Method for estimating the future value of a physical quantity in an industrial system such as a nuclear reactor

Abstract

A method for estimating a quantity of a system (1) comprising the following steps: - for each variable of a plurality of variables, obtaining a sequence of successive measurements, - determining a sequence of successive values ​​of a stability parameter, the parameter being a weighted sum of the rates of change of the variables, - identifying a time interval in which the stability parameter is less than or equal to a predetermined threshold for a duration greater than or equal to a predetermined duration, - estimating a sequence of successive estimates of a particular variable, a temporal start of the sequence being included in the time interval, - comparing the sequence of estimates and the sequence of measurements to determine a calibration parameter, and - estimating the physical quantity of the system (1) using the calibration parameter. Figure for the abstract: Fig. 10

Description

Title of the invention: Method for estimating the future value of a physical quantity of an industrial system such as a nuclear reactor FIELD OF INVENTION

[0001] The invention relates to the estimation of a future state of a complex industrial system and in particular of a nuclear power plant and a nuclear reactor. STATE OF THE ART

[0002] A complex industrial system is understood here as a system whose state depends on numerous physical quantities and whose evolution is highly dependent on its initial conditions. Characterizing such a system requires the measurement or knowledge of numerous variables. The evolution of such a system also depends on numerous variables, whose evolutions may be interdependent.

[0003] Industrial systems are difficult to control in the sense that in order to maneuver them from one operating point to another, the system can become highly unstable so that it is very difficult to predict its evolution.

[0004] A nuclear reactor is an example of a complex industrial system. The physical phenomena that occur within it fall under the domains of nuclear physics, particularly neutronics, thermodynamics, particularly thermohydraulics, and thermomechanics. In a reactor, the neutron population governs the initial generation of energy, which is then transferred to a coolant fluid via various transfer mechanisms. This population and its spatial distribution vary over time according to its interaction with the surrounding matter, including the fuel and the coolant fluid. This variation is primarily a function of neutron absorption (fissile, fertile, and sterile), scattering, reflection, and leakage. It should be noted that a variation in the neutron population can alter the properties of the surrounding matter, and these variations can then have a feedback effect on the neutron population.

[0005] For example, nuclear fission following an interaction between a heavy nucleus and a neutron produces energy that can be absorbed by the coolant. The temperature of the surrounding environment is thus modified, and this temperature variation in turn modifies the nuclear reactivity in the reactor, such as the probability of absorption of a free neutron by a heavy nucleus. The phenomena involved in the reactor are of diverse natures and interact with each other. Furthermore, it should be noted that the neutron population depends on the kinetics of the processes neutron production by fission is distributed between prompt neutrons and delayed neutrons, the latter being in the minority and appearing several seconds after the prompt neutrons, being essential to allow the control of the reactor.

[0006] In the case of nuclear reactors, there is a growing need for maneuverability linked to the emergence of so-called renewable energies. Systems producing so-called renewable energies are increasingly present on the electrical grid. By their very nature, these systems are linked to climatic conditions and therefore produce intermittent and fluctuating electrical power. The electrical grid cannot function properly when it also exhibits such fluctuations. Unlike the production of so-called renewable energy systems, electricity production by nuclear power plants can be controlled. Nuclear power plants can therefore be used to complement renewable energy systems to ensure overall stability of the electrical grid resulting from a balance between electricity production and consumption.To integrate more so-called renewable energy into the grid, while adapting to demand, it becomes necessary for the operation of a nuclear reactor to be adjusted more frequently, that is to say, it must be able to be adapted over shorter and more numerous time windows than in previous practice.

[0007] There is therefore a need to better predict the evolution of a complex industrial system, and in particular that of a nuclear reactor.

[0008] In order to introduce the invention and its application to a nuclear reactor more precisely, the following additions are specified.

[0009] The invention relates in particular to the field of nuclear reactor operation in a power plant. More specifically, it is used for neutron simulation of operating nuclear reactor cores, in order to allow the operator of a nuclear unit to assess foreseeable safety margins.

[0010] By nuclear unit we mean an electricity production unit, schematically composed of a nuclear boiler, the circuits supplying the current-generating turbine, and the control system allowing an operator to control the electricity production.

[0011] Compliance with these margins, which ensures safe operation, is currently guaranteed by the control rules. The simulation tool does not replace them, but allows, while respecting the control rules, for optimized control and better adaptation to the constraints of the electrical grid. In particular, when renewable energies produce electricity on the grid, the simulation tool allows the nuclear unit to adapt more quickly, all within the framework of an overall approach to reducing CO2 emissions.

[0012] More particularly, the field of the present invention relates to the numerical simulation of complex physical phenomena as they occur during the operation of a nuclear reactor.

[0013] Furthermore, the control of a nuclear reactor requires knowledge of the present state of the reactor in order to reach a predetermined operating point using control means available to the operating operator.

[0014] However, the physical phenomena occurring within a nuclear reactor are highly complex, requiring the application of multiple branches of physics, including nuclear physics, neutronics (e.g., the laws governing the neutron population), thermohydraulics (e.g., fluid mechanics and heat transfer), thermodynamics, and thermomechanics (the effect of stresses on materials subjected to heat). These sciences are applied here to the core of nuclear reactors. As can be seen, the phenomena that occur are of diverse natures; moreover, they interact with one another. Thus, for example, fission (nuclear physics) occurs through the interaction of a heavy nucleus with neutrons (neutronics). This fission produces heat that propagates through the surrounding matter (thermal physics), which in turn transfers its heat to the water that carries it away (thermohydraulics).Here, neutronics is the central branch because it governs the generation of the aforementioned phenomena. It allows for the characterization of the neutron population, distributed spatially and temporally according to an energy spectrum that depends on the interaction with matter. These interactions are absorption (fissile, fertile, and sterile), scattering, reflection, and neutron leakage. A kinetic component must also be considered, linked to the actual production of neutrons by fission, distributed into prompt and delayed neutrons. The latter, being less numerous and appearing several seconds after the prompt neutrons, are essential for controlling a nuclear reactor.

[0015] Thus, it is clear that neutron behavior cannot be fully grasped by a nuclear reactor operator who wishes to make a decision based on easily interpretable indicators. Ultimately, the operator lacks an assessment of a reactor's stability state that would allow them to predict, based on the observed state, a neutron simulation trajectory for the operating point that is as close as possible to reality. If the initial state is incorrect, the trajectory for the operating point will also be incorrect. This is the more general problem of knowing the "initial conditions" of a physical system before applying a physical simulation model to it.

[0016] It is therefore understood that the development of renewable energies leads to a need for increased maneuverability of nuclear power plants. Responding to grid demands requires the ability to simulate

[0017] - on the one hand, quickly, for example in less than two minutes of calculation for a power transient extending over a period of ten hours and

[0018] - on the other hand, precisely the steering actions to be undertaken.

[0019] Indeed, compliance with operating and safety margins must always be ensured, and these must therefore be anticipated precisely. Furthermore, accurate simulation allows the operator to refine their piloting strategy to optimize their driving.

[0020] This optimization can take many forms, and the action of the operator is fundamental to making the choices that lead to it.

[0021] Let us take, for example, the case of an increase in power by diluting the borated water present in the core. This dilution reduces the concentration of boron, which is a neutron absorber. This decrease in boron concentration therefore automatically increases the neutron population, and thus the number of fissions, and therefore the power level.

[0022] This dilution generates water movements leading to the discharge of borated water, which constitutes an effluent, a waste that penalizes the operator, even if this waste is recoverable.

[0023] The operator can act on the boron or the absorber bars to control the reactor power. Each method has different characteristics: • Absorbent bars have an immediate effect on reactor power, but destabilize it axially and create a time oscillation which will be compensated with boron (hence production of effluents in a second stage), and above all a risk of leaving the operating range if the compensation is not sufficient. Boron has a geometrically homogeneous effect on power, but its action is slower than that of absorber rods, and takes a certain amount of time (transporting boron from its storage reservoir to the core takes about ten minutes). Each introduction of borated water into the primary circuit that supplies the nuclear core must be compensated by withdrawing an equivalent volume (since the circuit volume is constant), in other words, by producing liquid effluents. However, the axial power distribution is less disturbed than when using absorber rods, which will require fewer compensations in the future.

[0024] It is therefore clear that the choice of control method requires reconciling two technically contradictory requirements: the control of the axial power distribution and the production of liquid effluents. The phenomena involved are complex, and the use of a simulation tool is therefore essential to evaluate the axial power imbalance and the volume of effluents produced. This knowledge will allow the operator, depending on external constraints (demand for rapid power variation, financial or technical constraints related to effluent production), to choose the good compromise.

[0025] Already known solutions / existing products or processes

[0026] The operator schematically has two main means of action to control the nuclear reactor:

[0027] • absorbent bars whose insertion or removal allows the power to be modulated and the reactor temperature,

[0028] • boron diluted in water used to absorb neutrons

[0029] In order to make good use of the control means at its disposal, i.e. efficiently and in compliance with safety criteria, the operator must know the present state of the nuclear unit, and satisfactorily understand the transition to be made in order to reach the desired future state.

[0030] It is primarily the evolution of xenon in the reactor core that impacts the procedures to be followed. Xenon-135, one of the fission products, produced directly or by the decay of other fission products, is a very significant neutron poison. Its production is linked to the current state of the nuclear reactor unit, as well as its past states, and strongly impacts future states. Produced and consumed under neutron flux, xenon thus reaches an equilibrium level linked to the current power output. In the event of a power decrease, xenon production continues, but its consumption does not due to the decrease in neutron flux. This leads to an accumulation of xenon that must be compensated for in order to stabilize a new power level.The physics of xenon causes oscillatory phenomena, known as xenon oscillations, which are temporal and spatial. If these oscillations are not controlled, they can lead to a departure from the acceptable operating range of the nuclear reactor. In practical terms, this can result in an automatic core shutdown due to a failure to meet safety margins. Therefore, accurate simulation of the future state of the core is fundamental for the operation of a maneuvering nuclear reactor.

[0031] Within the reactor, the axial direction z can be defined as the vertical direction given by a plumb line. This direction z is oriented vertically upwards within the reactor. To characterize the difference between a power P1, in the upper half of the reactor core and a power P2, in the lower half of the reactor core.

[0032] In particular, it is useful to have an accurate simulation of the axial offset (which can be designated by its English name "Axial Offset", abbreviated as AO) or the axial power imbalance (which can be designated by the notations "AI" or "DPAX"). These two quantities characterize the power difference between the top and bottom of the reactor and are defined by the following formulas:

[0033] AO DPAX = V4

[0035] with P„ the nominal power of the reactor.

[0036] The term P / , + P^ = Pt denotes the total instantaneous power in the reactor. The axial offset is a relative power value with respect to the total instantaneous power.

[0037] The axial imbalance “DPAX” is a relative value of power with respect to the nominal power P„, nominal power which is not necessarily equal to the total instantaneous power Pt.

[0038] The relationship between the two quantities is as follows: DPAX = AO x = AO x Pre /

[0039] With Prei which is the relative power, expressed as a percentage of the nominal power.

[0040] This simulation ensures that temperature thresholds that could compromise the integrity of the nuclear fuel are not exceeded. The axial offset parameter cannot be determined without a simulation tool over periods of several hours by even a highly experienced operator.

[0041] This simulation is performed using digital tools (Control Support Tools, or CSTs). The CST is a digital tool that operates continuously while the nuclear reactor is running. It is fed in real time by experimental data from the sensors and by essential data calculated from the sensor measurements. Its role is to provide the operator with a real-time overview of the core's state. This state is determined by:

[0042] • to the graphical representation of experimental quantities, associated with their limits theoretical or calculated (for example, the position of group R - absorbent bars known as regulating bars - associated with its insertion limit), and

[0043] • to the graphical representation of quantities important for piloting, but not measurable. These are therefore obtained by calculation (example: AO Iodine; axial imbalance of the distribution of Iodine between the upper half of the heart and the lower half of the heart).

[0044] In the text, the terms "clusters", "groups", or "bars", in the plural or singular, will be used to designate neutron-absorbing core insertion devices.

[0045] An absorbent rod, as its name suggests, is a long, pencil-like piece of absorbent material. In the core, the rods are grouped into clusters. Each cluster is inserted into a fuel assembly. They are grouped into clusters of equivalent neutralizing weight, distributed homogeneously within the core, so that their insertion does not destabilize the radial power distribution.

[0046] The role of the simulation is also to allow the operator to prepare a transient power. For this purpose, the OAP allows for simulations. The results of these simulations are also presented graphically to the user (control diagram, group positions, boron concentration curve, etc.) to allow them to visualize the evolution of the core's state.

[0047] The process is iterative: the operator formulates hypotheses (for example, obtaining a state by placing the bars in a given position), visualizes their impact, then modifies them and restarts a simulation if necessary, until his requirements in terms of control are met (for example, minimizing the volume of liquid effluents, obtaining core stability as soon as possible, ramping up power faster...).

[0048] Most current OAPs use 0D or 1D neutronics codes. For 0D codes, access to the simulation of the axial magnitude of the AO is impossible by design. The assistance provided is therefore limited, and the power transients to be performed cannot be optimized by the simulation tool.

[0049] 1D codes provide useful axial parameters for control, but modeling a complex entity like a nuclear core as a simple "1D wire" is limited for large cores. In these reactors, because neutrons have a small mean free path compared to the overall geometry, distant parts of the core will behave in a physically decoupled manner. To correctly simulate these "decoupled sectors" of the core, it is necessary to introduce 3D modeling.

[0050] Tools equipped with 3D code exist today, but most often correspond to "core monitoring" type aids, where the emphasis is on visualizing the current state of the unit, but not on simulating the future. The specific feature of the solution we are considering is having a 3D model of the core, combined with models we have developed to both represent the state of the core in real time and accurately, and also to be able to perform predictive simulations of the future state of the nuclear reactor.

[0051] Thus, the main objective of the invention is to determine the stability state of a core that coincides with the initial state—that is, the initial conditions—of the nuclear reactor core for which one wishes to simulate the trajectory of the operating point from said initial (stable) state to a final state, the desired target. Description of the invention

[0052] One object of the invention is to propose a method for estimating the future values ​​of a physical quantity of a complex system.

[0053] The objective is achieved within the framework of the present invention by means of a method for estimating a future value of a physical quantity of an industrial system, the method comprising the following steps:

[0054] - for each variable of a plurality of variables of the system, obtaining a sequence of successive measurements of the variable, each variable being associated with a normalized variable defined as the ratio of the variable to a reference value, and normalization of the sequence of successive measurements so as to obtain a sequence of successive measurements of the normalized variable,

[0055] - determination of a system stability parameter so as to obtain a sequence of successive values ​​of the stability parameter, the stability parameter being a sum weighted by predetermined coefficients of the rates of change of the normalized variables,

[0056] - identification of the most recent stability time interval in which for At each point in the interval, the stability parameter is less than or equal to a predetermined threshold for a period greater than or equal to a predetermined period.

[0057] - estimation for a particular variable of the plurality of variables of a part of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable, a temporal start of the sequence of successive measurements corresponding to an instant within the temporal interval, the estimation being carried out by an estimator,

[0058] - comparison of the sequence of successive estimates and the part of the sequence of successive measurements to determine a value for a calibration parameter, the calibration parameter being configured to be used by the estimator to improve the estimate, and

[0059] - estimation of the future value of the physical quantity of the system by the estimator using the value of the recalibration parameter.

[0060] Such a method is advantageously and optionally complemented by the following various features taken alone or in combination:

[0061] - a step of recording over time successive measurements of variables so to obtain the sequences of successive measurements;

[0062] -the industrial system being a nuclear reactor, the plurality of physical variables includes a reactor power, an average temperature of a reactor vessel, an axial power imbalance, a concentration of a chemical species in a heat transfer fluid circulating in the reactor, the chemical species being configured to absorb neutrons in the reactor, and a position of a device configured to absorb neutrons in the reactor, the particular variable being the axial power imbalance;

[0063] - the estimation of the future value of the physical quantity of the system takes into account includes a reactor control scenario corresponding to a sequence of successive values ​​of at least one variable controllable by an operator;

[0064] - a step of estimating future values ​​of the variables of the plurality of variables of the system and a step of determining a future value of the stability parameter from the future values ​​of the variables of the plurality of variables of the system;

[0065] - a step of determining a score for the control scenario, the score being a value of a quantity chosen from among a volume of effluent produced, an average deviation from the reference axial imbalance, and an average distance to the limits of a reactor operating domain; and

[0066] - the control scenario is a first control scenario, the estimation of the future value of the physical quantity of the system being realized a second time by replacing the first scenario with a second reactor control scenario corresponding to another sequence of successive values ​​of at least one controllable variable, the process preferably including a step of comparing the scores of the first scenario and the second scenario.

[0067] The invention also relates to a computer program comprising instructions adapted to the implementation of at least one of the steps of the process as presented above when said program is executed on a computer.

[0068] The invention also relates to a device for estimating a future value of a physical quantity of an industrial system, the device being configured to implement the process as described above, the device comprising a processing unit configured for

[0069] - for each variable of the plurality of variables of the system, obtain the time series porelles of measurements of system variables, a sequence of successive measurements of the variable, each variable being associated with a normalized variable defined as the ratio of the variable to a reference value, and normalizing the sequence of successive measurements so as to obtain a sequence of successive measurements of the normalized variable,

[0070] - determine the system stability parameter so as to obtain a sequence of successive values ​​of the stability parameter, the stability parameter being a sum weighted by predetermined coefficients of the rates of change of the normalized variables,

[0071] - identify the most recent stability time interval in which for each at the point in the interval where the stability parameter is less than or equal to a predetermined threshold for a period greater than or equal to a predetermined duration

[0072] the device further comprising the estimator configured to estimate for a particular variable of the plurality of variables a part of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable, a temporal start of the sequence of successive measurements corresponding to an instant within the temporal interval,

[0073] the processing unit being configured to compare the sequence of estimates suc cessives and the portion of the sequence of successive measurements so as to determine a value for a recalibration parameter, the recalibration parameter being configured to be used by the estimator to improve the estimate,

[0074] the estimator being configured to estimate the future value of the physical quantity using the value of the recalibration parameter. DESCRIPTION OF THE FIGURES

[0075] Other features and advantages of the invention will become apparent from the following description, which is purely illustrative and not limiting, and should be read in conjunction with the accompanying drawings on which:

[0076] [Fig.1] [Fig.1] is a schematic representation of a device for estimating a future value of a physical quantity of an industrial system according to an embodiment of the invention;

[0077] [Fig.2] [Fig.2] is a schematic representation of DPAX axial imbalances over time.

[0078] [Fig.3] [Fig.3] is a schematic representation of a power during the time.

[0079] [Fig.4] [Fig.4] is a schematic representation of a temperature during the time.

[0080] [Fig. 5] [Fig. 5] is a schematic representation of a boron concentration at the course of time.

[0081] [Fig.6] [Fig.6] is a schematic representation of an axial imbalance of power over time.

[0082] [Fig.7]

[0083] [Fig. 8] Figures 7 and 8 are schematic representations of a position of dis positive neutron absorption over time.

[0084] [Fig.9] [Fig.9] is a schematic representation of a stability parameter at the course of time.

[0085] [Fig. 10] [Fig. 10] is a schematic representation of reactor variables at the course of time. DETAILED DESCRIPTION OF THE INVENTION

[0086] With reference to [Fig.1], a method and device for estimating a future value of a physical quantity of an industrial system 1 is presented.

[0087] In a first step of the process, for each variable of a plurality of variables of the system, a sequence of successive measurements of the variable is obtained.

[0088] The variables used can be denoted

[0089] [Math.l]

[0090] with an index j varying from 1 to p.

[0091] The variables can be measured by sensors 2 of system 1 so as to generate measurement sequences. The sensors 2 are connected via a connection 4 to a processing unit 6 of an estimation device 20. The connection 4 allows the measurement sequences to be transmitted to the processing unit 6.

[0092] The measurement can be the signal directly from the sensor or a value deduced from one or more of these signals.

[0093] Some of the variables can be controlled via a control center 3 of the system 1. The sequence of successive values ​​of the command can then itself constitute a measurement sequence of the controlled variable. The control center can be connected by a connection 5 to the processing unit 6 so as to transmit the sequences of commanded values ​​to the processing unit 6.

[0094] The variables of the plurality of variables are chosen for their relevance to the stability of system 1.

[0095] A sequence of successive values ​​taken by a quantity can also be referred to as a "time series." Such a sequence shows the evolution of the quantity over time. Each measurement in the sequence is time-stamped, meaning it is associated with a specific measurement time. It is therefore possible to plot a graph of the quantity's evolution over time from the sequence of successive values. The time step between two successive values ​​is preferably constant within a sequence. Two distinct variables can be measured so that their time steps are equal or different.

[0096] During the first step, the processing unit 6 performs a normalization of each sequence of successive measurements.

[0097] Each variable is associated with a normalized variable defined as the ratio of the variable to a reference value, for example, the value taken by the variable at a nominal operating point of the system. The reference value is predefined before the implementation of the process. During normalization for each variable, a sequence of successive measurements of the normalized variable is produced.

[0098] In a second step, the processing unit 6 determines a system stability parameter so as to obtain a sequence of successive values ​​of the stability parameter.

[0099] The stability parameter is a sum weighted by predetermined coefficients of the rates of change of the normalized variables.

[0100] The stability parameter, which can also be designated by the expression "Composite Stability Criterion" abbreviated as "CSC", can be written in the following form:

[0101] [Math.2]

[0102] The CCS parameter appears here as the sum of p rates of change

[0103] [Math.3] APjl Ar

[0104] each rate being weighted by a coefficient or relative weight

[0105] [Math.4]

[0106] .

[0107] The variables

[0108] [Math.5] Pj

[0109] are normalized and the coefficients

[0110] [Math.6] aj

[0111] are dimensionless, so CCS has the dimensions of the inverse of time. CCS can be expressed in "dP," which has the dimensions of the inverse of time. The unit called "dP" was named "de Penguem" in honor of Lionel de Penguem, a physicist at EDF, some of whose work concerned the stability of nuclear reactors. Due to the orders of magnitude involved, CCS is often expressed in millidP, i.e., 10³ dP or mdP.

[0112] The variable

[0113] [Math.7] Pj

[0114] varies by quantity

[0115] [Math. 8] AP,

[0116] during the time interval

[0117] [Math.9] A t

[0118] . These variations are deduced from the sequence of successive measurements of the variable

[0119] [Math. 10] Pj

[0120] .

[0121] The rates of change of two different variables are not necessarily determined over the same time intervals. This may be the case, in particular, if the two variables are not measured with the same time step. But this can This may also be the case if a characteristic time of variation of the first variable is very different from a characteristic time of variation of the second variable.

[0122] We can impose that the sum of the coefficients be equal to 1:

[0123] [Math. 11] L> / =i

[0124] The coefficients are predetermined from historical data during a preliminary data analysis phase.

[0125] More specifically, a historical database of the operation of industrial systems is available, the behavior of which is sufficiently close to the system that one wishes to simulate. The database includes, in particular, time-stamped measurements of variables

[0126] [Math. 12] Pj

[0127] of these other industrial systems. Within this basis, stable operating ranges and unstable operating ranges can be defined for each industrial system. Coefficients are then sought

[0128] [Math. 13] "I

[0129] verifying that

[0130] [Math. 14]

[0131] and a threshold "s" such that: [Math. 15]

[0132] [Math. 15] CCS = / at < s - for all stable operating ranges; and [Math. 16]

[0133] [Math. 16] CCS= fiAPi I >S *" / =1 JJ / - for all unstable operating ranges.

[0134] At the end of the preliminary data analysis phase, a threshold s – hereinafter referred to as the “predetermined threshold” – and coefficients are therefore available.

[0135] [Math. 17] has,

[0136] which are the predetermined coefficients.

[0137] In a third step, the processing unit 6 determines a most recent stability time interval in which for each point of the interval the stability parameter is less than or equal to a predetermined threshold for a duration greater than or equal to a predetermined duration.

[0138] At the end of the second step, a sequence of successive values ​​of the stability parameter is obtained. These values ​​are compared to the predetermined threshold s. From the sequence of successive values ​​of the stability parameter, initial time ranges are defined where

[0139] [Math. 18] CCS < 5

[0140] and second time slots where

[0141] [Math. 19] CCS >s

[0142] .

[0143] We determine for each instant the first time ranges where

[0144] [Math.20] CCS < 5

[0145] , the duration since the inequality

[0146] [Math.21] CCS

[0147] is verified and this duration is compared to a predetermined duration.

[0148] This makes it possible to identify within the first time ranges where

[0149] [Math.22] CCS ​​

[0150] , the time intervals of stability in which for each point of the interval the stability parameter is less than or equal to a predetermined threshold for a period greater than or equal to the predetermined period.

[0151] We then identify among these time intervals of stability the one that is the most recent.

[0152] The predetermined duration can be established from historical data during a preliminary data analysis phase. The predetermined duration can, in particular, be given by a fraction of a time characteristic of an instability phenomenon affecting the industrial system, such as the most significant or slowest-acting instability phenomenon.

[0153] In a fourth step, an estimator 7 of the estimation device 20 is used to make an estimate for a particular variable of the plurality of variables of a part of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable.

[0154] The particular variable can be chosen as the most relevant variable among the plurality of variables with respect to the stability of system 1. The estimator 7 can optionally produce an estimate of all the variables of the plurality of variables.

[0155] The estimator here estimates a history of system 1, that is, the evolution of system 1 over a time range that has already occurred, since measurements of system 1 are available for this time range. The estimator 7 provides a sequence of successive estimates of the particular variable, which can therefore be compared to the measurements of that particular variable.

[0156] The time range which is the subject of this estimation corresponds to the most recent stability time interval, and more precisely a time start of the sequence of successive measurements which is estimated is an instant included in the most recent stability time interval.

[0157] Thus, only a part of the part of the sequence of successive measurements of the particular variable is estimated by the estimator 7.

[0158] This situation assumes a communication channel 9 from the processing unit 6 to the estimator 7. Data necessary for the estimation, such as in particular the most recent stability time interval, are transmitted from the processing unit 6 to the estimator 7.

[0159] Estimator 7 is a simulation tool for system 1 that provides predictions as close as possible to reality. However, such an estimator 7 necessarily has limitations. For example, it may not take into account the piloting history, which is not necessarily known to the simulation code developer. Furthermore, estimator 7 also has limitations because certain phenomena occurring in system 1 are not taken into account in the simulation code. This is the case for phenomena that are not modeled because they are either too complex or too computationally intensive (the code must be fast enough for industrial applications), or because they are related to imperfections in experimental measurements (bias, noise, uncertainties).

[0160] In a fifth step, the processing unit 6 performs a comparison of the sequence of successive estimates produced by the estimator 7 and the part of the sequence of successive measurements.

[0161] This situation assumes a communication channel 8 from the estimator 7 to the processing unit 6. Data necessary for the comparison, such as in particular the sequence of successive estimates, are transmitted from the estimator 7 to the processing unit 6.

[0162] Processing unit 6 determines the difference between the estimate made and the measurements taken for the particular variable. Processing unit 6 is configured to de To conclude, based on this comparison and, for example, the difference, a value for a calibration parameter is obtained. The calibration parameter is configured to be used by estimator 7 to improve the estimate already made and, more generally, the estimates made for the corresponding time period.

[0163] The comparison carried out by the processing unit 6 gives information related to the simulation limits of the estimator 7, and in particular information not taken into account by the estimator 7 such as the piloting history or unmodeled phenomena.

[0164] The recalibration parameter can be of a different nature, but it is a parameter used by the estimator 7 to produce its estimates and whose value can be adjusted.

[0165] In a sixth step, the estimator 7 produces the estimate of the future value of the physical quantity of the system using the value of the recalibration parameter.

[0166] For this purpose, the recalibration parameter is transmitted from the processing unit 6 to the estimator 7. The value of the recalibration parameter is readjusted in the code of the estimator 7. In this way, the estimates produced by the estimator 7 are improved, because information not taken into account by the estimator 7 such as the piloting history or unmodeled phenomena are taken into account, at least partially, via the recalibration parameter.

[0167] The method is based on the use of a stability parameter that identifies the most recent period of stable system operation. With respect to this stability range, the accuracy of an estimate produced by the estimator is evaluated by comparing it to an actual measurement. Based on this comparison, a calibration parameter is determined which, once transmitted to the estimator, corrects subsequent estimates produced by the estimator. Estimates of a future value of a system quantity can be improved, particularly if the future value corresponds to a period of system instability.

[0168] This makes it possible to better predict the evolution of a complex industrial system, and in particular that of a nuclear reactor.

[0169] Optionally, the method includes a step of recording successive measurements of the variables over time in order to obtain the sequences of successive measurements. This step is prior to or simultaneous with the first step mentioned above.

[0170] We are now interested in the situation where the industrial system is a nuclear reactor.

[0171] A nuclear reactor is the site of a fission chain reaction, primarily of uranium-235 nuclei by neutrons. Since the neutrons emitted during fission are too energetic to produce further fissions, it is necessary to slow them down to reach lower energy levels, where the probability of producing fissions is stronger. It is the water circulating in the nuclear core that serves to slow down the neutrons (acting as a moderator), and also to transport the heat produced during fissions (acting as a heat transfer fluid).

[0172] An absorbent is a compound, introduced voluntarily, or produced by reactions in the core, which captures neutrons preventing them from producing fissions.

[0173] This could be: - boron (introduced in a diluted form into the moderator, in order to limit the chain reaction); the operator can act on the concentration of boron in the moderator; - Absorber bars: this is a device introduced into the assemblies to absorb neutrons and reduce the chain reaction; the operator can act on the absorber bars; for example on the 1300 MWe stage, an R group, for regulation, is intended for the control of the core temperature, and a power compensation group (PCG), itself made up of several subgroups of different absorption bars, allows the power produced in the core to be modulated; - Xenon 135: a very efficient neutron absorber produced during the operation of the core; variations in reactor power involve complex physics which cause spatial and temporal oscillations in xenon concentration; the operator cannot act directly on the xenon, but must implement means of controlling it when operating the reactor; - samarium 149: also a neutron absorber, but less effective than xenon: samarium accumulates, unlike xenon which decays with a period of about 9h; - of various fission products generated during the operation of the core with effects less significant than Xenon or samarium; the operator cannot act on these elements.

[0174] The chain reaction can therefore be controlled by means of the absorber bars, or clusters, and the boron. Controlling these parameters makes it possible to establish the desired conditions in the core in order to produce the desired power.

[0175] However, the actions performed by the operator must never lead, even transiently, to a state that could compromise the integrity of the core. To this end, an operating range has been predefined. This range is established in consultation with the country's security authorities.

[0176] An operating domain consists of a set of operating points; these points represent physical quantities associated with the normal operation of the reactor. The objective of a nuclear reactor pilot is to maintain the state of the nuclear unit so that all predefined operating points remain within said operating range.

[0177] This allows, by monitoring the axial power imbalance (quantities AO, DI or DPAX mentioned above), to guarantee compliance with the integrity of the fuel.

[0178] When a power transient (i.e. a request for modulation of the power produced) is requested by the electrical network, the nuclear unit concerned may maneuver in such a way as to satisfy the request, or refuse to do so for safety reasons, if its condition does not allow it in compliance with safety criteria.

[0179] The method allows for the simulation of the power transient to be performed, starting from the reactor's current state, in order to verify that the safety criteria are met. This is to avoid incorrectly refusing to perform a maneuver when conditions permit. The method also allows for refining the control strategy in order to optimize the operating phase with full knowledge.

[0180] In the case where the industrial system is a nuclear reactor, and in relation to the first step, the plurality of physical variables may include in particular the following variables:

[0181] - a reactor power,

[0182] -an average temperature of a reactor vessel,

[0183] - an axial power imbalance DPAX as already introduced pre previously,

[0184] A fourth variable is the concentration of a chemical species in the heat transfer fluid, the chemical species being configured to absorb neutrons in the reactor. For example, this chemical species is boron.

[0185] A fifth variable is the position of a device configured to absorb neutrons in the reactor. More precisely, it is the axial position of this device, which can be maintained above the zone where the neutrons circulate or lowered into this zone. The lower the device is positioned in the structure, the more neutrons it captures. This allows the power and temperature of the reactor to be modulated.

[0186] Such a device can be divided into different units which can be activated, i.e. brought down into the area where neutrons circulate, independently of each other.

[0187] A first unit may be a group of busbars forming a power compensation group (which may be abbreviated as PCG). The purpose of the PCG is to compensate for the power defect, that is to say, the anti-reactivity due to power variation. The PCG itself may consist of several subunits. of different absorption bars. The GCP consists of multiple absorbent bars that are inserted into the heart. They are not all introduced into the heart at the same time, in order to preserve the gradual nature of the absorption action. They are divided into several subgroups (G1, G2, NI, and N2), each containing several clusters of absorbent bars with different absorption levels, and which are distributed homogeneously throughout the heart to avoid radial imbalances in the neutron flux.

[0188] A second unit may be a regulating group (which may be designated by the abbreviation group R) of absorber bars. The group R is a group intended for core temperature control.

[0189] More generally, in the case where the industrial system is a nuclear reactor, the plurality of variables could also include the following variables: the average temperature of the reactor core, the temperature of the hot branch, the temperature of the cold branch, the density of the water in the core, the effective temperature of the fuel and the associated Doppler effect, the temperature of the water in the core and the associated vacuum effect, the distribution of boron in the core, the distribution of xenon 135 in the core, the distribution of samarium 149 in the core, chemical measurements, an azimuthal imbalance of power according to the radial dimension of the core (also known by the English expression "tilt").

[0190] The Doppler effect, mentioned above, corresponds to a negative feedback related to the temperature increase in nuclear reactors. In a nuclear reactor, uranium-238 is the primary site of this effect. Indeed, the neutron absorption cross-section of uranium-238 varies greatly in the epithermal region of the neutron energy spectrum, the site of numerous resonances whose amplitude varies with the fuel temperature. When the fuel temperature increases, the resonances broaden, and therefore the neutron captures that do not immediately generate fission increase, leading to a decrease in the neutron flux and thus a decrease in the number of fissions, resulting in a reduction in power output. The Doppler effect is one of the self-stabilizing effects in the case of a power transient, and therefore a temperature transient (like the moderator effect, see below), in PWRs.This is a desirable beneficial effect for the intrinsic safety of nuclear reactors. This effect has a rapid temporal kinetic; it is one of the first effects to act when a power transient occurs.

[0191] The void coefficient, mentioned above, characterizes the evolution of the reactor's reactivity in the event of a decrease in the density of the heat transfer fluid, which in this case is water. In the sense of reactor physics, reactivity measures the tendency of the reactor to increase its power (supercritical state), decrease it (subcritical state), or Maintaining a stable (critical) state. The void ratio is a key factor in the reactivity of a nuclear reactor core. Pressurized water reactors are designed so that the water acts as both a coolant and a neutron moderator (in other words, a neutron retarder), giving the reactors a negative void ratio, which is desirable. Indeed, a negative void ratio corresponds to a self-stabilizing effect of the nuclear reaction: if the neutron power increases, the density of the water decreases, which in turn decreases the density of the moderator (water is both coolant and moderator), thus reducing the probability of neutron collision by diffusion with a water molecule, resulting in less neutron retardation, a decrease in the number of fissions, and consequently, a decrease in power.

[0192] The use of the five power variables—average temperature of a reactor vessel, axial power imbalance DPAX, concentration in the coolant of a chemical species configured to absorb neutrons in the reactor, and position of a device configured to absorb neutrons in the reactor—allows for the description of the reactor's overall stability. The use of the other variables mentioned above, in addition to or instead of the five variables, allows for the description of the stability of other reactor components, such as the reactor core, the steam generator, the primary circuit, or the secondary circuit.

[0193] For the example of the Active Core, the quantities of interest to be stabilized are the three-dimensional fields of the usual neutron feedback parameters, namely the physical quantities that act on the macroscopic cross sections used in reactor physics calculations (including the calculation of the three-dimensional power in the core). It should be recalled that the cross section is a quantity whose unit is the bam (1024 cm²) measuring the capacity or reaction rate of a nucleus (microscopic cross section) or a set of nuclei (macroscopic cross section; product of the microscopic cross section and the density of the nuclei considered to be of the same element) to interact with a neutron. These interactions are absorption, fission, and scattering. For a given interaction, the larger the cross section of a nucleus, the greater the reaction rate with a neutron.The parameters in question are the following three-dimensional fields: the water density in the core, the effective fuel temperature, the water temperature in the core, the boron distribution in the core, the xenon-135 distribution in the core, and the samarium-149 distribution in the core. These quantities are normalized to a typical order of magnitude such that a homogeneous reference increment of the 3D field produces a typical CCS value (e.g., 30 millidP of stability on the axial offset). Weights derived from basic reactor core operating database analysis allow the calculation of a composite criterion. same way as the CCS previously described in relation to the stability of the reactor in general and the five variables.

[0194] The use of the stability parameter in relation to the active core allows, among other things: - to indicate in a very visual way to the operator that the observed component is "stable"; the core is in particular stable when the operator does not change anything and the characteristic quantities keep the same value over a few hours or even a few days, that is to say when there is no transient in progress, and in particular no power transient or xenon oscillation; - to use the numerical values ​​of the stability parameter to accelerate calculations, for example by increasing the time step of the core code based on stability criteria; to avoid recalculating the cross sections, a very significant calculation step in the overall core calculation; to avoid unnecessary, or even costly in terms of calculation time, printouts based on stability criteria; - to contribute to the development of a stability parameter corresponding to the entire nuclear unit; and - to indicate which component is unstable in a multi-component set...

[0195] In the case where the industrial system is a nuclear reactor, and in relation to the second step, the stability parameter or Composite Stability Criterion is then written in the form of a sum of 5 rates of variation of the 5 variables.

[0196] In the case where the industrial system is a nuclear reactor, and with regard to the third step, the predetermined duration involved in the application of the process can be given based on the phenomenon of xenon oscillation instability. Since the characteristic oscillation time is between 15 and 35 hours, a predetermined duration of between 3 and 15 hours can be chosen.

[0197] In the case where the industrial system is a nuclear reactor, and in relation to the fourth stage, the estimator 7 includes a neutronics calculation code in order to simulate the behavior of the nuclear core.

[0198] This code integrates implementations of nuclear physics equations and numerical solvers to solve them. The code uses a 3D model of the reactor to perform the calculations. This 3D aspect can correspond either to an explicit three-dimensional model of the core in the calculation code, as is the case in the process presented here, or to a two-dimensional model (generally, a core reduced to a radial plane) followed by a deployment of the results in the axial direction (we then have a 2D x 1D approach).

[0199] The physical parameters introduced are calculated by the neutronics calculation code which is more precisely a nuclear reactor physics code whose main principles are now explained.

[0200] The 3D code for nuclear reactor physics is understood here as software configured to calculate the three-dimensional distribution of power (in Watts) in a nuclear reactor core from structural data (geometry, chemical composition, heavy nucleus composition, etc.).

[0201] To do this, the software must be able, for example, to calculate the temperature distribution of the coolant in the reactor core using three-dimensional geometry. The coolant is the fluid that removes the heat produced by nuclear fission. This calculation can be performed, in particular, by a module of the code called the "thermohydraulic module".

[0202] The software must also be able to calculate the temperature distribution of the nuclear fuel. This calculation can be performed in particular by a module of the code called the "thermal module" or "thermomechanical module" if mechanical aspects are also taken into account (for example, pellet-cladding interaction).

[0203] The software must also be able to calculate the neutron flux distribution from which the power is derived. This calculation can be performed, in particular, by a module called a "neutron module".

[0204] It is the coupled interaction of these three modules above that makes it possible to calculate the 3D power in the reactor core.

[0205] Neutron physics modifies the coolant temperature, and the coolant temperature modifies the fuel temperature. The temperatures of the moderator and the fuel modify neutron physics. Indeed, fission (nuclear physics) is caused by an interaction of a heavy nucleus with neutrons (managed by the neutron module). This fission produces heat that propagates through the surrounding matter (managed by the thermal module). This matter then transfers its heat to the water, the coolant, which carries it away, raising its temperature (managed by the thermohydraulic module) and consequently modifying the fuel temperature.

[0206] Counter-reaction phenomena are then observed. Since the coolant water is also the moderator by which neutrons are slowed down to promote fission, its temperature variation will alter its density and therefore the neutron slowing, which in turn impacts the next generation of fission reactions. At the same time, a change in fuel temperature increases neutron absorption reactions, particularly in Uranium-238, which alters the neutronics.

[0207] Thus, the aforementioned physical phenomena occurring within a nuclear reactor are highly complex, requiring the use of core physics calculation codes. For example, it is possible to use a calculation code that includes: an axial 1D thermohydraulic module in the channel containing the water heat transfer fluid, - a 1D radial thermal module in the cylindrical fuel rod, and - a 3D neutronics module in neutron scattering theory.

[0208] Neutron scattering is a theoretical model used to treat a simplified form of the Boltzmann equation that governs the behavior of neutrons in matter. Descriptions of the theoretical models can be found in the French-language reference "La physique des nucléaires nucléaires, 3ème édition" (author Serge Marguet, ISBN 978-2-7430-1105-5, Lavoisier edition) and the English-language reference "The physics of nuclear reactors" (author Serge Marguet, ISBN 978-3-319-59558-7, Springer edition). Other, older reference works present the fundamental principles of nuclear reactor physics, such as the "Traité de Neutronique" by Jean Bussac and Paul Reuss, published by Hermann, ISBN 2-705-6011-9 - second edition, 1985.

[0209] To obtain satisfactory estimates of the variables, it may be useful to perform several iterations to evaluate the set at equilibrium. Once convergence is achieved, the reactor physics code can produce results necessary for the operation and safety of the reactor, providing considerable assistance in the control of a nuclear reactor core, and in particular for the location and values ​​of hot spots, an "intelligent" distribution of power, calculated responses of the various core instrumentations, control and stability of the reactor as a function of time, irradiation of nuclear fuels (evaluation of bum-ups), and the three-dimensional distribution of Xenon 135.

[0210] In the case where the industrial system is a nuclear reactor, and in relation to the fourth step, the particular variable chosen to implement the process is the axial power imbalance, which is particularly relevant for estimating the stability of the reactor.

[0211] In the case where the industrial system is a nuclear reactor, and with regard to the fifth step, the calibration parameter can be chosen as a parameter ô0 involved in defining the diffusion coefficients of fast neutrons within the reactor at the upper and lower limits along the vertical axis z of the reactor. Fast neutrons correspond to the most energetic population of neutrons generated in the reactor.

[0212] In relation to these coefficients, we define the upper limit diffusion coefficient of the reactor Di>supreffourni by the estimator during a first estimation, the lower limit diffusion coefficient of the reactor Dijinfrefourni by the estimator during the first estimation, upper and lower limit diffusion coefficients of the reactor Di>sup Di,inf corrected using the calibration parameter ô0.

[0213] The correction is based on the following relationships:

[0214] Di>sup=(2 - Ô0 - |3p(P-Pnom)) Di>supref

[0215] D1>inf=(ô0+[3p(P-Pnom)) Dljnfref

[0216] P denotes the power of the reactor and Pnom the nominal power of the reactor, the parameter [3P is a predetermined correction coefficient which depends only on the power P.

[0217] By comparing the estimated axial imbalance with the measured axial imbalance, the value of the parameter ô0 can be found which, by correcting the diffusion coefficients according to the preceding relationships, allows the estimator to be recalibrated. The recalibrated estimator then provides estimates of the axial imbalance that better match the measurements.

[0218] The reflection of neutrons in the upper and lower limits is part of the modeling limits of neutron scattering in the reactor.

[0219] The error is all the greater when the deviation from the reactor's nominal power is large. The portion of the error in the estimates that arises from this deviation from the nominal power is taken into account by the term |3p(P-Pnom), which vanishes when the deviation from the nominal power is zero.

[0220] The remainder of the error committed is taken into account by the term ô0, a term which is more related to the piloting history not taken into account in the modeling otherwise.

[0221] Figure 2 illustrates the effect of the process on the accuracy of the estimates provided. [Fig. 2] shows three curves 21, 22, and 23 of the axial imbalance DPAX over time. Curve 21 represents a measured axial imbalance DPAX. Curves 22 and 23 represent an estimated axial imbalance DPAX calculated by the estimator at t=0.

[0222] For curve 22, the estimation was carried out by the estimator using the value of a calibration parameter estimated by comparing a sequence of successive estimates and measurements of axial imbalance with a time start of the sequence of successive measurements corresponding to an instant which is not included in a time interval of stability.

[0223] For curve 23, the estimation corresponds to the process described above. It was carried out by the estimator using the value of a recalibration parameter estimated by comparing a sequence of successive estimates and measurements of the axial imbalance with a time start of the sequence of successive measurements corresponding to an instant which is included in a time interval of stability, namely the most recent time interval of stability.

[0224] The estimation curve 23 is significantly closer to the measurement curve 21 than the estimation curve 21. In practice, curve 23 provides a sufficiently accurate estimate to allow the operator to confidently determine the appropriate course of action, whereas curve 21 does not provide sufficient accuracy.

[0225] The use of the recalibration parameter determined on the basis of the time interval The most recent stability parameter therefore provides better accuracy in the estimation. This is referred to as the 'green' recalibration parameter.

[0226] Applying the process to the case of the nuclear reactor allows: - a better response rate to network demands (better accuracy in calculations making it possible not to discard power transients due to uncertainty about compliance with safety margins), - improved optimization of management strategies (greater accuracy in calculations allows for more precise anticipation of future situations and for developing more relevant strategies to best address economic or environmental constraints), and - economic gains linked to achieving cardiac stability as early as possible when necessary (trials).

[0227] The general energy context is leading to increased demand for the maneuverability of nuclear reactors. Indeed, the installed capacity of renewable energies is increasing. As these are by nature intermittent, nuclear reactors will be increasingly required to modulate their output.

[0228] Figures 3 to 9 illustrate a calculation of the stability parameter for a time range of approximately 15 hours. Each figure represents the variations of a variable over this time range so that time is plotted on the x-axis on an axis graduated in hours.

[0229] Fig. 3 represents a normalized power curve 24, the vertical axis is graduated in percentage of the nominal power.

[0230] Fig. 4 represents a temperature curve 25, the vertical axis is graduated in degrees Celsius.

[0231] Fig. 5 represents a Boron concentration curve 26, the vertical axis is graduated in ppm.

[0232] Fig. 6 represents a curve 27 of axial power imbalance, the vertical axis is graduated in percentage (DPAX is calculated as a ratio of powers, therefore unitless. It is expressed as a percentage).

[0233] Figure 7 represents a position curve 28 of a first device, which is a first unit "G1" of the GCP neutron absorption group. The vertical axis is graduated in steps corresponding to a length: one step equals 1.5 centimeters. The insertion and extraction of the rods are implemented discretely so that the rods can only assume a series of discrete axial positions. Each insertion (respectively extraction) of a rod is accomplished by moving it downwards (respectively upwards) by a fixed length of approximately 1.5 cm, which defines one step.

[0234] Figure 8 represents a position curve 29 of a second neutron absorption device, the R control group; the vertical axis is graduated in steps. The step is defined here in the same way as before.

[0235] Figure 9 represents a scatter plot of points 30 representing the stability parameter as a function of time. The stability parameter is determined from the preceding variables; the vertical axis is graduated in dP. Line 31 represents the predetermined threshold, equal here to 30 mdP.

[0236] Figure 9 shows a high volatility in the numerical value of the stability parameter as soon as actions are performed by the operator (for example, the movement of absorbent clusters from 6 hours onwards on the preceding curves). This was intentional in the design of the stability parameter. Conversely, falling below a given threshold, for a given chosen duration, ensures the stability of the reactor.

[0237] In the case presented in these figures, there is no stable time interval, that is, no time phase where the stability parameter remains below the predetermined threshold for longer than the predetermined duration. In other words, there is insufficient stability for determining the recalibration parameter. If the process is to be implemented, the recalibration parameter to be used for estimation must therefore be derived from a previous stability phase. If there has not yet been one, for example in the case of core start-up, it is possible to use a default parameter that is equivalent to no recalibration, thus allowing for accurate prediction. Indeed, the main objective of the recalibration model is to compensate for historical effects, which are, in fact, limited when the reactor has been operating for a short time. "Mandatory" stability phases are required for carrying out periodic tests.The calibration parameters will thus have a minimum range of suitable values ​​to be determined.

[0238] The stability parameter can be provided to the operator to indicate the current reactor stability level or past stability levels. Referring to [Fig. 1], a human-machine interaction device 10 displays the stability parameter calculated by the processing unit 6. A connection 11 between the human-machine interaction device 10 and the processing unit 6 allows this information to be transferred.

[0239] Figure 10 illustrates a calculation of the stability parameter over a time period of approximately 16 hours. Time is plotted on the x-axis of a graph graduated in hours. Curve 34 represents the power, curve 37 the axial power imbalance, curve 38 the position of a first neutron absorption device, and curve 39 the position of a second neutron absorption device, the regulating group R. Finally, band 40 represents the stability parameter. The gray level of the band 40 is determined as follows: - Grey level 41 corresponds to the time intervals of stability, where the stability parameter is below the threshold for a duration greater than or equal to the predetermined duration; grey level 41 corresponds to a stable state of the reactor; - Grey level 42 corresponds to intervals where the stability parameter is below the threshold but for a duration shorter than the predetermined duration; grey level 42 corresponds to a state in the process of stabilizing the reactor; and - Grey level 43 corresponds to the intervals where the stability parameter is greater than the threshold; grey level 43 corresponds to an unstable state of the reactor.

[0240] Such a display of the stability parameter on the human-machine interaction device 10 allows the operator, according to the grayscale level or the green (stable) / orange (stabilizing) / red (unstable) color code, to immediately assess the state of the slice. This can be helpful, for example, when launching certain tests requiring a sufficiently stable starting point.

[0241] In this example, a reactor maneuver is performed in the first part, to the left of the curve, corresponding to a safety parameter that most often indicates an unstable reactor state (areas with a gray level of 42). The second part, once the maneuver is completed, allows visualization of a reactor stabilizing state (areas with a gray level of 42) and then a stable reactor state. This visualization allows the operator to ensure that a planned test or a new maneuver is initiated when the reactor is in a stable state.

[0242] Estimating the future value of the physical quantity of the system can also take into account a reactor control scenario corresponding to a sequence of at least one variable controllable by an operator.

[0243] A controllable variable can be a power setpoint, the position of a device configured to absorb neutrons in the reactor, a flow rate of the heat transfer fluid, or the concentration of a chemical species in the heat transfer fluid. A power setpoint or load program is a time-dependent change in the reactor's electrical power. This change can be expressed, in particular, as a sequence of successive power values ​​to be maintained, i.e., power setpoint values, as a function of time. A power setpoint is therefore understood here as a physical quantity with the dimensions of power, and which is the power to be maintained.

[0244] The operator can define a piloting scenario, that is, the different actions he plans to carry out. These are related to controllable variables such as For example, the positions of the control bars or the boron concentration. The operator chooses, for at least one controllable variable, a sequence of successive values ​​that the variable takes over the next few hours.

[0245] The operator then enters the data characterizing the scenario into the input of the estimator 7 via a connection 12 between the estimator 7 and the human-machine interface device, as illustrated in [Fig.1].

[0246] For example, the operator can base their calculations on a first estimate of the evolution of the axial power imbalance without specifying a scenario, and on the basis of this simulation, determine a scenario and run a second estimate of the evolution of the axial power imbalance, this time taking into account the determined scenario. The first and second estimates are transmitted via a connection 13 from the estimator 7 to the human-machine interface device 10, as illustrated in [Fig. 1].

[0247] The process as presented may also include a step of estimating future values ​​of the variables of the plurality of variables of the system and a step of determining a future value of the stability parameter from the future values ​​of the variables of the plurality of variables of the system.

[0248] For this, it is each variable of the plurality of variables which is the physical quantity whose future value the process estimates.

[0249] A first advantage of the modified process is the identification of possible future reactor stability ranges. When the estimated stability parameter is less than or equal to the predetermined threshold for a predetermined period, a stability interval begins. It lasts as long as the estimated stability parameter remains less than or equal to the predetermined threshold.

[0250] The second advantage of this method is to solve the problem of the first value of the estimates being different from the corresponding measurement.

[0251] As mentioned previously, when time t=0, corresponding to the moment when the estimate is produced, does not fall within a stable time interval, there is a risk that the initial values ​​of the estimates will be significantly different from the initial measurements. To ensure that the initial values ​​of the estimates are substantially identical to the initial measurements, the procedure can be modified as follows.

[0252] The recalibration parameter for estimating the first value is the local recalibration parameter, i.e., associated with the estimation time t=0. The local recalibration parameter is determined by comparing a sequence of successive estimates and measurements of the axial imbalance with a temporal start of the sequence of successive measurements corresponding to a time just preceding time t=0.

[0253] To estimate the following values, the local calibration parameter is retained as long as that the estimated stability parameter remains significantly constant. The stability parameter used for this purpose is estimated from the estimated values ​​of the variables in the plurality of variables.

[0254] If the stability parameter varies, then the local recalibration parameter is replaced by a mixed recalibration parameter which is calculated by taking into account the local recalibration parameter and the 'green' recalibration parameter. In other words, the local recalibration parameter is based both on a comparison of a sequence of successive estimates and measurements of the axial imbalance with a temporal start of the sequence of successive measurements corresponding to a time just before time t=0, and on a comparison of a sequence of successive estimates and measurements of the axial imbalance with a temporal start of the sequence of successive measurements corresponding to an instant which is included in the most recent stability time interval.

[0255] Therefore, as the estimation iterations progress, a transition is made from the local calibration parameter to the 'green' calibration parameter when the stability parameter varies. Indeed, the local calibration parameter, adapted to a reactor situation exhibiting a certain level of instability, is no longer valid when this level of instability changes, and it is then necessary to recover the green calibration parameters. This 'level of instability' is naturally accessible by calculating the CCS.

[0256] Thanks to this modification, the process makes it possible to ensure a prospective simulation: • whose start-up is in agreement with experimental measurements • whose progress remains predictable thanks to the 'green' recalibration parameter ideally determined using CCS.

[0257] The method may further include a step for determining a score for the control scenario. This score may be a value of a quantity chosen from among a volume of effluent produced, a mean deviation from the reference axial imbalance, and a mean distance to the limits of a reactor operating domain.

[0258] Operators rely on this control diagram representing the axial imbalance as a function of the total instantaneous power. This diagram illustrates, in particular, a safety zone or operating range of the reactor, which is an area around a reference line. The reference axial imbalance corresponds to this reference line. The reactor must operate in the area close to the reference line.

[0259] The reactor's operating range, and therefore its limits, are also defined in relation to the control diagram. This operating range is a safety zone located around the reference line. It can, for example, be defined by limiting axial imbalance lines in the diagram. The average distance to the boundaries can be evaluated as the distance to one of these limiting axial imbalance lines.

[0260] The scenario score provides the operator with an assessment of the estimated scenario against effluent production criteria that will require special treatment or reactor stability criteria.

[0261] It is possible to implement an estimation of two scenarios, which the operator then compares based on their respective scores. In this case:

[0262] - the command scenario mentioned above is then a first scenario of order,

[0263] - a second scenario is obtained and transmitted as input to the process, this second scenario corresponds to a sequence of at least one variable controllable by an operator, a sequence different from that which corresponds to the first scenario, and

[0264] - the step of estimating the future value of the physical quantity of the system being carried out a second time by replacing the first scenario with the second scenario.

[0265] Advantageously, the process includes a step of comparing the scores of the first scenario and the second scenario.

[0266] It should be noted that the different curves corresponding to the different estimated scenarios can be displayed on the same graph so that they can be compared visually by the operator.

[0267] When the process includes a step of comparing the scores of the first scenario and the second scenario, the process may further provide a ranking of the scenarios in ascending order for the selected criterion or, if several criteria have been calculated, for each of these criteria.

[0268] With reference to [Fig. 1], it is the decision support device 20 which can provide this classification of scenarios whose estimates are made by estimator 7.

[0269] Once the scenario is chosen by the operator, the operator can send the corresponding instructions to the control center 3 of the system 1, for example via the connection 14 between the human-machine interface device 10 and the control center 3, as illustrated in [Fig.1].

[0270] An object of the invention is a computer program comprising instructions adapted to the implementation of at least one of the steps of the process as it has been presented so far when said program is executed on a computer.

[0271] With reference to [Fig. 1], an object of the invention is a device 20 for estimating a future value of a physical quantity of an industrial system 1, the device 20 being configured to implement the process as described so far, the device 20 comprising a processing unit 6 configured to

[0272] - for each variable of the plurality of variables of the system, obtain the time series porelles of measurements of system variables, a sequence of successive measurements of the variable, each variable being associated with a normalized variable defined as the ratio of the variable to a reference value, and normalizing the sequence of successive measurements so as to obtain a sequence of successive measurements of the normalized variable,

[0273] - determine the system stability parameter so as to obtain a sequence of successive values ​​of the stability parameter, the stability parameter being a sum weighted by predetermined coefficients of the rates of change of the normalized variables,

[0274] - identify the most recent stability time interval in which for each at the point in the interval where the stability parameter is less than or equal to a predetermined threshold for a period greater than or equal to a predetermined duration

[0275] the device further comprising the estimator 7 configured to estimate for a particular variable of the plurality of variables a part of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable, a temporal start of the sequence of successive measurements corresponding to an instant within the temporal interval,

[0276] the processing unit 6 being configured to compare the sequence of successive estimates and the portion of the sequence of successive measurements so as to determine a value of a recalibration parameter, the recalibration parameter being configured to be used by the estimator to improve the estimate,

[0277] the estimator 7 being configured to estimate the future value of the physical quantity using the value of the recalibration parameter.

Claims

Demands

1. A method for estimating a future value of a physical quantity of an industrial system (1), the method comprising the following steps implemented by an estimation device: - for each variable of a plurality of variables of the system, obtaining a sequence of successive measurements of the variable, each variable being associated with a normalized variable defined as the ratio of the variable to a reference value, and normalizing the sequence of successive measurements so as to obtain a sequence of successive measurements of the normalized variable, - determining a stability parameter of the system so as to obtain a sequence of successive values ​​of the stability parameter, the stability parameter being a sum weighted by predetermined coefficients of the rates of change of the normalized variables,- identification of the most recent stability time interval in which, for each point in the interval, the stability parameter is less than or equal to a predetermined threshold for a duration greater than or equal to a predetermined duration; - estimation, for a particular variable, of the plurality of variables in a portion of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable, a temporal start of the sequence of successive measurements corresponding to an instant within the time interval, the estimation being performed by an estimator (7) of the estimation device; - comparison of the sequence of successive estimates and the portion of the sequence of successive measurements so as to determine a value of a recalibration parameter, the recalibration parameter being configured to be used by the estimator to improve the estimation.and - estimation of the future value of the physical quantity of the system (1) by the estimator (7) using the value of the calibration parameter.

2. A method according to claim 1 further comprising a step of recording successive measurements of the variables over time so as to obtain the sequences of successive measurements.

3. A method according to any one of claims 1 to 2, wherein, the industrial system being a nuclear reactor, the plurality of physical variables includes - a reactor power, - an average temperature of a reactor vessel, - an axial power imbalance - a concentration of a chemical species in a heat transfer fluid circulating in the reactor, the chemical species being configured to absorb neutrons in the reactor, and - a position of a device configured to absorb neutrons in the reactor, the particular variable being the axial power imbalance.

4. A method according to claim 3 wherein the estimation of the future value of the physical quantity of the system takes into account a reactor control scenario corresponding to a sequence of successive values ​​of at least one variable controllable by an operator.

5. Method according to claim 4 comprising a step of estimating future values ​​of the variables of the plurality of variables of the system and a step of determining a future value of the stability parameter from the future values ​​of the variables of the plurality of variables of the system.

6. A method according to any one of claims 4 to 5 further comprising a step of determining a score of the control scenario, the score being a value of a quantity chosen from a volume of effluent produced, an average deviation from the reference axial imbalance and an average distance to limits of a reactor operating domain.

7. A method according to any one of claims 4 to 6, wherein the control scenario is a first control scenario, the estimation of the future value of the physical quantity of the system being carried out a second time by replacing the first scenario with a second reactor control scenario corresponding to another sequence of successive values ​​of at least one controllable variable, the method preferably comprising a step of comparing the scores of the first scenario and the second scenario.

8. Computer program comprising instructions adapted to carry out at least one of the steps of the process according to any one of claims 1 to 7 when said program is executed on a computer.

9. Device (20) for estimating a future value of a physical quantity of an industrial system (1), the device (20) being configured To implement the method according to any one of claims 1 to 7, the device comprising a processing unit (6) configured to - for each variable of the plurality of variables of the system, obtain the time series of measurements of variables of the system (1), a sequence of successive measurements of the variable, each variable being associated with a normalized variable defined as the ratio of the variable to a reference value, and normalize the sequence of successive measurements so as to obtain a sequence of successive measurements of the normalized variable, - determine the system stability parameter so as to obtain a sequence of successive values ​​of the stability parameter, the stability parameter being a sum weighted by predetermined coefficients of the rates of change of the normalized variables, - identify the most recent stability time interval in which, for each point in the interval, the stability parameter is less than or equal to a predetermined threshold for a period greater than or equal to a predetermined duration the device further comprising the estimator (7) configured to estimate for a particular variable of the plurality of variables a part of the sequence of successive measurements of the particular variable, so as to obtain a sequence of successive estimates of the particular variable, a temporal start of the sequence of successive measurements corresponding to an instant within the temporal interval, the processing unit being configured to compare the sequence of successive estimates and the part of the sequence of successive measurements so as to determine a value of a calibration parameter, the calibration parameter being configured to be used by the estimator (7) to improve the estimation, the estimator (7) being configured to estimate the future value of the physical quantity using the value of the calibration parameter.

Images