Intelligent management system of signal processor hardware platform

Abstract

The application discloses an intelligent management system of a signal processor hardware platform, and belongs to the technical field of intelligent management of signal processors, and comprises the following: a data acquisition module, which is used for acquiring the running state data of all signal processors in real time, and obtaining the time sequence multidimensional data of each signal processor according to the running state data; a health assessment module, which is used for assessing each signal processor according to the time sequence multidimensional data to obtain a health degree score, and generating an aggregated health view according to the health degree score; a fault diagnosis module, which is used for judging whether the signal processor has a fault according to the health degree score and the running state data; if the signal processor has a fault, a root cause analysis result is generated; and an automatic processing module, which is used for matching and executing an automatic processing script according to the root cause analysis result.

Description

Technical Field

[0001] This invention belongs to the field of intelligent management technology for signal processors, and specifically relates to an intelligent management system for a signal processor hardware platform. Background Technology

[0002] Signal processing systems are widely used in radar signal processing, base station data processing, high-performance computing, and image processing. These systems typically consist of multiple chassis, each containing multiple signal processing boards (signal processors). This hardware usually requires continuous operation, and its stability and reliability are the cornerstones of ensuring the stability and reliability of ships.

[0003] Currently, the maintenance and management of such signal processing hardware platforms mainly rely on two modes: one is the periodic manual inspection mode, which refers to maintenance personnel entering the computer room or equipment site according to a predetermined cycle to check the working status of the hardware by observing equipment indicator lights, logging into the operating system to view logs, or running diagnostic tools; the other is the reactive maintenance mode, which is that maintenance personnel only intervene to troubleshoot and repair after the system has failed, services have been interrupted, or performance has been severely degraded. However, both of these methods have problems such as high operation and maintenance costs, delayed fault response, and difficulty in fault location.

[0004] Therefore, there is an urgent need for an intelligent management system for signal processor hardware platforms to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention provides an intelligent management system for a signal processor hardware platform to solve the problems of high operation and maintenance costs, delayed fault response, and difficulty in fault location in the prior art.

[0006] To achieve the above objectives, the present invention provides an intelligent management system for a signal processor hardware platform, comprising: The data acquisition module is used to acquire the operating status data of all signal processors in real time, and obtain the time-series multidimensional data of each signal processor based on the operating status data; The health assessment module is used to evaluate each signal processor based on time-series multidimensional data to obtain a health score, and generate an aggregated health view based on the health score. The fault diagnosis module is used to determine whether the signal processor has malfunctioned based on the health score and operating status data; if the signal processor malfunctions, it generates root cause analysis results. The automation processing module is used to match and execute automated processing scripts based on the root cause analysis results.

[0007] As an embodiment of the present invention, the data acquisition module performs the following operations: Obtain the device identification information and driver configuration parameters of each signal processor; The operating status data of each signal processor is acquired through a preset acquisition frequency; the operating status data includes sensor data and operating logs. The device identification information, drive configuration parameters, and operating status data are uniformly encapsulated and timestamped to obtain the time-series multidimensional data of each signal processor.

[0008] As an embodiment of the present invention, the health assessment module performs the following operations: Acquire time-series multidimensional data from the signal processor within a preset time period; The time-series multidimensional data is periodically read to obtain multidimensional indicator data, which includes: temperature, voltage, power consumption, processor utilization, load rate, memory utilization, link bit error rate, and mean time between failures. Based on the analytic hierarchy process, the weight vector of each indicator in the multidimensional indicator data is determined, and the initial health score of the signal processor is obtained by weighted summation of the weight vector of each indicator and the multidimensional indicator data. The penalty factor is calculated as follows: In the formula, Indicates the penalty factor. Indicates the first A weight vector of multidimensional indicator data, Indicates the number of multidimensional indicator data. Indicates the first Multidimensional indicator data, Indicates the first Alarm thresholds for multi-dimensional indicator data Indicates the first The limit threshold of multidimensional indicator data; The signal processor's health score is obtained by multiplying the penalty factor by the initial health score. A visual topology map is constructed based on the health score of the signal processor to obtain an aggregated health view.

[0009] As an embodiment of the present invention, the fault analysis module performs the following operations: The health score and operating status data are compared with the alarm threshold range in the preset alarm trigger rule base; if no health score or operating status data does not fall within the alarm threshold range, it is determined that the signal processor has not malfunctioned. If the health score and operating status data do not meet the alarm threshold range, the signal processor is determined to have malfunctioned and the comparison result is output. The comparison result includes: time of occurrence, location, and alarm type. The comparison results and operational status data are matched with historical fault cases in the case library, and the historical fault case with the highest matching degree is selected as the target case. The causes of failure in the target case are used as the results of root cause analysis.

[0010] As an embodiment of the present invention, the automated processing module performs the following steps: Based on the root cause analysis results and the preset script matching index, an automated processing script is obtained by matching in the script library; Execute the automated processing script and determine whether the signal processor fault has been eliminated; if the fault has not been eliminated, generate a fault problem package based on the operating status data, target case, health score, root cause analysis results, and time series multidimensional data, and send an alarm message to the administrator.

[0011] As an embodiment of the present invention, the fault analysis module also performs the following operations: After the administrator has repaired the faulty signal processor, they obtain the cause of the fault, maintenance method, comparison results, and time series multidimensional data to generate historical fault cases and store them in the case library.

[0012] As one embodiment of the present invention, it also includes: The fault impact prediction module performs the following operations: When a signal processor malfunctions, the malfunctioning signal processor will be used as the target signal processor. Obtain the target signal processor's operating status data and device relationship diagram; the device relationship diagram is a graph structure describing the physical connection relationships and task dependencies between each chassis and signal processor, and each node is associated with task flow path information; the task flow path information includes the task flow identifier, task flow priority, real-time traffic and service quality requirements passing through the node; The fault source nodes corresponding to the root cause analysis results are marked as abnormal nodes. Based on the equipment relationship graph, a graph convolutional network model is used to deduce the fault propagation path and predict the set of secondary nodes that may be affected by the fault and the fault probability. The probability of interruption of the affected task flow is calculated based on the set of secondary nodes, the failure probability, and the task flow path information, as shown below: In the formula, Indicates the first The probability of interruption of the affected task flow. Indicates the first The set of all nodes in the affected task flow. Indicates the first The first of the affected task flows The probability of failure of each node; The overall impact on tasks is calculated based on the interruption probability of all affected task flows, as shown below: In the formula, Indicates the overall impact of the task. Indicates the number of task flows affected. Indicates the first The priority weights of the affected task flows, Indicates the first Real-time traffic of the affected task flow, Indicates the first The service quality requirements of the affected task flow Represents the mapping loss function; Determine whether the overall impact of the task exceeds the preset impact threshold; if it does, directly send an alarm message and comparison results to the administrator.

[0013] The beneficial effects of this invention are as follows: by using a health assessment module and a fault diagnosis module, the health status and fault conditions of each signal processor are monitored and managed, and the faulty signal processor can be quickly located after a fault occurs; at the same time, the faulty signal processor is automatically intervened, and if the faulty signal processor still has a fault after the automatic intervention, a fault problem package is generated and an alarm message is sent to the administrator; this solves the problems of high operation and maintenance costs, delayed fault response and difficulty in fault location of existing management methods.

[0014] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0015] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a schematic diagram of the modules of the present invention; Figure 2 This is a schematic diagram of the execution flow of the data acquisition module of the present invention; Figure 3 This is a schematic diagram of the execution flow of the health assessment module of the present invention. Detailed Implementation

[0016] like Figures 1-3 As shown, the present invention provides an intelligent management system for a signal processor hardware platform, comprising: The data acquisition module is used to acquire the operating status data of all signal processors in real time, and obtain the time-series multidimensional data of each signal processor based on the operating status data; The health assessment module is used to evaluate each signal processor based on time-series multidimensional data to obtain a health score, and generate an aggregated health view based on the health score. The fault diagnosis module is used to determine whether the signal processor has malfunctioned based on the health score and operating status data; if the signal processor malfunctions, it generates root cause analysis results. The automation processing module is used to match and execute automated processing scripts based on the root cause analysis results.

[0017] The working principle of the above technical solution is as follows: The signal processing hardware platform includes several chassis, each containing multiple signal processors of the same or different models, forming an array to process signals generated by equipment such as ships. Current methods for maintaining this signal processing platform often rely on manual, periodic inspections, which are prone to numerous problems. Therefore, this technical solution, during the management of the signal processing platform, uses a data acquisition module to acquire the real-time operating status data of each signal processor, and then obtains the corresponding time-series multidimensional data for each signal processor based on this data. Finally, a health assessment module evaluates the health status of each signal processor based on its time-series multidimensional data, thus obtaining a health score for each signal processor. Subsequently, a 3D aggregated health view is constructed based on the health score of each signal processor and its chassis location. Simultaneously, the fault diagnosis module determines whether a signal processor has malfunctioned based on its health score and operational status data. If a signal processor malfunctions, the assessment is repeated every preset time interval (10 minutes). When a signal processor malfunctions, its operational status data is analyzed to obtain root cause analysis results. After confirming a malfunction, the automation module matches an automation script based on the root cause analysis results and controls the corresponding module of the malfunctioning signal processor to execute, performing automated intervention. If the malfunction persists after automated intervention, a fault problem package is generated and an alarm message is sent to the administrator. The beneficial effects of the above technical solution are as follows: Through the established health assessment and fault diagnosis modules, the health status and fault conditions of each signal processor are monitored and managed. After a signal processor malfunctions, the faulty signal processor can be quickly located. Simultaneously, automated intervention is performed on the faulty signal processor. If the faulty signal processor still malfunctions after automated intervention, a fault problem package is generated and an alarm message is sent to the administrator. This solves the problems of high maintenance costs, delayed fault response, and difficulty in fault location associated with existing management methods.

[0018] In one embodiment, the data acquisition module performs the following operations: Obtain the device identification information and driver configuration parameters of each signal processor; The operating status data of each signal processor is acquired through a preset acquisition frequency; the operating status data includes sensor data and operating logs. The device identification information, drive configuration parameters, and operating status data are uniformly encapsulated and timestamped to obtain the time-series multidimensional data of each signal processor.

[0019] The working principle and beneficial effects of the above technical solution are as follows: The data acquisition module first scans and establishes a communication link with each signal processor to obtain device identification information and corresponding drive configuration parameters; then, the module collects the operating status data of each signal processor in real time through in-band or out-of-band management channels according to a preset acquisition frequency. The operating status data includes at least sensor data and operating logs; finally, the module uniformly encapsulates the device identification information, drive configuration parameters, and collected operating status data, and marks each set of data with a precise timestamp to generate time-series multidimensional data with a unified format and time dimension, and outputs it to the upper-layer module to provide a standardized data foundation for subsequent health assessment and fault diagnosis. By automating data collection at preset frequencies, traditional manual inspections have been replaced, enabling 24 / 7 real-time monitoring of hardware status and significantly improving the timeliness and accuracy of data acquisition. By standardizing and encapsulating multi-source heterogeneous data and marking it with timestamps, high-quality time-series multidimensional data has been constructed, laying a solid data foundation for upper-level modules to perform health assessments, trend analysis, and intelligent fault diagnosis. This effectively reduces the need for manual intervention and improves overall operation and maintenance efficiency.

[0020] In one embodiment, the health assessment module performs the following operations: Acquire time-series multidimensional data from the signal processor within a preset time period; The time-series multidimensional data is periodically read to obtain multidimensional indicator data, which includes: temperature, voltage, power consumption, processor utilization, load rate, memory utilization, link bit error rate, and mean time between failures. Based on the analytic hierarchy process, the weight vector of each indicator in the multidimensional indicator data is determined, and the initial health score of the signal processor is obtained by weighted summation of the weight vector of each indicator and the multidimensional indicator data. The penalty factor is calculated as follows: In the formula, Indicates the penalty factor. Indicates the first A weight vector of multidimensional indicator data, Indicates the number of multidimensional indicator data. Indicates the first Multidimensional indicator data, Indicates the first Alarm thresholds for multi-dimensional indicator data Indicates the first The limit threshold of multidimensional indicator data; The signal processor's health score is obtained by multiplying the penalty factor by the initial health score. A visual topology map is constructed based on the health score of the signal processor to obtain an aggregated health view.

[0021] The working principle and beneficial effects of the above technical solution are as follows: The health assessment module first obtains time-series multidimensional data of the signal processor within a preset time window from the data acquisition module, and periodically reads it to extract multidimensional indicator data including temperature, voltage, power consumption, processor utilization, load rate, memory occupancy rate, link bit error rate, and fault-free operation time. Subsequently, the module constructs a judgment matrix based on the analytic hierarchy process (AHP), calculates the eigenvector corresponding to the largest eigenvalue and normalizes it to determine the weight vector corresponding to each indicator, and then performs a weighted summation of the current multidimensional indicator data according to the weight vector to obtain the initial health score of the signal processor. Before the weighted summation of the multidimensional indicator data, the data needs to be preprocessed, such as unifying the dimensions. It is worth noting that indicator data such as load rate and link bit error rate, which require additional calculation, have a different sampling frequency than data such as temperature and voltage obtained directly from sensors (sometimes requiring manual calculation). Therefore, each time the health score is calculated, the specific weights of the obtained indicator data are re-determined using the AHP. Based on this, the module further calculates the penalty factor. The penalty factor is calculated as follows: for any multi-dimensional indicator data, when its current value exceeds the preset alarm threshold, the individual penalty contribution is calculated according to the proportional relationship between its degree of exceedance and the limit threshold. The penalty contributions of all key indicators are then weighted and summed according to their corresponding weights, and the value is subtracted from 1 to obtain the final penalty factor. Finally, the module multiplies the penalty factor with the initial health score to obtain the corrected final health score, and constructs a visual topology map based on the health scores of all signal processors to generate a chassis-level aggregated health view. That is, based on the existing BIM model construction technology, a three-dimensional signal processor platform model is constructed, and then the color of the signal processor is determined according to the health score of the signal processor, and then the three-dimensional signal processor platform model is updated.

[0022] By introducing a penalty factor mechanism, abnormal indicators that exceed the alarm threshold are dynamically penalized, enabling the score to truly reflect the current operational risk status of the equipment and avoiding the problem that simple weighted summation may mask serious exceedances of individual indicators. The final generated visual aggregated health view transforms complex multidimensional data into an intuitive graphical display, allowing maintenance personnel to clearly grasp the health status of the entire chassis and each board, significantly improving the efficiency and accuracy of status perception and providing a reliable decision-making basis for subsequent fault early warning and automated handling.

[0023] In one embodiment, the fault analysis module performs the following operations: The health score and operating status data are compared with the alarm threshold range in the preset alarm trigger rule base; if no health score or operating status data does not fall within the alarm threshold range, it is determined that the signal processor has not malfunctioned. If the health score and operating status data do not meet the alarm threshold range, the signal processor is determined to have malfunctioned and the comparison result is output. The comparison result includes: time of occurrence, location, and alarm type. The comparison results and operational status data are matched with historical fault cases in the case library, and the historical fault case with the highest matching degree is selected as the target case. The causes of failure in the target case are used as the results of root cause analysis.

[0024] The working principle and beneficial effects of the above technical solution are as follows: First, the health score and real-time operating status data output by the health assessment module are acquired, and then compared item by item with the alarm threshold range in the preset alarm triggering rule base. If all data are within the normal threshold range, it is determined that the current signal processor has not malfunctioned. If any indicator (health score or operating status data) exceeds the alarm threshold range, it is immediately determined that the signal processor has malfunctioned, and the comparison result is output. The comparison result includes at least the time of occurrence of the fault, the specific location, and the alarm type (such as CPU chip overheating). Subsequently, the module uses the comparison result and the current operating status data as query conditions to perform similarity matching with historical fault cases stored in the case base, calculates the matching degree between the current fault and each historical fault case, and selects the historical fault case with the highest matching degree as the target case. Finally, the module extracts the historical fault causes recorded in the target case and outputs them as the root cause analysis result of the current fault. The threshold comparison mechanism enables real-time automatic fault diagnosis of signal processors, allowing for immediate detection of anomalies and recording of precise fault occurrence time, location, and type. This completely replaces the passive discovery mode of manual inspection, significantly shortening fault response time. By intelligently matching the current fault with a historical case database, the system leverages accumulated fault handling experience to provide root cause inferences for the current fault, significantly improving the accuracy and efficiency of fault diagnosis. The final root cause analysis results provide clear and traceable decision-making basis for subsequent automated handling or manual maintenance, effectively solving the problems of difficult fault location and reliance on expert experience in complex hardware platforms, further reducing operation and maintenance costs.

[0025] In one embodiment, the automated processing module performs the following steps: Based on the root cause analysis results and the preset script matching index, an automated processing script is obtained by matching in the script library; Execute the automated processing script and determine whether the signal processor fault has been eliminated; if the fault has not been eliminated, generate a fault problem package based on the operating status data, target case, health score, root cause analysis results, and time series multidimensional data, and send an alarm message to the administrator.

[0026] The working principle and beneficial effects of the above technical solution are as follows: The automated processing module, as the final execution unit of the intelligent management system, is responsible for converting the diagnostic results output by the fault analysis module into specific handling actions. First, it obtains the root cause analysis results, extracts key features such as fault type codes, fault source device identifiers, and suggested operation types from the results, and constructs a matching feature vector. Then, it compares the matching feature vector with a preset script matching index and retrieves the corresponding automated processing script from the script library according to a preset multi-level matching strategy. The script library pre-stores various handling scripts categorized by operation object and risk level, including reset scripts, process restart scripts, configuration reload scripts, and chassis-level scripts. The module includes power supply master / slave switching scripts, fan speed control scripts, backplane bus reset scripts, and link-level port disable / enable scripts and link switching scripts. After successful matching, the module sends and executes automated processing scripts to the faulty device through the remote management interface. After execution, the module re-acquires the operating status data of the signal processor to determine whether the fault has been eliminated. If the fault has not been eliminated, the module automatically collects all information related to the current fault, including real-time operating status data, matched target cases, health scores, root cause analysis results, and time-series multidimensional data, integrates them to generate a complete fault problem package, and sends alarm information to the administrator through a preset alarm channel to prompt manual intervention.

[0027] Through pre-defined multi-level matching strategies and a rich script library, the system automates the handling of various faults, directly translating root cause analysis results into executable operations. This completely changes the traditional maintenance method of manually logging into devices to execute commands, significantly improving fault recovery speed. For complex faults that cannot be recovered through automated handling, the module automatically generates a fault problem package containing complete fault information, providing comprehensive data support for manual intervention and avoiding redundant troubleshooting and information omissions. This greatly reduces reliance on real-time on-site maintenance personnel, effectively minimizing business downtime and improving the availability and operational efficiency of the entire signal processing hardware platform. In one embodiment, the fault analysis module also performs the following operations: After the administrator has repaired the faulty signal processor, they obtain the cause of the fault, maintenance method, comparison results, and time series multidimensional data to generate historical fault cases and store them in the case library.

[0028] The working principle and beneficial effects of the above technical solution are as follows: Based on automatic fault diagnosis, it further possesses the ability to learn from cases and update the knowledge base. When the administrator completes maintenance operations on the faulty signal processor, the module automatically triggers the case generation process: First, it acquires complete data throughout the entire lifecycle of the fault, including the fault cause determined in the root cause analysis results, the actual maintenance methods used by the administrator (such as replacing boards, upgrading firmware, cleaning heat sinks, etc.), the comparison results recorded when the fault occurred (occurrence time, location, alarm type), and time-series multi-dimensional data (change curves of indicators such as temperature, voltage, and power consumption) within a preset time window before and after the fault occurred. Subsequently, the module integrates and structures this multi-source data, generating standardized historical fault cases according to preset case templates. Finally, the module stores the generated new cases in the case library and updates the case index, enabling the case to be retrieved and matched in future fault diagnosis processes.

[0029] In one embodiment, it also includes: The fault impact prediction module performs the following operations: When a signal processor malfunctions, the malfunctioning signal processor will be used as the target signal processor. Obtain the target signal processor's operating status data and device relationship diagram; the device relationship diagram is a graph structure describing the physical connection relationships and task dependencies between each chassis and signal processor, and each node is associated with task flow path information; the task flow path information includes the task flow identifier, task flow priority, real-time traffic and service quality requirements passing through the node; The fault source nodes corresponding to the root cause analysis results are marked as abnormal nodes. Based on the equipment relationship graph, a graph convolutional network model is used to deduce the fault propagation path and predict the set of secondary nodes that may be affected by the fault and the fault probability. The probability of interruption of the affected task flow is calculated based on the set of secondary nodes, the failure probability, and the task flow path information, as shown below: In the formula, Indicates the first The probability of interruption of the affected task flow. Indicates the first The set of all nodes in the affected task flow. Indicates the first The first of the affected task flows The probability of failure of each node; The overall impact on tasks is calculated based on the interruption probability of all affected task flows, as shown below: In the formula, Indicates the overall impact of the task. Indicates the number of task flows affected. Indicates the first The priority weights of the affected task flows, Indicates the first Real-time traffic of the affected task flow, Indicates the first The service quality requirements of the affected task flow Represents the mapping loss function; Determine whether the overall impact of the task exceeds the preset impact threshold; if it does, directly send an alarm message and comparison results to the administrator.

[0030] The working principle and beneficial effects of the above technical solution are as follows: The fault impact prediction module starts after a signal processor malfunctions. First, it marks the faulty device as the target signal processor and acquires its operating status data and a pre-generated device relationship diagram. This diagram structurally describes the physical connections and task dependencies between various chassis and signal processors. Each node is also associated with task flow path information, including the task flow identifier, priority, real-time traffic, and quality of service requirements passing through that node. Subsequently, the module marks the fault source node corresponding to the root cause analysis results as an abnormal node and, based on the device relationship diagram, uses a graph convolutional network model to deduce the fault propagation path, predicting the set of potentially affected secondary nodes and their failure probabilities. Based on this, the module, combined with the task flow path information, calculates the interruption probability of each affected task flow, taking into account the failure risk of all nodes on the task flow path. Then, the module comprehensively calculates the overall task impact degree based on the interruption probability, priority weight, real-time traffic, and quality of service requirements of each task flow using a mapping loss function. Finally, the overall impact of the task is compared with a preset threshold. If the threshold is exceeded, an alarm message is immediately sent to the administrator and the comparison results are output. This module accurately predicts fault propagation paths through graph convolutional networks, enabling cross-level quantitative assessment of the impact of equipment failures on tasks. By combining task flow path information to calculate the probability of interruption, it can finely identify the affected business flows and their risk levels. Through multi-dimensional comprehensive calculation of the overall task impact and the introduction of a threshold alarm mechanism, it ensures that manual intervention is triggered only when a fault may cause serious business disruption.

[0031] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. An intelligent management system for a signal processor hardware platform, characterized in that, include: The data acquisition module is used to acquire the operating status data of all signal processors in real time, and obtain the time-series multidimensional data of each signal processor based on the operating status data; The health assessment module is used to evaluate each signal processor based on time-series multidimensional data to obtain a health score, and generate an aggregated health view based on the health score. The fault diagnosis module is used to determine whether the signal processor has malfunctioned based on the health score and operating status data; If the signal processor malfunctions, root cause analysis results are generated. The automation processing module is used to match and execute automated processing scripts based on the root cause analysis results.

2. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, The data acquisition module performs the following operations: Obtain the device identification information and driver configuration parameters of each signal processor; The operating status data of each signal processor is acquired through a preset acquisition frequency; the operating status data includes sensor data and operating logs. The device identification information, drive configuration parameters, and operating status data are uniformly encapsulated and timestamped to obtain the time-series multidimensional data of each signal processor.

3. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, The health assessment module performs the following operations: Acquire time-series multidimensional data from the signal processor within a preset time period; The time-series multidimensional data is periodically read to obtain multidimensional indicator data, which includes: temperature, voltage, power consumption, processor utilization, load rate, memory utilization, link bit error rate, and mean time between failures. Based on the analytic hierarchy process, the weight vector of each indicator in the multidimensional indicator data is determined, and the initial health score of the signal processor is obtained by weighted summation of the weight vector of each indicator and the multidimensional indicator data. The penalty factor is calculated as follows: In the formula, Indicates the penalty factor. Indicates the first A weight vector of multidimensional indicator data, Indicates the number of multidimensional indicator data. Indicates the first Multidimensional indicator data, Indicates the first Alarm thresholds for multi-dimensional indicator data Indicates the first The limit threshold of multidimensional indicator data; The signal processor's health score is obtained by multiplying the penalty factor by the initial health score. A visual topology map is constructed based on the health score of the signal processor to obtain an aggregated health view.

4. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, The fault diagnosis module performs the following operations: The health score and operating status data are compared with the alarm threshold range in the preset alarm trigger rule base; if no health score or operating status data does not fall within the alarm threshold range, it is determined that the signal processor has not malfunctioned. If the health score and operating status data do not meet the alarm threshold range, the signal processor is determined to have malfunctioned and the comparison result is output. The comparison result includes: time of occurrence, location, and alarm type. The comparison results and operational status data are matched with historical fault cases in the case library, and the historical fault case with the highest matching degree is selected as the target case. The causes of failure in the target case are used as the results of root cause analysis.

5. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, The automated processing module performs the following steps: Based on the root cause analysis results and the preset script matching index, an automated processing script is obtained by matching in the script library; Execute the automated processing script and determine whether the signal processor fault has been eliminated; If the fault is not eliminated, a fault problem package is generated based on the operating status data, target case, health score, root cause analysis results, and time series multidimensional data, and an alarm message is sent to the administrator.

6. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, The fault analysis module also performs the following operations: After the administrator has repaired the faulty signal processor, they obtain the cause of the fault, maintenance method, comparison results, and time series multidimensional data to generate historical fault cases and store them in the case library.

7. The intelligent management system for a signal processor hardware platform according to claim 1, characterized in that, Also includes: The fault impact prediction module performs the following operations: When a signal processor malfunctions, the malfunctioning signal processor will be used as the target signal processor. Obtain the target signal processor's operating status data and device relationship diagram; the device relationship diagram is a graph structure describing the physical connection relationships and task dependencies between each chassis and signal processor, and each node is associated with task flow path information; the task flow path information includes the task flow identifier, task flow priority, real-time traffic and service quality requirements passing through the node; The fault source nodes corresponding to the root cause analysis results are marked as abnormal nodes. Based on the equipment relationship graph, a graph convolutional network model is used to deduce the fault propagation path and predict the set of secondary nodes that may be affected by the fault and the fault probability. The probability of interruption of the affected task flow is calculated based on the set of secondary nodes, the failure probability, and the task flow path information, as shown below: In the formula, Indicates the first The probability of interruption of the affected task flow. Indicates the first The set of all nodes in the affected task flow. Indicates the first The first of the affected task flows The probability of failure of each node; The overall impact on tasks is calculated based on the interruption probability of all affected task flows, as shown below: In the formula, Indicates the overall impact of the task. Indicates the number of task flows affected. Indicates the first The priority weights of the affected task flows, Indicates the first Real-time traffic of the affected task flow, Indicates the first The service quality requirements of the affected task flow Represents the mapping loss function; Determine whether the overall impact of the task exceeds the preset impact threshold; if it does, directly send an alarm message and comparison results to the administrator.

Images