Method for evaluating abnormal discharge hazard degree of space ion electric propulsion

Abstract

The application relates to the technical field of ion electric propulsion, in particular to a space ion electric propulsion abnormal discharge hazard degree evaluation method, which comprises the following steps: step 1, data preliminary screening; step 2, comprehensive evaluation; step 3, threshold setting; step 4, matrix interpretation; step 5, grade division; and step 6, hazard evaluation. The application extracts key electric parameters through data preliminary screening, thereby avoiding calculation redundancy of full-parameter monitoring; the abnormal discharge characteristics are comprehensively described by using multi-dimensional time-frequency domain evaluation indexes, thereby overcoming the limitation of single parameter interpretation; a hierarchical evaluation domain is established based on the analytic hierarchy process, thereby realizing quantitative grading of the hazard degree; and real-time monitoring and hierarchical disposal are realized by combining a neural network and a satellite-borne control system.

Description

Technical Field

[0001] This application relates to the field of ion electric propulsion technology, and more specifically, to a method for assessing the degree of hazard caused by abnormal discharge in space ion electric propulsion. Background Technology

[0002] When ion electric propulsion is in operation, the voltage exceeds 1500V and the surrounding area is filled with plasma and neutral gas. If the environmental conditions change, it is very easy to cause abnormal discharge, which may result in the space ion electric propulsion system shutting down and restarting, or even causing system damage or failure.

[0003] Therefore, finding a method to assess the severity of abnormal discharges in ion electric propulsion and to precisely control its operating status can significantly improve the ease of use, usability, and availability of ion electric propulsion, while also enhancing its lifespan and reliability.

[0004] Based on existing experimental data from tens of thousands of hours of ion electric propulsion on the ground and in orbit, this invention proposes a data mining-based method for assessing the severity of abnormal discharges in space ion electric propulsion. This method provides application guidance for the efficient on-orbit operation of space ion electric propulsion and has significant engineering application value. Summary of the Invention

[0005] This application provides a method for assessing the severity of abnormal discharge in space ion electric propulsion, aiming to solve the problems in the prior art of making it difficult to quantitatively assess the severity of abnormal discharge, and of being unable to achieve refined interpretation and real-time processing.

[0006] To achieve the above objectives, this application provides a method for assessing the hazard level of abnormal discharge in space ion electric propulsion, comprising the following steps: Step 1: Initial data screening. Based on the operational life test database of the ion electric propulsion system (30,000 hours on the ground and 20,000 hours in orbit), a prediction set of abnormal discharge and electrical parameters is established. Using the partial least squares regression algorithm, with the variable projection importance index being greater than 1.2 times as the criterion, peak parameters closely related to abnormal discharge of ion electric propulsion are initially screened from stress parameters and reliability parameters. Step 2: Comprehensive evaluation. In the time and frequency domains, the peak parameters selected in Step 1 are correlated with the multidimensional factors of fault hazard assessment. The multidimensional factors include peak power, energy deposition, impact, volatility, and volatility intensity, thereby constructing a multidimensional anomaly evaluation index based on electrical parameters and carrying out a comprehensive evaluation of multidimensional discharge anomalies. Step 3: Threshold setting. Based on the normal operating parameters of the highest power of space ion electric propulsion design, all parameters are increased by 20% as the overpressure limit. The threshold at which each evaluation index in Step 2 is judged to be abnormal is calculated, and the abnormal judgment boundary of each index is determined. Step 4: Matrix interpretation. Based on the multi-dimensional evaluation indicators constructed in Step 2, an anomaly interpretation matrix closely related to each indicator is established. This matrix has evaluation indicators as rows and hazard levels as columns, with each matrix element representing the interpretation rule of the corresponding indicator under that hazard level. Step 5: Classification. The peak parameters selected in Step 1 are normalized. The normalized parameters are divided into four weight intervals using the analytic hierarchy process: no impact interval, abnormal recoverable interval, performance degradation interval, and irreversible failure interval, thereby establishing an evaluation domain for the degree of hazard of abnormal discharge. Step 6: Hazard assessment. The discharge parameters are monitored in real time using a neural network and onboard control system. The real-time collected electrical parameters are processed sequentially through steps 1 to 5 to determine the hazard level of the current abnormal discharge. Based on the severity of the hazard, one or more of the following treatment methods are automatically adopted: shutdown, partial shutdown, downgrading, or parameter adjustment.

[0007] Furthermore, in step 1, the stress parameters include grid voltage, grid current, anode voltage, anode current, accelerating grid voltage, accelerating grid current, neutralizer heating voltage and current, neutralizer holding voltage and current, cathode heating voltage and current, and cathode holding voltage and current; the reliability parameters include accelerating grid power, grid power, cathode power, and neutralizer power.

[0008] Furthermore, in step 1, the operational life test database for 30,000 hours on the ground and 20,000 hours in orbit is constructed by synchronously collecting and recording electrical parameter data, including various stress parameters and reliability parameters, during the continuous operation of the ion electric propulsion system under simulated space environment conditions on the ground for 30,000 hours and the actual operation in orbit for 20,000 hours. The data is then cleaned and normalized.

[0009] Furthermore, in step 1, the specific process of using the partial least squares regression algorithm is as follows: taking the abnormal discharge event as the dependent variable and each stress parameter and reliability parameter as the independent variables, a regression model is established by extracting principal components, and the variable projection importance index value of each parameter is calculated. When the index value is greater than 1.2, the parameter is determined to be an important stress parameter closely related to the abnormal discharge.

[0010] Furthermore, in step 2, the peak power is obtained by multiplying the voltage by the maximum value of the instantaneous current; the energy deposition is obtained by integrating the product of voltage and current over time; the impact is obtained by the difference between the peak power and the root mean square power; the volatility is obtained by dividing the difference between the maximum and minimum power values ​​by the average power value; and the volatility intensity is obtained by the root mean square of the sum of the squares of the differences between the power and the average value at each sampling point.

[0011] Furthermore, in step 3, the specific method for setting the threshold is as follows: based on the normal operating parameters of the highest power of ion electric propulsion design, each parameter is multiplied by 1.2 as the upper boundary. At the same time, considering the instrument measurement error and the normal fluctuation range, the larger value of 1.2 times the normal operating parameter and the upper limit of the measurement error is taken as the abnormal threshold.

[0012] Furthermore, in step 5, the specific implementation of the analytic hierarchy process is as follows: construct a hierarchical structure with the degree of hazard of abnormal discharge as the target layer, the peak parameters selected in step 1 as the criterion layer, and the four hazard levels as the scheme layer. Construct a judgment matrix through expert scoring, calculate the weight vector of each parameter, and then map the comprehensive score after weighting the real-time parameters to the interval between 0 and 1 to divide the four level intervals.

[0013] Furthermore, in step 6, the neural network adopts a three-layer backpropagation neural network. The number of input layer nodes corresponds to the number of peak parameters selected in step 1, the number of hidden layer nodes is 1.5 times the number of input layer nodes, and the number of output layer nodes is 4. The network is trained with ground test data to achieve rapid classification of real-time monitoring data.

[0014] Furthermore, in step 6, the specific handling measures taken according to the degree of harm are as follows: when it is determined that there is no impact, only log is recorded and no action is taken; when it is determined that the abnormality is recoverable, parameter adjustment is automatically performed to attempt recovery; when it is determined that the performance is degraded, downgrading is performed and a warning is issued; when it is determined that the failure is irreversible, shutdown or partial shutdown is immediately performed and the backup thruster is switched.

[0015] The method for assessing the hazard level of abnormal discharge in space ion electric propulsion provided in this application has the following beneficial effects: This application extracts key electrical parameters through initial data screening, avoiding computational redundancy in full-parameter monitoring; it employs multi-dimensional time-frequency domain evaluation indicators to comprehensively characterize abnormal discharge features, overcoming the limitations of single-parameter interpretation; it establishes a hierarchical evaluation domain based on the analytic hierarchy process (AHP), achieving quantitative classification of hazard levels; and it combines neural networks and a spaceborne control system to realize real-time monitoring and hierarchical response. This method has high engineering application value and strong targeting, and can extend the service life of ion electric propulsion by more than 10%, significantly improving on-orbit usability and reliability. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of the application and to make other features, objects, and advantages of the application more apparent. The illustrative embodiments and descriptions of this application are used to explain the application and do not constitute an undue limitation of the application. In the drawings: Figure 1 This is a flowchart illustrating the method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to an embodiment of this application. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] In this application, the terms "upper," "lower," "left," "right," "front," "rear," "top," "bottom," "inner," "outer," "middle," "vertical," "horizontal," "lateral," and "longitudinal" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are primarily for the purpose of better describing this application and its embodiments, and are not intended to limit the indicated device, element, or component to having a specific orientation, or to be constructed and operated in a specific orientation.

[0020] Furthermore, in addition to indicating location or positional relationship, some of the aforementioned terms may also have other meanings. For example, the term "above" may also be used in some cases to indicate a certain dependency or connection relationship. Those skilled in the art can understand the specific meaning of these terms in this application based on the specific circumstances.

[0021] In addition, the term "multiple" should mean two or more.

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] like Figure 1As shown, this application provides a method for assessing the severity of abnormal discharges in space ion electric propulsion. This method improves the interpretation of abnormal discharges in space ion electric propulsion from a coarse, partial parameter-based approach to a multi-dimensional, refined method, enabling immediate identification of anomalies and simultaneous implementation of appropriate measures. Specifically, it includes the following steps: Step 1: Initial data screening. Based on the operational life test database of the ion electric propulsion system (30,000 hours on the ground and 20,000 hours in orbit), a prediction set of abnormal discharge and electrical parameters is established. Using the partial least squares regression algorithm, with the variable projection importance index being greater than 1.2 times as the criterion, peak parameters closely related to abnormal discharge of ion electric propulsion are initially screened from stress parameters and reliability parameters. Furthermore, in step 1, the stress parameters include n items such as grid voltage, grid current, anode voltage, anode current, accelerating grid voltage, accelerating grid current, neutralizer heating voltage and current, neutralizer holding voltage and current, cathode heating voltage and current, and cathode holding voltage and current; the reliability parameters include m items such as accelerating grid power, grid power, cathode power, and neutralizer power. This covers all the key electrical characteristics of the ion electric propulsion system, ensuring the comprehensiveness of the initial data screening.

[0024] Furthermore, in step 1, the operational life test database for 30,000 hours on the ground and 20,000 hours in orbit is constructed by synchronously collecting and recording electrical parameter data, including various stress parameters and reliability parameters, during the continuous operation of the ion electric propulsion system under simulated space environment conditions on the ground for 30,000 hours and the actual operation in orbit for 20,000 hours. The data is then cleaned and normalized.

[0025] Specifically, in the ground test portion, the ion electric propulsion system was placed inside a vacuum chamber to simulate a space environment (vacuum level better than 10). -5 The system operates continuously for 30,000 hours (Pa, heat sink temperature 100K), recording all electrical parameters every millisecond during this period. In-orbit portion: Utilizing the onboard data recording system, 20,000 hours of electrical parameter data are collected during actual flight missions. After data acquisition, a moving average filter is used to remove high-frequency noise, and maximum-minimum normalization is used to map each parameter to the [0, 1] interval, thereby constructing a standardized database.

[0026] Furthermore, in step 1, the specific process of using the partial least squares regression algorithm is as follows: taking the abnormal discharge event as the dependent variable and each stress parameter and reliability parameter as the independent variables, a regression model is established by extracting principal components, and the variable projection importance index value of each parameter is calculated. When the index value is greater than 1.2, the parameter is determined to be an important stress parameter closely related to the abnormal discharge.

[0027] Specifically, using abnormal discharge events (labeled as 1) and normal events (labeled as 0) as dependent variables, and all stress parameters and reliability parameters as independent variables, the first 3-5 principal components were extracted. Peak parameters with a VIP value > 1.2 were selected by calculating the VIP value of each parameter. Experiments showed that 8-12 parameters, including grid voltage, accelerating grid current, and anode power, were typically selected, forming the input feature set for subsequent analysis.

[0028] More specifically, in step 1, a prediction set between abnormal discharges and electrical parameters is first established based on the operational life test database of the ion electric propulsion system (30,000 hours on the ground and 20,000 hours in orbit). This database records various abnormal discharge events (such as voltage breakdown and current spikes) that occurred during the long-term operation of the ion electric propulsion system, as well as hundreds of electrical parameters collected synchronously.

[0029] Secondly, partial least squares regression (PLR) is used for data mining. The core idea of ​​this algorithm is to use the abnormal discharge event as the dependent variable and each electrical parameter as the independent variable, and to establish a regression model by extracting principal components, thereby solving the problem of collinearity among high-dimensional parameters. The specific selection criterion uses the variable projection importance index. For the i-th parameter, its VIP value is calculated as follows: , Where p is the total number of independent variables, m is the number of principal components extracted, and n 𝑗𝑦 Let be the correlation coefficient between the j-th principal component and the dependent variable y, and let be the coefficient of correlation between y and y. 𝑖𝑗 Let Σ be the weight of the i-th independent variable on the j-th principal component. When VIP > 1.2, this parameter is determined to be an important stress parameter (i.e., peak parameter) closely related to abnormal discharge. This threshold of 1.2 is a recognized significance threshold determined through cross-validation.

[0030] Step 2: Comprehensive evaluation. In the time and frequency domains, the peak parameters selected in Step 1 are correlated with the multidimensional factors of fault hazard assessment. The multidimensional factors include peak power, energy deposition, impact, volatility, and volatility intensity, thereby constructing a multidimensional anomaly evaluation index based on electrical parameters and carrying out a comprehensive evaluation of multidimensional discharge anomalies. Furthermore, in step 2, the peak power is obtained by multiplying the voltage by the maximum value of the instantaneous current; the energy deposition is obtained by integrating the product of voltage and current over time; the impact is obtained by the difference between the peak power and the root mean square power; the volatility is obtained by dividing the difference between the maximum and minimum power values ​​by the average power value; and the volatility intensity is obtained by the root mean square of the sum of the squares of the differences between the power and the average value at each sampling point.

[0031] Specifically, peak power reflects the intensity of energy release at the moment of abnormal discharge, and the calculation formula is P. peak =V×I maxWhere V is voltage (e.g., grid voltage), I max The maximum instantaneous current during the abnormal discharge is represented by E. Energy deposition characterizes the total energy accumulation throughout the entire abnormal discharge process, calculated using the formula E=∫VIdt, with the integration interval from the start point to the end point of the abnormal discharge. The impact characterizes the disturbance magnitude of the abnormal discharge on the power supply system, calculated using the formula ΔP=P peak -P rms , where P rms The steady-state root-mean-square power before discharge. The volatility characterizes the relative amplitude of power fluctuations, calculated using the formula η = (P...). max -P min ) / P avg , where P max P min P avg These represent the maximum, minimum, and average power values ​​during the abnormal discharge period, respectively. The fluctuation intensity characterizes the dispersion of the power fluctuation, and is calculated using the following formula: Where n is the number of sampling points, P i The power values ​​at each sampling point are given.

[0032] More specifically, step 2, through the comprehensive evaluation of these five dimensions, can fully characterize the time-domain energy characteristics (peak power, energy deposition), frequency-domain disturbance characteristics (impact, volatility), and statistical discrete characteristics (fluctuation intensity) of abnormal discharge, avoiding misjudgment of a single parameter and improving the robustness of the evaluation.

[0033] Step 3: Threshold setting. Based on the normal operating parameters of the highest power of space ion electric propulsion design, all parameters are increased by 20% as the overpressure limit. The threshold at which each evaluation index in Step 2 is judged to be abnormal is calculated, and the abnormal judgment boundary of each index is determined. Furthermore, in step 3, the specific method for setting the threshold is as follows: Based on the normal operating parameters of the highest power of the ion electric propulsion design, each parameter is multiplied by 1.2 as the upper boundary. Simultaneously considering instrument measurement error and normal fluctuation range, the larger value between 1.2 times the normal operating parameter and the upper limit of the measurement error is taken as the abnormal threshold. This balances safety margin and false alarm rate, ensuring that abnormal readings are not falsely triggered within the allowable normal fluctuation range. Taking the accelerating gate current as an example, its rated value is 10mA, and after a 20% bias, it becomes 12mA. If the measurement error is ±0.1mA, then the final abnormal threshold is taken as 12mA.

[0034] Specifically, in step 3, for each peak parameter, its design rated value (e.g., rated grid voltage 1200V) is taken and multiplied by 1.2 to obtain the upper boundary (1440V). Simultaneously, considering instrument measurement error (e.g., ±0.5%) and normal fluctuation range (e.g., ±3%), the larger of 1.2 times the normal operating parameter and the upper limit of measurement error is taken as the abnormal threshold. For example, if the rated power is 1000W, the abnormal threshold is the larger of 1200W and (1000W × upper limit of measurement error).

[0035] Step 4: Matrix interpretation. Based on the multi-dimensional evaluation indicators constructed in Step 2, an anomaly interpretation matrix closely related to each indicator is established. This matrix has evaluation indicators as rows and hazard levels as columns, with each matrix element representing the interpretation rule of the corresponding indicator under that hazard level. Specifically, in step 4, an anomaly identification matrix closely related to each indicator is established based on the five-dimensional evaluation index constructed in step 2. The rows of the matrix correspond to the five evaluation indicators: peak power, energy deposition, impact, volatility, and volatility intensity. The columns correspond to four hazard levels: no impact, recoverable anomaly, performance degradation, and irreversible failure. Each matrix element represents the condition range for the corresponding indicator falling within that level. For example: Peak power: <1.2 times the rated value indicates no impact; 1.2~1.5 times indicates recoverable anomaly; 1.5~2.0 times indicates performance degradation; >2.0 times indicates irreversible failure. Energy deposition, impact, volatility, and volatility intensity indicators also have corresponding threshold ranges.

[0036] Step 5: Classification. The peak parameters selected in Step 1 are normalized. The normalized parameters are divided into four weight intervals using the analytic hierarchy process: no impact interval, abnormal recoverable interval, performance degradation interval, and irreversible failure interval, thereby establishing an evaluation domain for the degree of hazard of abnormal discharge. Furthermore, in step 5, the specific implementation of the analytic hierarchy process is as follows: construct a hierarchical structure with the degree of hazard of abnormal discharge as the target layer, the peak parameters selected in step 1 as the criterion layer, and the four hazard levels as the scheme layer. Construct a judgment matrix through expert scoring, calculate the weight vector of each parameter, and then map the comprehensive score after weighting the real-time parameters to the interval between 0 and 1 to divide the four level intervals.

[0037] Specifically, in step 5, the peak parameters selected in step 1 are normalized, and then the normalized parameters are divided into four weight intervals using the Analytic Hierarchy Process (AHP). The core of AHP is to construct a judgment matrix and determine the relative importance of each parameter through pairwise comparisons. A three-layer hierarchical structure is constructed: the target layer is the "degree of hazard of abnormal discharge", the criterion layer is the peak parameters selected in step 1 (such as grid voltage, accelerating grid current, etc.), and the scheme layer is the four hazard levels. The judgment matrix is ​​constructed through expert scoring, the weight vector of each parameter is calculated, and a consistency check is performed. The weighted sum of the peak parameters collected in real time is used to obtain a comprehensive score, which is mapped to the interval [0, 1] and divided into four levels: 0~0.25 is no impact, 0.25~0.5 is recoverable abnormality, 0.5~0.75 is performance degradation, and 0.75~1 is irreversible failure.

[0038] Step 6: Hazard assessment. The discharge parameters are monitored in real time using a neural network and onboard control system. The real-time collected electrical parameters are processed sequentially through steps 1 to 5 to determine the hazard level of the current abnormal discharge. Based on the severity of the hazard, one or more of the following treatment methods are automatically adopted: shutdown, partial shutdown, downgrading, or parameter adjustment.

[0039] Furthermore, in step 6, a three-layer backpropagation neural network is used. The number of input layer nodes corresponds to the number of peak parameters selected in step 1, the number of hidden layer nodes is 1.5 times the number of input layer nodes, and the number of output layer nodes is 4. The network is trained using ground-based experimental data to achieve rapid classification of real-time monitoring data. The three-layer backpropagation neural network has the same number of input layer nodes as the number of peak parameters selected in step 1 (e.g., 10), the same number of hidden layer nodes as 1.5 times the number of input layer nodes (i.e., 15), and 4 output layer nodes (corresponding to the probabilities of four hazard levels). Labeled data from the ground-based experimental database is used for training, with a training objective of classification accuracy >95%. The trained network can run on the spaceborne control system, achieving millisecond-level real-time classification.

[0040] Furthermore, in step 6, the specific handling measures taken according to the degree of harm are as follows: when it is determined that there is no impact, only log is recorded and no action is taken; when it is determined that the abnormality is recoverable, parameter adjustment is automatically performed to attempt recovery; when it is determined that the performance is degraded, downgrading is performed and a warning is issued; when it is determined that the failure is irreversible, shutdown or partial shutdown is immediately performed and the backup thruster is switched.

[0041] Specifically, the handling measures are as follows, depending on the severity of the hazard: No impact (0~0.25): Only log to the telemetry system, no action is taken. Recoverable anomaly (0.25~0.5): Automatic parameter adjustments are performed, such as reducing the grid voltage by 5% or decreasing the gas supply flow by 10%, attempting to restore normal operation. Performance degradation (0.5~0.75): Downgraded operation is implemented, such as switching from full power mode to 80% power mode, and a warning signal is issued to the ground. Irreversible failure (0.75~1): Immediate shutdown or partial shutdown (e.g., only shutting down the high-voltage power supply) is performed, and switching to the backup thruster is initiated, while an emergency fault code is sent.

[0042] More specifically, in step 6, the discharge parameters are monitored in real time using a neural network and the onboard control system. The real-time collected electrical parameters are processed sequentially through steps 1 to 5: first, peak parameters are obtained through VIP screening; then, five-dimensional evaluation indicators are calculated and compared with thresholds; then, the results are input into the judgment matrix to determine the level of each indicator; finally, the hazard level is derived by using the analytic hierarchy process (AHP). This grading strategy follows the principle of "prevention first, graded response," avoiding overreaction to minor anomalies while ensuring rapid protection in the event of serious faults. Using the space ion electric propulsion abnormal discharge hazard assessment method provided in this application embodiment, the service life of ion electric propulsion can be increased by more than 10%, significantly improving on-orbit reliability and ease of use.

[0043] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for assessing the severity of abnormal discharge hazards in space ion electric propulsion, characterized in that, Includes the following steps: Step 1: Initial data screening. Based on the operational life test database of the ion electric propulsion system (30,000 hours on the ground and 20,000 hours in orbit), a prediction set of abnormal discharge and electrical parameters is established. Using the partial least squares regression algorithm, with the variable projection importance index being greater than 1.2 times as the criterion, peak parameters closely related to abnormal discharge of ion electric propulsion are initially screened from stress parameters and reliability parameters. Step 2: Comprehensive evaluation. In the time and frequency domains, the peak parameters selected in Step 1 are associated with the multi-dimensional factors of fault hazard assessment. The multi-dimensional factors include peak power, energy deposition, impact, volatility, and volatility intensity, thereby constructing a multi-dimensional anomaly evaluation index based on electrical parameters and carrying out a comprehensive evaluation of multi-dimensional discharge anomalies. Step 3: Threshold setting. Based on the normal operating parameters of the highest power of space ion electric propulsion design, all parameters are increased by 20% as the overpressure limit. The threshold at which each evaluation index in Step 2 is judged to be abnormal is calculated, and the abnormal judgment boundary of each index is determined. Step 4: Matrix interpretation. Based on the multi-dimensional evaluation indicators constructed in Step 2, an anomaly interpretation matrix closely related to each indicator is established. This matrix has evaluation indicators as rows and hazard levels as columns, with each matrix element representing the interpretation rule of the corresponding indicator under that hazard level. Step 5: Classification. The peak parameters selected in Step 1 are normalized. The normalized parameters are divided into four weight intervals using the analytic hierarchy process: no impact interval, abnormal recoverable interval, performance degradation interval, and irreversible failure interval, thereby establishing an evaluation domain for the degree of hazard of abnormal discharge. Step 6: Hazard assessment. The discharge parameters are monitored in real time using a neural network and onboard control system. The real-time collected electrical parameters are processed sequentially through steps 1 to 5 to determine the hazard level of the current abnormal discharge. Based on the severity of the hazard, one or more of the following treatment methods are automatically adopted: shutdown, partial shutdown, downgrading, or parameter adjustment.

2. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 1, the stress parameters include grid voltage, grid current, anode voltage, anode current, accelerating grid voltage, accelerating grid current, neutralizer heating voltage and current, neutralizer holding voltage and current, cathode heating voltage and current, and cathode holding voltage and current; the reliability parameters include accelerating grid power, grid power, cathode power, and neutralizer power.

3. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 2, characterized in that, In step 1, the operational life test database for 30,000 hours on the ground and 20,000 hours in orbit was constructed by synchronously collecting and recording electrical parameter data, including various stress parameters and reliability parameters, during the continuous operation of the ion electric propulsion system under simulated space environment conditions on the ground for 30,000 hours and the actual operation in orbit for 20,000 hours. The data was then cleaned and normalized.

4. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 3, characterized in that, In step 1, the specific process of using the partial least squares regression algorithm is as follows: taking the abnormal discharge event as the dependent variable and each stress parameter and reliability parameter as the independent variables, a regression model is established by extracting principal components, and the variable projection importance index value of each parameter is calculated. When the index value is greater than 1.2, the parameter is determined to be an important stress parameter closely related to the abnormal discharge.

5. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 2, the peak power is obtained by multiplying the voltage and the maximum value of the instantaneous current; the energy deposition is obtained by integrating the product of voltage and current over time; the impact is obtained by the difference between the peak power and the root mean square power; the volatility is obtained by dividing the difference between the maximum and minimum power values ​​by the average power value; and the volatility intensity is obtained by the root mean square of the sum of the squares of the differences between the power and the average value at each sampling point.

6. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 3, the specific method for setting the threshold is as follows: based on the normal operating parameters of the highest power of ion electric propulsion design, each parameter is multiplied by 1.2 as the upper boundary. At the same time, considering the instrument measurement error and the normal fluctuation range, the larger value of 1.2 times the normal operating parameter and the upper limit of the measurement error is taken as the abnormal threshold.

7. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 5, the specific implementation of the analytic hierarchy process is as follows: construct a hierarchical structure with the degree of hazard of abnormal discharge as the target layer, the peak parameters selected in step 1 as the criterion layer, and the four hazard levels as the scheme layer. Construct a judgment matrix through expert scoring, calculate the weight vector of each parameter, and then map the comprehensive score after weighting the real-time parameters to the interval between 0 and 1 to divide the four level intervals.

8. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 6, the neural network adopts a three-layer backpropagation neural network. The number of input layer nodes corresponds to the number of peak parameters selected in step 1, the number of hidden layer nodes is 1.5 times the number of input layer nodes, and the number of output layer nodes is 4. The network is trained with ground test data to achieve rapid classification of real-time monitoring data.

9. The method for assessing the hazard level of abnormal discharge in space ion electric propulsion according to claim 1, characterized in that, In step 6, the specific handling measures taken according to the degree of harm are as follows: when it is determined that there is no impact, only log is recorded and no action is taken; when it is determined that the abnormality is recoverable, parameter adjustment is automatically performed to attempt recovery; when it is determined that the performance is degraded, downgraded operation is performed and a warning is issued; when it is determined that the failure is irreversible, the power is immediately shut down or partially shut down and switched to the backup thruster.

Images