Na fire barrier control system for ce fr demonstration fast reactor based on new fireproof thermal insulation material

Abstract

The application relates to the technical field of nuclear reactor safety protection, and discloses a CEFR demonstration fast reactor sodium fire blocking control system based on a novel fireproof heat insulation material, which comprises a multi-modal monitoring module and a blocking decision module. The former acquires multi-level data in real time, such as thermal radiation intensity (containing first thermal field distribution data of a distributed thermocouple array, second thermal conduction feedback data of optical fiber temperature measurement), heat insulation material deformation, sodium flow diffusion path and the like; the latter generates blocking regulation and control instructions through characteristic modeling of a blocking analysis processing layer containing data cleaning and blocking efficiency modeling. The system improves the accuracy and efficiency of sodium fire blocking through multi-dimensional monitoring, intelligent modeling and dynamic decision making, guarantees the safety of the fast reactor, and is suitable for the field of nuclear reactor sodium fire protection.

Description

Technical Field

[0001] This invention relates to the field of nuclear reactor safety protection technology, specifically to a sodium fire containment control system for the CEFR demonstration fast reactor based on novel fireproof and heat-insulating materials. Background Technology

[0002] In the field of nuclear energy, fast breeder reactors (CEFR demonstration fast reactors) have become an important development direction due to their high nuclear fuel utilization and breeding capacity. However, sodium fire risk remains a key challenge restricting their safe operation. Sodium, as a coolant in fast reactors, has high chemical reactivity and is prone to violent combustion upon contact with air at high temperatures, producing high-temperature flames and corrosive products, potentially leading to serious consequences such as reactor structural damage and radioactive material leakage. Traditional sodium fire protection technologies mainly rely on passive fire-resistant materials and simple monitoring and alarm systems, which have the following significant drawbacks:

[0003] Insufficient monitoring capabilities mean that traditional monitoring methods can typically only acquire single-dimensional temperature or flame signals, failing to comprehensively capture multi-level characteristic data during the spread of sodium fires. For example, key information such as the spatial distribution of thermal radiation intensity, the deformation behavior of insulation materials at high temperatures, and the dynamic path of sodium flow diffusion are difficult to collect accurately in real time. This results in delayed and one-sided judgments on the development trend of sodium fires, failing to provide comprehensive data support for containment decisions.

[0004] The scientific rigor and dynamic adaptability of containment decisions are lacking. Existing systems mostly employ preset fixed thresholds to trigger containment measures, failing to fully consider the dynamic changes in the thermal field distribution during sodium fire combustion and the characteristic differences in different containment regions. For example, during sodium fire propagation, factors such as the attenuation patterns of thermal radiation in different regions and the impact of material deformation on containment effectiveness have not been effectively modeled. This results in a lack of specificity in the generation of containment commands, potentially leading to over-protection or under-protection, affecting containment efficiency and the safe operation of the reactor.

[0005] The data processing and analysis capabilities are weak. Traditional systems lack effective cleaning and feature extraction methods for the collected raw data, and interference from noisy data can easily lead to biased analysis results. At the same time, it is difficult to perform in-depth modeling of barrier effectiveness based on historical data, and it is impossible to achieve intelligent prediction and optimization of the sodium fire barrier process, which limits the system's adaptability and protection capabilities in complex sodium fire scenarios.

[0006] The lack of synergy between material performance and system performance is a significant issue. Existing fireproof and heat-insulating materials have limited stability and insulation performance at high temperatures, and their collaborative working mechanism with monitoring and decision-making modules is inadequate, failing to form a complete closed-loop protection system. For example, there is a lack of effective technical connections in areas such as how to adjust the barrier strategy in real time after material deformation and how to feed monitoring data back to material performance optimization, resulting in the overall protective effectiveness of the system being difficult to fully realize.

[0007] As fast reactor technology advances towards higher power density and longer lifespan, higher demands are placed on the safety, reliability, and intelligence of sodium fire containment control systems. There is an urgent need to develop an intelligent control system capable of real-time and comprehensive monitoring of sodium fire characteristics, dynamically optimizing containment strategies, and efficiently coordinating with novel fire-resistant and heat-insulating materials. This would address the problems of traditional technologies, such as limited monitoring dimensions, delayed decision-making, coarse data processing, and poor system coordination, ensuring the safe and stable operation of the CEFR demonstration fast reactor. Summary of the Invention

[0008] The purpose of this invention is to provide a sodium fire containment control system for the CEFR demonstration fast reactor based on novel fire-resistant and heat-insulating materials, in order to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a sodium fire containment control system for a CEFR demonstration fast reactor based on novel fire-resistant and heat-insulating materials, the system comprising:

[0010] The multi-modal monitoring module is used to collect multi-level barrier data in real time during the sodium fire spread process. The multi-level barrier data includes a first monitoring sequence corresponding to thermal radiation intensity data, a second monitoring sequence corresponding to thermal insulation material deformation data, and a third monitoring sequence corresponding to sodium flow diffusion path data. The thermal radiation intensity data includes a first thermal field distribution data generated by a distributed thermocouple array and a second thermal conduction feedback data collected by a fiber optic temperature measuring device.

[0011] The barrier decision module is used to perform dynamic barrier performance processing on the multi-level barrier data and input it into the barrier analysis and processing layer for feature modeling, and generate sodium fire barrier control instructions based on the output results of the barrier analysis and processing layer.

[0012] The barrier analysis and processing layer includes a data cleaning module and a barrier performance modeling module. The data cleaning module is used to divide the original barrier data stream into thermal field partitions and filter out noise data. The barrier performance modeling module is obtained by joint training based on historical heat flow data and historical deformation data of multiple sodium fire suppression cycles. The barrier performance modeling module includes a thermal radiation response layer, a barrier compensation layer and a fire decision layer connected in sequence.

[0013] Preferably, the thermal radiation response layer is used to perform spatial domain correlation processing on different monitoring sequences in the original blocking data stream to generate thermal radiation coupling feature data; the blocking compensation layer is used to model the dynamic thermal attenuation relationship between the thermal radiation coupling feature data corresponding to each monitoring sequence to generate compensated blocking field data; and the fire decision layer is used to perform multi-level integration based on the compensated blocking field data and thermal radiation coupling feature data to generate sodium fire blocking control instructions.

[0014] Preferably, the step of modeling the dynamic thermal attenuation relationship between the thermal radiation coupling characteristic data corresponding to each monitoring sequence to generate compensated barrier field data includes:

[0015] A dynamic barrier efficiency algorithm is used to identify sodium fire spread nodes in the thermal radiation coupling characteristic data, and a thermal attenuation compensation sequence corresponding to each monitoring sequence is determined based on the barrier region type corresponding to each sodium fire spread node.

[0016] Calculate the deviation coefficient between thermal field nodes in the same barrier region in the thermal attenuation compensation sequence corresponding to any two monitoring sequences, and generate compensation barrier field data between the two monitoring sequences based on the deviation coefficient.

[0017] Preferably, the calculation of the deviation coefficient between thermal field nodes in the same barrier region in the thermal attenuation compensation sequences corresponding to any two monitoring sequences includes:

[0018] When the number of thermal field nodes in the thermal attenuation compensation sequences corresponding to any two monitoring sequences is inconsistent, virtual node interpolation is performed based on the barrier region parameters corresponding to the end thermal field node in the sequence with fewer thermal field nodes, and the deviation coefficient between thermal field nodes in the same barrier region is calculated based on the interpolated data.

[0019] Preferably, the data cleaning module is specifically used for:

[0020] The first monitoring sequence, the second monitoring sequence, and the third monitoring sequence are divided into equal gradients according to the preset thermal field partitions to generate standardized first thermal field distribution data, standardized second deformation data, and standardized third sodium flow data.

[0021] A dynamic weighting method is used to correct the standardized first thermal field distribution data and the standardized second deformation data in real time, and a fixed weighting method is used to perform steady-state optimization on the standardized third sodium flow data, outputting a first optimization sequence, a second optimization sequence, and a third optimization sequence; wherein, the first optimization sequence includes the optimized first thermal field distribution data and the optimized second heat conduction feedback data.

[0022] Preferably, the data cleaning module is further used for:

[0023] Calculate the thermal attenuation coefficient of the optimized first thermal field distribution data and the optimized second thermal conduction feedback data within the historical barrier period;

[0024] Based on the thermal attenuation coefficient and the optimized first thermal field distribution data, the expected heat flow value of the optimized second thermal conduction feedback data in the real-time isolation period is predicted according to the thermal field parameters in the real-time isolation period.

[0025] Based on the optimized second heat conduction feedback data and its expected heat flow value, target sodium flow compensation data is generated, and the monitoring sequence corresponding to the target sodium flow compensation data is used as the first optimization sequence.

[0026] Preferably, the barrier compensation layer specifically includes:

[0027] The thermal attenuation analysis unit is used to perform thermal radiation path tracing on each monitoring sequence in the thermal radiation coupling feature data, so as to extract the corresponding thermal gradient attenuation chain from each monitoring sequence.

[0028] The barrier field matching unit is used to spatially map the thermal gradient attenuation chain extracted from each monitoring sequence with the corresponding thermal radiation coupling feature data to generate compensated barrier field data.

[0029] Preferably, the barrier compensation layer further includes:

[0030] A thermal hysteresis compensation unit is used to perform thermal response delay correction processing on the compensation barrier field data.

[0031] Preferably, the fire decision-making layer specifically includes:

[0032] The multi-level integration unit contains multiple barrier control nodes, and each barrier control node is connected to each monitoring sequence in the compensation barrier field data and thermal radiation coupling characteristic data through association configuration.

[0033] The dynamic weight allocation unit is used to iteratively adjust the associated configuration through a dynamic weight optimization algorithm to minimize the deviation between the sodium fire barrier control command and the actual thermal field distribution.

[0034] The thermal imbalance positioning unit is used to locate the sodium fire anomaly area based on the compensation barrier field data and thermal radiation coupling characteristic data, and generate sodium fire barrier control commands.

[0035] Preferably, the dynamic weight allocation method specifically includes:

[0036] Adaptive barrier parameters are generated based on real-time sodium fire propagation characteristics.

[0037] The standardized first thermal field distribution data is segmented and optimized using a sliding window mechanism.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] In terms of monitoring capabilities, the multimodal monitoring module utilizes a distributed thermocouple array and fiber optic temperature measurement device to achieve multi-dimensional acquisition of thermal radiation intensity data. Simultaneously, by combining data on insulation material deformation and sodium flow diffusion paths, it constructs a multi-level data acquisition system encompassing first, second, and third monitoring sequences. This comprehensive monitoring mode can capture the dynamic characteristics of thermal field distribution, material deformation, and sodium flow movement during sodium fire propagation in real time, providing a rich and accurate data foundation for containment decisions. For example, the combination of first thermal field distribution data and second thermal conduction feedback data can accurately reflect the spatial distribution and conduction characteristics of thermal radiation, while sodium flow diffusion path data helps predict the direction of sodium fire propagation in advance, providing a forward-looking basis for the formulation of containment strategies.

[0040] At the data processing and modeling level, the design of the barrier analysis processing layer is highly innovative. The data cleaning module standardizes and optimizes the raw data through pre-defined gradient partitioning of the thermal field and dynamic / fixed weight allocation methods, effectively filtering out noise and improving data quality. For example, dynamic weighting is used to correct the first thermal field distribution data and the second deformation data in real time, which can adapt to the dynamic changes in the sodium fire combustion process, while the steady-state optimization of the third sodium flow data ensures its reliability in stable scenarios. The barrier effectiveness modeling module is jointly trained based on historical data. Through hierarchical processing of the thermal radiation response layer, barrier compensation layer, and fire decision layer, it achieves in-depth modeling of sodium fire barrier effectiveness. The spatial domain correlation processing of the thermal radiation response layer generates thermal radiation coupling feature data, the modeling of dynamic thermal attenuation relationships in the barrier compensation layer generates compensated barrier field data, and the multi-level integration of the fire decision layer ensures the scientific nature and pertinence of control commands. This hierarchical modeling mechanism can fully explore the potential patterns in the data and improve the system's adaptability and predictive ability to complex sodium fire scenarios.

[0041] The dynamic and precise nature of the blocking decision-making is a significant advantage of this invention. The blocking decision-making module, through dynamic processing and feature modeling of multi-level blocking data, can generate precise sodium fire blocking control commands based on real-time monitoring data. For example, in the calculation of the thermal attenuation compensation sequence, virtual node interpolation is used to address the inconsistency in the number of thermal field nodes, ensuring the accuracy of the deviation coefficient calculation and thus generating reliable compensation blocking field data. The dynamic weight allocation method generates adaptive blocking parameters based on real-time sodium fire spread characteristics and optimizes the data segmentally through a sliding window mechanism, enabling the system to dynamically adjust the blocking strategy according to the development stage of the sodium fire, achieving a shift from "passive response" to "active prediction." The dynamic weight optimization algorithm of the fire decision-making layer minimizes the deviation between the control commands and the actual thermal field distribution by iteratively adjusting the associated configuration, further improving the accuracy and effectiveness of the decision-making.

[0042] The system's coordination and intelligence levels have been significantly improved. A complete closed-loop collaborative mechanism has been formed between the multimodal monitoring module, the barrier decision-making module, and the new fireproof and heat-insulating materials. Monitoring data provides the basis for decision-making, decision commands drive the execution of barrier measures, and the performance of the new materials provides physical support for the barrier effect. For example, the thermal hysteresis compensation unit of the barrier compensation layer corrects the response delay of the compensation barrier field data, ensuring the synchronization of commands and actual thermal field changes; the thermal imbalance positioning unit can quickly locate the sodium fire anomaly area, realizing precise handling of local fire hazards. In addition, based on joint training and dynamic modeling of historical data, the system has self-learning and self-optimization capabilities, and can continuously improve barrier effectiveness with the accumulation of operating time, adapting to the sodium fire protection needs under different working conditions.

[0043] Through the aforementioned technological innovations, this invention effectively solves the problems of single monitoring dimensions, delayed decision-making, crude data processing, and poor system coordination in traditional sodium fire protection systems. It constructs an advanced protection system that integrates real-time monitoring, intelligent analysis, dynamic decision-making, and efficient isolation, providing a solid technical guarantee for the safe operation of the CEFR demonstration fast reactor. It has significant engineering application value and far-reaching technological innovation significance for improving the safety and reliability of fast reactor technology. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the working principle of the sodium fire containment control system for the CEFR demonstration fast reactor based on novel fireproof and heat-insulating materials described in this invention.

[0045] Figure 2 A flowchart of a multi-level collaborative execution system for sodium fire barrier control commands;

[0046] Figure 3 A flowchart for generating data to compensate for the barrier field;

[0047] Figure 4 A flowchart for calculating the thermal field nodal deviation coefficient;

[0048] Figure 5 This is a flowchart of the data cleaning module. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] Please see Figures 1-5The present invention relates to a sodium fire containment control system for a CEFR demonstration fast reactor based on a novel fire-resistant and heat-insulating material, the specific implementation steps of which are as follows:

[0051] The system comprises a multimodal monitoring module and a barrier decision module. The multimodal monitoring module collects multi-level barrier data in real time during the sodium fire spread process. This multi-level barrier data includes a first monitoring sequence corresponding to thermal radiation intensity data, a second monitoring sequence corresponding to insulation material deformation data, and a third monitoring sequence corresponding to sodium flow diffusion path data. The thermal radiation intensity data includes first thermal field distribution data generated by a distributed thermocouple array and second thermal conduction feedback data collected by a fiber optic temperature measurement device. The barrier decision module performs dynamic barrier effectiveness processing on the multi-level barrier data and inputs it into a barrier analysis processing layer for feature modeling. Based on the output of the barrier analysis processing layer, it generates sodium fire barrier control commands. The barrier analysis processing layer includes a data cleaning module and a barrier effectiveness modeling module. The data cleaning module partitions the original barrier data stream into thermal fields and filters out noise data. The barrier effectiveness modeling module is jointly trained based on historical heat flow data and historical deformation data from multiple sodium fire suppression cycles. The barrier effectiveness modeling module includes a thermal radiation response layer, a barrier compensation layer, and a fire decision layer connected in sequence.

[0052] The present invention will be further described below with reference to Examples 1 to 5:

[0053] Example 1: In the specific implementation of the multimodal monitoring module, a distributed thermocouple array covers the surface of the fireproof and heat-insulating material and key structural nodes around the sodium pool of the fast reactor in a grid-like layout. The spacing between the thermocouple nodes is set to 50-100 cm, forming a dense monitoring network for areas prone to sodium fires. Each thermocouple node is connected to the data processing unit via a wired transmission link to transmit temperature signals in real time. After analog-to-digital conversion, the signals are converted into a two-dimensional matrix to generate the first thermal field distribution data. Each element in the matrix corresponds to the real-time temperature value and three-dimensional spatial coordinates of a monitoring point, which can intuitively present the spatial distribution of the thermal field.

[0054] The fiber optic temperature measurement device employs distributed fiber optic sensors, laid along the heat conduction-sensitive paths such as the outer wall of the sodium pipe and the joints of the insulation layer. The sensors excite backscattered Raman light in the fiber by emitting laser pulses, and calculate the temperature value at each monitoring point based on the frequency shift of the scattered light. The device generates a second heat conduction feedback data output in the form of a continuous curve, accurately reflecting the temperature gradient changes along the fiber path and enabling continuous monitoring of the heat conduction process.

[0055] The second monitoring sequence relies on miniature deformation sensors, including strain gauges and displacement gauges, deployed at key stress-bearing locations within the insulation material. Strain gauges are attached to the surface of the fiber reinforcement layer of the insulation material to monitor the linear strain generated under thermal stress; displacement gauges are installed at the joints of the insulation material to monitor the relative displacement between adjacent panels. Both types of sensors are set to a sampling frequency of 100Hz to acquire deformation data in real time. After high-frequency noise is removed by a hardware filtering circuit, time-series data containing parameters such as strain value, displacement, and deformation rate are generated, reflecting the real-time deformation state of the insulation material.

[0056] The third monitoring sequence was acquired through a sodium flow diffusion path monitoring system, which consists of a flow sensor, a level sensor, and an image acquisition device. An electromagnetic flowmeter, installed at the outlet of the simulated sodium leakage pipeline, measures the volumetric flow rate of sodium in real time. Its measurement principle is based on Faraday's law of electromagnetic induction, providing high-precision flow data. Static pressure level sensors are distributed at different depths in the trench, calculating the diffusion height of sodium on the ground by measuring the hydrostatic pressure of the liquid. The sensor's measurement accuracy is ±1 mm. A high-temperature resistant camera captures video images of the sodium flow diffusion area at a rate of 25 frames per second. Image processing algorithms are used to analyze the video stream in real time, identifying the sodium flow front position and diffusion area. Combined with flow and level data, a three-dimensional dynamic model of the sodium flow diffusion path is constructed. This model is dynamically updated over time, visually displaying the diffusion trend of the sodium flow.

[0057] In terms of data transmission and synchronization, the various sensors in the multimodal monitoring module achieve time synchronization of data acquisition through a unified time synchronization protocol (such as the IEEE 1588 precision clock protocol), ensuring that different types of monitoring data have consistent timestamps. All monitoring data is transmitted to the input interface of the isolation decision module via industrial Ethernet. During transmission, the TCP / IP protocol is used to ensure data reliability, while a VPN encrypted channel is used to prevent data leakage. After the data arrives, it first enters a buffer queue, waiting for the isolation decision module to call and process it. The design of the entire multimodal monitoring module ensures comprehensive, real-time, and accurate monitoring of key parameters such as thermal radiation, material deformation, and sodium flow diffusion during sodium fire propagation, providing a rich and reliable data foundation for subsequent isolation decisions.

[0058] Example 2: The data cleaning module of the barrier analysis processing layer performs multi-stage processing on the raw barrier data stream. First, thermal field partitioning is performed. Based on preset temperature thresholds, the raw data corresponding to the first monitoring sequence (thermal radiation intensity data), the second monitoring sequence (insulation material deformation data), and the third monitoring sequence (sodium flow diffusion path data) are divided into three gradient intervals: high temperature zone (temperature ≥ 800℃), medium temperature zone (500℃ ≤ temperature < 800℃), and low temperature zone (temperature < 500℃). Differentiated filtering strategies are adopted for the data characteristics of different zones: in the high temperature zone, a 5×5 pixel sliding window is used for median filtering of the first thermal field distribution data to effectively remove abnormal temperature peaks caused by sensor malfunctions or external interference; in the medium and low temperature zones, a mean filtering algorithm is used to smooth the temperature field distribution and reduce the impact of random noise on the thermal field trend analysis. For the second deformation data, by setting physical thresholds for material elastic deformation (e.g., strain ≤ 0.3%, displacement ≤ 2 mm), abnormal data points exceeding the thresholds are automatically identified and removed. Subsequently, a cubic spline interpolation algorithm is used to fill in any missing points that may occur during data acquisition, ensuring the continuity and integrity of the deformation data. For the third sodium flow data, a benchmark model is established based on historical data accumulated under normal operating conditions of the fast reactor. Through comparative analysis of real-time data and the benchmark model, statistical testing methods (such as Z-test) are used to identify and remove noise signals caused by factors such as equipment vibration and electromagnetic interference, ensuring the reliability of the sodium flow diffusion path data.

[0059] A dynamic weighting method is applied to the real-time correction of the first thermal field distribution data and the second deformation data. The system generates adaptive barrier parameters by real-time monitoring of sodium fire spread characteristics (such as the rate of change of thermal radiation intensity and the deformation rate of the insulation material). Specifically, when the change in thermal radiation intensity per unit time exceeds a preset threshold, the system automatically adjusts the weight coefficients of the thermal field data and the deformation data: if the rate of increase in thermal radiation intensity accelerates, the weight coefficient of the thermal field data increases accordingly, while the weight coefficient of the deformation data decreases proportionally, and vice versa. The weight adjustment is achieved through a linear interpolation algorithm. For example, if the initial weight coefficients are 0.6 for thermal field data and 0.4 for deformation data, when the rate of change of thermal radiation intensity reaches +200℃ / 5min, the weight of the thermal field data increases to 0.8, and the weight of the deformation data decreases to 0.2. For the third sodium flow data, a fixed weight allocation strategy was adopted, setting the weights of the three parameters, flow rate, liquid level, and diffusion area, to 0.5, 0.3, and 0.2, respectively. A steady-state optimized sodium flow data sequence was generated through a weighted average algorithm. This strategy ensured the stability and repeatability of sodium flow diffusion path monitoring.

[0060] The data cleaning module also performs calculations of the thermal attenuation coefficient and predictions of the expected heat flow. By analyzing the time series of the first thermal field distribution data and the second thermal conduction feedback data in historical isolation periods, the thermal attenuation coefficient is obtained using the least squares method. This coefficient characterizes the propagation characteristics of thermal radiation in fireproof and heat-insulating materials (such as the attenuation of thermal radiation intensity per unit distance). In the real-time isolation period, the system calculates the expected heat flow value of the second thermal conduction feedback data using a linear extrapolation method based on the thermal field parameters (such as the highest temperature and temperature gradient distribution) of the current first thermal field distribution data and the historical thermal attenuation coefficient. If there is a deviation between the actual collected thermal conduction feedback data and the expected value, the system generates target sodium flow compensation data based on the magnitude and direction of the deviation. This data includes the correction amount and the corresponding time-space coordinate information, and is input into the subsequent processing module as supplementary content of the first optimization sequence to improve the synergy between thermal field data and deformation data.

[0061] The entire data cleaning process adopts a layered processing architecture: first, the quality of the original data is improved through partitioning and noise filtering; then, dynamic and fixed weight allocation are used to optimize and correct different types of data; finally, thermal decay analysis and expected value prediction are used to enhance the correlation between data. This module's design ensures that the data input to the barrier effectiveness modeling module has high accuracy, consistency, and synergy, providing a reliable data foundation for subsequent feature modeling and barrier decision-making.

[0062] Example 3: The thermal radiation response layer of the barrier performance modeling module performs spatial domain correlation processing on multimodal monitoring data. Specifically, the temperature matrix of the first thermal field distribution data, the strain-displacement matrix of the second deformation data, and the diffusion path model of the third sodium flow data are first uniformly mapped to a three-dimensional coordinate system. A spatial reference system containing X, Y, and Z axes is established with the geometric center of the fast reactor sodium pool as the origin. Through a coordinate transformation algorithm, the data points of each monitoring sequence are converted into coordinate values ​​with a unified spatial reference (e.g., metric units), ensuring that the data of different physical quantities correspond one-to-one in space. Subsequently, the spatial distance and temporal correlation of data points in each monitoring sequence are calculated: for any two data points (e.g., the temperature value of a thermocouple node and the strain value of an adjacent strain gauge), the spatial distance is calculated using the Euclidean distance formula, and the correlation of the time series is calculated using the Pearson correlation coefficient, generating thermal radiation coupling characteristic data. This data is stored in the form of multidimensional vectors, each containing correlation indices (e.g., correlation coefficient, covariance value) of physical quantities such as temperature, strain, and sodium flow velocity, characterizing the spatial coupling relationship and dynamic correlation characteristics between different monitoring sequences.

[0063] The thermal attenuation analysis unit of the barrier compensation layer performs thermal radiation path tracing for each monitoring sequence. Taking the first thermal field distribution data as an example, starting from the core node of the high-temperature zone (i.e., the node with the highest temperature value), it searches for adjacent nodes according to the direction of temperature gradient descent, constructing a tree-like thermal radiation propagation path. Each path node records the temperature value, propagation time, and path branch information, forming a thermal gradient attenuation chain. For example, a core node of a high-temperature zone (temperature 1000℃, coordinates X1, Y1, Z1) propagates thermal radiation to four adjacent nodes. The temperatures of each branch node are 900℃, 850℃, 880℃, and 920℃, respectively, and are recorded as the second layer nodes of the path, and so on until the temperature is below the low-temperature zone threshold. For the second deformation data, starting from the node with the maximum strain value, it traces the strain propagation path, establishes a deformation gradient attenuation chain, and records the strain value, deformation rate, and propagation direction of each node.

[0064] The barrier field matching unit spatially maps the thermal gradient attenuation chain of each monitoring sequence with the corresponding thermal radiation coupling characteristic data. Specifically, the coordinates of each node in the thermal gradient attenuation chain are matched with the spatial coordinates in the thermal radiation coupling characteristic data. A three-dimensional interpolation algorithm (such as trilinear interpolation) is used to map the node values ​​(temperature, strain, etc.) of the attenuation chain to a unified three-dimensional spatial grid. Each grid cell (voxel) contains real-time values ​​of multiple physical quantities (such as temperature, strain, and sodium flow velocity) and spatial location information, forming compensated barrier field data. This data is stored in volumetric data form and can be presented through a visualization interface as a three-dimensional field distribution including temperature contour maps, strain vector fields, and sodium flow streamlines, intuitively demonstrating the coupling relationship between thermal radiation propagation and material deformation and sodium flow diffusion.

[0065] The thermal hysteresis compensation unit of the barrier compensation layer performs time correction on the data of the compensated barrier field, taking into account the delay characteristics of sensor signal transmission and data processing. The system pre-measures the delay times of each monitoring sequence experimentally: the distributed thermocouple signal has a transmission delay of approximately 50 milliseconds due to cable length and analog-to-digital conversion; the fiber optic temperature measurement signal has a delay of approximately 100 milliseconds due to laser pulse emission and scattered light acquisition; and the signal delays of the deformation sensor and sodium flow monitoring device are approximately 30 milliseconds and 80 milliseconds, respectively. Based on the measured delay times of each monitoring sequence, the thermal hysteresis compensation unit performs time shifting processing on the compensated barrier field data, that is, it backscales the data collected at the current moment by the corresponding delay time, ensuring that the data accurately reflects the real-time state of the physical field. For example, for thermocouple data with a delay of 50 milliseconds, the data at time t is considered as the thermal field state at time t-50ms, ensuring that all physical quantities are synchronized in the time dimension.

[0066] The entire processing flow of the barrier compensation layer realizes the transformation from single-sequence data to multi-physics coupled data: the propagation characteristics of each monitoring sequence are extracted through thermal radiation path tracing, a unified three-dimensional field model is constructed using spatial mapping, and time delay correction ensures the real-time nature of the data. The compensation barrier field data generated by this module provides the fire decision-making layer with comprehensive information including spatial distribution, temporal evolution, and the coupling relationship of multiple physical quantities, laying the foundation for the generation of sodium fire barrier control commands.

[0067] Example 4: The multi-level integrated unit of the fire decision-making layer consists of multiple barrier control nodes. Each node corresponds to a specific fireproof and heat-insulating area around the fast reactor sodium pool, such as the sodium pump room, heat exchanger compartment, and pipeline corridor. Each node is connected to the monitoring sequences in the compensation barrier field data and thermal radiation coupling characteristic data through a preset association configuration, forming a "data input-command output" mapping relationship. Taking the sodium pump room node as an example, its input data includes the temperature data of the distributed thermocouples in the area, the deformation data of the strain gauges, the sodium flow rate data of the electromagnetic flowmeter, and the diffusion area data of the camera. The output command corresponds to starting the cooling spray system in the area, adjusting the pneumatic support device of the heat insulation material, or triggering the audible and visual alarm device. The association configuration of each node is predefined through a topology table, which clarifies the data input port, processing logic, and control signal interface of the actuator.

[0068] The dynamic weight allocation unit iteratively adjusts the input weights in the associated configuration using a dynamic weight optimization algorithm. This algorithm uses the deviation between the actual thermal field distribution after the execution of the sodium-fire barrier control command and the expected effect of the command as the optimization objective, and updates the input weight coefficients of each node using the gradient descent method. The specific process is as follows: First, based on the current compensation barrier field data and thermal radiation coupling characteristic data, the initial output command for each node is calculated (e.g., a node's command is to reduce the regional temperature to below 600℃). Then, within a preset time window after command execution (e.g., 10 minutes), actual thermal field data is collected and the deviation from the expected effect is calculated (e.g., if the actual temperature drops to 650℃, the deviation is +50℃). Next, based on the sign and magnitude of the deviation, the weight coefficients of the node's input data are adjusted—if the actual temperature is higher than expected, the weight of the thermal field data is increased to strengthen the priority of temperature control, and the weight of the deformation data is reduced to weaken the interference of secondary factors; otherwise, the adjustment is reversed. The weight adjustment step size is automatically set by the algorithm, typically a decimal between 0.01 and 0.1, ensuring the smoothness of the adjustment process.

[0069] The thermal imbalance location unit, based on compensation barrier field data and thermal radiation coupling characteristic data, locates sodium fire anomaly areas through multi-dimensional threshold comparison and cluster analysis. First, normal operating condition thresholds are set for each physical quantity (e.g., temperature ≤ 800℃, strain ≤ 0.3%, sodium flow velocity ≤ 0.1m / s). When any physical quantity in a region exceeds the threshold, it is marked as a potential anomaly point. Then, the DBSCAN density clustering algorithm is used to spatially cluster the potential anomaly points, aggregating adjacent points with reachable density into anomaly regions. A thermal imbalance index is calculated for each anomaly region. This index is a weighted sum of parameters such as temperature deviation, strain rate, and sodium flow diffusion velocity, with weights pre-set based on the degree of influence of each parameter on sodium fire barrier (e.g., temperature deviation weight 0.5, strain rate weight 0.3, sodium flow velocity weight 0.2). Based on the magnitude of the thermal imbalance index, the anomaly regions are divided into three levels: low risk (index ≤ 50), medium risk (50 < index ≤ 100), and high risk (index > 100). Corresponding sodium fire containment control instructions are generated for different risk levels: low-risk areas trigger early warning signals and increase monitoring frequency; medium-risk areas initiate local cooling measures (such as turning on fans or spray devices in the area); and high-risk areas trigger emergency containment measures, including releasing fire extinguishing dry powder, closing area isolation gates, or activating the backup insulation layer deployment mechanism.

[0070] The entire fire control decision-making process embodies the characteristics of hierarchical decision-making and dynamic optimization: multi-level integration units achieve precise regional control, dynamic weight allocation units improve the accuracy of command generation through feedback mechanisms, and thermal imbalance location units achieve rapid identification and graded response of abnormal areas through data-driven methods. The collaborative work of these functional units ensures that the system can generate targeted and timely sodium fire containment control commands based on real-time monitoring data, providing efficient decision support for fast reactor sodium fire prevention and control.

[0071] Example 5: Adaptive barrier parameter generation based on fuzzy logic control principle in dynamic weight allocation method. The system collects characteristic parameters of sodium fire propagation in real time, including the rate of change of thermal radiation intensity. (Unit: ℃ / s) Maximum strain rate of thermal insulation material (unit: Sodium diffusion rate (Unit: m / s). These parameters, after normalization, are input into the fuzzy logic controller. The controller's output is the weighting adjustment coefficients for the thermal field data and deformation data. The input and output variables of the fuzzy controller are both defined within the universe of discourse [-1, 1], and are divided into three fuzzy sets: "low," "medium," and "high," using a triangular membership function. For example, when... Belongs to the "high" fuzzy set Belongs to the "middle" fuzzy set When it belongs to the "high" fuzzy set, output according to the preset fuzzy rule table. That is, the weight of thermal field data increases by 0.2, and the weight of deformation data decreases by 0.2 accordingly.

[0072] The sliding window mechanism is applied to the piecewise optimization processing of standardized first thermal field distribution data. The window length is set to... Minutes, sliding interval is Minutes, forming overlapping time window sequences ( Within each window, the thermal field data matrix... ( For the number of rows, Perform principal component analysis (PCA) on the column number to calculate the covariance matrix. Extract the first three principal components As a feature vector, the angle between the feature vectors of the current window and the previous window is compared. ( To determine the trend of thermal field distribution. If Exceeding the preset threshold If a significant shift occurs at the center of the thermal field, then a weighted average is applied to the data within the window. The weighting coefficients are proportional to the contribution rate of the eigenvectors. ( , (Eigenvalues ​​corresponding to principal components) to enhance the data weight of regions of change.

[0073] In the steady-state optimization of the third sodium flow data in the data cleaning module, the fixed weight allocation method is implemented through a linear combination formula:

[0074]

[0075] in, The optimized sodium flow data sequence, The volumetric flow rate data measured by the electromagnetic flow meter. The diffusion height data is obtained from a hydrostatic level sensor. This refers to the diffusion area data identified by the image processing algorithm. In this formula, the flow rate parameter... The liquid level parameter is assigned the highest weight (0.5) because it directly reflects the sodium leakage rate. Spatial planning plays a key role in barrier measures, hence its weight is second (0.3), and the diffusion area parameter... The system with the greatest environmental interference has the lowest weight (0.2). Through this fixed weight combination, the system can stably output the third optimized sequence that characterizes the overall state of sodium flow diffusion, providing reference data with continuous time sequence and clear weight logic for blocking decisions.

[0076] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A sodium fire containment control system for a CEFR demonstration fast reactor based on a novel fire-resistant and heat-insulating material, characterized in that, include: The multi-modal monitoring module is used to collect multi-level barrier data in real time during the sodium fire spread process. The multi-level barrier data includes a first monitoring sequence corresponding to thermal radiation intensity data, a second monitoring sequence corresponding to thermal insulation material deformation data, and a third monitoring sequence corresponding to sodium flow diffusion path data. The thermal radiation intensity data includes a first thermal field distribution data generated by a distributed thermocouple array and a second thermal conduction feedback data collected by a fiber optic temperature measuring device. The barrier decision module is used to perform dynamic barrier performance processing on the multi-level barrier data and input it into the barrier analysis and processing layer for feature modeling, and generate sodium fire barrier control instructions based on the output results of the barrier analysis and processing layer. The barrier analysis and processing layer includes a data cleaning module and a barrier performance modeling module. The data cleaning module is used to divide the original barrier data stream into thermal field partitions and filter out noise data. The barrier performance modeling module is obtained by joint training based on historical heat flow data and historical deformation data of multiple sodium fire suppression cycles. The barrier performance modeling module includes a thermal radiation response layer, a barrier compensation layer, and a fire decision layer connected in sequence.

2. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 1, characterized in that, The thermal radiation response layer is used to perform spatial domain correlation processing on different monitoring sequences in the original blocked data stream to generate thermal radiation coupling feature data; the blocking compensation layer is used to model the dynamic thermal attenuation relationship between the thermal radiation coupling feature data corresponding to each monitoring sequence to generate compensated blocking field data. The fire decision-making layer is used to perform multi-level integration based on compensation barrier field data and thermal radiation coupling characteristic data to generate sodium fire barrier control commands.

3. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 2, characterized in that, The process of modeling the dynamic thermal attenuation relationship between the thermal radiation coupling characteristic data corresponding to each monitoring sequence and generating compensated barrier field data includes: A dynamic barrier efficiency algorithm is used to identify sodium fire spread nodes in the thermal radiation coupling characteristic data, and a thermal attenuation compensation sequence corresponding to each monitoring sequence is determined based on the barrier region type corresponding to each sodium fire spread node. Calculate the deviation coefficient between thermal field nodes in the same barrier region in the thermal attenuation compensation sequence corresponding to any two monitoring sequences, and generate compensation barrier field data between the two monitoring sequences based on the deviation coefficient.

4. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 3, characterized in that, The calculation of the deviation coefficient between thermal field nodes in the same barrier region in the thermal attenuation compensation sequences corresponding to any two monitoring sequences includes: When the number of thermal field nodes in the thermal attenuation compensation sequences corresponding to any two monitoring sequences is inconsistent, virtual node interpolation is performed based on the barrier region parameters corresponding to the end thermal field node in the sequence with fewer thermal field nodes, and the deviation coefficient between thermal field nodes in the same barrier region is calculated based on the interpolated data.

5. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 1, characterized in that, The data cleaning module is specifically used for: The first monitoring sequence, the second monitoring sequence, and the third monitoring sequence are divided into equal gradients according to the preset thermal field partitions to generate standardized first thermal field distribution data, standardized second deformation data, and standardized third sodium flow data. A dynamic weighting method is used to correct the standardized first thermal field distribution data and the standardized second deformation data in real time, and a fixed weighting method is used to perform steady-state optimization on the standardized third sodium flow data, outputting a first optimization sequence, a second optimization sequence, and a third optimization sequence; wherein, the first optimization sequence includes the optimized first thermal field distribution data and the optimized second heat conduction feedback data.

6. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 5, characterized in that, The data cleaning module is also used for: Calculate the thermal attenuation coefficient of the optimized first thermal field distribution data and the optimized second thermal conduction feedback data within the historical barrier period; Based on the thermal attenuation coefficient and the optimized first thermal field distribution data, the expected heat flow value of the optimized second thermal conduction feedback data in the real-time isolation period is predicted according to the thermal field parameters in the real-time isolation period. Based on the optimized second heat conduction feedback data and its expected heat flow value, target sodium flow compensation data is generated, and the monitoring sequence corresponding to the target sodium flow compensation data is used as the first optimization sequence.

7. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 2, characterized in that, The barrier compensation layer specifically includes: The thermal attenuation analysis unit is used to perform thermal radiation path tracing on each monitoring sequence in the thermal radiation coupling feature data, so as to extract the corresponding thermal gradient attenuation chain from each monitoring sequence. The barrier field matching unit is used to spatially map the thermal gradient attenuation chain extracted from each monitoring sequence with the corresponding thermal radiation coupling feature data to generate compensated barrier field data.

8. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 7, characterized in that, The barrier compensation layer further includes: A thermal hysteresis compensation unit is used to perform thermal response delay correction processing on the compensation barrier field data.

9. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 2, characterized in that, The fire control decision-making body specifically includes: The multi-level integration unit contains multiple barrier control nodes, and each barrier control node is connected to each monitoring sequence in the compensation barrier field data and thermal radiation coupling characteristic data through association configuration. The dynamic weight allocation unit is used to iteratively adjust the associated configuration through a dynamic weight optimization algorithm to minimize the deviation between the sodium fire barrier control command and the actual thermal field distribution. The thermal imbalance positioning unit is used to locate the sodium fire anomaly area based on the compensation barrier field data and thermal radiation coupling characteristic data, and generate sodium fire barrier control commands.

10. The CEFR demonstration fast reactor sodium fire containment control system based on novel fire-resistant and heat-insulating materials as described in claim 5, characterized in that, The dynamic weight allocation method specifically includes: Adaptive barrier parameters are generated based on real-time sodium fire propagation characteristics. The standardized first thermal field distribution data is segmented and optimized using a sliding window mechanism.

Images