Abstract
Thermal pollution from nuclear power plant thermal discharge is a significant environmental concern in the sustainable development of nuclear energy. Airborne infrared remote sensing technology, with its unique advantages such as high sensitivity, high resolution, and quantitative inversion capabilities, has become an important monitoring method in this field. Based on the analysis of key performance indicators of airborne infrared observation of thermal discharge, this paper systematically reviews the development history and typical applications of airborne infrared monitoring of thermal discharge both domestically and internationally, and elaborates on core aspects such as radiometric calibration, geometric correction, and temperature inversion. Building an integrated space-air-ground collaborative monitoring network and deeply integrating artificial intelligence technology are inevitable trends in promoting the operationalization and intelligentization of nuclear power plant thermal discharge monitoring in China.
Introduction
Energy is the cornerstone of national economic and social development. Guided by the strategic goal of "carbon peaking and carbon neutrality", China's nuclear power industry is entering a period of rapid development. Currently, operational, quasi-construction, and planned nuclear power plants cover almost all coastal provinces, forming a large-scale coastal nuclear power belt, making significant contributions to ensuring national energy security and optimizing the energy structure.
However, during normal operation, nuclear power units need to discharge waste heat from the turbine into the surrounding waters. The resulting large volume of warm wastewater will inevitably lead to an increase in the water temperature of the nearby sea area, forming a significant temperature rise zone. This "thermal effect" constitutes one of the most concerning environmental issues in the development of nuclear power [1]. Although warm wastewater is different from chemical pollution, its continuous heat input will change the physicochemical properties of the water body, thereby having a profound impact on the aquatic ecosystem [2-3]. Therefore, continuous and accurate monitoring and evaluation of nuclear power warm wastewater is an indispensable link in coordinating nuclear power development and environmental protection and achieving sustainable development.
The methods currently used for monitoring nuclear power plant thermal discharge mainly include traditional monitoring methods (fixed-point measurement, ship measurement) , airborne aerial platform monitoring (manned aircraft, unmanned aircraft) and satellite platform observation (see Table1) . The most widely used remote sensing monitoring of thermal discharge is spaceborne infrared imaging. Many infrared data such as NOAA (AVHRR) [4], MODIS, Landsat (TM/ETM+) [5-6], and SDGSAT-1 (TIS) [7] have been used by scholars at home and abroad. However, the resolution of tens of meters to kilometers is not enough to accurately depict the spatial morphological characteristics of thermal discharge, and the overhead time of sun-synchronous orbit satellites is fixed, which cannot meet the needs of high-efficiency monitoring. Therefore, since the United States began to use airborne infrared cameras to carry out nuclear power plant thermal discharge monitoring in the1970s [8-9], various airborne infrared payloads have emerged [10-12]. With the development of unmanned aerial vehicle technology, lightweight uncooled infrared cameras have also been used [13-15].
Table1Comparison of different monitoring methods
This article will systematically review the development history and current application status of nuclear power plant thermal discharge monitoring using manned airborne infrared technology at home and abroad, explain its basic principles and key technical aspects, summarize the achievements, shortcomings and future development trends in this field, and aim to provide scientific reference and technical support for the operationalization, refinement and intelligentization of nuclear power plant thermal discharge environmental supervision.
1 Analysis of the performance indicators of airborne infrared observation of thermal discharge
1.1 Noise Equivalent Temperature Difference and Dynamic Range
The noise equivalent temperature difference (NETD) determines the temperature sensitivity of the infrared system and directly affects the accuracy of the temperature discharge inversion; while the dynamic range defines the temperature range that the system can measure and determines its monitoring applicability. GB 3097—1997 "Seawater Quality Standard" stipulates that for Class I and II seawater, "the temperature rise of seawater caused by human activities shall not exceed 1℃ in summer and 2℃ in other seasons"; for Class III and IV seawater, "the temperature rise of seawater caused by human activities shall not exceed 4℃ in summer and 4℃ in other seasons". The water quality standards in the United States vary from state to state, but usually require that the upper limit of surface water temperature in summer be32℃, while the temperature discharge is allowed to exceed the maximum temperature by 2~3℃ [16]. The UK Environment Agency stipulates that the temperature difference between the inlet and outlet should be less than 8℃.
Wei J et al. [5] conducted long-term monitoring of the thermal discharge from 66 nuclear power plants worldwide, showing that the average maximum surface temperature increase of nuclear power plants was 4.80 K, while the highest temperature rise value of Tianwan Nuclear Power Plant in my country was 8.51 K.
Currently, the accuracy of surface water temperature measurement at fixed points is generally better than 0.1 K. Under conditions of an average temperature rise of 5 K, a 0.1 K resolution system can only record 50 temperature difference values. High-sensitivity infrared systems, however, can capture much more temperature detail. According to NETD's calculation formula:
(1)
In the formula, T and T0 are blackbody temperatures; VS is the pixel response voltage; and VN is the pixel noise voltage.
If NETD is increased to 20 mK, then the temperature difference of 5 K can capture at least 250 temperature variations more precisely. For high-precision cooled infrared systems, VN (i，j) can reach 6; considering a maximum temperature difference of at least 10 K, Vs (i，j) should be greater than 3000, that is, the dynamic range of the digital number ( DN) value should be greater than 3000; considering different seasons and temperatures, the actual dynamic range of the DN value should be even larger.
1.2 Spatial Resolution
Spatial resolution is a key indicator for measuring the detail rendering capability of remote sensing images, directly determining the accuracy of target feature identification and classification. For nuclear power plant thermal discharge monitoring applications, the choice of resolution needs to be closely integrated with the actual application scenario. The most intuitive aspects of nuclear power plant thermal discharge monitoring are the inlet and outlet. Measurement results using ArcGIS Earth software on targets of interest in typical domestic and international nuclear power plants show that the barrier net at the inlet is typically wider than 2 m, the width of the guide dike above the water surface is generally greater than 15 m, and the width of the inlet and outlet is typically greater than 50 m. According to the Johnson criterion, detection, identification, and confirmation of targets require at least 1.5, 6, and 12 pixels, respectively. To more finely distinguish mixed areas, the resolution should be better than 3 m; under the condition of an instantaneous field of view of 1 mrad, the flight altitude can be set to 2000－3000 m.
1.3 Temporal Resolution
Temporal resolution refers to the time interval between repeated observations of the same area by a remote sensing platform, which directly affects the accuracy of acquiring characteristic tidal images of thermal discharge. To assess the impact range, NB/T20299—2014 "Technical Specification for Environmental Impact Assessment of Thermal Discharge from Nuclear Power Plants" and HJ 1037—2019 "Guidelines for Environmental Impact Assessment of Intake and Discharge from Nuclear Power Plants (Trial) " stipulate that after nuclear power plant operation, monitoring of the flow field and temperature field (planar and vertical) distribution under typical tidal patterns (spring, mid, and neap tides) and semi-lunar tidal patterns in winter and summer should be conducted. Airborne remote sensing, with its flexible mobility, can dynamically monitor characteristic tidal times, effectively compensating for the limitations of fixed satellite overpass times and meeting the key requirement of temporal resolution.
The typical monitoring area for thermal discharge is greater than 200 km². At an altitude of 2000 m and a field of view of 70°, the swath width can reach 2800 m, and with 30% overlap, approximately 0.5 hours of coverage can be achieved. Considering that the characteristic tidal time observed is usually 1－2 hours, under these flight conditions, at least two coverages of each of the two characteristic tidal times can be achieved in a single sortie, ensuring backup of effective tidal times and efficient data acquisition.
2 Aerial cameras and applications for monitoring thermal drainage
2.1 Typical Infrared Imaging System
High-performance infrared imaging systems are indispensable for monitoring thermal discharge from nuclear power plants. Since the1970s, various types of airborne remote sensing equipment suitable for thermal infrared monitoring of oceans and water bodies have been developed both domestically and internationally. These systems have been continuously optimized in key performance aspects such as spatial resolution, temperature sensitivity, and field of view, propelling thermal discharge monitoring from qualitative description to quantitative inversion.
Table2Typical infrared imagers used for monitoring thermal drainage both domestically and internationally
* The OMIS-1 contains eight long-wave infrared bands, while the OMIS-2 contains one long-wave infrared band.
2.2 Operational Modular Imaging Spectrometer
The Operational Modular Imaging Spectrometer (OMIS) is the earliest imaging spectrometer in China: the OMIS-1 contains eight long-wave infrared bands, and the OMIS-2 contains one long-wave infrared band. From August to September 2001, the OMIS-1 imaging spectrometer was used to collect data on the thermal discharge from the Futsu Power Plant in Tokyo Bay, Japan (see Figure1) . Although this was conventional thermal power plant monitoring, it laid the foundation for subsequent monitoring of thermal discharge from nuclear power plants in China.
Fig.1: Temperature classification of thermal discharge from Fujin Power Plant [17].
2.3 Marine Airborne Multi-spectrum Scanner
The Marine Airborne Multi-spectrum Scanner (MAMS) is a professional imager for marine applications. Its infrared spectrum includes two channels: mid-wave infrared channel and long-wave infrared channel. MAMS has carried out thermal discharge monitoring tasks at Qinshan Nuclear Power Plant and Tianwan Nuclear Power Plant for several years. As shown in Figure2, the temperature rise in the core area of Tianwan Nuclear Power Plant is basically consistent with the research results of Wei J et al. [5]
Fig.2: Results of thermal discharge monitoring at Tianwan Nuclear Power Plant: (a) Infrared water temperature retrieval results; (b) Temperature rise extraction map.
2.4 Airborne Bispectral Scanner
Airborne Bispectral Scanner (ABS) is an ultraviolet/infrared dual-band imager introduced in China. The instrument has a large field of view of 90° and a built-in calibration blackbody. It has been used in the monitoring of thermal wastewater at Qinshan Nuclear Power Plant (see Figure3) . It acquired imaging data of eight tidal states in one observation cycle and realized dynamic monitoring [12].
Fig.3: The temperature rise during neap tides, which was inverted using ABS data [12].
2.5 Wide-field Scanning Infrared Water Temperature Imager
Wide-field Scanning Infrared Water Temperature Imager (WSIR-WTI) is a new generation infrared imager developed to meet the monitoring needs of nuclear power plant thermal discharge operations. It is a lightweight, compact, wide-field-of-view, and highly sensitive scanning imaging instrument, weighing only 5 kilograms. The instrument weighs approximately 5 kg (see Figure4) . It employs a planar scanning scheme, achieving large field-of-view and high-sensitivity imaging; its operating efficiency can reach 500 km2/h (2 m@2 km) , and its temperature sensitivity can reach up to 20 mK; combined with ground-based synchronous measurement data, the accuracy of sea surface temperature retrieval can reach 0.3℃. For several years, temperature monitoring has been conducted in the sea areas of nearly 10 nuclear power plants, including Tianwan, Qinshan, and Sanmen. Figure5 shows the effect of temperature monitoring on the Qinshan nuclear power plant. The monitoring area in the left figure is approximately 220 km², with 5 routes and a single coverage time of approximately 40 minutes using30% lateral overlap. The right figure shows a comparison of dynamic monitoring under different tidal conditions.
Fig.4: Actual image of WSIR-WTI
Fig.5: Application effect of WSIR-WTI in thermal drainage monitoring
3 Key Technologies
3.1 Wide field-of-view, high resolution imaging
A wide field-of-view and high resolution are fundamental to the efficient and accurate capture of spatial details in warm water plumes by airborne infrared surveillance. While early unit scanning methods offered a large field of view, their low resolution limited the operating altitude, severely impacting operational efficiency. Currently, mainstream area array scanning imaging achieves a balance between high resolution, high sensitivity, and a large field of view. However, ensuring image quality requires precise scan control and high-precision data post-processing as essential support.
3.2 Radiometric calibration
Radiometric calibration is a crucial step in establishing a quantitative relationship between the raw digital quantized values output by the sensor and the spectral radiance values at the entrance pupil, and it is a prerequisite for all quantitative remote sensing analyses. Airborne infrared systems commonly employ onboard blackbody calibration schemes to ensure absolute data accuracy. Traditional optomechanical scanning achieves real-time calibration by deploying low-temperature and high-temperature blackbodies on both sides; WSIR-WTI uses an internal surface-source variable-temperature blackbody, with blackbody temperature control employing a TEC heating/cooling composite control, thereby achieving high-precision internal calibration.
3.3 Correction and image stitching
Single infrared images have limited coverage and are susceptible to geometric distortion due to changes in flight attitude. Geometric correction is performed using high-precision position and attitude data acquired by an airborne Position and Orientation System (POS) . Pixel-level image stitching is then achieved using high-precision camera calibration data or a base map. However, overlapping areas often exhibit radiance differences, necessitating radiance fusion algorithms such as color balancing and feathering to eliminate seams and generate a seamless, large-area temperature radiance map.
3.4 Temperature retrieval
Temperature retrieval is the ultimate goal of the entire technology chain. Thermal infrared sensors measure the radiance of the top layer of the atmosphere. To obtain sea surface temperature, the influence of the atmosphere must be removed from this radiance. Atmospheric correction is the most critical and complex step in temperature retrieval.
(2)
In the formula, λ is the center wavelength; L (λ) is the radiance received by the sensor; Ts is the physical temperature; B (Ts, λ) is the blackbody radiance corresponding to temperature Ts; ε (λ) is the emissivity at wavelength λ; t (λ) is the atmospheric transmittance at wavelength λ; L↓ (λ) is the downward atmospheric radiance; L↑ (λ) is the upward atmospheric radiance.
In the thermal infrared band, the sea surface emissivity can be a constant of 0.98 to 0.99. Based on the atmospheric conditions, the measured relative humidity, air temperature, and visibility atmospheric parameters are input into software such as MODTRAN to calculate t (λ) , L↓ (λ) , and L↑ (λ) . Then, the sea surface temperature value is retrieved pixel by pixel according to Equation (1) to generate an intuitive and quantifiable sea surface temperature distribution map, from which various characteristic parameters of thermal discharge can be accurately extracted.
4 Challenges, development trends and prospects
While airborne infrared remote sensing technology has demonstrated significant advantages in monitoring nuclear power plant thermal discharge, we must be soberly aware that no single technology can solve all problems. Currently, airborne infrared remote sensing of nuclear power plant thermal discharge is evolving from a "point" (ground-based) approach to a "surface" (airborne/satellite) approach, and further towards "real-time, precise, and flexible" monitoring, ultimately aiming to build a highly efficient and collaborative "integrated air-space-ground" monitoring network. In this process, we face both severe challenges and significant development opportunities.
4.1 Main challenges
With the increasing demand for refined and quantitative monitoring, the requirements for the core performance of airborne infrared sensors are becoming increasingly stringent. Currently, further improving the temperature sensitivity, spatial resolution, and dynamic range of infrared imagers while maintaining a wide field-of-view still faces technical challenges.
The near-surface atmosphere is extremely complex, and the radiation signals received by airborne infrared sensors are greatly affected by the atmosphere. Therefore, breaking through the bottleneck of high-precision atmospheric correction is the core challenge to improving the accuracy of temperature inversion.
The research and development and procurement costs of manned airborne remote sensing platforms and high-performance research-grade infrared sensors are relatively high. The airspace application and flight support involved in mission execution place enormous pressure on routine, high-frequency operational monitoring, limiting the wider adoption and application of this technology. Therefore, there is an urgent need to develop lightweight, low-power, high-performance cooled infrared sensors to achieve low-cost, high-frequency operational monitoring.
4.2 Development trends and prospects
To address these challenges, future technological development will focus on instrument upgrades, intelligent collaboration, and in-depth applications.
The future monitoring network will inevitably be a collaborative application of "air, space, and ground integration". However, as the primary component of the entire monitoring process, the performance of high-performance infrared sensors is a prerequisite for determining the quality of the final data products. On the one hand, for manned platforms, we should focus on high-end research-grade sensors to further break through the limits of sensitivity and resolution. On the other hand, we need to vigorously develop lightweight, low-cost, and intelligent infrared payloads adapted to UAVs to achieve synergistic optimization between the platform and the sensor.
Artificial intelligence (AI) technology will revolutionize the processing and analysis of massive amounts of remote sensing data. Utilizing intelligent models (such as machine learning and deep learning) will drive the development of high-precision atmospheric correction models. Combined with hydrological and meteorological data, AI-based warm discharge diffusion prediction models can be constructed to forecast the dynamics of plumes over the next few hours, providing forward-looking information for intelligent power plant control. Based on data assimilation technology, monitoring results will be transformed from isolated "temperature image data" into fused information revealing the intrinsic connections between "thermal discharge transport, water quality parameter changes, and ecological responses, " and further elevated into a systematic knowledge base capable of quantitatively assessing risks and supporting management decisions. This transformation enables monitoring work to directly serve nuclear power environmental safety early warning, ecological damage assessment, and marine environmental protection management, thereby achieving harmonious coexistence between nuclear power and the environment.
5 Conclusion
Airborne infrared remote sensing has become a core technology for the refined and operational monitoring of nuclear power plant thermal discharge. This article systematically reviews the application and development of airborne infrared remote sensing technology in nuclear power plant thermal discharge monitoring. Currently, airborne infrared imaging technology has entered a development stage characterized by large-area arrays and high resolution. Under this trend, the core performance of several systems independently developed in China has reached international advanced levels. However, high-precision radiometric calibration and atmospheric correction remain key technical challenges for achieving accurate temperature field inversion. In the future, China's airborne infrared monitoring of nuclear power plant thermal discharge should focus on developing high-sensitivity, high-resolution, large-field-of-view infrared cameras, building a multi-platform collaborative monitoring network, promoting intelligent technology, deepening data fusion and application, and ultimately providing stronger technical support for nuclear power plant environmental safety and management.