function of the openloop generated fuel flow rate command. This makes it possible to reduce the compl'exity and the cost associated with implementing the,invention. Saturation of the common integrator also makes it possible to limit the correction signals delivered by the corrector networks. The common integrator may also be used by the openloop so as to limit discontinuities that might appear between fuel flow rate commands. Correspondingly, in a particular implementation, the method of generation further includes a saturating step for saturating the open-loop generated command or the open-loop generated command as corrected using the 5 correction signal, which saturation depends on the nominal relationship. By way of example, this saturation is defined from a determined percentage of the nominal relationship. This saturation step serves to limit the fuel flow 10 rate commands used for regulating the turbine engine on starting. This saturation step may be envisaged in par~icuiar when it is desired to limit the flow rate of fuel injected into the combustion chamber, e.g. in order to 15 remain within fuel injection limits specified by the metering device of the turbine engine. This saturation step also makes it possibleto guarantee that the commands used for regulating the turbine engine in fuel flow rate are not divergent or 20 aberrant, in particular in the event of a failure of the turbine engine. Regardless of whether the open-loop generated command is subjected to an adjustment step, this saturation step may force the open-loop generated command 25 to take one or the other of a first limit value and a second limit value corresponding respectively to a minimum percentage and to a maximum percentage of the open-loop generated command that has not been subjected to said adjustment step, whenever the current value of 30 the open-loop generated command is respectively less than the first limit value or greater than the,second limit value. In a particular implementation, the various steps of the generator method are determined by computer program 35 instructions. Consequently, the invention also provides a computer program on a data medium, the program being suitable for being implemented in a generator device or more generally in a computer, the program including instructions adapted for performing steps of a method of generation as defined above. The program may use any programming language, and be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form. The invention also provides a computer readable data medium that includes instructions of a computer program as mentioned above. The data medium may be any entity or device capable of storing the program. For example, the medium may comprise storage means such as a read only memory (ROM), e.g. a compact disk (CD) ROM or a microelectronic circuit ROM, or indeed magnetic recording means, such as a floppy disk or a hard disk. Furthermore, the data medium may be a transmissible medium such as an electrical or optical signal, suitable for being conveyed via an electrical or optical cable, by radio, or by other means. The program of the invention may in particular be downloaded from an Internet type network. Alternatively, the data medium 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 provides a turbine engine including a generator device of the invention. The generator device is preferably iqcorporated in the full authority control system of the aircraft. The turbine engine of the invention benefits from the same advantages as those mentioned above for the method of generation and the generator device. In other implementations or embodiments, it is also possible to envisage that the method of generation, the generator device, and the turbine engine of the invention present in combination all or some of the above-mentioned characteristics. 5 Brief description of the drawings Other characteristics and advantages of the present invention appear from the following description made with reference to the accompanying drawings, which show an implementation having no limiting character. In the 10 figures: Figure 1 shows a turbine engine and a generator device in accordance with the invention, in a parLicular embodiment; Figure 2 is a diagram representing the hardware 15 architecture of the generator device of Figure 1; Figure 3 is in the form of a flow chart showing the main steps of the method of generation implemented by the generator device of Figure 1; Figure 4 shows control architecture that can be 20 used by the Figure 1 generator device for implementing the steps shown in Figure 3; and Figures 5A and 5B represent examples of corrector networks that can be used in the generator device. 25 Detailed description of the invention Figure 1 is a diagram showing a turbine engine 1 in accordance with the invention in its environment, in a particular embodiment. In this embodiment, the turbine engine 1 is a bypass 30 two-spool turbojet for propelling an airplane. Nevertheless, the invention applies to otQer turbine engines such as, for example: a single-spool turbojet or a turboprop, and also to other types of aircraft. In known manner, the turbojet 1 has a fuel metering 35 device, also called a fuel meter, that is suitable for adjusting the quantity of fuel coming from the fuel circuit of the airplane and delivered by the fuel injector system of the combustion chamber of the turbojet. The fuel injector, the fuel circuit, and the fuel injector system of the combustion chamber of the turbojet 1 are omitted in Figure 1 for simplification purposes. In this example, the fuel meter of the turbojet 1 has a fuel metering valve (FMV) of position that varies as a function of the fuel flow rate to be injected into the combustion chamber. The fuel flow rate to be injected into the combustion chamber is transmitted to the fuel meter in the form of a command WFCmd, via a servo-control-loop. This fuel flow rate command WFCmd is established by a generator device 2 in accordance with the invention, which device is incorporated in the presently-described embodiment in the FADEC system 3 of the airplane. In order to establish this command, the regulator device 2 relies on two main functional entities: a generator module 2A suitable for operating in an open-loop to generate a fuel flow rate command WF OL on the basis of a relationship or a network of regulation relationships pre-established as a function of the current reduced speed of rotation of the turbojet 1; and a monitor module 2B suitable for operating in a closed-loop to monitor operating parameters of the turbojet 1, and for acting via said closed-loop to maintain these operating parameters within a predetermined range of values by means of various corrector networks referenced R1, R2, and R3. These corrector networks are suitable, where appropriate, for delivering correction signals that enable,the monitor module 2B to modulate (i.e. adjust or correct) the command WE - OL as generated by the module 2A, so that the current values of the operating parameters of the turbojet that result from the metering device applying the command as adjusted remains contained within the above-specified range of values. In the presently-described example, provision is made for the module 2B to use the corrector networks Rl, R2, and R3 to monitor two operating parameters of the turbojet 1, namely: the rate of acceleration, written (dN2/dt), of the high-pressure compressor of the turbojet 1, which rate is obtained by taking the time derivative of the speed of rotation N2 of the high-pressure compressor; and the exhaust gas temperature, written EGT, at the outlet from the turbine of the turbojet 1. Nevertheless, no limit is put on the number of operating parameters of the turbojet that may be monitored in accordance with the invention, and in other implementations, it is possible to envisage monitoring only the rate of acceleration of a compressor of the turbojet 1, or in a variant monitoring other operating parameters in addition to the rate of acceleration of the compressor of the turbojet and/or the gas temperature at the outlet from the turbojet turbine. In the presently-described implementation, the above-described functional modules 2A and 2B are software modules implemented by the generator device 2 in the context of the logic applied by the FADEC 3 to regulating the turbojet I. For this purpose, the generator device 2 possesses the hardware architecture of a computer, as shown diagrammatically in Figure 2. In particular, it comprises a processor 4, a random access memory (RAM) 5, a ROM 6, a non-volatile flash memory 7, and communication means 8, possibly shared with other regulator units of the FADEC 3. I The communication means 8 comprise means for communicating with various sensors 9 of the airplane, and suitable for providing the generator device 2 with measurements of the current values for the speed of rotation N2 of the high-pressure compressor of the turbojet 1, of the temperature EGT of the gas at the outlet from the turbine of the turbojet 1, and also of the static pressure PS32 in the combustion chamber and the total pressure Pt at the inlet to the fan of the turbojet 1. By way of example, the sensors 9 may comprise a speed sensor, a temperature sensor, and pressure sensors positioned so as to measure the parameters N2, EGT, PS32, and Pt, in conventional manner. The measurements delivered by these sensors 9 enable the generator device 2 specifically to estimate a current value for the rate of acceleration (dN2/dt), in conventional manner, by differentiating the speed of rotation N2, and to monitor the parameters (dN2/dt) and EGT in accordance with the invention. The ROM 6 of the generator device 2 constitutes a data medium in accordance with the invention that is readable by the processor 4 and that stores a computer program in accordance with the invention, including instructions for executing steps of a method of generation in accordance with the invention and as described below with reference to Figure 3. Figure 3 is a flow chart showing the main steps of the method of generation of the invention in a particular implementation in which it is performed by the generator device 2 of Figure 1 for the purpose of regulating the fuel flow rate of the turbojet 1. Such a method applies in preferred manner during a stage of starting the turbojet 1. It is assumed in this example that the turbojet 1 is in a starting stage (step EO). This starting stage results from a specific command being appl,ied to the turbojet 1, and it can easily be detected in conventional manner. It should be observed that the invention applies to any type of starting of the turbojet 1: it may be starting the turbojet 1 on the ground after it has been shut down for a long period, or equally well restarting in flight, or to reigniting the turbojet 1 after a flameout of short duration. In accordance with the invention, the generator device 2 of the FADEC 3 acts during this starting stage 5 of the turbojet 1 to perform "main" regulation of the fuel flow rate to be injected into the combustion chamber of the turbojet 1 in an open-loop and using the module 2A (step E10). More specifically, during this step E10, the module 10 2A generates a fuel flow rate command WF-OL on the basis of a relationship or of a network of pre-established relationships LN. This network of relationships establishes a command (i.e. a value) for the fuel flow rate to be sent to the fuel meter of the turbojet 1, 15 which command is established as a function of the reduced speed of rotation of the turbojet 1. Such a network of relationships is itself known and has already been described. It applies in particular both to a first relationship for ensuring ignition of the 20 combustion chamber, and delivering a fuel flow rate command as a function of the reduced speed of the compressor, and also to a C/P second limit for managing the spin-up stage of the turbojet up to idling speed. The way in which such command relationships are prepared 25 and taken into account is known to the person skilled in the art and is not described further herein. In the prior art, the command WF - OL is to be delivered to the fuel meter of the turbojet 1 directly. In contrast, in accordance with the invention, in 30 parallel with this open-loop regulation implemented using the command WF - OL, the generator device 2 ,uses the module 2B to monitor the current values of the rate of acceleration (dN2/dt) of the high-pressure compressor of the turbojet 1 and of the temperature EGT at the outlet 35 from the turbine (step E20). These current values are the values of the rate of acceleration dN2/dt and of the temperature EGT that results from the fuel flow rate regulation performed by the FADEC on the basis of the command WF - OL generated by the module 2A without correction, at least while the method is starting. These current values are obtained by the module 2B from measurements taken by the speed and temperature sensors 9 of the airplane, e.g. periodically, giving current values for the speed of rotation N2 of the highpressure spool and for the temperature EGT at the outlet from the turbine. Thereafter, the module 2B differentiates the current measurement of the speed N2 relative to time in order to estimate a current value of the rate of acceleration dN2/dt. In accordance with the invention, these current values for the rate of acceleration dN2/dt and for the temperature EGT are monitored by the module 2B, i.e. they are analyzed, and where appropriate they are processed. More specifically, during the monitoring step E20, the module 2B acts via a closed-loop making use of the corrector networks R1, R2, and R3 with suitable parameters and interconnected with one another to maintain the current values of the rate of acceleration dN2/dt and of the temperature EGT within determined ranges of values (also referred to in the present description as the "control corridor"). These values are maintained by using correction signals delivered by the corrector networks R1, R2, and R3, which signals are used by the monitoring module 2B to adjust (i.e. to correct or to modulate) the command WL OL as generated by the module 2A (step E30). It should be observed that the commaqd WL - OL is not adjusted all the time (which is why this is drawn in dashed lines in Figure 3): such adjustment is undertaken only when it is found to be necessary in order to maintain the rate of acceleration dN2/dt and the temperature EGT within the ranges of values that have been set in order to ensure operability of the turbojet 1. In other words, the closed-loop put into place by the monitoring module 2B is not a full authority closedloop: the main command for regulating the fuel meter is the command WF - OL delivered by the module 2A, which command is modulated in ancillary manner by the monitoring module 2B in order to maintain the values of the monitored operating parameters within the desired ranges of values. In the presently-described embodiment, the ranges of values under consideration for the rate of acceleration (dN2/dt) and for the temperature EGT are defined as described below. The range of values under consideration for the rate of acceleration is defined by a minimum setpoint value THRl that is set (e.g. by the service in charge of the operability of the turbojet 1) so as to avoid the turbojet 1 stagnating (i.e. the setpoint THRl represents a value for the rate of acceleration below which the turbojet 1 is considered to be abnormally stagnant), and by a maximum setpoint value THR2 that represents a value for the rate of acceleration above which it is considered that the turbojet 1 is accelerating too fast and runs the risk of the turbojet surging. It should be observed that the surging limit of the turbojet is difficult to transpose into a maximum setpoint value for the rate of acceleration, such that in the presently-described implementation, this maximum setpoint value THR2 is determined by training. For this purpose, a surging detector is used that qtores each surging event of the turbojet 1 together with the conditions under which such surging takes place, and for each event that is detected in this way it updates the threshold value THR2 as a function of the corresponding conditions. Such a mechanism for determining the threshold value THR2 is described in greater detail in as yet unpublished French patent application No. 11/51778. The range of values under consideration for the temperature EGT is defined by an upper limit only, i.e. 5 by a maximum temperature setpoint value THR3. This setpoint value is determined for example by the service in charge of the operability of the turbojet 1 so as to limit any risk of forced interruption of the turbojet 1 as a result of a temperature that is too high. 10 As described above, in the presently-described embodiment, the command WF - OL is adjusted by the monitoring module 2B on the basis of correction signals delivered by the corrector networks Rl, R2, and R3, which networks are of the proportional integral (PI) type or of 15 the proportional double integral (PI-I) type (referred to more generally in the present description as corrector networks of the proportional integral type). More specifically, the corrector network R1 in this example is for correcting the command WE - OL for fuel flow 20 rate in such a manner as to maintain the current value of the rate of acceleration of the high-pressure compressor above the minimum setpoint value THR1, below which there is a risk of the turbojet 1 stagnating. To this end, the corrector network R1 is suitable 25 for delivering a correction signal referenced SIGl that is to be added to the command WL - OL. In the present example, the corrector network R1 is a proportional double integral network (or PI-I integrator) having the transfer function Cl(p), where -p 30 is the Laplace variable, as follows: where K1 and ~1 are respective parameters of the corrector network R1. These parameters K1 and ~1 of the corrector network R1 depend on the state of the turbojet 35 1; more particularly, in this example, the gain K1 depends on the static pressure PS32 in the combustion chamber of the turbojet and on the total pressure Pt at the inlet to the fan of the turbine engine, whereas the parameter zl is set as a function of the inertia of the turbojet 1 in response to a flow rate of fuel injected into the combustion chamber. The current values for the pressures PS32 and Pt are delivered to the corrector network R1 by the sensors 9. The correction signal SIGl is obtained by applying the transfer signal Cl(p) to an error signal, written 61, between the minimum setpoint value THRl and the current value of dN2/dt, in other words: 61 = THRl - (dN2/dt) Naturally, other parameters may be taken into account for setting K1 and 71. Thus, the corrector network R1 causes an error signal 61 between the current value of the rate of acceleration and its minimum setpoint THR1, to correspond to a fuel flow rate increment SIGl for causing this error 61 to disappear, in other words for enabling the rate of acceleration of the high-pressure compressor of the turbojet 1 to return to a "normal" value (i.e. within the limits of the control corridor that has been set for the turbojet) . In similar manner, the corrector network R2 in this example is for correcting the fuel flow rate command WF - OL so as to maintain the current value of the rate of acceleration of the high-pressure compressor below the maximum setpoint value THR2, above which there exists a risk of the turbojet 1 surging. For this purpose, the corrector netwqrk R2 is suitable for delivering a correction signal written SIG2 for adding to the command WL - OL. In the presently-described example, the corrector network R2 is also a proportional double integral (PI-I) network having its transfer function C2(p) given by: where K2 and 22 are respective parameters of the corrector network R2. These parameters K2 and 22 of the corrector network R2 depend on the state of the turbojet 1; more particularly in this example, the gain K2 depends on the static pressure PS32 and on the total pressure Pt at the inlet to the fan of the turbine engine, while 22 is set as a function of the inertia of the turbojet 1 in response to a flow rate of fuel injected into its combustion chamber. The correction signal SIG2 is obtained by applying the transfer function C2(p) to an error signal written 62 between the maximum setpoint value THR2 and the current value of dN2/dt, in other words: 62 = THR2 - (dN2/dt) Naturally, other parameters could be taken into account for setting K2 and 22 Thus, the corrector network R2 causes an error signal 62 between a current value of the rate of acceleration and its maximum setpoint THR2 to correspond to a fuel flow rate increment SIG2 for causing the error 62 to disappear, in other words for enabling the rate of acceleration of the high-pressure compressor of [.he turbojet 1 to return to a "normal" value (i.e. a value within the limits of the control corridor that has been set for the turbojet) . Finally, the corrector network R3 in this example is for correcting the fuel flow rate command WF - OL so as to maintain the current value of the temperature EGT below the maximum temperature setpoint value THR3, above which there exists a non-negligible risk of it being necessary to interrupt starting of the turbojet 1. For this purpose, the corrector network R3 is suitable for delivering a correction signal written SIG3 that is to be added to the command WL - OL. In the presently-described example, the corrector network R3 is likewise a proportional integral network where K3 and 23 designate respective parameters of the corrector network R3. These parameters K3 and 23 of the corrector network R3 depend on the state of the turbojet 1; more particularly in this example, the gain K3 depends on the static pressure PS32 and on the total pressure Pt at the inlet to t.he fan of the turbine engine, while 23 is set as a function of the inertia of the turbojet 1 in response to a fuel flow rate injected into its combustion chamber. The correction signal SIG3 is obtained by applying the transfer function C3(p) to an error signal written 63 between the maximum setpoint value THR3 and the current value EGT, in other words: A3 = THR3 - EGT Naturally, other parameters may be taken into account for setting K3 and 23. Thus, the corrector network R3 causes an error signal 63 between the current value of the temperature EGT and its maximum setpoint THR3 to correspond lLo a fuel flow rate increment SIG3 that is to cause the error 63 to disappear, in other words that is to enable the temperature EGT of the turbojet 1 to be maintained at a value that is "normal" (i.e. within the limits of the control corridor that has been set for that value). In the presently-described implementation, the monitor module 2B establishes a hierarchy'between the correction signals SIG1, SIG2, and SIG3 delivered by the corrector networks R1, R2, and R3. In other words, at any given instant, it selects the correction signal from among the correction signals SIG1, SIG2, and SIG3 that is to be used for adjusting the command WF - OL. In this example, this selection is performed by a chain of functions of the min/max type (i.e. minimum or maximum types), which functions are applied to the output of the corrector networks in pairs. An example of such functions is described in greater detail below with reference to Figure 4. These functions serve to give precedence to one correction signal relative to another in order to adjust the command WF - OL. Preferably, precedence is given to the monitored operating parameters that comply with high setpoints, i.e. that comply with the setpoints THR2 and THR3. In other words, this means that if both a correction signal SIGl is delivered by the corrector network Rl and a correction signal SIG3 is delivered by the corrector network R3, then precedence is given to selecting the correction signal SIG3 for adjusting the command WF- O L. The fuel flow rate command that results from the adjustment step E30 is written WFCmd whether or not there is any adjustment (WFCmd=WF- O L if no adjustment is needed) . In the presently-described implementation, the generator device 2 performs a step of saturating the command WFCmd prior to delivering it to the fuel meter. This saturation depends on the command WF - OL established by the module 2A, and it is established by applying respective gains Gmin and Gmax to the command WF - OL (step E40). This saturation seeks to ensure that the command WFCmd does indeed lie between two limit values derived from the command WE - OL (these limit valueg corresponding for example to a percentage of the command WE - OL as defined by the gains Gmin and Gmax). For this purpose, the command WFCmd is saturated where appropriate to GminxWF - OL or to GmaxxWF - OL as a function of its current value, i.e. if the command WFCmd is less than GminxWF - OL, its value is forced to GminxWF - OL; on the contrary, if the command WFCmd is greater than GmaxxWF - OL, then its value is forced to the value GmaxxWE - OL. This serves to ensure in particular that no command having an aberrant (or "outlier") value is transmittedto the fuel meter (which might happen for example if the turbojet 1 has failed), or quite simply to confine the command that is transmitted to the fuel meter to within a determined range of values, e.g. corresponding to the fuel injection limits specified by the fuel meter. Where applicable, the saturated command WFCmd is then delivered to the fuel meter (step E50). There follows a more detailed description given with reference to Figure 4 of the control architecture implemented in the presently-described embodiment by the monitoring module 2B for maintaining the current values of the rate of acceleration (dN2/dt) and ofthe temperature EGT in the above-mentioned value ranges. This control architecture serves to perform steps E20, E30, and E40 as described above, which consist in monitoring the operating parameters of the turbojet 1, in adjusting the command WF - OL as generated in an open-loop by the module 2A, and in saturating the command WFCmd as delivered to the fuel meter, where necessary. In this architecture, the three corrector networks R1, R2, and R3 that are used by the module 2B for determining the appropriate correction, if any, that needs to be applied to the command WE - OL all share a common integrator I that is saturated as a function of the current value of the command WF - OL. This is possible given the transfer Sunctions C1, C2, C3 defining the networks, which functions can be written in the form of a product of a first transfer function Cl', C2', C3' implemented by a respective module 9, 10, or 11 as multiplied by a l/p integration second function that is performed by the saturated integrator I. The use of a common saturated integrator advantageously makes it possible to limit discontinuities in the flow rate setpoints delivered by the corrector networks R1, R2, and R3, and makes it easy to saturate commands coming from the closed-loop (cf. step E40). The modules 9 and 10 that implement the transfer functions C1 and C2 respectively also include respective second integrators as shown in Figure 5A (integrator element 9J) as described in greater detail below. The second integrator may suffer from problems known as "wind-up" or as "drift" (or indeed runaway), that are well known to the person skilled in the art. In order to manage these problems, the architecture shown in Figure 4 proposes activating the second integrator only when the current value of the rate of acceleration dN2/dt is close to its setpoint, in other words close to the value THRl for the module 9 or the value THR2 for the module 10. This activat.ion or deactivation of the integrators of the modules 9 and 10 is managed respectively by modules 12 and 13. More precisely: The module 12 compares the estimated difference 61 between the setpoint THRl and the current value of (dN2/dt) relative to a chosen negative threshold S1. If 61