Multi-medium tritium monitoring system outside fusion reactor plant and monitoring method thereof

Abstract

The application provides a fusion reactor off-site multi-medium tritium monitoring system and a monitoring method thereof. The system comprises an air tritium monitoring subsystem, a water body tritium monitoring subsystem, a soil / biological tritium monitoring subsystem, an environmental monitoring data center, and an alarm and emergency response module. The system can comprehensively and intelligently guarantee off-site environmental safety in all directions, all-weather, low detection limit and linkage.

Description

Technical Field

[0001] This technical solution relates to the application field of nuclear fusion devices, and in particular to a multi-medium tritium monitoring system and its monitoring method outside the fusion reactor plant. Background Technology

[0002] Tritium is a radioactive isotope of hydrogen with a half-life of approximately 12.3 years, emitting low-energy beta rays. Fusion reactors (such as tokamak devices) typically use tritium as one of their fuels, and its management and emissions must be strictly controlled during operation. Even though tritium's radiation hazard is relatively low, it can still spread through air and water once leaked into the environment, requiring sufficient attention. Internationally, tritium emissions from nuclear facilities are strictly regulated: for example, in operating nuclear fission power plants, the emission of a certain amount of tritium into the environment within permissible limits is considered normal. Taking a nuclear power plant in China as an example, since its commissioning, the annual emission of gaseous tritium from its chimney has been monitored, and its annual emission is typically within the range of 10^11–10^12 becquerels; in 2010, the annual emission of gaseous tritium was approximately 1.26 × 10^12 becquerels, only 5.3% of the limit of 2.4 × 10^13 becquerels approved by the nuclear safety regulatory authorities. Environmental monitoring results indicate that the tritium concentration in the air surrounding the power plant remains at background levels, and its impact on public health is negligible. This demonstrates that tritium emissions are controllable and safe under normal operating conditions, provided that strict monitoring and management are implemented.

[0003] However, abnormal tritium leaks can still occur during the operation of nuclear facilities. For example, an anomaly was detected during routine groundwater monitoring at a nuclear power plant in the United States. Tracing the leak confirmed that approximately 1.51 million liters (about 400,000 gallons) of tritium-containing radioactive water had leaked into the plant's groundwater. Fortunately, timely monitoring detected the leak and prevented further spread of the contaminated water. The leak was largely contained within the plant area, without entering the nearby Mississippi River or contaminating surrounding drinking water sources. Approximately 25% of the leaked tritium was recovered and treated by extracting groundwater, and the remaining tritium concentration was diluted and attenuated, posing no measurable risk to the environment or the public. This case highlights the importance of environmental tritium monitoring: conventional environmental monitoring networks can detect abnormal tritium levels early, allowing for timely measures to prevent the spread of tritium into the environment. For fusion reactors, since the amount of tritium stored and used on-site far exceeds that of fission nuclear power plants (where tritium is the main fusion fuel), and tritium is mainly in gaseous form (such as tritium molecule gas or tritized water vapor), a more sophisticated external environmental monitoring system is needed to quickly detect abnormal tritium signals in the event of a leak, ensuring the safety of the surrounding environment and people.

[0004] Currently, nuclear power plants and other nuclear facilities generally have radiation environment monitoring systems in place to continuously track radioactivity levels around the plant site. For example, a nuclear power base in my country has nine radiation environment monitoring substations deployed within a 20-kilometer radius. However, for a considerable period in the past, environmental tritium monitoring primarily relied on manual sampling and laboratory analysis. For instance, the traditional method for airborne tritium involved monthly on-site air sample collection (continuous sampling for several hours), followed by chemical processing and liquid scintillation counting analysis in the laboratory before data reporting. This method is highly sensitive but has significant limitations: the low sampling frequency (e.g., monthly) fails to reflect instantaneous changes in tritium concentration in real time; the sampling and analysis process is cumbersome, involving significant manual labor and delays. When tritium levels in the environment rise rapidly (e.g., in a leak accident), it may not be detected until the next sampling cycle, missing the optimal response time.

[0005] Some existing invention patents (such as CN108982643A) propose online monitoring schemes for tritium-containing gases, such as improving the oxidation and separation efficiency of tritides through closed-loop sampling, thereby improving the accuracy of tritium concentration measurement. CN115015988A also mentions an online monitoring method for tritized water (HTO) and hydrogen-tritium (HT) in nuclear power plants, especially heavy water reactors and molten salt reactors, which generate tritium.

[0006] However, a complete system solution for off-site multi-medium tritium monitoring for fusion reactors is still lacking. Fusion devices differ from traditional pressurized water reactor nuclear power plants in the magnitude and form of tritium production and release: fusion reactors may release trace amounts of tritium into the environment during normal operation and shutdown maintenance. Therefore, not only is highly sensitive real-time monitoring of atmospheric tritium necessary, but also sampling and analysis of soil and water environments that may be affected by tritium deposition. While existing nuclear power plant environmental monitoring networks provide valuable experience, they have not yet been specifically optimized and integrated for the characteristics of tritium leakage from fusion reactors. For example, how to simultaneously handle online monitoring of gaseous tritium and low-background analysis of tritium in water and soil within the same system, and how to set monitoring point layouts and alarm thresholds based on the operational characteristics of fusion reactors, all require further technological innovation. Summary of the Invention

[0007] This invention aims to at least partially address one of the technical problems in related technologies. To this end, one objective of this invention is to establish a complete multi-media tritium monitoring system outside fusion reactor facilities to overcome the shortcomings of traditional monitoring in terms of media coverage, detection timeliness, and sensitivity. Specific problems include: 1) how to achieve real-time continuous monitoring of tritium in the air to promptly detect possible transient leaks; 2) how to effectively sample and accurately measure low-concentration tritium in media such as water and soil, including tritium accumulation under normal emissions and tritium contamination under abnormal conditions; 3) how to integrate and analyze data from different monitoring points and different media to promptly identify and locate the source of problems when tritium levels are abnormal; and 4) how to reduce human intervention and improve system reliability to meet the needs of long-term stable operation. Ultimately, this invention provides a tritium monitoring solution tailored to the characteristics of fusion reactors, capable of comprehensively, 24 / 7, with low detection limits, and intelligently interconnected to ensure the safety of the external environment.

[0008] To address the aforementioned technical problems, this invention provides a multi-media tritium monitoring system outside a fusion reactor plant. Its structure includes: an air tritium monitoring subsystem, a water tritium monitoring subsystem, a soil / biological tritium monitoring subsystem, an environmental monitoring data center, and an alarm and response module. All components are connected via wired or wireless networks to form a complete monitoring network.

[0009] In one aspect of the invention, a multi-media monitoring system for tritium outside a fusion reactor plant is proposed. According to an embodiment of the invention, the system includes: an air tritium monitoring subsystem, which is used to collect tritium-containing water and send it to a liquid scintillation counting module to continuously measure the tritium activity concentration online, with a typical sampling interval of several minutes, and automatically transmits the data to an environmental monitoring data center after each measurement;

[0010] A water body tritium monitoring subsystem is used to automatically collect water samples and use electrolytic enrichment-liquid flashover technology to achieve continuous water sample monitoring and transmit the data to the environmental monitoring data center.

[0011] A soil / biological tritium monitoring subsystem is used to periodically collect samples of soil, plants, aquatic organisms, etc. around the plant site, analyze the tritium activity concentration in them, and transmit the data to the environmental monitoring data center.

[0012] An environmental monitoring data center is used to receive and store data uploaded by the air tritium monitoring subsystem, the water tritium monitoring subsystem, and the soil / biological tritium monitoring subsystem, and then convert the collected data into a unified risk indicator (TRI).

[0013] An alarm and emergency response module is used to trigger a tiered early warning and push handling suggestions when the TRI exceeds a set threshold; wherein, the calculation formula for the TRI index is:

[0014]

[0015] in:

[0016] C air C soil C gw These represent the measured values ​​of tritium in air, soil gas, and groundwater, respectively.

[0017] M is a set of meteorological parameters (including wind speed, wind direction, atmospheric stability, etc.);

[0018] θ represents the soil moisture content;

[0019] h represents the groundwater level;

[0020] w air w soil w gw Let w be the weighting coefficients corresponding to the three media, and satisfy w air + w soil + w gw = 1;

[0021] f air (·), f soil (·), f gw (·) is a cross-medium conversion function that converts tritium concentrations in various media into dimensionless risk indicators.

[0022] The present invention discloses a multi-medium monitoring system for external tritium in a fusion reactor plant, characterized in that the system comprises:

[0023] a) Several off-site sampling nodes, including:

[0024] a1) Air node: Equipped with a switchable pretreatment module (humidity control and / or catalytic conversion unit) and a tritium detection module to distinguish or simultaneously measure HT and HTO;

[0025] a2) Soil gas node: Deployed at a depth of 0.3–1.5 m from the ground surface, equipped with soil gas sampling, humidity / temperature measurement and tritium detection modules;

[0026] a3) Shallow groundwater node: installed in a (2–30 m) monitoring well, equipped with a water sampling / online analysis module and water level and temperature sensors;

[0027] Each of the above nodes includes an environmental parameter module (wind speed, wind direction, temperature and humidity, rainfall / groundwater level, or at least one of these), a time synchronization module, and a communication module.

[0028] b) A unified monitoring platform, including:

[0029] b1) Quality control and calibration module: performs time synchronization, zero-point / span calibration, missing measurement compensation and anomaly removal for multi-node data;

[0030] b2) Cross-media normalization module: converts air tritium concentration, soil tritium concentration and groundwater tritium activity into a unified risk index TRI;

[0031] b3) Spatial Analysis and Backpropagation Module: Based on meteorological / hydrological parameters, spatial interpolation and source term location / impact range estimation are performed on anomaly points;

[0032] b4) Adaptive node placement module: Updates suggestions for adding / migrating nodes based on TRI and boundary conditions;

[0033] b5) Early Warning and Linkage Module: Triggers tiered early warnings based on set thresholds and pushes handling suggestions; wherein, the TRI is calculated according to the formula:

[0034] in:

[0035] C air C soil C gw These represent the measured values ​​of tritium in air, soil gas, and groundwater, respectively.

[0036] M is a set of meteorological parameters (including wind speed, wind direction, atmospheric stability, etc.);

[0037] θ represents the soil moisture content;

[0038] h represents the groundwater level;

[0039] w air w soil w gw Let w be the weighting coefficients corresponding to the three media, and satisfy w air + w soil + w gw = 1;

[0040] f air (·), f soil (·), f gw (·) is a cross-medium conversion function that converts tritium concentrations in various media into dimensionless risk indicators.

[0041] According to an embodiment of the present invention, the main feature of the system is "cross-media normalized TRI + adaptive point placement + three types of node collaboration", which is different from the "single medium + single instrument" in the prior art.

[0042] In another aspect, the present invention also proposes a method for multi-medium monitoring of tritium outside a fusion reactor plant. According to an embodiment of the present invention, the method for multi-medium monitoring of tritium outside a fusion reactor plant includes:

[0043] S1: Simultaneously collect tritium content in air, water, soil, and biological environments, and label them as C. air C gw And C soil ;

[0044] S2: According to the formula Calculate TRI;

[0045] S3: Based on wind field, rainfall, and groundwater level, spatiotemporal interpolation and source term inverse estimation are performed on TRI to obtain the affected area;

[0046] S4: Based on the TRI gradient and uncertainty, output the point placement optimization suggestions;

[0047] S5: If TRI exceeds the threshold, trigger a tiered early warning and emergency response mechanism.

[0048] In another aspect, the present invention also proposes a method for multi-medium monitoring of tritium outside a fusion reactor plant. According to an embodiment of the present invention, the method for multi-medium monitoring of tritium outside a fusion reactor plant includes:

[0049] S1: Tritium and environmental parameters are collected simultaneously at three types of off-site nodes;

[0050] S2: The platform performs quality control / calibration and utilizes f air f soil f gw Normalize multi-media data into TRI;

[0051] S3 performs spatiotemporal interpolation and source term inverse estimation on TRI based on wind field / rainfall / groundwater level to obtain the affected area;

[0052] S4: Based on the TRI gradient and uncertainty, output the point placement optimization suggestions;

[0053] S5: If TRI exceeds the threshold, trigger a tiered early warning and emergency response mechanism.

[0054] In another aspect, the present invention also proposes a unified monitoring platform device for external tritium media in fusion reactor plants. According to an embodiment of the present invention, the unified monitoring platform device for external tritium media in fusion reactor plants includes a processor and a memory, wherein the memory stores a computer program for executing the steps S1–S5 described above.

[0055] In another aspect, the present invention also provides a computer-readable storage medium. According to an embodiment of the present invention, the computer-readable storage medium stores a program that, when executed by a processor, implements the aforementioned method.

[0056] The system of this invention achieves significant technical effects through the above solutions: First, in terms of monitoring timeliness, online air tritium monitoring compensates for the low frequency of traditional manual sampling, capturing tritium concentration fluctuations on a minute-level timescale. For example, if a minor tritium leak causes a rise in the reading at a monitoring point downwind of the plant boundary within one hour, this system will immediately record this change and issue an alarm, far faster than traditional methods that require monthly sampling. Second, in terms of spatial comprehensiveness, the multi-media monitoring design ensures that there are corresponding monitoring methods for tritium whether it diffuses through the atmosphere or settles into soil and water, leaving no monitoring blind spots. For example, while air monitoring detects an anomaly, the system can simultaneously monitor whether tritium levels rise in adjacent well samples, thereby determining whether tritium has entered groundwater runoff. Third, in terms of quantitative accuracy, the system employs advanced low-background measurement technology and automated quality control processes, capable of distinguishing subtle changes in extremely low concentrations of tritium. Extensive long-term monitoring data shows that, under accident-free conditions, the tritium activity concentration in the environmental media surrounding nuclear facilities has remained within the range of background fluctuations for decades. Therefore, if any anomaly exceeds the background level, even if the absolute dose impact remains below regulatory standards (such as the public annual effective dose limit of 100 millirems), this system can sensitively detect trend changes and trace their source. Finally, regarding operational reliability, the system is highly automated, reducing manual labor intensity and errors, while regular calibration and comparative experiments (such as environmental tritium laboratory control measurements) ensure the authenticity and accuracy of the data. In summary, this solution demonstrates excellent performance in ensuring monitoring is "true, accurate, comprehensive, fast, and up-to-date": the data is authentic and reliable, highly accurate, comprehensively covered, responds quickly, and updates information in real time, providing a stable and reliable safety barrier for tritium levels in the environment surrounding the fusion reactor.

[0057] According to an embodiment of the present invention, TRI itself is a dimensionless "risk normalization index," and the threshold is not a unique fixed number, but rather "reverse-engineered" based on regulations and baseline data. A more natural engineering approach is to... TRI = 1 The baseline is designed to be "close to the regulatory reference level," and a 2-3 level graded early warning system is established based on this.

[0058] Preferably, TRI=1 is set as the baseline value for the overall exposure level equivalent to the regulatory reference concentration, and the TRI threshold adopts a dimensionless hierarchical index:

[0059] The threshold T1 for Level 1 (yellow) warning is set at 0.3–0.5;

[0060] The threshold T2 for Level 2 (orange) warning is set to 1.0;

[0061] The threshold T3 for Level 3 (Red) early warning is set to 3.0;

[0062] T2 corresponds to a combined risk level of approximately 4000 Bq / m³ for air tritium concentration and approximately 100 Bq / L for water tritium concentration, referencing relevant standards such as the EU drinking water tritium indication value of 100 Bq / L.

[0063] The multi-medium tritium monitoring system and method for fusion reactor plants proposed in this invention have at least one of the following beneficial effects, addressing the shortcomings of the prior art:

[0064] 1) Multi-media integrated monitoring: Enables comprehensive monitoring of tritium activity in various media such as ambient air, water bodies (surface water / groundwater) around the plant, soil and biological samples, covering all environmental pathways through which tritium may migrate, and greatly improving the comprehensiveness of tritium leakage detection.

[0065] 2) Combining real-time and delayed monitoring: The system uses online continuous monitoring of tritium in the air to obtain tritium concentration change curves in real time; for media such as water and soil, it uses periodic automatic sampling + rapid analysis methods. The combination of the two modes ensures that both instantaneous leakage is captured and the cumulative impact is assessed.

[0066] 3) High-sensitivity, low-background detection: Employing highly sensitive tritium monitoring instruments and ultra-low background technology, the detection limit is low, enabling the detection of abnormal increases in trace tritium levels in the environment. This allows for early warning even when tritium concentrations are far below regulatory limits, providing ample reaction time.

[0067] 4) Automation and Remote Data: The monitoring device operates automatically and transmits data wirelessly, requiring no frequent manual intervention. Monitoring data is aggregated in real time to the central environmental monitoring data platform, and automatically analyzed, stored, and triggers alarms for exceeding thresholds via software. Once the tritium concentration at a monitoring point exceeds a preset threshold, the system automatically issues an alarm and notifies relevant personnel, enabling rapid response.

[0068] 5) Enhance safety oversight and public trust: The deployment of this system will provide fusion facility operators and regulatory authorities with scientifically detailed tritium emission data. On the one hand, it ensures that emissions comply with the "lowest possible and reasonably feasible" (ALARA) principle; on the other hand, transparent environmental monitoring information will help enhance public confidence in the safety of fusion energy.

[0069] The system or method of the present invention is significantly superior to traditional technologies in terms of monitoring medium range, time response, sensitivity and intelligence, and can more effectively ensure the stability and safety of tritium background in the environment surrounding the fusion reactor. Attached Figure Description

[0070] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0071] Figure 1 This is a schematic diagram of a multi-medium tritium monitoring system outside a fusion reactor plant;

[0072] Figure 2 This is a conceptual diagram of tritium source inversion near the plant area. The red square represents the nuclear power plant site, the blue triangle represents the tritium monitoring well (which detected a high tritium concentration), the orange area represents the putative tritium source area, the green arrow indicates the prevailing wind direction (e.g., from southeast to northwest), and the blue arrow indicates the tritium plume transport path.

[0073] Figure 3 This is a schematic diagram of the sampling points before and after 2002 (A) and after (B, including 2002). Among them, W1-W18 are the sampling points for seawater monitoring in the bay of the D1 nuclear power plant in 2000, and L1-L10 are the sampling points for seawater monitoring in the bay of the D1 nuclear power plant in 2002 (because the emission outlet of the nuclear power plant was changed in 2002). Detailed Implementation

[0074] The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0075] As attached Figure 1 As shown below, the structure and function of each subsystem are described in detail:

[0076] 1. Air Tritium Monitoring Subsystem: Several air tritium monitoring devices are deployed at key locations around the fusion plant and at the plant boundary (such as downwind of the prevailing wind direction, near exhaust ducts, etc.). Each device includes components such as a sampling pump, dryer, catalytic oxidation furnace, gas-liquid separator, and low-background liquid scintillation counting unit. Its working principle is as follows: Ambient air is continuously drawn into the device at a certain flow rate. First, dust and interfering gases are filtered out. Then, hydrogen tritium (HT) in the air is converted into tritium (HTO) in water vapor through a catalyst. The tritium-containing water is collected by the gas-liquid separator and sent to the liquid scintillation counting module to measure the tritium activity concentration. The device adopts a closed-loop circulation design, with uncaptured inert gases circulated back to the system inlet to reduce sample loss and exhaust emissions. Each air tritium monitor can achieve continuous online measurement, with a typical sampling interval of several minutes. After each measurement, the data is automatically transmitted to the monitoring center. Under normal circumstances, the air tritium readings at each monitoring point should be close to the environmental background level (usually < a few BK per cubic meter); once the reading at a certain point rises significantly and exceeds the threshold (based on the background statistical settings, such as the background mean + 3σ), the system will immediately mark it as abnormal.

[0077] 2. Water Tritium Monitoring Subsystem: Several water sampling and monitoring devices are deployed around the fusion device, including monitoring of surface water and groundwater. Surface water monitoring points can be located at stormwater outfalls within the plant area, downstream of surrounding streams, etc.; groundwater monitoring wells are located downstream of the plant perimeter. The monitoring device consists of an automatic sampler and a laboratory analysis unit. The automatic sampler can collect water samples at a preset frequency (e.g., daily or weekly) and perform necessary pretreatment (e.g., filtration to remove impurities). The water monitoring system is equipped with an online monitoring module: continuous water sample monitoring is achieved using electrolytic enrichment-liquid scintillation technology. When the tritium concentration in the water exceeds the management limit (e.g., drinking water standards or environmental benchmarks), the system triggers an alarm and can increase the sampling frequency or initiate emergency sampling.

[0078] 3. Soil / Biotritium Monitoring Subsystem: Primarily used to assess the long-term impacts of atmospheric tritium deposition on soil and the biological environment. Monitoring includes periodically collecting samples from soil, plants, and aquatic organisms around the plant site to analyze tritium activity concentrations. Since soil and biotritium content is typically extremely low, a combustion oxidation-water absorption method is used to convert tritium in the samples into water, which is then measured using liquid scintillation counting. The system is equipped with mobile sampling devices and central laboratory analytical equipment. The specific implementation method is as follows: Monitoring personnel (or automated sampling robots) collect representative surface soil samples, samples of typical local crops (such as vegetables and rice), and samples of high-level organisms in the local food chain (such as fish) quarterly or semi-annually. The samples are sealed and sent to the laboratory for drying, grinding, and high-temperature combustion to convert bound tritium (organic tritium) in the samples into tritized water, which is then quantitatively analyzed. If the monitoring results show no significant change compared to historical baselines, it indicates that the environmental tritium level is stable; if an upward trend is found, further investigation into the sources of tritium (atmospheric deposition, irrigation water, etc.) is needed and control measures should be taken.

[0079] 4. Environmental Monitoring Data Center: Located in the central monitoring room of the nuclear facility or regulatory station, this center receives and stores data uploaded by various monitoring substations. The data center is equipped with a monitoring software platform with functions including real-time data display, historical data storage, anomaly trend analysis, and map location display. On the large screen in the data center, users can intuitively view the instantaneous air tritium values, cumulative dose estimates, water tritium concentration time series, and periodic monitoring results of soil biotritium from each monitoring station. The system features multi-level access control, allowing different users (operators, regulators, and the public to access corresponding data views). The data center also integrates a meteorological data interface to obtain information such as wind direction, wind speed, and rainfall to assist in the analysis of tritium migration and diffusion. When any monitoring indicator exceeds a set threshold, the monitoring software will automatically record the event and enter an alert state.

[0080] 5. Alarm and Emergency Response Module: This module is linked to the data center. When monitoring data triggers early warning conditions, the alarm module sends signals through various means, such as audible and visual alarms and notification of relevant management personnel. Alarm information includes the time, location, and extent of exceedance. Upon receiving the alarm, operators will activate the emergency plan: first, verify whether it is an instrument malfunction or a false alarm (this can be verified by calibrating the instrument and repeating sampling); if a tritium leak is confirmed, the fusion device control room will be promptly notified to take control measures (such as cutting off the leak source and activating the plant's ventilation and purification system), while simultaneously reporting to regulatory authorities and initiating environmental emergency monitoring. Emergency monitoring measures include increasing the frequency of sampling in the surrounding area, expanding the monitoring range downwind, and assessing the impact range of the tritium leak. Through the above-mentioned linked response mechanism, timely early warning, cause location, pollution containment, and protection of public safety are ensured in the event of abnormal tritium emissions.

[0081] Without departing from the spirit of the present invention, some modifications or substitutions can be made to the above technical solutions. First, in the air tritium monitoring section, if it is temporarily impossible to obtain a continuous online tritium monitor, an intermittent sampling + portable analysis scheme can also be adopted: for example, an air sample is automatically collected every few hours, the moisture is collected by condensation, and then measured by a mobile liquid scintillation counting device. Although the timeliness is slightly worse than that of real online monitoring, it has been greatly improved compared with the traditional sampling by day / week. Second, in the water body monitoring aspect, it can be replaced by the method of passive sampler combined with laboratory analysis. A passive sampler, such as an adsorption device inserted into a groundwater well, can accumulate tritium in the environment for a long time, so as to provide a time-integrated concentration value for analysis. This method has low cost and is power-free, and is suitable as a supplement or backup for online monitoring. Third, if the network signal is poor in remote areas in the data transmission part, the local storage + manual inspection and meter reading mode can be adopted to temporarily replace the remote transmission, that is, each monitoring station stores data, and the operation and maintenance personnel go to download the data and update the system regularly. Finally, for fusion facilities of different scales, the system architecture can also be simplified: for example, for small devices with extremely low tritium usage, only air tritium monitoring can be arranged; for fusion reactors close to the coast, the tritium monitoring of the marine path (seawater, seafood) needs to be strengthened. The above replacement schemes are all within the scope of the concept of the present invention and can be combined and adjusted according to actual needs to achieve the optimal monitoring effect.

[0082] 1. Overall architecture of the off-site multi-media tritium monitoring system for fusion reactors: A monitoring network architecture covering three media of air, water, and soil organisms is proposed, which is particularly suitable for the safety monitoring of the peripheral environment of high-tritium operating fusion devices. The system consists of multiple monitoring substations and a central data platform. Each substation operates independently and is centrally networked to achieve real-time data sharing and linkage alarm. This overall solution ensures a comprehensive perception of tritium in the environment and is one of the core innovation points of the present invention.

[0083] 2. Complete process of the monitoring method: The present invention also protects a monitoring method for tritium outside the fusion reactor, and its steps include: planning the monitoring plan according to the characteristics of the fusion reactor → deploying the monitoring network and establishing the background database → collecting samples of each medium in real time / regularly → respectively measuring the tritium activity in air, water, soil organisms → transmitting the data to the central platform for comparison and analysis → starting the alarm and emergency procedures when the detected value exceeds the warning threshold → regularly outputting the monitoring report and publicly releasing information. This method covers the entire cycle of planning, implementation, response, and evaluation, and is the guiding process to ensure the effective operation of the monitoring system.

[0084] 3. Multi-source Data Integration and Intelligent Alarm System: This invention integrates multi-media monitoring data at the software level, establishing a tritium monitoring database and intelligent analysis model. When a tritium anomaly occurs at a monitoring point, the system combines data from other points and meteorological data to comprehensively determine the nature of the event and automatically triggers an alarm and pushes information through preset logic. This intelligent linkage design ensures timely and reliable tritium leak early warning. This protection point encompasses innovative hardware and software integration, including data acquisition and communication protocols, threshold determination algorithms, and alarm linkage rules.

[0085] I. TRI Unified Index Calculation Formula

[0086] The tritium concentration in each medium is uniformly converted into a risk indicator:

[0087]

[0088] in:

[0089] Concentration of tritium in the air (unit: Bq / m³)

[0090] Concentration of tritium in soil atmosphere (unit: Bq / m³)

[0091] Concentration of tritium in groundwater (unit: Bq / L)

[0092] The risk normalization function for the corresponding medium is defined as follows:

[0093]

[0094] Reference concentration thresholds for each medium, such as:

[0095]

[0096]

[0097]

[0098] Weighting coefficients, satisfying:

[0099]

[0100] II. Core formula for the inverse derivation of the Gaussian plume model (used for air tritium inversion)

[0101] Predict the tritium concentration at point (x, y, z) :

[0102]

[0103] in:

[0104] Emission source strength (Bq / s)

[0105] Average wind speed (m / s)

[0106] Lateral and vertical diffusion coefficients (m) are related to atmospheric stability level.

[0107] Effective emission height (m)

[0108] III. Infiltration-Lag Groundwater Tritium Migration Model (for predicting post-rainwater well water response)

[0109] Considering the tritium concentration transport delay from the surface to the well water, it can be simplified into a first-order response model:

[0110]

[0111] in:

[0112] Groundwater tritium concentration

[0113] The cumulative time (preferably in hours / days) since the selected start time represents the current calculated groundwater tritium concentration. The moment

[0114] In time Rainfall intensity (mm / h) or infiltration flux

[0115] : Integral variable, representing a historical moment

[0116] The infiltration-response time constant (hours to days) is related to permeability, soil moisture content, etc.

[0117] Infiltration coefficient: Reflects the proportion of tritium carried into the aquifer during the infiltration process.

[0118] IV. Dynamic Weight Adjustment Mechanism

[0119] In the event of extreme weather, the platform automatically adjusts its weighting:

[0120]

[0121] in:

[0122] : Cumulative rainfall in the past 24 hours (mm)

[0123] Trigger threshold (e.g., 30 mm)

[0124] Rainfall – Weighting Gain Coefficient

[0125] Maximum weight adjustment limit (e.g., +0.4)

[0126] Air weight decreases synchronously:

[0127]

[0128] Soil air weight may be slightly adjusted or kept unchanged depending on the circumstances.

[0129] The system is as follows:

[0130] a) Several off-site sampling nodes, including:

[0131] a1) Air node: Equipped with a switchable pretreatment module (humidity control and / or catalytic conversion unit) and a tritium detection module to distinguish or simultaneously measure HT and HTO;

[0132] a2) Soil gas node: Deployed at a depth of 0.3–1.5 m from the ground surface, equipped with soil gas sampling, humidity / temperature measurement and tritium detection modules;

[0133] a3) Shallow groundwater node: installed in a (2–30 m) monitoring well, equipped with a water sampling / online analysis module and water level and temperature sensors;

[0134] Each of the above nodes includes an environmental parameter module (wind speed, wind direction, temperature and humidity, rainfall / groundwater level, or at least one of these), a time synchronization module, and a communication module.

[0135] b) A unified monitoring platform, including:

[0136] b1) Quality control and calibration module: performs time synchronization, zero-point / span calibration, missing measurement compensation and anomaly removal for multi-node data;

[0137] b2) Cross-media normalization module: converts air tritium concentration, soil tritium concentration and groundwater tritium activity into a unified risk index TRI;

[0138] b3) Spatial Analysis and Backpropagation Module: Based on meteorological / hydrological parameters, spatial interpolation and source term location / impact range estimation are performed on anomaly points;

[0139] b4) Adaptive node placement module: Updates suggestions for adding / migrating nodes based on TRI and boundary conditions;

[0140] b5) Early Warning and Linkage Module: Triggers tiered early warnings based on set thresholds and pushes handling suggestions; wherein, the TRI is calculated according to the formula:

[0141] in:

[0142] C air C soil C gw These represent the measured values ​​of tritium in air, soil gas, and groundwater, respectively.

[0143] M is a set of meteorological parameters (including wind speed, wind direction, atmospheric stability, etc.);

[0144] θ represents the soil moisture content;

[0145] h represents the groundwater level;

[0146] w air w soil w gw Let w be the weighting coefficients corresponding to the three media, and satisfy w air + w soil + w gw = 1;

[0147] f air (·), f soil (·), f gw (·) is a cross-medium conversion function that converts tritium concentrations in various media into dimensionless risk indicators.

[0148] Example 1: Analysis of tritium source inversion and monitoring point optimization at a certain base.

[0149] Principle introduction: Physical constraint coupling inverse reasoning and uncertainty-driven point placement strategy.

[0150] Coupled Physical Model Back-Depth Analysis: This application integrates physical models of atmospheric and subsurface media migration to perform back-depth analysis on high-TRI (tritium index) areas detected by monitoring. When a significant increase in tritium index (“TRI hotspot”) occurs at a monitoring point or in a certain area, the system first calls the Gaussian plume model to constrain the possible source locations. The Gaussian plume model is a classic method that assumes continuous point-source release of pollutants and a normal distribution of pollutants in the atmosphere. It can use measured wind direction, wind speed, atmospheric stability, and other data to back-calculate possible source areas upwind. For example, if a downwind air monitoring station shows an abnormally high TRI, combined with the meteorological data at the time, the model can deduce which area upstream along the wind direction might have a release source and estimate the source strength range. At the same time, the patent considers the migration lag characteristics of tritium in the soil-groundwater system and provides an infiltration lag model analysis for groundwater anomalies. This model integrates the precipitation-infiltration-groundwater recharge process: after tritium settles from the air, it can accumulate in the soil and enter the aquifer with infiltrated rainwater. The entire process involves a certain delay and diffusion. The model uses physical parameters (such as soil permeability, porosity, and unconfined aquifer thickness) as constraints and utilizes rainfall and groundwater flow velocity to estimate tritium transport time.

[0151] The required constraints for inverse calculations are as follows: For the atmospheric model, the system calls upon real-time meteorological data, including wind speed, wind direction, turbulent diffusion parameters, and atmospheric stability, as well as radioactive decay and wet / dry deposition parameters for source strength estimation. For the groundwater model, hydrogeological parameters such as soil permeability coefficient, permeability coefficient, groundwater flow velocity, and aquifer thickness are required, along with rainfall and water level data (e.g., 24-hour rainfall and well water level time series). These physical constraints ensure that the inverse calculation results are reasonable and reliable, preventing source term locations from exceeding physical possibilities. For example, the Gaussian model considers the diffusion coefficient of atmospheric diffusion (determining the plume's dispersion range) and the impact of building wakes, while the soil-water model considers the hysteresis coefficient of infiltration delay (determining the time it takes for tritium to reach the aquifer after rainfall), to avoid misjudgments caused by simple interpolation. The patented inverse calculation process not only relies on current monitoring values ​​but also utilizes historical monitoring sequences and empirical parameters to correct the model, making the location more accurate. Compared to traditional nuclear power plants that rely on fixed monitoring points pre-positioned downwind and downstream to infer leak locations, this patent achieves proactive inversion through model coupling: once an anomaly is detected, the potential source area can be immediately inferred by combining wind and water flow fields, improving the speed and accuracy of source tracing. Traditional deployments often involve empirically selecting monitoring locations (such as setting up monitoring stations in the prevailing wind direction) and manually analyzing the source after an anomaly occurs. In contrast, this strategy uses a physical model to automatically complete the preliminary diagnosis, narrowing down the investigation scope and assisting staff in quickly locating the problem area.

[0152] Uncertainty-Driven Optimization of Monitoring Site Layout: After initial reverse-engineering location, the patent introduces geostatistical and information entropy methods to assess the coverage of the current monitoring network over the pollution field, thereby intelligently recommending new monitoring sites or optimizing the existing site layout. Specifically, using existing monitoring data and model predictions, the Kriging interpolation surface and its standard deviation (uncertainty field) of tritium distribution within the region are calculated. The Kriging method can provide the concentration estimate and variance for each unmonitored location; a large variance indicates high prediction uncertainty at that location. The patent overlays the uncertainty field with the TRI risk field to identify areas with high uncertainty and high risk potential as candidate monitoring sites. For example, if reverse-engineering indicates a possible tritium migration channel downstream of a plant, but the current monitoring stations in that area are sparse and the predicted concentration has significant uncertainty, the system will suggest adding monitoring wells or soil gas sampling points to reduce uncertainty. Similar methods have achieved good results in radiation monitoring network optimization research: using a Gaussian process / Kriging model to capture pollution heterogeneity based on a small number of observation points and selecting new monitoring locations by maximizing information gain. The patent references these methods to compare multiple candidate site selection schemes (e.g., comparing the magnitude of reduced prediction errors) to select the sites that best improve the monitoring network's effectiveness. Maximum entropy estimation also employs information theory, selecting sampling locations that minimize scene entropy (uncertainty), thus making the overall pollution distribution clearer. With an uncertainty-driven mechanism, the monitoring network can "adaptively" improve: as operational data accumulates, the system continuously identifies monitoring blind spots and weak points and proposes optimization suggestions. Compared to current nuclear power plant site selection, which mainly relies on experience and fixed patterns (such as placing stations at regular intervals or focusing on residential areas and downwind areas), this dynamic optimization strategy can more flexibly cope with complex terrain and changing environmental conditions, improving the spatial representativeness of monitoring. Especially in monitoring pollutants like tritium, which have complex migration paths and low background levels, introducing spatial uncertainty-guided site selection ensures that the risk of missed detection is minimized, and the overall network has a higher probability of capturing anomalies.

[0153] Example 2: Tritium Source Inversion and Monitoring Point Optimization Analysis at a Certain Base

[0154] 1. Tritium Leakage Source Inversion Model and Results

[0155] Model Construction and Parameter Setting: Based on tritium monitoring data and environmental characteristics around the plant area, a Gaussian plume-infiltration hysteresis coupling model was established to invert the possible tritium leakage source location and emission characteristics. The model assumes that the nuclear power plant continuously emits tritized water vapor (HTO) gaseous effluent. First, a Gaussian plume model was used to simulate the diffusion concentration field of HTO in the atmosphere. Key parameters included: wind speed of typical values ​​of approximately 2–3 m / s (based on common wind speeds under static and stable atmospheric conditions); atmospheric stability of E–F classes to generate higher near-ground concentrations (consistent with the situation where high tritium levels in monitoring wells usually occur under low wind speed and stable conditions); emission source height of approximately 50–60 m (the height of the nuclear power plant's exhaust stack); and diffusion parameters estimated according to stability classification using the Pasquill-Gifford formula. The model outputs the instantaneous average concentration distribution C(x,y,z), where the ground concentration at a downwind distance x can be expressed as:

[0156]

[0157] Where Q is the tritium release rate, u is the wind speed, y and z are the horizontal and vertical diffusion parameters, and H is the emission height. The model considers the deposition and re-volatilization of HTO by vegetation and the ground surface, but mainly uses rainfall leaching as the mechanism for tritium migration into soil and groundwater. By introducing an infiltration hysteresis module, the model simulates the delayed process of rainfall depositing atmospheric HTO into the soil and infiltrating into the monitoring well. The inventors assume that at the time of rainfall, atmospheric tritium is released at a rate of leaching coefficient... Settlement occurs, and the time delay of tritium infiltration from the soil to groundwater increases. It can be estimated based on parameters such as rainfall intensity and soil permeability, and usually ranges from several hours to several days.

[0158] The tritium elution rate (i.e., effective settling velocity) used in the model to estimate the tritium deposition rate from the air to the Earth's surface can be taken on the following orders of magnitude:

[0159]

[0160] This coefficient reflects the proportion of tritium carried to the surface by raindrops per unit time under effective rainfall conditions. The values ​​are referenced from empirical values ​​for aerosol wet deposition parameters at nuclear power plants, and the physical solubility of HTO in air-raindrop mixtures is considered. Soil moisture content and permeability coefficients are referenced from local soil properties to determine the transport delay and dilution rate of tritium from the surface into well water.

[0161] Inversion Process and Key Monitoring Point Selection: Within the model framework, the monitoring point with the highest tritium concentration around the plant site was selected as the starting point for inversion. According to monitoring results, the PR1 monitoring well at the D1 nuclear power plant recorded the highest tritium level: the peak tritium concentration in the PR1 well water reached 22.5 Bq / L (occurring in June 2005); tritium was detected in all samples from well C from 2005 to 2014, with an annual average of 6.8–14.2 Bq / L and a maximum single sample concentration of 15.6 Bq / L. The PR1 well showed a single anomalous increase, while well C showed a persistently high level, indicating that they may correspond to different source characteristics: the PR1 peak may be related to a one-time release scenario, while well C reflects a long-term cumulative effect. Therefore, the PR1 monitoring well data was selected for inverting a single anomalous source event, and the C monitoring well was selected for inverting a persistent low-intensity source. The model uses the tritium concentration at the monitoring well location to calculate possible source parameters upwind: the monitoring well is considered to be affected by a single upwind source, and the tritium concentration in the well water is known. In this case, the atmospheric deposition flux corresponding to the flux is estimated using a rainfall infiltration model, and then the source strength Q and the distance from the source to the well are derived using the Gaussian plume formula. To constrain the inversion solution, the location of the monitoring well relative to the nuclear power plant and the prevailing wind direction are comprehensively utilized—for example, if well PR1 is located southwest of a region of the nuclear power plant, the tritium source is likely located upwind (northeast) of the well, close to the nuclear power plant. Figure 2 As shown.

[0162] Estimated Source Area: Based on inversion calculations, the source location corresponding to the anomalous tritium concentration at well PR1 is likely located in a localized area near the nuclear power plant site, roughly within a few hundred meters upwind of well PR1. Specifically, well PR1 is located near the boundary of the D1 nuclear power plant. The prevailing wind direction and topographical conditions when the high value of 22.5 Bq / L was detected in June 2005 indicated a possible anomalous tritium emission source within approximately 0.1–1 km south or southwest of the nuclear power plant site. This range encompasses the soil area near the nuclear power plant's emission pipelines and a localized area downwind of the plant site. The estimated source area's latitude and longitude coordinates roughly fall within the area inside the nuclear power plant's perimeter wall, closer to well PR1 (a few degrees south and west relative to the nuclear power plant's reference point, at a distance of several hundred meters). In general, the spatial range of the tritium source indicated by the high tritium values ​​at the D1 nuclear power plant monitoring wells is limited to the soil-atmosphere exchange zone within a few hundred meters of the nuclear power plant site. This is consistent with the actual monitoring results showing that the tritium impact mainly occurs around the reactor. No obvious signs of tritium sources were found at greater distances, indicating that the tritium source area is relatively limited.

[0163] Estimated emission intensity and duration: Based on Gaussian inversion calculations and monitoring well concentration levels, the estimated emission intensity of the anomalous tritium source is approximately every 10 11 -10 12The magnitude is in the Bq range, and the duration could range from months to years. Specifically, the source strength that caused the PR1 well to reach a peak of 22.5 Bq / L in 2005 is estimated to be around 10. 11 A tritium release event of approximately Bq could occur within a rainfall period prior to sampling (on a timescale of several weeks). This magnitude corresponds to 1%–10% of the total annual gaseous tritium emissions from the nuclear power plant that year, representing a small or background level release. It is noteworthy that the annual gaseous tritium emissions from the two units at Site D1 between 2001 and 2009 ranged from 7.2 × 10⁻⁶. 11 Up to 1.67×10 12 Bq, approximately 1.26 × 10⁻⁶ in 2010 12 Bq represents only about 5% of the permitted limit. Therefore, it can be inferred that whether it's a single anomaly or a long-term source, the tritium emission intensity is at a low level within the normal permitted range for nuclear power plants. Regarding duration, PR1 corresponds to a transient / short-term event (a peak value caused by a single rainfall event), while Well C corresponds to long-term continuous emissions (slow releases accompanying the unit's operating cycle). Model inversion supports this temporal characteristic: short-term sources will cause a significant spike, while continuous sources cause background elevation over many years without extreme peaks. Therefore, it is inferred that the possible duration of anomalous emissions ranges from a single rainfall event (day-level) to years of continuous low-intensity releases. Current monitoring data is more consistent with a long-term, stable, small-scale release scenario than a large-scale, instantaneous leak.

[0164] Inversion Reliability Analysis: The reliability of the inversion results is moderately high. Firstly, the source region obtained from the inversion (near the plant area) closely matches the actual tritium emission source location of the nuclear power plant: the tritium in the monitoring wells mainly originates from the normal emission of gaseous tritium from the nuclear reactor and its auxiliary facilities, which then settles and infiltrates. This conclusion is also supported by previous investigations: it was suspected that the high tritium concentration in the PR1 well originated from a leak in the effluent pipeline, but detailed investigation confirmed that the tritium in the well water originated from nuclear current effluent deposited in the atmosphere, not from a direct pipeline leak. This conclusion is consistent with our inverted source region (soil upwind of the plant area), increasing the reliability of the inversion location. Secondly, we verified this by comprehensively analyzing the spatial distribution and concentration trends of multiple monitoring points: comparing the results of the two adjacent monitoring wells, PR1 and P5, shows that the tritium levels in the two wells have similar ranges over many years and are not significantly correlated with the operation of Unit E, indicating that the tritium source of PR1 / P5 is mainly the D1 nuclear power plant itself. This is consistent with the inverted source location of "near the D1 plant area." Therefore, given the existing monitoring point layout, the qualitative source location retrieval is reliable. However, uncertainties remain in the quantitative retrieval: variations in model parameters (wind speed, stability, etc.) can cause an uncertainty of approximately ±50% in the source intensity estimation. The sparse distribution of monitoring wells also limits the accuracy of the solution for the exact source location, potentially limiting it to a range of several hundred meters rather than a precise point. Furthermore, the single anomaly in well PR1 may have been influenced by meteorological chance factors, and the model failed to fully reconstruct the short-term atmospheric process. These factors introduce moderate uncertainty into the retrieval results. Overall, multiple corroborations improve reliability: locating the source area near the plant is reliable, but specific values ​​regarding emission intensity and duration should be considered as orders of magnitude estimates. Further improvements in retrieval accuracy can be achieved by increasing the monitoring point density and using a more refined atmospheric numerical model.

[0165] 2. Uncertainty Analysis and Optimization of Tritium Monitoring Network Layout

[0166] Current Monitoring Coverage and Blind Spot Assessment: A series of environmental tritium monitoring points have been deployed around the D1 nuclear power plant site, including monitoring wells, water sampling points, and a few meteorological sampling points. Currently, there are 8 land-based monitoring wells (PR1 and P5 at the D1 nuclear power plant, A, B, and C at Phase I of a certain E plant, and P1, P2, and P3 at Phase II of a certain E plant), mainly distributed near the reactors; there are only 2 atmospheric water vapor sampling points (F city near the plant and Z city approximately 4 km away); there is 1 rainwater sampling point (G city); and there were previously 5 drinking water sampling points (Z city, D1 nuclear power plant, S city area, Y area, and X area), but some remote points were reduced after 2006. Overall, the current monitoring network has achieved basic coverage within a certain range around the nuclear power plant, but blind spots and areas with high uncertainty still exist on a larger spatial scale.

[0167] Uncertainty Distribution Analysis: The uncertainty of current monitoring data can be assessed using the Kriging spatial interpolation method. Due to the sparse distribution of monitoring points, uncertainty increases rapidly between points. Especially in areas far from any monitoring point, the interpolation standard deviation increases significantly. For example, interpolating the groundwater tritium concentration field centered on the nuclear power plant reveals a sharp increase in uncertainty outside the area enclosed by the monitoring wells, indicating a monitoring blind zone. Furthermore, west of the Z-city gas station and further away from the plant area, atmospheric tritium concentration can only be extrapolated due to the lack of monitoring, leading to significant uncertainty in the background value assumption. Therefore, the current network has limited predictive ability for tritium distribution within a 10km radius of the plant site, and the reliability of predictions decreases beyond a few kilometers. Considering the prevailing wind direction (southeast in summer, northerly in winter): in winter, tritium drifts southward towards the sea, with minimal land impact; in summer, sea breezes blow tritium towards the northern inland areas, where tritium deposition and accumulation may occur in the mountains north of the site, but currently there are no monitoring points in this direction, creating a potential blind zone. Monitoring wells are primarily located near the plant perimeter, covering an area of ​​only about 1–2 km. There are almost no tritium monitoring points in land areas exceeding 2 km (especially within the 5–10 km range), with only one gas sampling point 4 km from Z city. For example, there are no monitoring points in the land areas northeast and southeast of the nuclear power plant (if there are, the current network does not cover residential areas or environmentally sensitive areas there). While monitoring points are deployed for seawater in the ocean direction, there is a lack of monitoring further downwind (such as inland areas or islands further up the prevailing wind direction). This means that if tritium drifts to an unconventional direction with abnormal winds (e.g., the prevailing wind reverses), the current network may not be able to detect it in time. Existing monitoring focuses more on water samples (well water, seawater, rainwater) and a small amount of atmospheric wet deposition, without direct monitoring of soil gaseous media. However, tritium water vapor in soil pores is a crucial link between the atmosphere and groundwater, and is also the first medium affected by potential pipeline leaks. If tritium accumulates in the soil, the existing network can only detect it indirectly through well water or air sampling, which involves a certain lag and uncertainty. Based on the above analysis, several areas with high uncertainty or lack of representativeness in tritium monitoring can be identified: (1) Downwind extension of the plant area: Northwest-North side in summer and South side in winter, based on the prevailing wind direction. There is no monitoring coverage in these directions within 2–5 km of the plant area, which are areas where the potential changes in tritium concentration are unknown. (2) Missing key media: Soil along the underground drainage ditch / pipeline network of the plant area. If a micro-leak occurs in the pipeline, tritium may be confined to the nearby soil and shallow groundwater, which currently lacks dedicated detection. (3) Surface runoff collection areas: For example, low-lying areas around the plant area or downstream rainwater collection points, where rainwater may carry tritium and accumulate, and there is currently no monitoring in these areas. (4) Weak areas for residential drinking water: Although there is monitoring at the water intake of S City (40 km away) and it has been normal, there are no monitoring points in the local groundwater or small water sources within 5–10 km of the plant area. Once tritium migrates out of the plant, it may affect these water sources, but it is difficult to detect in time.

[0168] Recommendations for the deployment of new monitoring points: To address the aforementioned blind spots and deficiencies, the following recommendations are made to optimize the monitoring network deployment, using uncertainty to drive improvements in monitoring capabilities:

[0169] Adding atmospheric water vapor monitoring points: A tritium water vapor sampling point will be added approximately 2 km downwind of the prevailing wind direction at the nuclear power plant site. For example, an air wet deposition monitoring station could be deployed 2 km north of the plant area (or added 2–3 km west of the other major prevailing wind direction). This will enhance coverage of the tritium diffusion path in the perennial prevailing wind direction, reducing the uncertainty of tritium concentration prediction in that direction by an estimated 30–50%. Simultaneously, since the Z-city monitoring point is 4 km from the plant area and only slightly above the background level, adding points closer to the site will help capture changes in the tritium concentration gradient outside the plant boundary, improving the sensitivity of the TRI response to abnormal emissions (early warning capability for tritium events).

[0170] Additional groundwater monitoring wells: In addition to existing monitoring wells, it is recommended to install 1–2 shallow monitoring wells in key directions, 500 m–1 km outside the plant boundary. For example, one groundwater monitoring well could be installed at the foot of the mountain to the northeast of the plant area and another in the plain to the northwest, to monitor tritium migration outside the plant area. Site selection should consider topography and groundwater flow direction, prioritizing wells in low-lying areas and catchment areas to increase the probability of tritium interception. These wells can supplement the limitations of existing wells, which are mainly located within the plant, and improve monitoring coverage of tritium migration outside the plant. Once the new wells detect tritium anomalies, the uncertainty of risk in the area will be significantly reduced. According to Kriging uncertainty analysis, each additional well can reduce the interpolation variance within a 1–2 km radius by approximately 20–40%, and multiple wells can significantly narrow the prediction confidence interval.

[0171] Establish additional soil gas monitoring points: Deploy monitoring points in soil media where tritium may accumulate near nuclear power plants. For example, bury soil gas sampling pipes every 100 m along the liquid effluent discharge pipeline of the nuclear power plant to periodically measure tritium in the soil air. Soil gas monitoring has high sensitivity and can detect minor pipeline leaks or anomalies in underground tritium accumulation at an early stage. It is recommended to focus on deploying monitoring points in the soil near well PR1 (as a pipe leak was previously suspected there) and the soil area around well E. This way, even if tritium has not yet entered the groundwater, early warning can be provided at the soil stage. The newly added soil gas monitoring will improve the entire media chain of atmosphere-soil-water, significantly reducing the uncertainty of unknown tritium sources. Especially for the "source term localization" model, it can provide more direct constraints and improve the accuracy of anomaly source identification.

[0172] Optimize marine and runoff monitoring: Given that liquid effluents primarily enter the sea, it is recommended to add a seawater sampling point near the southern shore of the plant area to specifically monitor the immediate impact of abnormal emissions on nearshore seawater, filling the blind spot in the near-plant sea area beyond the current 10 points at West D1. Additionally, simple sampling should be added at land-based rainwater runoff inlets, such as collecting samples of accumulated water at the plant's drainage outlets where they flow into the external environment, to monitor whether rainfall carries tritium out of the plant area. These measures will further close the monitoring loop for all possible tritium migration paths around the nuclear power plant.

[0173] Expected Enhancement in Monitoring Capabilities: Through the above optimizations, the spatial representativeness and early warning sensitivity of the monitoring network will be significantly improved. The newly added downwind air monitoring station can detect concentration increases within hours of an abnormal tritium emission, providing an earlier warning than the previous 4 km-away station and reducing uncertainty in concentration distribution estimation. The addition of groundwater wells and soil gas monitoring points expands the monitoring network's coverage of the nuclear power plant's periphery from the current approximately 1–2 km to a range of 3–5 km, significantly reducing interpolation errors in far-field tritium concentrations. In particular, soil gas monitoring enables direct detection of underground leakage pathways, issuing alarms before tritium enters groundwater, and triggering the early warning mechanism earlier when the response threshold of the TRI risk index is exceeded. Overall, the optimized network improves the comprehensive detection probability of abnormal tritium emission events from nuclear power plants, and the average relative uncertainty of tritium concentration field estimation is expected to decrease by more than 40%. A more comprehensive monitoring network will also continuously correct the tritium migration model (through a dynamic weight update mechanism combined with new data to correct predictions), forming a virtuous cycle of monitoring and prediction. Ultimately, this will enhance insights into the environmental behavior of tritium, ensuring that tritium levels around nuclear power plants remain under control in a way that the public can be assured of.

[0174] 3. Relevant calculation data are shown in Tables 1-7 and Figure 3

[0175]

[0176]

[0177]

[0178]

[0179]

[0180]

[0181]

[0182] 4. Calculation model code

[0183] # D1 Tritium Inversion and Monitoring Point Optimization Analysis:

[0184] import numpy as np

[0185] import matplotlib.pyplot as plt

[0186] from scipy.integrate import quad

[0187] from scipy.spatial.distance import cdist

[0188] from sklearn.gaussian_process import GaussianProcessRegressor

[0189] from sklearn.gaussian_process.kernels import RBF, ConstantKernel as C

[0190] # -------------------------

[0191] # 1. Gaussian plume model (ground concentration)

[0192] # -------------------------

[0193] def gaussian_plume_ground(Q, u, H, sigma_y, sigma_z):

[0194] return Q / (2 * np.pi * u * sigma_y * sigma_z) * np.exp(-H**2 / (2 *sigma_z**2))

[0195] # Example parameters (units are consistent)

[0196] Q = 1e11 / (24 * 3600)# Bq / s, releasing 1e11 Bq per day.

[0197] u = 2.0# Wind speed (m / s)

[0198] H = 50# Emission height (m)

[0199] sigma_y = 30# Diffusion coefficient m

[0200] sigma_z = 15

[0201] C_air_peak = gaussian_plume_ground(Q, u, H, sigma_y, sigma_z)

[0202] # -------------------------

[0203] # 2. Infiltration hysteresis model (groundwater response)

[0204] # -------------------------

[0205] def R_example(tau):

[0206] return 3.0# mm / h (simplified to a constant)

[0207] def gw_concentration(t, alpha=0.5, lambd=48):

[0208] integrand = lambda tau: R_example(tau) * np.exp(-(t - tau) / lambd)

[0209] result, _ = quad(integrand, 0, t)

[0210] return alpha * result

[0211] C_gw = gw_concentration(72) # Well water response after 72 hours

[0212] # -------------------------

[0213] #3. TRI Index

[0214] # -------------------------

[0215] def TRI(C_air, C_soil, C_gw, C_ref_air=4000, C_ref_soil=4000, C_ref_gw=100,

[0216] w_air=0.5, w_soil=0.25, w_gw=0.25):

[0217] f_air = C_air / C_ref_air

[0218] f_soil = C_soil / C_ref_soil

[0219] f_gw = C_gw / C_ref_gw

[0220] return w_air * f_air + w_soil * f_soil + w_gw * f_gw

[0221] C_soil = 0.5 * C_air_peak

[0222] tri_value = TRI(C_air_peak, C_soil, C_gw)

[0223] # -------------------------

[0224] # 4. Simplified Spatial Layout Optimization (Kriging Interpolation + Variance Evaluation)

[0225] # -------------------------

[0226] # Simulated monitoring point

[0227] X_known = np.array([[0, 0], [0, 500], [500, 0], [500, 500]])

[0228] y_known = np.array([C_gw, 8.5, 5.0, 7.0])

[0229] # Establishing a Gaussian process

[0230] kernel = C(1.0) * RBF(length_scale=500)

[0231] gp = GaussianProcessRegressor(kernel=kernel, alpha=0.5)

[0232] gp.fit(X_known, y_known)

[0233] # Grid Prediction

[0234] x = y = np.linspace(-500, 1000, 50)

[0235] X_grid = np.array([[i, j] for i in x for j in y])

[0236] y_pred, sigma = gp.predict(X_grid, return_std=True)

[0237] # Visualizing Prediction Uncertainty

[0238] plt.figure(figsize=(6,5))

[0239] plt.tricontourf(X_grid[:,0], X_grid[:,1], sigma, levels=14, cmap="YlOrRd")

[0240] plt.plot(X_known[:,0], X_known[:,1], 'ko', label='Existing monitoring points')

[0241] plt.colorbar(label='Predicted Standard Deviation (Bq / L)')

[0242] plt.title("Uncertainty Distribution of Monitoring Blind Zones")

[0243] plt.xlabel("x (m)")

[0244] plt.ylabel("y (m)")

[0245] plt.legend()

[0246] plt.tight_layout()

[0247] plt.show().

[0248] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0249] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A multi-media monitoring system for external tritium in a fusion reactor plant, characterized in that, The system includes an air tritium monitoring subsystem, which is used to collect tritium-containing water and send it to a liquid scintillation counting module to continuously measure the tritium activity concentration online, and automatically transmit the data to the environmental monitoring data center after each measurement. A water body tritium monitoring subsystem is used to automatically collect water samples and use electrolytic enrichment-liquid flashover technology to achieve continuous water sample monitoring and transmit the data to the environmental monitoring data center. A soil / biological tritium monitoring subsystem is used to periodically collect samples of soil, plants, aquatic organisms, etc. around the plant site, analyze the tritium activity concentration in them, and transmit the data to the environmental monitoring data center. An environmental monitoring data center is used to receive and store data uploaded by the air tritium monitoring subsystem, the water tritium monitoring subsystem, and the soil / biological tritium monitoring subsystem, and then convert the collected data into a unified risk indicator (TRI). An alarm and emergency response module is used to trigger a tiered early warning and push handling suggestions when the TRI exceeds a set threshold; wherein, the calculation formula for the TRI index is: in: C air C soil C gw These represent the measured values ​​of tritium in air, soil gas, and groundwater, respectively. M is a set of meteorological parameters; θ represents the soil moisture content; h represents the groundwater level; w air w soil w gw Let w be the weighting coefficients corresponding to the three media, and satisfy w air + w soil + w gw = 1; f air (·), f soil (·), f gw (·) is the cross-medium transformation function that converts tritium concentrations in various media into dimensionless risk indicators; In the event of extreme weather, the platform automatically adjusts its weighting: in: : Cumulative rainfall in the past 24 hours (mm) Trigger threshold (e.g., 30 mm) Rainfall – Weighting Gain Coefficient Maximum weight adjustment limit (e.g., +0.4) Air weight decreases synchronously: Soil air weight may be slightly adjusted or kept unchanged depending on the circumstances.

2. A multi-media monitoring system for external tritium in a fusion reactor plant, characterized in that, The system includes: a) Several off-site sampling nodes, including: a1) Air node: Equipped with a switchable pre-processing module and a tritium detection module to distinguish or simultaneously measure tritium in hydrogen and water vapor; a2) Soil gas node: Deployed at depths below the surface, equipped with soil gas sampling, humidity / temperature measurement and tritium detection modules; a3) Shallow groundwater node: installed in the monitoring well, equipped with a water sampling / online analysis module and water level and temperature sensors; b) A unified monitoring platform, including: b1) Quality control and calibration module: performs time synchronization, zero-point / span calibration, missing measurement compensation and anomaly removal for multi-node data; b2) Cross-media normalization module: converts air tritium concentration, soil tritium concentration and groundwater tritium activity into a unified risk index TRI; b3) Spatial Analysis and Backpropagation Module: Based on meteorological / hydrological parameters, spatial interpolation and source term location / impact range estimation are performed on anomaly points; b4) Adaptive node placement module: Updates suggestions for adding / migrating nodes based on TRI and boundary conditions; b5) Early Warning and Linkage Module: Triggers tiered early warnings based on set thresholds and pushes handling suggestions; wherein, the TRI is calculated according to the formula: The formula for calculating the TRI index is: in: C air C soil C gw These represent the measured values ​​of tritium in air, soil gas, and groundwater, respectively. M is a set of meteorological parameters; θ represents the soil moisture content; h represents the groundwater level; w air w soil w gw Let w be the weighting coefficients corresponding to the three media, and satisfy w air + w soil + w gw = 1; f air (·), f soil (·), f gw (·) is the cross-medium transformation function that converts tritium concentrations in various media into dimensionless risk indicators; In the event of extreme weather, the platform automatically adjusts its weighting: in: : Cumulative rainfall in the past 24 hours (mm) Trigger threshold (e.g., 30 mm) Rainfall – Weighting Gain Coefficient Maximum weight adjustment limit (e.g., +0.4) Air weight decreases synchronously: Soil air weight may be slightly adjusted or kept unchanged depending on the circumstances.

3. A method for multi-media monitoring of tritium outside a fusion reactor plant, applied to the multi-media monitoring system for tritium outside a fusion reactor plant as described in claim 1, characterized in that, include: S1: Simultaneously collect tritium content in air, water, soil, and biological environments, and label them as C. air C gw And C soil ; S2: According to the formula Calculate TRI; S3: Based on wind field, rainfall, and groundwater level, spatiotemporal interpolation and source term inverse estimation are performed on TRI to obtain the affected area; S4: Based on the TRI gradient and uncertainty, output the point placement optimization suggestions; S5: If TRI exceeds the threshold, trigger a tiered early warning and emergency response mechanism.

4. A unified monitoring platform device for multi-media tritium outside a fusion reactor plant, characterized in that, The unified monitoring platform device for tritium multi-media outside the fusion reactor plant includes a processor and a memory, wherein the memory stores a computer program that executes steps S1–S5 as described in claim 3.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method of claim 3.

Images