Patent Application
20220115155
The present invention relates in general to the protection and/or monitoring of the core of a pressurized water nuclear reactor against a boiling crisis.
Maintaining the integrity of the first safety barrier (fuel cladding) is a major objective in the design of nuclear reactors, and is the subject of a safety demonstration carried out, among other things, by simulating the behavior of the reactor during families of accidental transients, using numerical calculation codes.
All this analysis is recorded in the final safety report for the facility, which is required by the nuclear safety authority to grant operating authorization for the reactor.
The integrity of the fuel cladding is guaranteed by the absence of fusion in the center of the fuel pellets, and by the absence of a boiling crisis (Departure from Nucleate Boiling, DNB) on the surface of the cladding. These conditions must be met at all points in the core.
The boiling crisis is characterized by a degradation of the thermal exchange between the heat transfer fluid of the primary circuit and the surface of the fuel cladding, which can lead to a calefaction phenomenon and a possible loss of integrity of the first barrier.
Each nuclear reactor is equipped with a protection system, whose role is to guarantee the safety of the reactor for a certain number of accident situations. The protection system typically includes a protection chain known as low CTFR (Critical Thermal Flux Ratio). In reactors with four primary loops such as the 1300 MWe, N4 and EPR, the low CTFR protection chain ensures the protection of the nuclear reactor against the boiling crisis phenomenon during category 2 accidental transients, characterized by an insertion of reactivity with moderate dynamics (uncontrolled withdrawal of control units, untimely dilution of the primary fluid leading to a decrease in the concentration of boron, untimely cooling of the primary circuit, etc.).
To this end, the protection system includes an information processing unit (called a digital integrated protection and control system or DIPC) on 1300 MWe and N4 reactors), which integrates a functional unit known as the “low CTFR algorithm”. This algorithm calculates the safety margins online with respect to the boiling crisis. To do so, it calculates the CTFR, and compares the calculated value to a limit value taking into account the reactor design uncertainties and the uncertainties in reconstruction of the CTFR value. The low CTFR protection chain causes an automatic shutdown of the reactor if the reconstructed value is below the limit.
The critical thermal flux ratio or CTFR is defined as the ratio between the critical heat flux and the local heat flux. The local heat flux corresponds to the heat flux emitted by the fuel rods through the cladding. The critical heat flux is the value of the heat flux leading to damage to the fuel rod cladding, which depends on the local thermo-hydraulic conditions of the fluid.
The low CTFR algorithm uses a simplified thermo-hydraulic model of the core, adjusted at the design stage by corrective functions (bias curves) acting on the local thermo-hydraulic quantities, in order to ensure a conservative response to the CTFR calculated by the 3D thermo-hydraulic reference code.
This simplified thermo-hydraulic modeling enables calculation of the local thermal flux from the measured quantities characterizing the power distribution of the reactor and the local thermo-hydraulic conditions allowing the evaluation of the critical thermal flux.
In addition, the low CTFR algorithm uses a predetermined critical flux correlation to calculate the critical thermal flux. This correlation is determined empirically from experimental data. The design of this correlation shows a complex and costly development. Its multidimensional polynomial-type analytical formula is postulated a priori and manually adjusted in successive steps during the development process in order to meet certain statistical criteria.
The low CTFR calculation algorithm is not very precise, so that it is necessary to allow large margins to guarantee safety during nuclear reactor operations. This is very penalizing for reactor control. In addition, although simplified, the thermo-hydraulic modeling requires sophisticated numerical processing using microprocessor systems that allow the algorithm to converge with times compatible with protection of the nuclear reactor.
In this context, the invention aims to propose a method for operating a nuclear reactor that does not have the above defects.
To this end, the invention relates to a method for operating a nuclear reactor comprising a core, the method containing the following steps:
The low CTFR algorithm has the major defect that the local thermo-hydraulic quantities are reconstructed with a simplified physical model: single channel, single phase model, without thermo-hydraulic correlation other than the critical flux correlation (pressure drop, void rate, . . . ).
The use of a deep neural network enables determination of the critical heat flux ratio as a function of the current values of the limiting quantities characterizing the operation of the nuclear reactor, much more precisely and much faster than with the low CTFR algorithm implemented today.
The values predicted by the deep neural network are much closer to the values calculated by the reference thermo-hydraulic code. The reconstruction error of the CTFR is greatly reduced. As a consequence, the operating margins of the nuclear reactor are much better.
The operating method may also have one or more of the following characteristics, considered individually or in any technically possible combination:
Y(X0)=[Πi=1i=nH+1σi·Ai]·X0
where Y is the calculated critical thermal flux ratio
According to a second aspect, the invention relates to a nuclear reactor comprising:
Other features and advantages will emerge from the detailed description, given below as an indication and in no way limiting, with reference to the appended figures, among which are the following:
The operating method shown schematically in
This nuclear reactor is typically a Pressurized Water Reactor (PWR).
Alternatively, it is intended to be used in any other suitable nuclear reactor.
The nuclear reactor 1 has a core 3.
The core 3 typically consists of a plurality of prismatic nuclear fuel assemblies arranged one against the other. Each nuclear fuel assembly comprises a plurality of nuclear fuel rods, held in position relative to each other by grids.
Each rod has a tubular cladding made of a metal such as zircaloy, with nuclear fuel pellets stacked inside the cladding.
The nuclear reactor 1 also includes a pressure vessel 5 containing a core 3 and at least one primary loop 7.
Typically, the nuclear reactor has several primary loops 7, for example three or four primary loops 7.
The or each primary loop 7 includes hot and cold branches 9, 11 connected fluidically directly to the vessel 5, a pressurizer 13 controlling the pressure in the primary loop 7, and a primary pump 15 providing circulation of primary fluid in the primary loop 7. The pressurizer 13 is generally common to the different primary loops.
Typically, the primary loop 7 also has a steam generator 17, located upstream of the primary pump 15.
The hot branch 9 fluidly connects a primary fluid outlet manifold from the vessel 5 to the primary fluid inlet water box of the steam generator 17. The pressurizer 13 is connected to hot branch 9.
An intermediate branch 19 of the primary circuit, often referred to as the “U branch”, connects the primary fluid outlet water box of the steam generator 17 to the suction of the primary pump 15. The cold branch 11 connects an outlet of the primary pump 15 to a primary fluid inlet of the tank 5.
The nuclear reactor 1 also has core reactivity control devices 21.
These control devices 21 are typically designated by the name of “control clusters” and include a plurality of rods made of a neutron-absorbing material.
The nuclear reactor 1 also includes mechanisms, not shown, configured to selectively move the control members 21 inside the vessel 5, so as to insert or extract them from core 3.
The nuclear reactor 1 also has a control system 23.
This control system 23 is typically a protection system, configured to cause an emergency shutdown of the nuclear reactor in the event of an accident situation.
To do this, the control system 23 is configured to send an automatic reactor shutdown command to the drive mechanism of the control bodies 21. Such an order causes the rapid insertion of the control bodies into the core 3 of the nuclear reactor.
Alternatively, the control system 23 is a monitoring system, configured to generate an alarm when the nuclear reactor approaches its operational limits.
According to another variant, the control system 23 acts as both a monitoring system and a protection system.
The nuclear reactor 1 is also equipped with a plurality of sensors, configured to measure a plurality of quantities characterizing the operation of the nuclear reactor 1.
Thus, the nuclear reactor 1 has temperature sensors 23, 25 measuring the temperature of the primary fluid in the hot and cold branches 9, 11 respectively.
The nuclear reactor 1 includes a sensor 27, measuring the pressure in the pressurizer 13.
The nuclear reactor also has a sensor 29 measuring the rotation speed of the primary pump 15.
The nuclear reactor 1 also has neutron flux detectors 31.
According to a first embodiment of the invention, these detectors are ex-core detectors arranged outside the core 3.
The detectors 31 are distributed around the vessel 5.
Typically, they are multi-stage ex-core chambers, with 6 sections, for example, arranged one above the other along the central axis of the vessel.
The detectors 31 are used to measure the neutron flux emitted by the core 3 at different axial levels.
The nuclear reactor 1 includes a sensor 33 of the position of the core reactivity control devices 21.
The nuclear reactor also includes an information processing unit 35, configured to implement an operating method that will be described below.
The information processing unit is, for example, an element of the control system 23. It typically consists of a processor and a memory related to the processor. Alternatively, the data processing unit is made in the form of programmable logic components, such as FPGAs (Field-Programmable Gate Array), or in the form of dedicated integrated circuits, such as ASICs (Application-Specific Integrated Circuits).
The various sensors 23, 25, 27, 29, 31, 33 provide information to the data processing unit 35.
The method for operating the nuclear reactor 1 according to the first embodiment will now be described, with reference to
The method consists of the following steps:
The acquired quantities include at least neutron flux measurements made by the neutron flux detectors 31, allowing reconstruction of the axial power distribution P(z) in core 3, as described below.
The acquired quantities also include, for the single or each primary loop 7, one or more of the following quantities: pressure p of the pressurizer, primary fluid flow Q in the primary loop, primary fluid temperature in the hot and cold branches Tbc and Tbf.
Typically, all of the above quantities are acquired.
The pressurizer pressure p and primary fluid temperatures in the hot and cold branches Tbc and Tbf are supplied directly from sensors 27, 23, 25.
The primary fluid flow rate Q is reconstructed using the speed of rotation of the primary pumps, measured by the sensor 29. It is determined by calculation, using for example a relation directly indicating the primary fluid flow rate from the rotation speed.
The acquired quantities also include the current insertion position of the core reactivity control bodies 21, provided by the sensor 33
Thus, as illustrated in
The acquisition step S10 also includes a sub-step S22, during which at least one other quantity characterizing the operation of the nuclear reactor is determined by calculation from the measured values. This typically relates to the reconstruction of the primary fluid flow Q, or the application of filters on the quantities acquired in step S20.
The step of calculation of at least one CTFR S12 comprises a sub-step S24 of determination of the entries of the deep neural network 37 using the acquired quantities, and a sub-step S25 of calculation of the at least one CTFR using the deep neural network 37 by using the entries determined in sub-step S24.
The entries of the neural network 37 include at least the axial power distribution P(z) in the core 3 of the nuclear reactor, reconstructed from the acquired neutron flux measurements.
The axial power distribution P(z) is a vector having a plurality of components, each component corresponding to a mean power level of the nuclear reactor at a given axial position. Each component defines one of the entries of the deep neural network.
The number of components depends on the degree of accuracy desired for the reconstruction of the power distribution, and thus on the desired accuracy of the CTFR calculation. For example, the axial power distribution has as many components as there are sections in each detector 31: the signals from the different detector sections define the deep neural network entries.
Alternatively, the axial power distribution has more components than the number of sections in each detector 31. A module for reconstructing the axial power distribution from the values coming from the different sections of the detectors provides from a dedicated algorithmic processing (such as a polynomial interpolation) the more finely meshed axial power distribution, each component of which defines one of the deep neural network entries.
In sub-step S24, the power of the nuclear reactor 1 at a given axial level is thus calculated from the measured neutron fluxes, using for example analytical relationships directly giving the power level as a function of the neutron flux and the thermal power released by the Pth core, reconstructed as described below.
The entries of the neural network 37 also include the thermal power released by the Pth core.
In sub-step S24, the heat power output Pth is determined by calculation, using at least the pressurizer pressure p, the primary fluid flow rate Q, and the primary fluid temperatures in the hot and cold branches Tbc and Tbf.
The calculation method is known and will not be detailed here.
Typically, a calculation of the thermal power Pth released for each primary loop 7 is performed, using the quantities acquired for said primary loop 7.
Entries to the neural network 37 also include the enthalpy elevation factor FΔH.
In sub-step S24, the enthalpy rise factor FΔH is determined by implementing the following operations:
The radial power peak factor Fxy(z) and the enthalpy rise factor FΔH are calculated using known formulas, which will not be detailed here.
Preferably, an enthalpy rise factor calculation FΔH is performed for each primary loop 7, using the released thermal power Pth calculated for said primary loop 7.
When the nuclear reactor 1 comprises multiple primary loops 7, a critical thermal flux ratio CTFR is preferably calculated in sub-step S25 for each primary loop 7, using the deep neural network 37. For this purpose, the deep neural network 37 uses the entries determined with at least the acquired quantities for the corresponding primary loop 7.
Alternatively, a single CTFR critical thermal flux ratio is calculated using the deep neural network 37. It is calculated, for example, using entries determined by any processing of the acquired quantities for the primary loops, adapted for the CTFR evaluation.
As can be seen in
This makes it possible to account for a large number of different situations in the nuclear reactor.
Each hidden layer 43 comprises at least five neurons 45, preferably at least seven neurons 45, and preferably at least ten neurons 45.
The input layer 39 has as many input neurons as there are entries.
For example, input layer 39 has one input neuron for each of the entries p, Q, Tbf, Pth, FΔH. It also has one input neuron for each of the components of the axial power distribution P(z).
For example, input layer 39 has eleven input neurons, if the axial power distribution P(z) is a six-component vector.
The output layer 41 has a single output neuron, corresponding to the minimum value of the PCR calculated for the corresponding primary loop.
Advantageously, the entries are subjected to a first post-processing before being introduced into the neural network 37 in step S22 of
The first post-processing consists, for example, in applying dynamic compensation terms, typically advance-delay filters, allowing each acquired quantity to be the image of the corresponding physical parameter. This post-processing compensates for the instrumentation response time (cold and hot branch temperature measurements of the vessel), the dropping time of the core reactivity control devices, and the response time of the processes and processing algorithms. The functional nature of information processing by the neural network 37 allows the need to use filters to be considerably reduced, compared with conventional algorithms.
The entries are preferably the subject of a second post-processing before being introduced into the neural network, by normalization of each of the quantities or components.
The entries constitute a vector, called an input vector and noted X0. Each hidden layer i of the neural network elaborates a vector of values Xi, by processing the vector of values Xi-1 from the immediately preceding layer i−1.
Thus, for each hidden layer i, i being between 1 and nH,
X
i=σi·Ai·Xi-1
where Ai is a predetermined matrix of dimensions Ni*Ni-1, Ni, where Ni is the number of neurons in layer i and Ni−1 is the number of neurons in layer i−1;
In other words, the critical thermal flux ratio is calculated by the deep neural network 37 using the following formula:
Y(X0)=[Πi=1i=nH+1σi·Ai]·X0
where Y is the calculated critical thermal flux ratio.
Each matrix Ai acts as an affine transformation on the components of the vector Xi−1. In other words:
A
i
X
i-1
=W
i
X
i-1
+b
i
where Wi is the matrix of synaptic weights that link the Ni neurons of layer i to the Ni-1 neurons of layer i−1, and bi is the vector of the Ni dimensional biases of layer i.
For the last layer we note that the matrix AnH+1 is of dimension 1×NnH and that σnH+1=1: there is no application of a non-linear transfer function to establish the Y signal of the last layer.
The non-linear operators σi each act as a non-linear function, for example sigmoid, on all the components of the AiXi−1 vector. A sigmoid function is a function expressed as follows:
where 1 is a specially parameterized constant.
In step S14, for each primary loop, a set of deviations is calculated between the critical heat flux ratio calculated for said primary loop and the plurality of predetermined reference values. Thus, as many deviation calculations are performed as there are primary loops.
These values are operating reference thresholds, called protection (or monitoring) thresholds, which incorporate the uncertainties of the method, that is, the measurement uncertainties and the calculation uncertainties.
Typically, the calculated deviations are displayed alphanumerically or graphically on at least one screen for real-time monitoring of the protection and operating margins of the unit for each primary loop.
When a single CTFR is calculated, a single set of deviations is calculated in step S14.
Step S16 is used to interpret the sets of deviations found in step S14.
To do this, a provisional control signal is elaborated for each primary loop, using the set of deviations calculated for that loop.
When the control system is a protection system, the provisional control signal is chosen from the list containing at least the following values: automatic reactor shutdown, do nothing.
When the control system is a monitoring system, the provisional control signal is chosen from the list containing at least the following values: alarm, do nothing.
When the control system is a protection and monitoring system, the temporary control signal is chosen from the list containing at least the following values: automatic reactor shutdown, alarm, do nothing.
The control signal sent to the control system of reactor 23 is elaborated using the provisional control signals of all primary loops.
After having interpreted the deviations noted in step S14, step S16 consists of applying voting logics for the signals elaborated for each loop of the reactor. For example, if the majority of the provisional control signals have the value “automatic reactor shutdown”, then the control signal sent to control system 23 is “automatic reactor shutdown”.
In step S18, the control system 23 receives the control signal developed in step 16, and acts accordingly. For example, it triggers an automatic shutdown of the reactor 1 by insertion of the reactivity control devices of the core 21 if the control signal is “automatic reactor shutdown”, or the emission of an alarm signal if the control signal elaborated has the value “alarm”.
Preferably, the method also includes a step S26 for the design of the deep neural network 37. The design step S26 consists of the following sub-steps:
Advantageously, the database includes at least 500,000 reactor states, preferably at least 1,000,000 reactor states.
Each state is defined by a set of values of the quantities characterizing the operation of the nuclear reactor, and by a critical thermal flux ratio value. The said critical thermal flux ratio value is calculated by a 3D thermo-hydraulic reference code using the corresponding set of values of the quantities characterizing the operation of the nuclear reactor.
The 3D core thermal-hydraulic calculation code is a high-precision code, for example the FLICA code. Typically, Monte-Carlo sampling techniques are used to generate the database.
The states are chosen to cover a plurality of predetermined accident situations leading to trigger the intervention of the reactor's generic low-level CTFR protection chain.
Predetermined accident situations are, for example, those referred to as Category 2 in the nuclear reactor safety report, which require the intervention of the generic low-level CTFR protection chain.
The ranges of variation of the quantities characterizing the operation of the nuclear reactor are also chosen in such a way that they cover the range of validity of the critical thermal flux correlation.
The data are generated in such a way that they optimize the detection of physical symmetry during the learning phase by the neural network.
Preferably, the input variables are normalized.
The learning sub-step S30 is performed on only a part of the database created in sub-step S28. For example, it uses between 50% and 90% of the database, typically 80%.
The learning phase consists of the following operations:
This is achieved, for example, by developing an adaptive batch stochastic gradient descent algorithm. Updating of the synaptic weight and bias tensor learning at each cycle is performed on data batches of evolving size by scanning the entire learning base, parameterized by the cycle number.
In addition, inertia methods on the adaptation step of the learning algorithm are advantageously used, in order to improve the efficiency and robustness of the learning phase.
Preferably, step S26 also includes a sub-step S32 of validation of the deep neural network determined in sub-step S30. This validation phase is performed on a second part of the database, which was not used in sub-step S30. Typically, the entire part of the database that was not used in step S30 is used in step S32. This enables validation of the deep neural network predictions developed at the end of step S30 by comparison with the predictions generated with the reference code, and to characterize the generalizability of the predictive model thus developed. Sensitivity studies are preferably carried out during sub-step S32. If the validation phase is unsatisfactory, we return to sub-step S30, then modify the neuronal structure (that is, the number of hidden layers, number of neurons per layer), or the hyperparameters of the learning algorithm, in order to restart the learning step aimed at determining the matrices of the synaptic weights and bias vectors of the neural network thus defining the previously defined Ai matrices.
Preferably, step S26 also includes a sub-step S34 of phenomenological validation (testing) of the deep neural network on database-independent data determined in sub-step S28.
During sub-step S34, the deep neural network is implemented in coupling with calculation codes simulating accidental transients. These accidental transients are typically those defined in the nuclear reactor safety report.
If this final test phase is not satisfactory, sub-step S30 is repeated.
Thus, the method described above is particularly adapted to be implemented by a reactor of the type described above.
Conversely, the nuclear reactor 1 described above is particularly adapted to implement the method just described.
In particular, as shown in
The reactor control system 23 is configured to cause an automatic shutdown of the reactor when the elaborated control signal has the “automatic reactor shutdown” value. The reactor control system 23 is configured to generate an alarm when the processed control signal has the value “alarm”.
In other words, the module 47 is configured to implement step S10 of the method described above. The module 49 is configured to carry out step S12 of the method described above. The module 51 is configured to perform step S14 of the method described above and the module 53 is configured to perform step S16 of the method described above.
A second embodiment of the invention will now be described, with reference to
In the second embodiment, the neutron flux detectors 31 are not ex-core detectors arranged outside the core 3. The neutron flux 31 detectors are in-core detectors, permanently inserted in the core 3.
They are distributed in several sets, the detectors of the same set being staggered vertically above each other.
The detectors are colons, for example, vertically arranged on rods in the core. Each electron rod allows the neutron flux to be measured at several axial heights in the core, such as six axial heights for a vertical arrangement of 6 electrons. The nuclear reactor is typically equipped with 12 rods of 6 electrons each, distributed in the core.
The thermal power Pth released is determined by calculation, using the acquired neutron flux measurements. This calculation no longer uses the pressurizer pressure p, the primary fluid flow rate Q, the temperatures of the primary fluid in the hot and cold branches Tbc, Tbf. The calculation formulas used are known and will not be detailed here.
The enthalpy rise factor FΔH is reconstructed using the axial power distribution P(z). It is not necessary to reconstruct the factor Fxy(z) from the measurement provided by the sensor 33 of the position of the reactivity control bodies of the core 21.
In step S12, a critical thermal flux ratio CTFR is calculated for each set of 31 in-core neutron flux detectors, using the deep neural network 37, with the entries determined using the neutron flux measurements made by said 31 in-core neutron flux detectors.
From the neutron flux measurements at several axial heights provided by the detector set, a set of parameters P(z), Pth and FΔH, specific for this detector set, is determined. These parameters are used as entries for the neural network 37.
Other entries are also used, such as pressurizer pressure p, primary fluid flow Q, primary fluid temperature in the cold branch Tbc, Tbf. These input quantities are the same for all sensor sets.
In the deviation calculation step S14, a set of deviations is calculated between each previously determined critical heat flux ratio and the plurality of predetermined reference values.
In the control signal generation step S16, a provisional control signal is generated for each calculated critical heat flux ratio using the corresponding set of deviations.
The control signal is elaborated using at least several of the provisional control signals according to the voting logic described above.
A third embodiment of the invention will now be briefly described. Only the points by which the third embodiment differs from the second will be detailed below. Identical elements or elements performing the same functions will be designated by the same references.
In the third embodiment, the data processing unit 35 has a continuous 3-dimensional mapping of the power in the reactor core.
This mapping is one of the quantities characterizing the operation of the reactor. The other quantities, for example Tbf, p and Q, are acquired as in the first and second modes.
Some entries of the deep neural network 37 are determined by computation, using the 3D mapping of power distribution in the core, thus replacing the quantities P(z), Pth and FΔH previously described for the first embodiment of the invention.
Advantageously, the components of the 3D power distribution constitute the entries of a dedicated neural structure, called convolution layer, which enables optimization of the extraction of characteristics of the power distribution in order to elaborate the deep neural network, and in particular to optimize its structure for a hardware implementation. The output of this convolution layer defines part of the entries of the deep neural network 37. In particular, this output replaces the vector of the components of the axial power distribution at the input of the deep neural network 37, as well as the size FΔH. This neural structure, adapted to process spatially structured data, replaces matrix multiplication by a mathematical convolution operation between the input power distribution of the structure and a series of kernels (or filters) whose degrees of freedom are adjusted during the training of step S30. The convolution layer also includes a non-linear processing of the convolution outputs through non-linear transfer functions, as well as a third step allowing to introduce invariants under local geometric transformations (such as translations). These steps can be repeated in the elaboration of the optimal neuronal structure.
The method for the exploitation of the invention may present multiple variants.
The quantities characterizing the operation of the nuclear reactor, acquired in step S10, and the entries of the deep neural network may not correspond exactly to the list described above. Other quantities may be acquired. Some quantities may not be acquired. Some entries may not be used. Other entries may be added.
In any case, the power distribution has an important impact on the location in the core and the value of the minimum margin of the CTFR.
Taking this into account improves the accuracy of the CTFR reconstruction. The quantities characterizing the operation of the nuclear reactor are not necessarily acquired in the way described above and shown in
The convolution layer described in the third embodiment could also be implemented in the first two realization modes in order to process the axial power distribution P(z) in the neuronal structure. The other entries of the neural network being unchanged.
The method described above has multiple advantages.
The response time is extremely short, for example in the millisecond range. This is obtained in particular because the method does not require the implementation of convergence loops, only the parameters specific to the deep neural network being necessary, the unit calculation operations are simple and easily parallelizable if the hardware solution allows it.
Consequently, it frees up computing capacity for other parts of the nuclear reactor control and instrumentation system.
This performance is obtained by the fact that the complexity of the underlying physics is encoded by the parameterization of the deep neural network from the database. This allows the deep neural network to reconstruct the CTFR value of the thermo-hydraulic reference code, by simply using the synaptic weights and transfer functions of the neural structure.
The CTFR computation provided by the deep neural network is very reliable, and the response is close to the one provided by the 3D thermo-hydraulic reference code (here FLICA). The conservatism of the response is adjustable.
The quality of the response provided by the deep neural network is notably due to the fact that no reconstruction of intermediate local thermo-hydraulic variables is performed without fine physical modeling. The reliability of the response provided is a function of the fineness of the mesh of the database used for learning the neural network. This only costs computing time, before commissioning in the nuclear reactor. The method is also more robust to the propagation of uncertainties on the input data (initiated by random fluctuations in the measurements from the core instrumentation) which improves the validation and hardware qualification stage.
| Number | Date | Country | Kind |
|---|---|---|---|
| 1858157 | Sep 2018 | FR | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2019/074373 | 9/12/2019 | WO | 00 |
