An ultra-junction rad-fugal nuclear battery performance data processing system based on an iterative algorithm

Abstract

The application provides an ultra-junction rad-fugal nuclear battery performance data processing system based on an iterative algorithm, and relates to the technical field of data processing.The method comprises the following steps: obtaining a radiation source layer basic parameter, a transducer device layer basic parameter and performance layer data of an ultra-junction rad-fugal nuclear battery; organizing the performance layer data into a data distribution structure; processing a data point set based on the data distribution structure through a minimum circumscribed rectangle algorithm to obtain a minimum circumscribed rectangle covering all data points, and taking a geometric center of the minimum circumscribed rectangle as a reference data node; determining two main data change directions according to the reference data node, and calculating data correlation between the two directions to obtain a core data interval.The application realizes regulation and control and long-term stable optimization of the performance of the ultra-junction rad-fugal nuclear battery.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a superjunction radiation-voltaic nuclear battery performance data processing system based on an iterative algorithm. Background Technology

[0002] Superjunction radiation-voltaic nuclear batteries, relying on the core advantages of trench structure to optimize electric field distribution and suppress carrier recombination, have become a key power supply solution for long-life extreme scenarios such as deep space exploration and implantable medical devices. They require nuclear batteries to have high energy conversion efficiency and extremely low long-term decay rate. Furthermore, they need to achieve precise parameter matching between typical radiation sources and suitable transducer devices through data processing. However, existing data processing methods mostly rely on static physical models, which can only fit the correlation between energy and power based on initial performance data, without considering the coupling analysis of dynamic decay process and process boundary conditions.

[0003] For example, when a certain aerospace team was developing a superjunction radiation-voltaic nuclear battery for a Mars probe, they used traditional processing methods to establish a linear model based solely on initial measured energy deposition and output power to determine the trench structure and doping parameter combination of the transducer. However, they did not take into account the natural decay of the radioactive source activity over time, nor did they consider the deviation between the actual process precision and design requirements during MEMS fabrication. Ultimately, this resulted in good performance stability in the laboratory during short-term verification, but after the probe had been running in orbit for a period of time, the actual performance of the battery deviated significantly from the design expectations, and the carrier collection efficiency also decreased, failing to meet the long-term stable power supply requirements during the mission cycle. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a superjunction radiation-voltaic nuclear battery performance data processing system based on iterative algorithm, so as to realize the regulation and long-term stable optimization of the performance of superjunction radiation-voltaic nuclear batteries.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] Firstly, a method for processing performance data of superjunction radiation-voltaic nuclear batteries based on an iterative algorithm, the method comprising:

[0007] To obtain the basic parameters of the radiation source, the basic parameters of the transducer, and the performance data of the superjunction radiation-voltaic nuclear battery;

[0008] The performance layer data is organized into a data distribution structure. Based on the data distribution structure, the data point set is processed by the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the reference data node, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval.

[0009] Based on the core data range, basic parameters at the radiation source level, and basic parameters at the transducer level, a nonlinear coupling relationship is established to describe the complete link of radiation source ray energy distribution, trench electric field distribution, carrier collection efficiency, and output power.

[0010] Based on the nonlinear coupling relationship, dynamic performance data of the long-term decay process of nuclear batteries are introduced as iterative variables for iterative optimization, and multiple dynamic performance feature points are selected based on the core data interval.

[0011] A dynamic performance trajectory is obtained based on the evolution relationship of each feature point in the time series, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained.

[0012] The final parameter combination is taken as input, and the process boundary conditions for MEMS miniaturization are introduced as constraints for adaptation correction to obtain the final optimized parameter combination.

[0013] Furthermore, the fundamental parameters at the radiation source level, the fundamental parameters at the transducer level, and the performance data of the superjunction radiation-voltaic nuclear battery will be obtained, including:

[0014] Call the communication interface with the radioactive source characteristic database to obtain basic parameters at the radioactive source level, including isotope type, emitted particle type, energy spectrum distribution, and source activity;

[0015] Based on the fundamental parameters at the radiation source level, the transducer device structure definition file that matches the characteristics of the radiation source is received and parsed, and the fundamental parameters at the transducer device level, including trench geometry, doping profile, material composition and electrode configuration, are extracted from the definition file.

[0016] By combining the fundamental parameters at the radiation source level and the fundamental parameters at the transducer level, and through an integrated data acquisition service, performance layer data corresponding to the parameter combinations is selectively obtained from the historical database.

[0017] Furthermore, the performance layer data is organized into a data distribution structure. Based on this structure, the data point set is processed using the minimum bounding rectangle algorithm to obtain a minimum bounding rectangle covering all data points. The geometric center of this minimum bounding rectangle is used as the reference data node. Two main data change directions are determined based on the reference data node, and the data correlation between these two directions is calculated to obtain the core data interval, including:

[0018] Receive performance layer data and organize the performance layer data into a data distribution structure in a multi-dimensional space, where each data point corresponds to a set of performance measurement values;

[0019] Based on the data distribution structure, the data point set is processed by the minimum bounding rectangle algorithm to construct a minimum bounding rectangle that covers all data points, and the geometric center of the minimum bounding rectangle is used as the reference data node.

[0020] Using the baseline data node as a reference, the two main data change directions with the largest variance in the data distribution are determined by the eigenvector analysis method;

[0021] Calculate the data correlation between two main data change directions, and determine the core data interval of the data distribution based on the correlation metric.

[0022] Furthermore, based on the core data range, fundamental parameters at the radiation source level, and fundamental parameters at the transducer level, a nonlinear coupling relationship describing the complete link of radiation source energy distribution, trench electric field distribution, carrier collection efficiency, and output power is established, including:

[0023] Based on the core data range and combined with the basic parameters at the radioactive source level, the energy deposition distribution relationship of radioactive source rays in the transducer is established.

[0024] Based on the energy deposition distribution relationship, and combined with the trench geometry and doping profile in the basic parameters of the transducer, the electric field distribution relationship in the trench region is established.

[0025] Based on the electric field distribution relationship, the relationship between the electric field and the carrier transport and collection efficiency is established.

[0026] By integrating the relationships of energy deposition distribution, electric field distribution, and carrier collection efficiency, a complete nonlinear coupling relationship from the radiation source's ray energy to the output power is constructed.

[0027] Furthermore, based on the nonlinear coupling relationship, dynamic performance data of the long-term decay process of the nuclear battery is introduced as an iterative variable for iterative optimization. Multiple dynamic performance feature points are selected based on the core data interval, including:

[0028] The complete nonlinear coupling relationship is used as the basic physical relationship framework. Under the basic physical relationship framework, dynamic performance data of the long-term decay process of nuclear battery is introduced. The dynamic performance data includes the decay sequence of performance parameters at different time points.

[0029] Based on the core data range, the introduced dynamic performance data is filtered and processed to obtain multiple dynamic performance feature points that characterize different decay stages.

[0030] Multiple dynamic performance feature points are used as iterative variables and input into the basic physical relationship framework to establish an iterative optimization mechanism;

[0031] By executing the established iterative optimization mechanism, multiple rounds of performance optimization search are performed within the framework of basic physical relationships, and the optimized set of dynamic performance feature points is output.

[0032] Furthermore, a dynamic performance trajectory is obtained based on the evolution relationship of each feature point over time, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that satisfies the preset target performance parameters is obtained, including:

[0033] Receive the optimized set of dynamic performance feature points, and sort and organize the feature points according to the time series.

[0034] Based on the sorted feature point sequence, a dynamic performance trajectory reflecting the evolution of nuclear battery performance over time is constructed using curve fitting methods.

[0035] Analyze the changing patterns of dynamic performance trajectories, and based on these patterns, obtain performance correction factors for revising theoretical models.

[0036] The performance correction factor is applied to the complete nonlinear coupling relationship to perform the first correction on the nonlinear coupling relationship, resulting in the corrected nonlinear coupling relationship.

[0037] Based on the corrected nonlinear coupling relationship, the iterative optimization process is repeated until the output parameter combination meets the preset target performance parameter requirements, thus obtaining the final parameter combination.

[0038] Furthermore, using the final parameter combination as input, the process boundary conditions for MEMS miniaturization are introduced as constraints for adaptation correction, resulting in the final optimized parameter combination, including:

[0039] Based on the final parameter combination, process boundary conditions for MEMS miniaturization are introduced, including minimum processing size, maximum aspect ratio and material compatibility requirements.

[0040] The final parameter combination is compared and analyzed with the process boundary conditions to identify parameter items that exceed the process range, so as to obtain the comparison and analysis results.

[0041] Based on the comparative analysis results, the parameters that are outside the feasible range of the process are modified for adaptability, and the modified parameter combinations are obtained.

[0042] The parameter combination after adaptation correction is verified to meet the preset target performance requirements while satisfying the process boundary conditions, thus obtaining the final optimized parameter combination.

[0043] Secondly, a superjunction radiation-voltaic nuclear battery performance data processing system based on an iterative algorithm includes:

[0044] The acquisition module is used to acquire basic parameters of the radiation source layer, basic parameters of the transducer layer, and performance layer data of the superjunction radiation photovoltaic cell.

[0045] The calculation module organizes the performance layer data into a data distribution structure. Based on the data distribution structure, it processes the data point set using the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the reference data node, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval.

[0046] A module is established to create a nonlinear coupling relationship that describes the complete link of radiation source energy distribution, trench electric field distribution, carrier collection efficiency, and output power, based on core data ranges, basic parameters at the radiation source level, and basic parameters at the transducer level.

[0047] The optimization module is used to iteratively optimize the nonlinear coupling relationship by introducing the dynamic performance data of the long-term decay process of the nuclear battery as the iterative variable. It selects multiple dynamic performance feature points based on the core data interval. According to the evolution relationship of each feature point in the time series, a dynamic performance trajectory is obtained, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained.

[0048] The processing module takes the final parameter combination as input, introduces the process boundary conditions for MEMS miniaturization as constraints for adaptation correction, and obtains the final optimized parameter combination.

[0049] Thirdly, a computing device includes:

[0050] One or more processors;

[0051] A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method.

[0052] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0053] The above-described solution of the present invention has at least the following beneficial effects:

[0054] By employing multi-source data fusion to acquire data from the radiation source, transducer, and performance layer, and extracting core data intervals through minimum bounding rectangle and eigenvector analysis, a full-link nonlinear coupling relationship from the radiation source's ray energy distribution to its output power is constructed. Furthermore, long-term decay dynamic data is introduced as an iterative variable for trajectory correction and optimization. Simultaneously, the technique of combining MEMS process boundary conditions for adaptability constraints overcomes the technical problems of existing data processing methods that rely on static models, lack dynamic decay and process limitations, leading to large deviations between actual nuclear battery performance and design expectations, insufficient long-term stability, and difficulty in engineering parameter combinations. This improves the targeting and accuracy of data processing, enables precise prediction and dynamic optimization of long-term nuclear battery performance, and ensures that the final parameter combination meets both preset performance targets and adapts to actual process conditions, thereby enhancing the performance stability and engineering application reliability of superjunction radiation-voltage nuclear batteries in long-life extreme scenarios. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating a method for processing performance data of a superjunction radiation-voltaic nuclear battery based on an iterative algorithm, as provided in an embodiment of the present invention.

[0056] Figure 2 This is a schematic diagram of a superjunction radiation-voltaic nuclear battery performance data processing system based on an iterative algorithm, provided by an embodiment of the present invention. Detailed Implementation

[0057] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0058] like Figure 1 As shown, an embodiment of the present invention proposes a method for processing performance data of superjunction radiation-voltage nuclear batteries based on an iterative algorithm. The method includes the following steps:

[0059] Step 1: Obtain the basic parameters of the radiation source, the basic parameters of the transducer, and the performance data of the superjunction radiation-voltaic nuclear battery.

[0060] Step 2: Organize the performance layer data into a data distribution structure. Based on the data distribution structure, process the data point set using the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. Use the geometric center of the minimum bounding rectangle as the reference data node. Determine two main data change directions based on the reference data node and calculate the data correlation between the two directions to obtain the core data interval.

[0061] Step 3: Based on the core data range, basic parameters at the radiation source level, and basic parameters at the transducer level, establish a nonlinear coupling relationship describing the complete link of radiation source ray energy distribution, trench electric field distribution, carrier collection efficiency, and output power.

[0062] Step 4: Based on the nonlinear coupling relationship, the dynamic performance data of the long-term decay process of the nuclear battery is introduced as the iterative variable for iterative optimization, and multiple dynamic performance feature points are selected based on the core data interval.

[0063] Step 5: Based on the evolution relationship of each feature point in the time series, a dynamic performance trajectory is obtained, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained.

[0064] Step 6: Using the final parameter combination as input, introduce the process boundary conditions for MEMS miniaturization as constraints for adaptation correction to obtain the final optimized parameter combination.

[0065] In this embodiment of the invention, because multi-source data is used to acquire basic data on the radiation source, transducer, and performance layer, a data distribution structure is constructed and core data intervals are extracted through the minimum bounding rectangle algorithm and eigenvector analysis, and a full-link nonlinear coupling relationship from radiation source ray energy to output power is established based on the core intervals, and then the long-term decay dynamic data of the nuclear battery is introduced as an iterative variable for optimization and trajectory correction, and finally, the parameter combination is adapted to the boundary conditions of MEMS miniaturization process. Therefore, this invention overcomes the technical problems of existing superjunction radiation-voltaic nuclear batteries, which rely on static models for performance data processing, lack coupling analysis of dynamic decay processes and actual process limitations, resulting in large deviations between actual battery performance and design expectations, insufficient long-term operational stability, and difficulty in implementing parameter combinations in engineering applications. This achieves the technical effect of improving the targeting of data processing and the accuracy of models, realizing the prediction and dynamic optimization of the long-term performance of nuclear batteries, ensuring that the final parameter combination meets both the preset performance target and adapts to the actual processing technology, and enhancing the performance reliability and engineering application value of superjunction radiation-voltaic nuclear batteries in long-life extreme scenarios.

[0066] In a preferred embodiment of the present invention, step 1 above may include:

[0067] Step 1.1: Call the communication interface with the radioactive source characteristic database to obtain basic parameters at the radioactive source level, including isotope type, emitted particle type, energy spectrum distribution, and source activity. Specifically, this includes: calling the radioactive source characteristic database interface, which stores various isotope characteristic data, through a pre-configured communication link. The database contains detailed physical characteristic records of different isotopes. After the interface connection is established, send a parameter query request adapted to the Mars exploration scenario to the database, focusing on obtaining basic parameters at the radioactive source level that match the long-life power supply requirements, including isotope types suitable for long-term energy release, the type of emitted particles of the isotope, the energy spectrum distribution of the emitted particles, and the source activity value that meets the long-term power supply requirements of the detector.

[0068] Step 1.2: Based on the fundamental parameters at the radiation source level, receive and parse the transducer device structure definition file that matches the characteristics of the radiation source. Extract the fundamental parameters at the transducer device level, including trench geometry, doping profile, material composition, and electrode configuration, from the definition file. Specifically, after obtaining the fundamental parameters at the radiation source level, and considering the characteristics of the radiation source, receive the transducer device structure definition file that matches the characteristics of the radiation source from the transducer device design storage unit. The file is pre-compiled by the transducer device design team according to the energy conversion requirements of different radiation sources and contains complete structural information of the transducer device. The structure definition file is disassembled and analyzed using a dedicated file parsing tool to extract the fundamental parameters at the transducer device level, specifically including the trench geometry adapted to β-ray energy deposition, the doping profile conducive to carrier transport, the material composition suitable for energy conversion, and the electrode configuration to ensure charge output.

[0069] Step 1.3: Combining the fundamental parameters at the radioactive source level and the fundamental parameters at the transducer level, the integrated data acquisition service is used to selectively acquire performance layer data corresponding to the parameter combinations from the historical database. Specifically, this includes: after acquiring the fundamental parameters at the radioactive source level and the transducer level, the two types of parameters are correlated and integrated to form parameter combinations that include specific isotope types, particle types, source activity, specific trench sizes, doping profiles, and material compositions; then, the integrated data acquisition service is activated. The service has a parameter combination matching function and uses a preset correlation algorithm to search the historical database storing past research and testing data of superjunction radiation-voltaic nuclear batteries; during the search, the integrated parameter combinations are used as filtering conditions to selectively search for past case data that are consistent or highly similar to the parameter combinations in terms of radioactive source characteristics and device structure, and finally, the performance layer data corresponding to these cases are extracted, including key performance indicators such as battery output power, carrier collection efficiency, and energy conversion efficiency.

[0070] In this embodiment of the invention, by employing technical means such as calling the radioactive source characteristic database interface to obtain basic parameters of the radioactive source, extracting device parameters from the transducer structure file based on the radioactive source characteristic analysis and matching, and combining the two types of parameters to obtain corresponding performance layer data from the historical database through integrated data acquisition services, the technical problems of lack of correlation and matching between radioactive sources, transducers and performance data, scattered and inefficient parameter acquisition, and insufficient data targeting in the existing superjunction radiation-voltaic nuclear battery data acquisition process are overcome. Thus, the accurate correlation and targeted acquisition of multi-source data are achieved, improving the efficiency and accuracy of data acquisition.

[0071] In a preferred embodiment of the present invention, step 2 above may include:

[0072] Step 2.1: Receive performance layer data and organize it into a multi-dimensional data distribution structure. Each data point corresponds to a set of performance measurement values. Specifically, this includes: first, receiving past case performance data obtained from historical databases, as well as measured performance data under the current R&D plan. The performance layer data covers key indicators such as battery output power, carrier collection efficiency, and energy conversion efficiency. Then, based on the performance requirements of the Mars probe for long-term stable power supply from the nuclear battery, organize this performance layer data into a multi-dimensional data distribution structure. For example, construct a three-dimensional data space with output power as the first dimension, carrier collection efficiency as the second dimension, and energy conversion efficiency as the third dimension. Each data point corresponds to a set of performance measurement values ​​obtained from a complete performance measurement. For example, a certain data point corresponds to a specific combination of output power, carrier collection efficiency, and energy conversion efficiency measured under a certain test condition.

[0073] Step 2.2: Based on the data distribution structure, the data point set is processed using the minimum bounding rectangle algorithm to construct a minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the benchmark data node. Specifically, this includes: based on the constructed multidimensional spatial data distribution structure, the minimum bounding rectangle algorithm is used to process the data point set consisting of all performance data points contained therein; taking a three-dimensional data space as an example, the algorithm calculates the minimum rectangular space that completely covers all three-dimensional data points, i.e., a three-dimensional cuboid. Each side of this minimum rectangular space is parallel to the coordinate axes of the data space and just touches the outermost data point in the data point set; after determining this minimum bounding rectangle covering all data points, the geometric center of the rectangle is calculated. The set of performance parameter values ​​corresponding to the geometric center in the multidimensional space is the benchmark data node for performance analysis. For example, the output power, carrier collection efficiency, and energy conversion efficiency corresponding to this center can be used as a reference benchmark to judge the degree of deviation of the performance of other data points.

[0074] Step 2.3, using the baseline data node as a reference, determines the two main data change directions with the largest variance in the data distribution through eigenvector analysis. Specifically, this includes: using the determined baseline data node as a reference, analyzing the data distribution in multidimensional space using eigenvector analysis; firstly, calculating the deviation of all performance data points from the baseline data node, constructing the data covariance matrix, and finding the two eigenvectors with the largest eigenvalues ​​by solving for the eigenvalues ​​and eigenvectors of the covariance matrix; the directions corresponding to the two eigenvectors are the two main data change directions with the largest variance in the data distribution. For example, one direction may correspond to the trend of output power decreasing with the decay of radioactive source activity, and the other direction may correspond to the trend of carrier collection efficiency decreasing with the deviation of MEMS processing trench size. These two directions reflect the most important change patterns of performance data.

[0075] Step 2.4: Calculate the data correlation between the two main data change directions, and determine the core data interval of the data distribution based on the correlation metric. Specifically, this includes: for the two determined main data change directions, calculating the correlation between the performance data in the two directions. For example, by analyzing whether there is a synchronous increase or decrease relationship between the change in output power in one direction and the change in carrier collection efficiency in another direction, a correlation metric method, such as correlation coefficient calculation, is used to quantify the degree of correlation between the two. Based on the correlation metric results, the data range with strong correlation between the two directions and performance parameters close to the preset target performance of the Mars probe's nuclear battery, such as high energy conversion efficiency and low performance decay rate, is selected and determined as the core data interval of the data distribution.

[0076] In this embodiment of the invention, because it employs the technique of receiving performance layer data and organizing it into a data distribution structure in a multi-dimensional space, processing the data point set using the minimum bounding rectangle algorithm to construct a minimum bounding rectangle covering all data points and using its geometric center as the reference data node, and then using the reference data node as a reference to determine the two main data change directions with the largest variance in the data distribution through eigenvector analysis, and finally calculating the correlation between the two directions and determining the core data interval based on the correlation metric, it overcomes the shortcomings of existing superjunction radiation-voltaic nuclear battery performance data processing, such as the lack of structured integration of performance layer data, unclear key data change patterns, and difficulty in defining the core effective data range. This achieves a systematic and structured analysis of performance data, accurately locating the key change dimensions and core effective range of the data, and providing a focused and high-quality data foundation for subsequently establishing the nonlinear coupling relationship between the radiation source and the transducer.

[0077] In a preferred embodiment of the present invention, step 3 above may include:

[0078] Step 3.1: Based on the core data interval and combined with the basic parameters at the radioactive source level, establish the energy deposition distribution relationship of the radioactive source rays in the transducer device. Specifically, this includes: using a defined core data interval as the analysis scope, and combining the acquired basic parameters at the radioactive source level, such as the β-ray type, energy spectrum distribution, and source activity of Nickel-63, analyzing the interaction between the radioactive source rays and the transducer device; determining the energy loss of rays in different regions of the transducer device, such as inside the trench, on the trench sidewall, and near the electrode, through the correlation between energy deposition and performance parameters in the core data interval, and establishing the energy deposition distribution relationship; for example, based on the energy range of β-rays and the material composition of the transducer device, such as the atomic number and density of diamond, analyzing the correspondence between ray penetration depth and energy deposition density, clarifying the energy deposition differences at different depths in the trench, and ensuring that the relationship reflects the compatibility between ray energy and the transducer device material and structure within the core data interval.

[0079] Step 3.2: Based on the energy deposition distribution relationship, and combined with the trench geometry and doping profile in the basic parameters of the transducer device, establish the electric field distribution relationship within the trench region. Specifically, this includes: based on the established energy deposition distribution relationship, and combined with the trench geometry in the basic parameters of the transducer device, such as depth, width, aspect ratio, and doping profile, such as the concentration gradient of n-type and p-type doping within the trench and the boundary of the doped region, analyze the influence of charge distribution in the energy deposition region on the electric field; for example, based on the depth-to-width ratio of the trench and the doping concentration differences at different locations, determine the formation mechanism of the electric field inside the trench, how the charge balance between the highly doped and low-doped regions generates the transverse and longitudinal electric fields, and how the carrier generation in the concentrated energy deposition region affects the local electric field intensity; and through the correlation between the electric field distribution and performance parameters within the core data interval, establish the distribution relationship of electric field intensity within the trench region as a function of location and energy deposition amount.

[0080] Step 3.3 establishes the relationship between the electric field distribution and the carrier transport and collection efficiency. Specifically, this includes: analyzing the impact of different electric field intensities and directions on carrier transport based on the established electric field distribution; for example, how a higher electric field intensity within the trench accelerates the separation of electrons and holes, reducing the recombination probability of carriers on the trench sidewalls; how the degree of matching between the electric field direction and the carrier motion direction affects the transport path length, thus changing the collection time; and, combining the correlation between carrier collection efficiency and electric field parameters within the core data interval, establishing the relationship between the electric field distribution and carrier transport velocity, recombination rate, and collection probability, clarifying the range of electric field intensity at which carrier collection efficiency can meet the long-term stability requirements of the Mars probe's nuclear battery.

[0081] Step 3.4 integrates the energy deposition distribution relationship, electric field distribution relationship, and carrier collection efficiency relationship to construct a complete nonlinear coupling relationship from the radiation source's X-ray energy to the output power. Specifically, this includes: integrating the energy deposition distribution relationship, electric field distribution relationship, and carrier collection efficiency relationship to construct a complete link from the radiation source's X-ray energy input to the nuclear battery's output power; for example, the radiation source's X-ray energy is converted into the amount of carriers generated in the transducer through the energy deposition relationship. Under the action of the electric field described by the electric field distribution relationship, the carriers are converted into collectable charge through the transport and collection efficiency relationship, and finally converted into output power through electrode configuration; the coupling relationship must reflect nonlinear characteristics. For example, when the X-ray energy is too high, the concentrated energy deposition may lead to local carrier saturation, which may reduce the collection efficiency; when the electric field strength exceeds a certain value, the tunneling effect is triggered, leading to leakage current and a decrease in output power.

[0082] In this embodiment of the invention, a technical approach is adopted that uses a core data interval to establish the energy deposition distribution relationship of radiation in the transducer device based on the basic parameters at the radiation source level. Then, the relationship of electric field distribution within the trench is established by combining the trench geometry and doping profile of the transducer device. Based on this electric field distribution relationship, the influence of the electric field on carrier transport and collection efficiency is established. Finally, the relationships are integrated to construct a complete nonlinear coupling relationship from radiation source radiation energy to output power. This overcomes the technical problem in existing superjunction radiation-voltaic nuclear battery performance data processing where there is a lack of full-link correlation analysis of physical processes such as radiation source energy transfer, device electric field distribution, and carrier collection. The relationships between each link are isolated and do not form a closed-loop coupling, leading to large deviations between performance prediction and reality. This achieves a complete physical link mapping from radiation source radiation energy input to output power, clarifies the influence mechanism of each link parameter on the final performance, provides an accurate and complete physical relationship framework for state iteration optimization, and improves the reliability of performance analysis and parameter optimization.

[0083] In a preferred embodiment of the present invention, step 4 above may include:

[0084] Step 4.1: The complete nonlinear coupling relationship is used as the basic physical framework. Under this framework, dynamic performance data of the long-term degradation process of the nuclear battery is introduced. The dynamic performance data includes the performance parameter degradation sequence at different time points. Specifically, the complete nonlinear coupling relationship from the radiation source energy to the output power is used as the basic physical framework for analyzing the long-term performance of the nuclear battery. Considering that the nuclear battery will naturally decay due to the activity of the radiation source, such as the long half-life of nickel-63, it will still slowly decay and experience performance degradation. Therefore, dynamic performance data of the long-term degradation process of the nuclear battery is introduced under the basic framework. The dynamic performance data comes from the long-term operation test records of similar nuclear batteries in the past and includes the performance parameter degradation sequence at different time points. For example, the specific values ​​of parameters such as output power, carrier collection efficiency, and energy conversion efficiency at the time when the nuclear battery is first put into use, after 1 year of operation, after 3 years of operation, and after 5 years of operation, so as to reflect the performance change law of the nuclear battery over the time of operation.

[0085] Step 4.2: Based on the core data range, the introduced dynamic performance data is filtered to obtain multiple dynamic performance feature points representing different decay stages. Specifically, this includes: based on the determined core data range, i.e., the effective data range reflecting the key performance changes of the nuclear battery, the introduced long-term decay dynamic performance data is filtered; during the filtering, the focus is on retaining data points that fall within the core data range and can reflect the characteristics of different decay stages of the nuclear battery, while excluding abnormal data outside the range, such as data on sudden performance drops caused by accidental failures; for example, the long-term decay process of the nuclear battery is divided into three stages: the initial stable period, the linear decay period with small performance fluctuations within 1 year after commissioning, the slow decay period with linear performance decline over time from 1 to 5 years of operation, and the slow decay period after 5 years of operation. Two to three representative data points are selected from each stage: the peak performance point is selected for the initial stable period, the performance point at the intermediate time node is selected for the linear decay period, and the stable performance point after the decay rate slows down is selected for the slow decay period. Finally, multiple dynamic performance feature points representing different decay stages are obtained.

[0086] Step 4.3 involves using multiple dynamic performance characteristic points as iterative variables and inputting them into the basic physical relationship framework to establish an iterative optimization mechanism. Specifically, this includes using the selected dynamic performance characteristic points as iterative variables to adjust the basic physical relationship framework. Combined with the long-term performance requirements of the nuclear battery for the Mars probe, such as a carrier collection efficiency decrease of no more than 10% and an output power decay rate of no more than 0.3% per year throughout the mission cycle, these variables are input into the basic physical relationship framework to establish an iterative optimization mechanism. The core logic of this mechanism is to adjust the iterative variables, i.e., the performance parameters corresponding to the characteristic points at each decay stage, so that the performance decay curve output by the basic framework closely matches the actual long-term decay pattern. For example, if the framework calculates an output power decay rate of 0.5% after 3 years of operation, which is higher than the preset 0.3%, then by adjusting the parameters corresponding to the characteristic points, such as the correlation radioactive source activity decay coefficient and the transducer carrier recombination rate, the gap between the framework's calculated value and the target decay rate is gradually reduced.

[0087] Step 4.4 involves executing the established iterative optimization mechanism to perform multiple rounds of performance optimization search within the framework of basic physical relationships, outputting an optimized set of dynamic performance feature points. Specifically, this includes: following the established iterative optimization mechanism, performing multiple rounds of performance optimization search within the framework of basic physical relationships; after the first round of optimization, comparing the deviation between the decay curve output by the framework and the actual long-term dynamic data; if the deviation between the calculated carrier collection efficiency after 5 years of operation and the actual test value is 6%, then adjusting the iterative variables based on this deviation, such as correcting the performance decline slope of the feature points during the linear decay period; after entering the second round of optimization, recalculating the deviation; if the deviation drops to within 3%, continuing to fine-tune the variables for the third round of optimization until the deviation is less than 1%, and the performance parameters corresponding to the feature points at each decay stage meet the long-term performance goals of the Mars probe's nuclear battery, such as maintaining more than 95% of the initial value of the output power after 5 years of operation; after completing multiple rounds of optimization, outputting the optimized set of dynamic performance feature points.

[0088] In this embodiment of the invention, by employing a complete nonlinear coupling relationship as the basic physical relationship framework, and introducing long-term degradation dynamic performance data of nuclear batteries containing performance parameter decay sequences at different time points within this framework, multiple dynamic performance feature points characterizing different decay stages are selected based on the core data interval. These feature points are then used as iterative variables to input the framework to establish an iterative optimization mechanism and perform multiple rounds of performance optimization search to output an optimized set of feature points. Therefore, this invention overcomes the technical problem of existing superjunction radiation-voltaic nuclear battery data processing ignoring the long-term degradation dynamic process, resulting in a large deviation between long-term performance prediction and actual operation. This achieves the goal of capturing the performance change patterns of nuclear batteries at different decay stages, making the basic physical relationship framework more consistent with long-term actual operation scenarios, and improving the targeted technical effect of long-term performance analysis and optimization.

[0089] In a preferred embodiment of the present invention, step 5 above may include:

[0090] Step 5.1: Receive the optimized set of dynamic performance feature points, and sort and organize the feature points according to the time series. Specifically, this includes: receiving the output optimized set of dynamic performance feature points, which covers key performance parameters of the nuclear battery from the initial commissioning to long-term operation, such as 1 year, 3 years, and 5 years of operation, such as output power and carrier collection efficiency; sorting these feature points in chronological order, placing the performance points at the initial commissioning stage at the beginning of the sequence, and then arranging the feature points of each subsequent stage in ascending order of operating time; and then organizing these sorted points in a structured manner to form a performance data sequence unfolded along the time axis, clearly presenting the change sequence of nuclear battery performance as operating time progresses.

[0091] Step 5.2: Based on the sorted feature point sequence, a dynamic performance trajectory reflecting the evolution of nuclear battery performance over time is constructed using curve fitting methods. Specifically, this includes: using the feature point sequence sorted by time series, curve fitting methods, such as polynomial fitting and exponential fitting, are employed to make these discrete feature points continuous. For example, with time as the horizontal axis and output power as the vertical axis, the power feature points at each time point are substituted into the fitting algorithm to generate a smooth curve. This curve reflects the dynamic performance trajectory reflecting the evolution of the nuclear battery's output power over time. Similarly, corresponding dynamic trajectories are constructed for other performance parameters such as carrier collection efficiency. The trajectory must accurately match the distribution trend of the feature points, truthfully reflecting the overall change process of the nuclear battery's performance from initial stability to gradual decay during long-term operation.

[0092] Step 5.3 involves analyzing the changing patterns of the dynamic performance trajectory and obtaining performance correction factors to modify the theoretical model based on these patterns. Specifically, this includes: analyzing the changing patterns of the constructed dynamic performance trajectory, focusing on the slope changes, inflection point positions, and decay rates; for example, it was found that the output power trajectory has a small slope and slow decay in the initial year, then the slope increases after one year, entering a linear decay stage, and after five years the slope gradually decreases again, with the decay slowing down. Simultaneously, the decay trajectory of carrier collection efficiency shows a strong correlation with the decay trajectory of radioactive source activity. Based on the decay patterns, performance correction factors are determined. For example, for the linear decay stage, a time-dependent decay coefficient correction factor is set to adjust the impact of radioactive source activity decay on energy deposition; for the initial stable period, a recombination rate correction factor is set to correct the theoretically calculated value of the carrier recombination probability.

[0093] Step 5.4 involves applying the performance correction factor to the complete nonlinear coupling relationship to perform the first correction, resulting in a corrected nonlinear coupling relationship. Specifically, this includes: applying the obtained performance correction factor to the constructed complete nonlinear coupling relationship to perform the first correction of the nonlinear coupling relationship system; for example, introducing the attenuation coefficient correction factor of the linear decay stage into the energy deposition distribution relationship to adjust the calculation method of the change of radioactive source activity over time, so that the theoretical value of energy deposition is closer to the actual attenuation situation; and introducing the recombination rate correction factor into the carrier collection efficiency relationship to correct the carrier recombination rate under the action of the electric field, so that the calculation of collection efficiency is closer to the actual data of the initial stable period.

[0094] Step 5.5: Based on the corrected nonlinear coupling relationship, repeat the iterative optimization process until the output parameter combination meets the preset target performance parameter requirements, thus obtaining the final parameter combination. Specifically, this includes: based on the obtained corrected nonlinear coupling relationship, re-execute the iterative optimization process: reintroduce dynamic performance data, screen feature points, and input the corrected framework for multiple rounds of optimization; after each round of optimization, compare the performance prediction value corresponding to the output parameter combination with the preset target performance parameters of the Mars probe, such as an output power attenuation rate of less than or equal to 0.3% and a carrier collection efficiency of greater than or equal to 85% within the mission cycle; if there is still a deviation after the first correction, such as a predicted attenuation rate of 0.4% over 5 years, then repeatedly adjust the performance correction factor, correct the coupling relationship, and execute the iterative optimization process until the performance prediction value corresponding to the output parameter combination fully meets the preset target requirements, ultimately obtaining a parameter combination suitable for a long-term Mars exploration mission.

[0095] In this embodiment of the invention, because it adopts the technical means of receiving and optimizing the set of dynamic performance feature points and organizing them in a time series, constructing the dynamic performance trajectory of the nuclear battery performance over time through curve fitting based on the sorted feature point sequence, analyzing the trajectory change law to obtain the performance correction factor to correct the complete nonlinear coupling relationship for the first time, and then repeatedly performing iterative optimization based on the corrected coupling relationship until the final parameter combination that meets the preset target performance parameters is output, it overcomes the technical problems of existing superjunction radiation-voltaic nuclear battery data processing that does not construct the dynamic performance evolution law based on the time series and lacks a targeted model correction mechanism, resulting in the nonlinear coupling relationship being out of touch with the long-term actual operating performance and the parameter combination being difficult to meet the preset target performance requirements of long-life scenarios. Thus, it achieves the technical effect of capturing the long-term performance evolution trend of nuclear batteries, making the nonlinear coupling relationship more in line with the actual decay law, and ensuring that the final parameter combination can stably match the target performance of long-cycle tasks through continuous iterative optimization, thereby improving the performance compliance rate of nuclear batteries in long-life extreme scenarios.

[0096] In a preferred embodiment of the present invention, step 6 above may include:

[0097] Step 6.1: Based on the final parameter combination, introduce process boundary conditions for MEMS miniaturization. These boundary conditions include minimum processing size, maximum aspect ratio, and material compatibility requirements. Specifically, based on the obtained final parameter combination, including key parameters such as trench geometry, doping concentration, and material composition of the transducer, and considering the miniaturization requirements of the nuclear battery for the Mars rover, MEMS technology is needed to miniaturize the device to fit the limited space of the rover. Therefore, process boundary conditions for MEMS miniaturization are introduced. These boundary conditions are determined based on the actual capabilities of current MEMS processing technology. The minimum processing size is the smallest structural size that can be stably achieved by the current process, such as the minimum linewidth commonly found in silicon-based MEMS processes. The maximum aspect ratio is the limit ratio to avoid structural collapse or accuracy degradation during processing, such as the maximum aspect ratio typically achievable by deep silicon etching processes. Material compatibility requirements address the compatibility of transducer materials, such as diamond and crystalline coordination polymers, with commonly used MEMS materials, such as photoresist and metal electrode materials, ensuring that the materials do not undergo chemical reactions or physical property degradation during processing, thus affecting the final performance of the nuclear battery.

[0098] Step 6.2 involves comparing and analyzing the final parameter combination with the process boundary conditions to identify parameters that exceed the process range, thus obtaining the comparison analysis results. Specifically, this includes comparing and analyzing each of the final parameter combination with the introduced MEMS process boundary conditions. For example, if the trench width setting of the transducer in the final parameter combination is less than the minimum processing size of the MEMS process, or the ratio of trench depth to width exceeds the maximum aspect ratio, or the material used for the transducer has compatibility issues with the photoresist in the MEMS lithography process (e.g., the material will be corroded by the photoresist), then the parameter is identified as a parameter that exceeds the process range. Through comprehensive comparison, all parameters that do not meet the process boundary conditions are identified, forming a comparison analysis result that includes the name of the parameter exceeding the limit, the current setting value, and the process allowable range. This clarifies the specific direction for subsequent corrections and avoids situations where parameters cannot be actually processed due to ignoring process limitations, as is the case with traditional processing methods.

[0099] Step 6.3: Based on the comparative analysis results, adaptability corrections are performed on parameters that exceed the process feasibility range to obtain a parameter combination for adaptability correction. Specifically, this includes: adaptability corrections are performed on parameters that exceed the process feasibility range based on the comparative analysis results. The correction process must ensure that the target performance of the nuclear battery is preserved as much as possible. For example, if the trench width is less than the minimum processing size, the trench width is appropriately increased to the minimum value allowed by the process, without exceeding the maximum aspect ratio, while the trench depth is finely adjusted to maintain the original aspect ratio range, avoiding significant fluctuations in the electric field distribution due to structural changes. If the transducer material is incompatible with the photoresist, it is replaced with a substitute material with similar chemical properties and that meets the material compatibility requirements. The carrier mobility change that may be caused by the material replacement is compensated by adjusting the doping concentration to ensure that the carrier collection efficiency is not affected. After each correction, an adaptability correction parameter combination is obtained in which all parameters meet the boundary conditions of the MEMS process.

[0100] Step 6.4 verifies the adapted parameter combination to ensure it meets both the process boundary conditions and the preset target performance requirements, thus obtaining the final optimized parameter combination. This includes: dual verification of the obtained adapted parameter combination; the first verification is process feasibility verification, confirming that all modified parameters, such as trench size and material selection, are within the MEMS process boundary conditions and can be stably produced using existing processing equipment; the second verification is performance compliance verification, substituting the modified parameter combination into the modified nonlinear coupling relationship to simulate the nuclear battery's operation during the Mars exploration mission cycle, checking whether performance indicators such as output power, carrier collection efficiency, and annual decay rate still meet the preset target requirements, such as a carrier collection efficiency of no less than 85% and an annual decay rate of no more than 0.3% during the mission cycle; if any indicator is found to be unmet, the parameter modification scheme is readjusted until the modified parameter combination simultaneously meets both the process boundary conditions and the preset target performance requirements, ultimately obtaining the final optimized parameter combination that can be directly used for the production of superjunction radiative-volt nuclear batteries for the Mars probe.

[0101] In this embodiment of the invention, by employing a technical approach that introduces MEMS miniaturization process boundary conditions based on the final parameter combination, including minimum processing size, maximum aspect ratio, and material compatibility requirements, and comparing and analyzing the final parameter combination with the process boundary conditions to identify parameters exceeding the process range, adaptability corrections are made to the out-of-limit parameters based on the comparison results, and the corrected parameter combination is then verified to ensure that both the process boundary conditions and the preset target performance are met, the technical problem of existing superjunction radiation-voltaic nuclear batteries ignoring the actual process limitations of MEMS processing, resulting in the designed parameter combination exceeding the processing capability and failing to be engineering-implemented, and further exacerbating the deviation between actual performance and design expectations due to process deviations, is overcome. This achieves the technical effect of ensuring that the final optimized parameter combination not only conforms to the actual feasible range of MEMS miniaturization processing, avoiding process-level implementation obstacles, but also stably achieves the preset target performance of the nuclear battery, realizing effective connection between design and process, and improving the reliability and feasibility of nuclear battery parameter combinations in engineering applications.

[0102] like Figure 2 As shown, embodiments of the present invention also provide a superjunction radiation-voltaic nuclear battery performance data processing system based on an iterative algorithm, comprising:

[0103] The acquisition module is used to acquire basic parameters of the radiation source layer, basic parameters of the transducer layer, and performance layer data of the superjunction radiation photovoltaic cell.

[0104] The calculation module organizes the performance layer data into a data distribution structure. Based on the data distribution structure, it processes the data point set using the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the reference data node, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval.

[0105] A module is established to create a nonlinear coupling relationship that describes the complete link of radiation source energy distribution, trench electric field distribution, carrier collection efficiency, and output power, based on core data ranges, basic parameters at the radiation source level, and basic parameters at the transducer level.

[0106] The optimization module is used to iteratively optimize the nonlinear coupling relationship by introducing the dynamic performance data of the long-term decay process of the nuclear battery as the iterative variable. It selects multiple dynamic performance feature points based on the core data interval. According to the evolution relationship of each feature point in the time series, a dynamic performance trajectory is obtained, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained.

[0107] The processing module takes the final parameter combination as input, introduces the process boundary conditions for MEMS miniaturization as constraints for adaptation correction, and obtains the final optimized parameter combination.

[0108] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing performance data of superjunction radiation-voltaic nuclear batteries based on an iterative algorithm, characterized in that, The method includes: To obtain the basic parameters of the radiation source, the basic parameters of the transducer, and the performance data of the superjunction radiation-voltaic nuclear battery; The performance layer data is organized into a data distribution structure. Based on the data distribution structure, the data point set is processed by the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the reference data node, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval. Based on the core data range, basic parameters at the radiation source level, and basic parameters at the transducer level, a nonlinear coupling relationship is established to describe the complete link of radiation source ray energy distribution, trench electric field distribution, carrier collection efficiency, and output power. Based on the nonlinear coupling relationship, the dynamic performance data of the long-term decay process of nuclear battery is introduced as the iterative variable for iterative optimization. Based on the core data interval, the long-term decay process of nuclear battery is divided into the initial stable period, the linear decay period and the slow decay period. Multiple dynamic performance feature points are selected from each decay stage. A dynamic performance trajectory is obtained based on the evolution relationship of each feature point in the time series, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained. The final parameter combination is taken as input, and the process boundary conditions for MEMS miniaturization are introduced as constraints for adaptation correction to obtain the final optimized parameter combination.

2. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 1, characterized in that, To obtain fundamental parameters at the radiation source level, fundamental parameters at the transducer level, and performance level data of the superjunction radiation-voltaic nuclear battery, including: Call the communication interface with the radioactive source characteristic database to obtain basic parameters at the radioactive source level, including isotope type, emitted particle type, energy spectrum distribution, and source activity; Based on the fundamental parameters at the radiation source level, the transducer device structure definition file that matches the characteristics of the radiation source is received and parsed, and the fundamental parameters at the transducer device level, including trench geometry, doping profile, material composition and electrode configuration, are extracted from the definition file. By combining the fundamental parameters at the radiation source level and the fundamental parameters at the transducer level, and through an integrated data acquisition service, performance layer data corresponding to the parameter combinations is selectively obtained from the historical database.

3. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 2, characterized in that, The performance layer data is organized into a data distribution structure. Based on the data distribution structure, the data point set is processed by the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle that covers all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the baseline data nodes, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval, including: Receive performance layer data and organize the performance layer data into a data distribution structure in a multi-dimensional space, where each data point corresponds to a set of performance measurement values; Based on the data distribution structure, the data point set is processed by the minimum bounding rectangle algorithm to construct a minimum bounding rectangle that covers all data points, and the geometric center of the minimum bounding rectangle is used as the reference data node. Using the baseline data node as a reference, the two main data change directions with the largest variance in the data distribution are determined by the eigenvector analysis method; Calculate the data correlation between two main data change directions, and determine the core data interval of the data distribution based on the correlation metric.

4. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 3, characterized in that, Based on the core data range, fundamental parameters at the radioactive source level, and fundamental parameters at the transducer level, a nonlinear coupling relationship is established to describe the complete link of radioactive source ray energy distribution, trench electric field distribution, carrier collection efficiency, and output power, including: Based on the core data range and combined with the basic parameters at the radioactive source level, the energy deposition distribution relationship of radioactive source rays in the transducer is established. Based on the energy deposition distribution relationship, and combined with the trench geometry and doping profile in the basic parameters of the transducer, the electric field distribution relationship in the trench region is established. Based on the electric field distribution relationship, the relationship between the electric field and the carrier transport and collection efficiency is established. By integrating the relationships of energy deposition distribution, electric field distribution, and carrier collection efficiency, a complete nonlinear coupling relationship from the radiation source's ray energy to the output power is constructed.

5. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 4, characterized in that, Based on nonlinear coupling, dynamic performance data of the long-term decay process of nuclear batteries are introduced as iterative variables for optimization. Based on the core data interval, the long-term decay process is divided into an initial stable period, a linear decay period, and a slow decay period. Multiple dynamic performance feature points are selected from each decay stage, including: The complete nonlinear coupling relationship is used as the basic physical relationship framework. Under the basic physical relationship framework, dynamic performance data of the long-term decay process of nuclear battery is introduced. The dynamic performance data includes the decay sequence of performance parameters at different time points. Based on the core data interval, the introduced dynamic performance data is filtered and processed to obtain multiple dynamic performance feature points that characterize the initial stable period, the linear decay period, and the slow decay period. Multiple dynamic performance feature points are used as iterative variables and input into the basic physical relationship framework to establish an iterative optimization mechanism; By executing the established iterative optimization mechanism, multiple rounds of performance optimization search are performed within the framework of basic physical relationships, and the optimized set of dynamic performance feature points is output.

6. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 5, characterized in that, A dynamic performance trajectory is obtained based on the evolution relationship of each feature point over time. A performance correction factor is then generated based on this trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, a final parameter combination that satisfies the preset target performance parameters is obtained, including: Receive the optimized set of dynamic performance feature points, and sort and organize the feature points according to the time series. Based on the sorted feature point sequence, a dynamic performance trajectory reflecting the evolution of nuclear battery performance over time is constructed using curve fitting methods. Analyze the changing patterns of dynamic performance trajectories, and based on these patterns, obtain performance correction factors for revising theoretical models. The performance correction factor is applied to the complete nonlinear coupling relationship to perform the first correction on the nonlinear coupling relationship, resulting in the corrected nonlinear coupling relationship. Based on the corrected nonlinear coupling relationship, the iterative optimization process is repeated until the output parameter combination meets the preset target performance parameter requirements, thus obtaining the final parameter combination.

7. The method for processing superjunction radiation-voltaic nuclear battery performance data based on an iterative algorithm according to claim 6, characterized in that, Using the final parameter combination as input, and incorporating the process boundary conditions for MEMS miniaturization as constraints for adaptation correction, the final optimized parameter combination is obtained, including: Based on the final parameter combination, process boundary conditions for MEMS miniaturization are introduced, including minimum processing size, maximum aspect ratio and material compatibility requirements. The final parameter combination is compared and analyzed with the process boundary conditions to identify parameter items that exceed the process range, so as to obtain the comparison and analysis results. Based on the comparative analysis results, the parameters that are outside the feasible range of the process are modified for adaptability, and the modified parameter combinations are obtained. The parameter combination after adaptation correction is verified to meet the preset target performance requirements while satisfying the process boundary conditions, thus obtaining the final optimized parameter combination.

8. A superjunction radiation-voltaic nuclear battery performance data processing system based on an iterative algorithm, the system implementing the method as described in any one of claims 1 to 7, characterized in that, include: The acquisition module is used to acquire basic parameters of the radiation source layer, basic parameters of the transducer layer, and performance layer data of the superjunction radiation photovoltaic cell. The calculation module organizes the performance layer data into a data distribution structure. Based on the data distribution structure, it processes the data point set using the minimum bounding rectangle algorithm to obtain the minimum bounding rectangle covering all data points. The geometric center of the minimum bounding rectangle is used as the reference data node. Based on the reference data node, two main data change directions are determined, and the data correlation between the two directions is calculated to obtain the core data interval. A module is established to create a nonlinear coupling relationship that describes the complete link of radiation source energy distribution, trench electric field distribution, carrier collection efficiency, and output power, based on core data ranges, basic parameters at the radiation source level, and basic parameters at the transducer level. The optimization module is used to iteratively optimize the long-term decay process of nuclear batteries by taking the nonlinear coupling relationship as the basis and introducing the dynamic performance data of the long-term decay process of nuclear batteries as the iterative variable. Based on the core data interval, the long-term decay process of nuclear batteries is divided into the initial stable period, the linear decay period and the slow decay period. Multiple dynamic performance feature points are selected from each decay stage. A dynamic performance trajectory is obtained based on the evolution relationship of each feature point in the time series, and a performance correction factor is generated based on the trajectory. Through continuous correction and optimization of the nonlinear coupling relationship, the final parameter combination that meets the preset target performance parameters is obtained. The processing module takes the final parameter combination as input, introduces the process boundary conditions for MEMS miniaturization as constraints for adaptation correction, and obtains the final optimized parameter combination.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. 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 as described in any one of claims 1 to 7.

Images