An active anti-vibration system for vehicle-mounted graphics servers in high-vibration environments

Abstract

The application relates to the technical field of vehicle-mounted equipment and discloses a vehicle-mounted graphic server active anti-seismic system for a high-vibration environment, which comprises the following steps: first, collecting real-time vibration data streams of a vehicle-mounted graphic server, constructing an energy atlas, analyzing resonance modal clusters, and generating a damping control path; then, analyzing dominant resonance modes, identifying critical oscillation nodes, and calculating an inertia compensation coefficient; then, performing nonlinear filtering on a vibration frequency spectrum based on the coefficient to obtain an anti-interference frequency band set, generating a servo driving instruction set through topological mapping, inquiring dynamic stiffness parameters, and marking a parameter distribution vector; then, loading the vector to an active memory controller, monitoring energy dissipation states, adaptively adjusting gain threshold combinations, and calculating an anti-seismic robustness index; finally, generating vibration characteristic lines, extracting steady-state, transient, and chaotic components, and constructing a multi-cascade active anti-seismic mechanism. The application can guarantee the anti-seismic stability of vehicle-mounted graphic calculation.

Description

Technical Field

[0001] The invention relates to an active anti-vibration system for a vehicle-mounted graphics server facing a high-vibration environment, belonging to the technical field of vehicle-mounted equipment. Background Art

[0002] An on-board graphics server refers to a high-performance computing device that processes graphics rendering, visual computing and other tasks in real time while the vehicle is driving. Its stability is directly affected by the vehicle's vibration environment.

[0003] Existing seismic design often relies on passive damping solutions, such as rubber pads for vibration isolation or spring-loaded mechanical buffers. These methods only absorb vibration energy within a specific frequency band and are unable to adapt to broadband vibrations found in complex road conditions (such as the combination of high-frequency bumps and low-frequency shakes). This can lead to loose hardware connections or storage device failure. Furthermore, traditional structures lack intelligent recognition of vibration direction (such as random impact in three-dimensional axial directions), which can easily cause graphics processing interruptions or heat dissipation anomalies during sudden braking or off-roading. Therefore, an active seismic mitigation system for in-vehicle graphics servers is needed. This system, through real-time vibration sensing and dynamic compensation mechanisms, can ensure the seismic stability of in-vehicle graphics computing. Summary of the Invention

[0004] The present invention provides an active anti-seismic system for an on-vehicle graphics server in a high-vibration environment, the main purpose of which is to ensure the anti-seismic stability of on-vehicle graphics computing.

[0005] To achieve the above-mentioned object, the present invention provides an active anti-seismic system for an in-vehicle graphics server in a high-vibration environment, comprising: a path generation module, a compensation coefficient calculation module, a vector marking module, an index calculation module, and a mechanism construction module;

[0006] The path generation module is configured to collect a real-time vibration data stream from a graphics server running in a vehicle-mounted environment, construct a vibration energy spectrum corresponding to the real-time vibration data stream, analyze the resonance mode clusters in the vibration energy spectrum, and generate a damping control path corresponding to the resonance mode clusters.

[0007] The compensation coefficient calculation module is used to analyze the dominant resonance mode corresponding to the damping control path, identify the critical oscillation node in the damping control path based on the dominant resonance mode, and calculate the inertia compensation coefficient corresponding to the critical oscillation node;

[0008] The vector marking module is configured to perform nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set, perform topological mapping on the anti-interference frequency band set to generate a servo drive instruction set, query dynamic stiffness parameters corresponding to the servo drive instruction set, and mark parameter allocation vectors corresponding to the dynamic stiffness parameters;

[0009] The index calculation module is configured to load the parameter allocation vector into a preset active memory controller, monitor the energy dissipation state in the active memory controller, adaptively adjust the gain threshold combination of the real-time vibration spectrum during the servo compensation process based on the energy dissipation state, and calculate the seismic robustness index corresponding to the gain threshold combination;

[0010] The mechanism construction module is used to generate the vibration characteristic pattern corresponding to the vehicle environment based on the seismic robustness index, extract the steady-state component, transient component and chaotic component in the vibration characteristic pattern, and construct a multi-cascade active seismic resistance mechanism corresponding to the graphics server in the vehicle environment based on the steady-state component, the transient component and the chaotic component.

[0011] Optionally, generating a damping control path corresponding to the resonant mode cluster includes:

[0012] Analyzing the modal frequency distribution in the resonant mode cluster;

[0013] identifying energy convergence nodes in the modal frequency distribution;

[0014] Mapping the phase offset interval corresponding to the energy convergence node;

[0015] Analyzing the damping compensation amount corresponding to the energy convergence node based on the phase offset interval;

[0016] Based on the damping compensation amount, a damping control path corresponding to the resonant mode cluster is generated.

[0017] Optionally, analyzing the dominant resonance mode corresponding to the damping control path includes:

[0018] Extracting path distribution characteristics in the damping control path;

[0019] Based on the path distribution characteristics, dividing the resonance frequency band intervals corresponding to the damping control path;

[0020] Marking the energy extreme value points in the resonance frequency band;

[0021] Analyzing the modal coupling relationship corresponding to the energy extreme point;

[0022] Based on the modal coupling relationship, a dominant resonance mode corresponding to the damping control path is determined.

[0023] Optionally, the identifying a critical oscillation node in the damping control path based on the dominant resonance mode includes:

[0024] Analyzing the resonance distribution energy corresponding to the dominant resonance mode;

[0025] Querying the energy diffusion period corresponding to the resonance distribution energy;

[0026] extracting a peak oscillation segment within the energy diffusion period;

[0027] generating a node oscillation label corresponding to the peak oscillation segment;

[0028] Based on the node oscillation label, identify the critical oscillation node in the damping control path

[0029] Optionally, calculating the inertia compensation coefficient corresponding to the critical oscillation node includes:

[0030] The inertia compensation coefficient corresponding to the critical oscillation node is calculated using the following formula:

[0031]

[0032] in, represents the inertia compensation coefficient corresponding to the critical oscillation node, and Represent the lower and upper limits of the vibration frequency, respectively. Indicates frequency The corresponding vibration amplitude function, represents the number of different frequency components in the critical oscillation node, Indicates the number index of different frequency components, Indicates the The amplitude of the frequency component, Indicates the The frequency of the frequency component, represents the sampling period of the critical oscillation node, Indicates time Vibration velocity function.

[0033] Optionally, the performing nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set includes:

[0034] Analyzing the compensation response component corresponding to the inertia compensation coefficient;

[0035] parsing spectrum response data corresponding to the real-time vibration spectrum in the vehicle environment based on the compensation response component;

[0036] Querying a frequency band response weight set corresponding to the spectrum response data;

[0037] screening an optimized frequency band group in the frequency band response weight set;

[0038] Nonlinear filtering is performed on the parameters within the optimized frequency band group to obtain an anti-interference frequency band set.

[0039] Optionally, the marking of the parameter allocation vector corresponding to the dynamic stiffness parameter includes:

[0040] Parsing parameter identifiers corresponding to the dynamic stiffness parameters;

[0041] Dividing the modal coupling domain corresponding to the dynamic stiffness parameter based on the parameter identifier;

[0042] locating dominant stiffness components in the modal coupling domain;

[0043] Analyzing the distribution offset value corresponding to the dominant stiffness component;

[0044] The parameter allocation vector corresponding to the dynamic stiffness parameter is marked based on the allocation offset value.

[0045] Optionally, adaptively adjusting a gain threshold combination of the real-time vibration spectrum during servo compensation based on the energy dissipation state includes:

[0046] Extracting energy consumption characteristic parameters corresponding to the energy dissipation state;

[0047] Generating a gain adjustment list corresponding to the energy consumption characteristic parameters;

[0048] Calculating, item by item, a modified gain threshold corresponding to the real-time vibration spectrum based on the gain adjustment list;

[0049] Performing frequency domain segmentation screening on the modified gain threshold to obtain an effective gain threshold item;

[0050] Based on the effective gain threshold item, the gain threshold combination of the real-time vibration spectrum in the servo compensation process is adaptively adjusted.

[0051] Optionally, calculating the seismic robustness index corresponding to the gain threshold combination includes:

[0052] The seismic robustness index corresponding to the gain threshold combination is calculated using the following formula:

[0053]

[0054] in, represents the seismic robustness index corresponding to the gain threshold combination, represents the number of gain parameters in the gain threshold combination, Indicates the number index of gain parameters, Indicates the The weight coefficient of the gain parameter, and They represent the start time and end time of the vibration time interval respectively. Indicates the The adjustment coefficient corresponding to the gain parameter is Indicates the Gain threshold over time The change function of represents the number of energy dissipation parameters in the gain threshold combination, represents the quantitative index of energy dissipation parameters, Indicates the The weight coefficient of the energy dissipation parameter, Indicates the The actual value of the energy dissipation parameter, Indicates the Baseline values ​​of energy dissipation parameters

[0055] Optionally, generating a vibration characteristic pattern corresponding to the vehicle environment based on the seismic robustness index includes:

[0056] Analyzing a stability interval threshold corresponding to the seismic robustness index;

[0057] extracting a vibration modal entropy value in the vehicle environment according to the stability interval threshold;

[0058] constructing a vibration primitive sequence corresponding to the vehicle environment based on the vibration modal entropy value;

[0059] Iteratively reconstructing the vibration primitive sequence to obtain a sequence vibration seed;

[0060] Based on the sequence vibration seeds, a vibration characteristic pattern corresponding to the vehicle environment is generated.

[0061] Compared with the problems described in the background technology, the present invention can accurately construct a vibration energy spectrum to identify resonance modes by collecting real-time vibration data streams of the graphics server running in the vehicle environment, provide data support for dynamic damping control, and realize inertia compensation and nonlinear filtering under broadband vibration by analyzing vibration characteristics, and then generate servo drive instructions to optimize anti-seismic parameters, and finally construct a multi-cascade active anti-seismic mechanism to ensure the stable operation of the equipment under complex road conditions. Furthermore, the present invention can accurately locate the key frequency range with the highest vibration energy proportion by analyzing the dominant resonance mode corresponding to the damping control path, provide a core basis for the identification of critical oscillation nodes, lay an efficient modal analysis foundation for subsequent nonlinear filtering and servo drive instruction generation, and effectively enhance the anti-resonance ability of the vehicle-mounted graphics server in a broadband vibration environment. Furthermore, the present invention performs nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set, which can accurately adapt to the characteristics of the critical oscillation nodes, dynamically optimize the frequency domain filtering parameters, and make the anti-interference The scrambled band set effectively shields resonance energy, providing a pure spectrum foundation for subsequent servo drive command generation, and improving the system's anti-vibration stability under complex road conditions. Furthermore, the present invention loads the parameter allocation vector into a preset active memory controller and monitors the energy dissipation state in the active memory controller. This enables the active memory controller to accurately execute the anti-vibration strategy based on the dynamic stiffness parameter, ensuring that the servo drive system responds to vibration interference with the optimal parameter combination. This provides data support for the system's adaptive adjustment of the parameter allocation vector, ensuring the stable operation of the vehicle-mounted graphics server in complex vibration environments. Finally, based on the anti-vibration robustness index, the present invention generates a vibration characteristic pattern corresponding to the vehicle-mounted environment, accurately extracts the vibration pattern under complex road conditions, and constructs a dedicated vibration pattern for the vehicle-mounted graphics server. This can help the system predict the resonance risk under different working conditions, provide intuitive feature guidance for dynamically optimizing anti-vibration strategies, promote the upgrade of anti-vibration control from experience-based adaptation to data-driven, and continuously improve the stability and reliability of vehicle-mounted hardware in various vibration scenarios. Therefore, the embodiments of the present invention provide an active anti-vibration system and system for a vehicle-mounted graphics server in a high-vibration environment, which can ensure the anti-vibration stability of vehicle-mounted graphics computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A schematic diagram of a process flow of an active anti-vibration system for an in-vehicle graphics server in a high-vibration environment provided by one embodiment of the present invention;

[0063] Figure 2 A schematic diagram of a task scheduling process in an active anti-seismic system for an on-board graphics server in a high-vibration environment provided by one embodiment of the present invention;

[0064] Figure 3A schematic diagram of a module for implementing an active anti-vibration system for an in-vehicle graphics server in a high-vibration environment, provided by an embodiment of the present invention.

[0065] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0066] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0067] In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0068] In practice, the server-side device deployed in an active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments may be composed of one or more devices. The above-mentioned active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments can be implemented as a service instance, a virtual machine, or hardware devices. For example, the active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments can be implemented as a service instance deployed on one or more devices in a cloud node. Simply put, the active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments can be understood as software deployed on a cloud node that provides an active vibration mitigation service for an in-vehicle graphics server designed for high-vibration environments to each client. Alternatively, the active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments can be implemented as a virtual machine deployed on one or more devices in a cloud node. Application software for managing each client can be installed in the virtual machine. Alternatively, the active vibration mitigation system for an in-vehicle graphics server designed for high-vibration environments can be implemented as a server-side device composed of multiple hardware devices of the same or different types, with one or more hardware devices configured to provide the active vibration mitigation service for an in-vehicle graphics server designed for high-vibration environments to each client.

[0069] In terms of implementation, the active vibration-resistant system for an in-vehicle graphics server designed for high-vibration environments and the user end are mutually compatible. Specifically, the active vibration-resistant system for an in-vehicle graphics server designed for high-vibration environments is implemented as an application installed on a cloud service platform, and the user end is implemented as a client that establishes a communication connection with the application. Alternatively, the active vibration-resistant system for an in-vehicle graphics server designed for high-vibration environments is implemented as a website, and the user end is implemented as a webpage. Alternatively, the active vibration-resistant system for an in-vehicle graphics server designed for high-vibration environments is implemented as a cloud service platform, and the user end is implemented as a mini-program within an instant messaging application.

[0070] Example 2:

[0071] Reference Figure 1 FIG. 1 is a functional module diagram of an active anti-vibration system for an in-vehicle graphics server in a high-vibration environment, provided by one embodiment of the present invention.

[0072] The active vibration-resistant system 100 for an on-board graphics server for high-vibration environments described in the present invention can be installed in a cloud server. In terms of implementation, it can be implemented as one or more service devices, or as an application installed in the cloud (e.g., a server or server cluster for online drinking water quality monitoring), or it can be developed as a website. Depending on the functionality implemented, the active vibration-resistant system 100 for an on-board graphics server for high-vibration environments includes a path generation module 101, a compensation coefficient calculation module 102, a vector labeling module 103, an index calculation module 104, and a mechanism construction module 105.

[0073] In an embodiment of the present invention, based on the tracking of an active anti-vibration system for an on-board graphics server for a high-vibration environment, each of the above modules can be independently implemented and called with other modules. The call here can be understood as a module that can connect to multiple modules of another type and provide corresponding services to the multiple modules connected to it. In an active anti-vibration system for an on-board graphics server for a high-vibration environment provided by an embodiment of the present invention, the scope of application of an active anti-vibration architecture for an on-board graphics server for a high-vibration environment can be adjusted by adding modules and directly calling them without modifying the program code, thereby realizing cluster-type horizontal expansion, so as to achieve the purpose of quickly and flexibly expanding an active anti-vibration system for an on-board graphics server for a high-vibration environment. In actual applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0074] In the following, various components and specific workflows of an active anti-vibration system for an in-vehicle graphics server in a high-vibration environment are described in conjunction with specific embodiments.

[0075] The path generation module 101 is used to collect the real-time vibration data stream of the graphics server running in the vehicle environment, construct a vibration energy spectrum corresponding to the real-time vibration data stream, and analyze the resonant mode clusters in the vibration energy spectrum to generate a damping control path corresponding to the resonant mode clusters.

[0076] By collecting real-time vibration data streams from a graphics server running in a vehicle-mounted environment, the present invention can accurately construct a vibration energy map to identify resonance modes, provide data support for dynamic damping control, and implement inertia compensation and nonlinear filtering under broadband vibration by analyzing vibration characteristics. This in turn generates servo drive instructions to optimize anti-seismic parameters, ultimately constructing a multi-stage active anti-seismic mechanism to ensure stable operation of the equipment under complex road conditions.

[0077] Among them, the on-board environment refers to the comprehensive physical space in which the vehicle is located during driving, covering the complex situations formed by multiple factors such as road conditions, driving speed, weather conditions, etc. For example, when a vehicle is driving on a highway, the on-board environment is relatively stable but with certain regular low-frequency vibrations; while on bumpy rural roads, the vibration frequency will increase significantly, reaching 20-50Hz, accompanied by dynamic changes such as sudden braking and sharp turns. These factors together constitute a unique environment that affects on-board equipment; the graphics server refers to a high-performance computing device installed inside the vehicle, which is specially used for real-time processing of graphics rendering, visual computing and other tasks. Taking an autonomous driving vehicle as an example, it needs to perform real-time analysis and processing of a large amount of image data collected by the on-board camera, and may have to process up to 30 frames per second with a resolution of 1920×1080. This relies on the powerful graphics processing capabilities of the graphics server. The real-time vibration data stream refers to a continuous data set collected in real time by sensors, reflecting the vibration state of the on-board graphics server during operation. For example, when an off-road vehicle is traveling, a three-axis acceleration sensor installed on the graphics server will collect acceleration data in the X, Y, and Z directions at a frequency of 100 times per second, forming a data stream containing information such as vibration amplitude, frequency, and duration. This data can accurately reflect the vibration condition experienced by the graphics server during vehicle driving, such as when passing through a gravel section (vibration amplitude can reach 5g and frequency is 30-80Hz), providing a basis for subsequent seismic analysis and processing. Optionally, the real-time vibration data stream collected when the graphics server is running in a vehicle environment can be achieved through a vibration signal sensing detection method, such as using a MEMS accelerometer in combination with an NI DAQ data acquisition card to capture high-frequency vibration signals, thereby obtaining a real-time vibration data stream.

[0078] Furthermore, the present invention constructs a vibration energy spectrum corresponding to the real-time vibration data stream and analyzes the resonant mode clusters in the vibration energy spectrum. It can accurately locate the distribution characteristics of vibration energy in the frequency dimension through visualization means, provide intuitive data support for dynamic damping control, and effectively identify the frequency interval combination where vibration energy is concentrated, so that the system can actively adapt to the broadband vibration environment based on the vibration characteristics, thereby improving the seismic stability of the on-board graphics server.

[0079] Among them, the vibration energy spectrum refers to the conversion of the real-time collected vibration data stream into a visual spectrum with frequency as the horizontal axis and energy intensity as the vertical axis through technical means such as spectrum analysis. It can intuitively present the distribution of vibration energy at different frequencies. For example, on an engineering vehicle traveling on a bumpy mountain road, the vibration data collected by the on-board graphics server is processed, and the generated vibration energy spectrum shows that the energy intensity reaches 80dB at a frequency of 30Hz and the energy intensity is 75dB at a frequency of 50Hz, which clearly reflects the main vibration energy concentration frequency band in this environment; the resonance mode cluster refers to a collection of multiple frequency intervals with high energy intensity and interrelated frequency intervals in the vibration energy spectrum. The vibrations in these frequency intervals are easy to cause equipment resonance. For example, in a truck traveling at high speed and encountering a crosswind, in the vibration energy spectrum of the on-board graphics server, the energy intensities in the frequency ranges of 20-25 Hz, 35-40 Hz, and 55-60 Hz all exceed 70 dB, which together constitute a resonant mode cluster. Optionally, the construction of the vibration energy spectrum corresponding to the real-time vibration data stream can be achieved by a spectrum estimation technology, such as using the Welch power spectrum density estimation method to analyze the frequency domain energy distribution of the vibration signal to obtain a vibration energy spectrum; the analysis of the resonant mode cluster in the vibration energy spectrum can be achieved by a modal parameter identification algorithm, such as using the random subspace identification (SSI) method to extract the natural frequency and damping ratio of the vibration system to obtain a resonant mode cluster.

[0080] Furthermore, the present invention can provide a clear adjustment direction for subsequent dominant resonance mode analysis by generating a damping control path corresponding to the resonance mode cluster, lay the foundation for nonlinear filtering and servo drive instruction generation, and help the system dynamically adjust the damping parameters according to the resonance characteristics, thereby achieving active adaptation to the broadband vibration environment and effectively improving the seismic reliability of the on-board graphics server under complex road conditions.

[0081] The damping compensation refers to the amount by which the system damping needs to be adjusted to offset vibration energy at the energy convergence node. For example, on an engineering vehicle traveling on a bumpy road, at the 50Hz energy convergence node, the onboard graphics server needs to increase the system damping from the initial 5N·s / m to 8N·s / m to effectively suppress vibration. This additional 3N·s / m is the damping compensation, which is used to reduce the impact of resonance on the equipment.

[0082] As an embodiment of the present invention, generating the damping control path corresponding to the resonance mode cluster includes: analyzing the modal frequency distribution in the resonance mode cluster; identifying the energy convergence node in the modal frequency distribution; mapping the phase offset interval corresponding to the energy convergence node; based on the phase offset interval, analyzing the damping compensation amount corresponding to the energy convergence node; based on the damping compensation amount, generating the damping control path corresponding to the resonance mode cluster.

[0083] Among them, the modal frequency distribution refers to the frequency value and distribution state of each vibration mode in the resonance mode cluster, which reflects the distribution characteristics of vibration energy at different frequencies. For example, in an off-road vehicle traveling on an undulating mountain road, the modal frequency distribution of the on-board graphics server in the resonance mode cluster shows obvious peaks at 20Hz, 35Hz and 50Hz, which correspond to the low-frequency shaking of the vehicle during bumps, the vibration of the suspension system and the impact frequency of the road surface. These frequency values ​​and distribution constitute the modal frequency distribution; the energy convergence node refers to the specific frequency point where the vibration energy is highly concentrated in the modal frequency distribution. These nodes are the key factors that cause equipment resonance. For example, when a heavy truck is driving at high speed over a speed bump, Analysis of the vibration data of the on-board graphics server shows that the vibration energy intensity reaches 90dB at a frequency of 40Hz, far exceeding that of other frequency points. This 40Hz is the energy convergence node. If left untreated, it can easily cause resonance damage to server components. The phase offset range refers to the phase difference range between the vibration signal corresponding to the energy convergence node and the standard sine wave signal, which reflects the phase change characteristics of the vibration signal. For example, on a bus with frequent starts and stops, the phase offset of the vibration signal of its on-board graphics server at the 30Hz energy convergence node relative to the standard sine wave signal fluctuates within a range of ±15°. This ±15° range is the phase offset range, which is important for analyzing the complex characteristics of vibration.

[0084] Furthermore, the analysis of the modal frequency distribution in the resonant modal cluster can be achieved through a spectrum peak extraction algorithm, such as: using fast Fourier transform (FFT) combined with peak search technology to identify the dominant frequency component, thereby obtaining the modal frequency distribution; the identification of the energy convergence node in the modal frequency distribution can be achieved through an energy density analysis method, such as: using short-time Fourier transform (STFT) to calculate the time-frequency energy distribution and locate the high-energy concentration area, thereby obtaining the energy convergence node; the mapping of the phase offset interval corresponding to the energy convergence node can be achieved through a phase demodulation technology, such as: extracting the instantaneous phase of the vibration signal based on the Hilbert transform and calculating the phase difference range, thereby obtaining the phase offset interval; the analysis of the damping compensation amount corresponding to the energy convergence node can be achieved through a modal damping fitting method, such as: using the half-power bandwidth method to calculate the damping ratio of each mode and derive the compensation coefficient, thereby obtaining the damping compensation amount; the generation of the damping control path corresponding to the resonant modal cluster can be achieved through an optimization control algorithm, such as: using particle swarm optimization (PSO) to solve the optimal damper parameter configuration scheme, thereby obtaining the damping control path.

[0085] The compensation coefficient calculation module 102 is used to analyze the dominant resonance mode corresponding to the damping control path, identify the critical oscillation node in the damping control path based on the dominant resonance mode, and calculate the inertia compensation coefficient corresponding to the critical oscillation node.

[0086] By analyzing the dominant resonance mode corresponding to the damping control path, the present invention can accurately locate the key frequency range with the highest vibration energy proportion, provide a core basis for the identification of critical oscillation nodes, lay an efficient modal analysis foundation for subsequent nonlinear filtering and servo drive instruction generation, and effectively enhance the anti-resonance capability of the on-board graphics server in a wide-band vibration environment.

[0087] Among them, the dominant resonance mode refers to the vibration mode that plays a decisive role in the vibration response of the equipment and has the highest energy proportion among the multiple resonance modes corresponding to the damping control path. For example, an analysis of a certain vehicle-mounted graphics server found that in its resonance mode, the energy of the 40Hz mode accounted for 45% of the total vibration energy, which is much higher than other modes. It plays a dominant role in the vibration process of the equipment. The 40Hz mode is the dominant resonance mode, and its characteristics directly affect the seismic design and control strategy of the equipment.

[0088] As an embodiment of the present invention, the analysis of the dominant resonance mode corresponding to the damping control path includes: extracting the path distribution characteristics in the damping control path; dividing the resonance frequency band interval corresponding to the damping control path based on the path distribution characteristics; marking the energy extreme points in the resonance frequency band interval; analyzing the modal coupling relationship corresponding to the energy extreme points; and determining the dominant resonance mode corresponding to the damping control path based on the modal coupling relationship.

[0089] The path distribution characteristics refer to the distribution form and change trend of the damping control path in the frequency-damping parameter space, including information such as the discrete or continuous distribution of frequency, the amplitude and interval of damping parameter adjustment, etc. For example, in the damping control path of a certain vehicle-mounted graphics server, the damping parameter increases by 0.5N·s / m every 2Hz in the 20-30Hz frequency range, and by 1N·s / m every 5Hz in the 30-40Hz range. These frequency and damping parameter change patterns constitute the path distribution characteristics. The resonant frequency band refers to the continuous frequency range that may cause device resonance, divided according to the distribution characteristics of the damping control path. For example, an analysis of a vehicle traveling on a rugged mountain road identified two frequency bands, 25-35Hz and 45-55Hz, based on the damping control path. In these two ranges, the vibration energy of the vehicle-mounted graphics server is easily concentrated, and the risk of device resonance is high. It is the resonance frequency band range; the energy extreme point refers to the specific frequency point where the vibration energy reaches the local maximum or minimum value within the resonance frequency band range. These points have a significant impact on the vibration state of the equipment. For example, in the 50-60Hz resonance frequency band range of a certain vehicle-mounted equipment, the vibration energy intensity at 53Hz reaches 85dB, which is the maximum energy point in this range; the energy intensity at 57Hz is only 60dB, which is the minimum point. Both are energy extreme points; the modal coupling relationship refers to the correlation characteristics of interaction and mutual influence between different resonance modes, which is manifested as the transfer and conversion of vibration energy between modes. For example, in a vehicle that frequently accelerates and decelerates, the 20Hz vibration mode of the on-board graphics server is coupled with the 30Hz mode. When the 20Hz modal energy is enhanced, the 30Hz modal energy will also increase through structural transmission. This energy transfer relationship between modes is the modal coupling relationship.

[0090] Furthermore, the extraction of path distribution characteristics in the damping control path can be achieved through a feature vector decomposition method, such as: using the principal component analysis (PCA) algorithm to reduce the dimension of the path parameter matrix, thereby obtaining the path distribution characteristics; the division of the resonance frequency band corresponding to the damping control path can be achieved through a frequency band energy integration method, such as: using 1 / 3 octave analysis to divide the vibration energy concentration frequency band, thereby obtaining the resonance frequency band; the marking of the energy extreme points in the resonance frequency band can be achieved through a gradient detection algorithm, such as: using the Canny edge detection operator to identify the peak position of the energy spectrum curve, thereby obtaining the energy extreme points; the analysis of the modal coupling relationship corresponding to the energy extreme points can be achieved through a coherence function calculation method, such as: establishing a modal transfer function matrix based on cross-spectral density analysis, thereby obtaining the modal coupling relationship; the determination of the dominant resonance mode corresponding to the damping control path can be achieved through an energy proportion statistical method, such as: using the modal participation factor method to calculate the energy contribution rate of each order mode, thereby obtaining the dominant resonance mode.

[0091] Specifically, to further understand the execution logic and data flow relationship corresponding to the task scheduling process in this solution, please refer to Figure 2 The task scheduling flowchart, Figure 2 As the core process framework of the vehicle graphics server's seismic resistance mechanism system, it clearly presents the complete chain from basic data input to seismic resistance strategy output: the input layer focuses on key information such as the vehicle graphics server topology network and target vibration level, which serves as the basis for subsequent analysis; the processing layer converts raw data into executable seismic resistance control principles through a step-by-step logic of "determining vibration uncertainty input → estimating unit vibration parameters → calculating failure probability → conducting simulation tests"; the output layer uses "server network performance loss ratio, fitted vulnerability curve" and other results as feedback. It should be noted that the relationship between the various links in the flowchart is essentially an abstract distillation of the vehicle graphics server's vibration-resistance logic. In actual scenarios, the complexity of parameter calculation (such as the dynamic relationship between modal coupling and energy proportion) and the diversity of link adaptation (differentiated seismic resistance strategy rules corresponding to different vibration levels) are far greater than what is presented in the diagram. This architecture only provides a concise display of the core logic, providing an intuitive reference for understanding the systematic approach to the vehicle graphics server's seismic resistance mechanism.

[0092] Based on the dominant resonance mode, the present invention identifies the critical oscillation nodes in the damping control path, accurately locates the key positions where vibration energy easily triggers hardware resonance, provides precise targets for the calculation of inertia compensation coefficients, improves the targeting and effectiveness of anti-seismic strategies, and lays a foundation for node-level precision control for subsequent nonlinear filtering and servo drive adjustment.

[0093] Among them, the critical oscillation node refers to the key frequency point or structural position in the damping control path where the vibration energy is concentrated and easily causes equipment resonance damage. For example, when the vibration energy of the on-board graphics server at a frequency of 60Hz reaches 85dB, and this frequency point coincides with the natural frequency of the server hard disk bracket, 60Hz is the critical oscillation node.

[0094] As an embodiment of the present invention, the identifying of the critical oscillation node in the damping control path based on the dominant resonance mode includes: analyzing the resonance distribution energy corresponding to the dominant resonance mode; querying the energy diffusion period corresponding to the resonance distribution energy; extracting the peak oscillation segment within the energy diffusion period; generating a node oscillation label corresponding to the peak oscillation segment; and identifying the critical oscillation node in the damping control path based on the node oscillation label.

[0095] Among them, the resonance distribution energy refers to the distribution state of vibration energy in the dominant resonance mode in the frequency dimension or time dimension, reflecting the strength of energy at different frequency points or time periods. For example, when the dominant resonance mode of a certain on-board graphics server is 40Hz, its resonance distribution energy shows that within 0.5 seconds after the vehicle passes through the speed bump, the energy at the 40Hz frequency gradually increases from 70dB to 85dB, and the energy at the 45Hz and 35Hz frequency points increases to 75dB and 72dB respectively, forming a specific energy distribution pattern; the energy diffusion period refers to the time period in which the vibration energy of the dominant resonance mode diffuses and transfers to other frequencies or structural components. During this period, the energy will deviate from the initial concentrated frequency band. For example, when a truck passes through a gravel section at a speed of 60km / h, the energy of its dominant resonance mode (30Hz) gradually diffuses from 30Hz to the 35-40Hz frequency band within 0.3 seconds to 0.8 seconds. The time period of 10 seconds is the energy diffusion period; the peak oscillation segment refers to the continuous data segment in which the vibration amplitude reaches the local maximum value during the energy diffusion period. This segment contains high-energy vibration characteristics. For example, when a certain engineering vehicle crosses a bumpy road, the vibration amplitude of the on-board graphics server at a frequency of 50 Hz reaches 9g in the 0.6-0.7 second time period within the energy diffusion period (0.5-1.0 seconds), far exceeding the average amplitude of 5g during this period. This 100-millisecond vibration data is the peak oscillation segment; the node oscillation label refers to a unique identifier generated by encoding and annotating the characteristic parameters of the peak oscillation segment, which includes information such as frequency, amplitude, and duration. For example, for a peak oscillation segment (frequency 40 Hz, amplitude 8g, and duration 0.2 seconds), the generated node oscillation label is "NODE_40Hz_8g_020ms". This label can accurately locate the vibration characteristics and provide a data index for subsequent node identification.

[0096] Furthermore, the analysis of the resonance distribution energy corresponding to the dominant resonance mode can be achieved through a modal energy integration method, such as: using the frequency response function area method to calculate the energy integral value of each mode in the frequency domain, thereby obtaining the resonance distribution energy; the query of the energy diffusion period corresponding to the resonance distribution energy can be achieved through a time-energy attenuation analysis method, such as: using an exponential attenuation model to fit the vibration signal envelope curve to calculate the energy half-life, thereby obtaining the energy diffusion period; the extraction of the peak oscillation segment within the energy diffusion period can be achieved through a dynamic threshold detection algorithm, such as: using a sliding window standard deviation method to capture the vibration waveform segment exceeding the 3σ threshold, thereby obtaining the peak oscillation segment; the generation of the node oscillation label corresponding to the peak oscillation segment can be achieved through a feature encoding method, such as: constructing a hash coding sequence of vibration features based on discrete wavelet coefficients, thereby obtaining a node oscillation label; the identification of the critical oscillation node in the damping control path can be achieved through a topological network analysis method, such as: using the betweenness centrality algorithm to calculate the key nodes in the vibration transmission path, thereby obtaining the critical oscillation node.

[0097] By calculating the inertia compensation coefficient corresponding to the critical oscillation node, the present invention can provide accurate dynamic compensation parameters for nonlinear filtering, making the generation of anti-interference frequency band sets more consistent with vibration characteristics, effectively reducing the risk of hardware failure caused by resonance, and providing a quantitative inertial force control basis for the stable operation of the on-board graphics server under complex road conditions.

[0098] The inertia compensation coefficient is calculated based on the following formula, taking into account the vibration frequency, amplitude, sampling period and speed of the critical oscillation node. , which is used to quantitatively compensate for the inertia influence of the critical oscillation node, reflects the quantitative index of regulating the inertia effect of the node, and guides the implementation of subsequent inertia compensation strategies.

[0099] As an embodiment of the present invention, the calculating the inertia compensation coefficient corresponding to the critical oscillation node includes:

[0100] The inertia compensation coefficient corresponding to the critical oscillation node is calculated using the following formula:

[0101]

[0102] in, represents the inertia compensation coefficient corresponding to the critical oscillation node, and Represent the lower and upper limits of the vibration frequency, respectively. Indicates frequency The corresponding vibration amplitude function, represents the time constant, represents the number of different frequency components in the critical oscillation node, Indicates the number index of different frequency components, Indicates the The amplitude of the frequency component, Indicates the The frequency of the frequency component, represents the sampling period of the critical oscillation node, Indicates time Vibration velocity function.

[0103] Specifically, the vibration frequency refers to the number of times the vibration is repeated per unit time during the vibration process, and is a physical quantity that describes the periodicity of the vibration. and Define the upper and lower limits of the critical oscillation node analysis. Different frequency components affect the calculation of the inertia compensation coefficient. The time constant can be set to =0.1s; the vibration amplitude function refers to the frequency is a variable that describes the function of how the vibration amplitude changes with frequency. , which reflects the distribution of vibration intensity at different frequencies, is the key basis for measuring the contribution of vibration energy at each frequency when calculating the inertia compensation coefficient, such as ,That is the reference amplitude (unit: m), is the attenuation coefficient (unit: s² / m). For example, in the implementation scenario, the vehicle environment vibration data can be set , The sampling period refers to the time interval between two adjacent samples when collecting the vibration signal of the critical oscillation node. , determines the time resolution of signal acquisition, affects the accuracy of calculations such as vibration velocity function integration, and is related to the accuracy of inertia compensation coefficient results; the vibration velocity function refers to the time is a variable, describing the function of the change of vibration velocity over time , reflects the dynamic characteristics of vibration in the time domain, and its integral operation participates in the calculation of inertia compensation coefficient, reflecting the effect of vibration motion state on inertia compensation, such as ,That is the maximum speed (unit: m / s), is the fundamental frequency (unit: Hz). In the implementation scenario, it can be set , .

[0104] For example, assuming the vehicle environment: , , ; Set function: , ; can be calculated , which can increase the vibration attenuation rate by about 30%.

[0105] The vector marking module 103 is used to perform nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set, perform topological mapping on the anti-interference frequency band set to generate a servo drive instruction set, query the dynamic stiffness parameters corresponding to the servo drive instruction set, and mark the parameter allocation vector corresponding to the dynamic stiffness parameters.

[0106] Based on the inertia compensation coefficient, the present invention performs nonlinear filtering on the real-time vibration spectrum in the vehicle environment to obtain an anti-interference frequency band set. The anti-interference frequency band set can accurately adapt to the characteristics of critical oscillation nodes and dynamically optimize frequency domain filtering parameters, so that the anti-interference frequency band set can effectively shield the resonance energy, provide a pure spectrum basis for subsequent servo drive command generation, and improve the system's anti-vibration stability under complex road conditions.

[0107] Among them, the anti-interference frequency band set refers to the frequency band set that can effectively resist interference and make the vibration spectrum purer after nonlinear filtering of the optimized frequency band group. For example, after filtering the selected optimized frequency band group, the interference in the 25-35Hz and 40-50Hz frequency bands is suppressed, forming an anti-interference frequency band set that allows the equipment to operate stably.

[0108] As an embodiment of the present invention, the real-time vibration spectrum in the vehicle environment is nonlinearly filtered based on the inertia compensation coefficient to obtain an anti-interference frequency band set, including: analyzing the compensation response component corresponding to the inertia compensation coefficient; parsing the spectrum response data corresponding to the real-time vibration spectrum in the vehicle environment based on the compensation response component; querying the frequency band response weight set corresponding to the spectrum response data; screening the optimized frequency band group in the frequency band response weight set; and performing nonlinear filtering on the group parameters in the optimized frequency band group to obtain the anti-interference frequency band set.

[0109] Among them, the compensation response component refers to the specific response part caused by the inertia compensation coefficient and the interaction with the vibration system, reflecting the specific performance of inertia compensation in the vibration response. For example, in the vibration compensation of the on-board graphics server, under the action of the inertia compensation coefficient, the vibration response change with an amplitude of 0.2g and a phase shift of 10° caused by compensation at a frequency of 30Hz is the compensation response component; the spectrum response data refers to the information set about the vibration response of different frequency components obtained by analyzing the real-time vibration spectrum, covering frequency, amplitude, phase, etc. For example, after the real-time vibration spectrum of a certain device is analyzed, a series of information such as 1.5g amplitude and 30° phase at 25Hz and 1.2g amplitude and 25° phase at 35Hz are obtained within the frequency range of 20-50Hz. The data constitutes the spectrum response data; the frequency band response weight set refers to a quantitative set of the "importance" of different frequency bands in the vibration response based on the spectrum response data, reflecting the degree of influence of each frequency band on the overall vibration. For example, when analyzing the vehicle vibration spectrum, the 0-20Hz band has a weight of 0.3 (due to its small impact on the equipment), and the 20-40Hz band has a weight of 0.6 (due to high resonance risk). These weight values ​​constitute the frequency band response weight set; the optimized frequency band group refers to the selection of frequency band combinations that need to be optimized and are critical to system anti-interference from the frequency band response weight set. For example, in vibration control, the two frequency bands of 25-35Hz and 40-50Hz that have a great impact on equipment stability and require key filtering are selected according to the weight set to form the optimized frequency band group.

[0110] Furthermore, the analysis of the compensation response components corresponding to the inertia compensation coefficients can be achieved through a frequency domain decomposition method, such as: using a complex modal analysis method to extract the real and imaginary response components of the compensation system, thereby obtaining the compensation response components; the analysis of the spectral response data in the real-time vibration spectrum can be achieved through a power spectrum estimation method, such as: using a Burg algorithm to calculate the parameterized spectral characteristics of the vibration signal, thereby obtaining spectral response data; the query of the frequency band response weight set corresponding to the spectral response data can be achieved through an energy proportion statistical method, such as: calculating the percentage of the energy of each frequency band to the total energy based on the frequency band energy integral, thereby obtaining a frequency band response weight set; the screening of the optimized frequency band group in the frequency band response weight set can be achieved through a genetic algorithm, such as: using the NSGA-II multi-objective optimization algorithm to select the optimal frequency band combination, thereby obtaining an optimized frequency band group; the nonlinear filtering of the intra-group parameters in the optimized frequency band group can be achieved through a wavelet threshold denoising method, such as: using the Daubechies wavelet basis function to perform soft threshold filtering, thereby obtaining an anti-interference frequency band set.

[0111] The present invention generates a servo drive instruction set by topologically mapping the anti-interference frequency band set, which can accurately adapt to vibration suppression requirements and convert frequency domain anti-interference characteristics into drive control logic. This allows the servo system to adjust its output in real time based on the optimized frequency band information, enhances the system's response adaptability to complex vibration environments, builds a precise and coordinated drive control system for the stable operation of the on-board graphics server, and improves the hardware's shock resistance and performance retention capabilities.

[0112] Among them, the servo drive instruction set refers to an instruction set for controlling the action of the servo system, which is generated based on the anti-interference frequency band set and converted through topological mapping. It contains parameters such as drive frequency, torque, displacement, etc., and guides the servo actuator to respond accurately. For example, in a vehicle-mounted scenario, for the 20-30Hz anti-interference frequency band, instructions are generated: the drive frequency is set to 25Hz, the torque is adjusted at a gradient of 0.8N·m, and the displacement is controlled at ±1mm. The servo is precisely controlled accordingly to offset vibration interference. Optionally, the topological mapping of the anti-interference frequency band set can be achieved through a graph neural network method, such as: using the GraphSAGE algorithm to construct a frequency band association graph and extract topological features to generate a servo drive instruction set.

[0113] Furthermore, the present invention can accurately match vibration suppression requirements with hardware adjustment capabilities by querying the dynamic stiffness parameters corresponding to the servo drive instruction set, providing a quantitative stiffness control basis for the active anti-seismic system. It can dynamically adjust the stiffness characteristics of components such as dampers and springs according to real-time vibration characteristics, thereby achieving precise dissipation of broadband vibration energy and inertia force compensation, thereby improving the anti-resonance stability and hardware protection performance of the on-board graphics server under complex road conditions.

[0114] The dynamic stiffness parameter refers to the stiffness characteristic value that can be adjusted in real time according to the servo drive instruction set. It is used to describe the parameter of the elastic element (such as a damper or piezoelectric ceramic) in the anti-seismic system, which changes with the vibration frequency and amplitude. For example, when the on-board graphics server is vibrating in the 40Hz frequency band, the dynamic stiffness parameter can instruct the piezoelectric actuator to adjust the stiffness from the initial 6N / mm to 9N / mm in real time to match the vibration resistance requirements of this frequency band. By dynamically changing the stiffness value, the resonant energy is effectively suppressed. Optionally, querying the dynamic stiffness parameter corresponding to the servo drive instruction set can be achieved through frequency response function inversion methods, such as using the H1 estimation method to calculate the transfer function between the instruction input and the system response to obtain the dynamic stiffness parameter.

[0115] Furthermore, by marking the parameter allocation vector corresponding to the dynamic stiffness parameter, the present invention can provide an accurate quantitative mapping basis for the parameter loading of the active memory controller, ensure that the dynamic stiffness parameter is allocated to each servo drive unit according to the preset logic, optimize the dynamic response efficiency during the servo compensation process, and build an orderly and accurate stiffness control system for the on-board graphics server in a wide-band vibration environment, thereby effectively improving the robustness and environmental adaptability of the anti-seismic system.

[0116] The parameter allocation vector is a multidimensional vector that combines the dynamic stiffness parameters, their corresponding allocated offset values, the effective frequency band, and other information according to a specific logic. This vector provides a quantitative parameter configuration scheme for the servo drive. For example, for the 30-40Hz modal coupling domain, the parameter allocation vector can be expressed as [35Hz, 12N / mm, +2N / mm], corresponding to the dominant frequency, baseline stiffness value, and offset, respectively. The servo system uses this vector to precisely adjust the stiffness parameters of the corresponding frequency band.

[0117] As an embodiment of the present invention, marking the parameter allocation vector corresponding to the dynamic stiffness parameter includes: parsing the parameter identifier corresponding to the dynamic stiffness parameter; dividing the modal coupling domain corresponding to the dynamic stiffness parameter based on the parameter identifier; locating the dominant stiffness component in the modal coupling domain; analyzing the allocation offset value corresponding to the dominant stiffness component; and marking the parameter allocation vector corresponding to the dynamic stiffness parameter based on the allocation offset value.

[0118] Among them, the parameter identifier refers to a coding symbol used to uniquely identify the dynamic stiffness parameter, which contains key information such as parameter type, effective frequency band, adjustment properties, etc., to facilitate rapid retrieval and matching by the system. For example, the identifier of a dynamic stiffness parameter is "KD_40Hz_±15%", where "KD" represents the stiffness damping parameter, "40Hz" identifies the effect on the 40Hz frequency band, and "±15%" indicates that the parameter can fluctuate by 15% based on the benchmark value, which is used to accurately locate the parameter application scenario; the modal coupling domain refers to the frequency range or structural area where multiple vibration modes affect each other under the action of dynamic stiffness parameters, reflecting the coupling characteristics of energy transfer between modes. For example, in the 30-40Hz frequency band of the on-board graphics server, the 35Hz mode and the 38Hz mode generate energy coupling due to structural resonance. When the dynamic stiffness parameter acts on this frequency band, the 30-40Hz range formed is the modal coupling domain. At this time The vibration responses of the two modes will be correlated with each other; the dominant stiffness component refers to the stiffness parameter component that plays a decisive role in vibration suppression in the modal coupling domain, and its energy proportion or control weight is significantly higher than other components. For example, in the 40-50Hz modal coupling domain, the dynamic stiffness parameter at the 45Hz frequency point is 10N / mm, accounting for 60% of the total stiffness control weight in this area, while the weights of other frequency points are all less than 20%. In this case, the 45Hz stiffness parameter corresponding to 10N / mm is the dominant stiffness component; the allocation offset value refers to the adjustment amount of the dominant stiffness component relative to the theoretical reference value during actual allocation, which is used to compensate for parameter deviations caused by modal coupling or environmental interference. For example, the theoretical reference value of the 40Hz dominant stiffness component is 8N / mm, but due to road impact during vehicle driving, the actual demand increases to 9.5N / mm. The difference between the two +1.5N / mm is the allocation offset value, which is used to correct the accuracy of parameter allocation.

[0119] Furthermore, the analysis of the parameter identifier corresponding to the dynamic stiffness parameter can be achieved through a feature encoding method, such as: using a hash algorithm to uniquely encode the stiffness parameter matrix to obtain a parameter identifier; the division of the modal coupling domain corresponding to the dynamic stiffness parameter can be achieved through a clustering analysis method, such as: using a K-means algorithm to similarity group the stiffness parameters to obtain a modal coupling domain; the positioning of the dominant stiffness component in the modal coupling domain can be achieved through a principal component analysis method, such as: extracting the maximum variance direction of the stiffness parameter matrix based on the PCA algorithm to obtain the dominant stiffness component; the analysis of the distribution offset value corresponding to the dominant stiffness component can be achieved through a residual calculation method, such as: using the least squares method to fit the stiffness distribution curve and calculating the deviation between the actual value and the fitted value to obtain the distribution offset value; the marking of the parameter allocation vector corresponding to the dynamic stiffness parameter can be achieved through a vector quantization method, such as: using the LVQ algorithm to vectorize the stiffness parameter to obtain a parameter allocation vector.

[0120] The index calculation module 104 is used to load the parameter allocation vector into a preset active memory controller, monitor the energy dissipation state in the active memory controller, adaptively adjust the gain threshold combination of the real-time vibration spectrum during the servo compensation process based on the energy dissipation state, and calculate the seismic robustness index corresponding to the gain threshold combination.

[0121] By loading the parameter allocation vector into a preset active memory controller and monitoring the energy dissipation state in the active memory controller, the present invention enables the active memory controller to accurately execute the anti-seismic strategy based on the dynamic stiffness parameters, ensuring that the servo drive system responds to vibration interference with the optimal parameter combination. This can provide data support for the system to adaptively adjust the parameter allocation vector, thereby ensuring the stable operation of the on-board graphics server in complex vibration environments.

[0122] Among them, the preset active memory controller refers to an intelligent control unit that integrates anti-seismic strategy parameters (such as stiffness and damping adjustment logic) in advance. It can memorize the optimal control plan based on historical vibration data and actively execute it. For example, a certain vehicle-mounted controller is preset to automatically adjust the stiffness of the piezoelectric actuator from 8N / mm to 10N / mm when 40Hz vibration is detected. This preset parameter combination is generated through 1000 times of road test data optimization and can quickly respond to similar vibration scenarios; the energy dissipation state refers to the real-time state of the anti-seismic system dissipating vibration energy through dampers, piezoelectric elements, etc. under the control of the active memory controller, which is usually measured in decibels of energy attenuation or power loss. For example, when a truck decelerates When the controller activates the preset parameters, the 50Hz vibration energy drops from 85dB to 60dB, corresponding to the piezoelectric element dissipating 2.5J of energy per second. The attenuation process and the energy consumption value are the energy dissipation state. Optionally, the parameter allocation vector is loaded into the preset active memory controller through a dynamic memory mapping method, such as using DMA (direct memory access) technology to realize the rapid transmission and loading of vector data, thereby completing the control parameter configuration; the monitoring of the energy dissipation state in the active memory controller can be achieved through a power integration algorithm, such as using the Joule integration method to calculate the accumulated energy consumption of each module of the controller in real time, thereby obtaining the energy dissipation state.

[0123] Furthermore, based on the energy dissipation state, the present invention adaptively adjusts the gain threshold combination of the real-time vibration spectrum in the servo compensation process, so that the system can dynamically optimize the amplification factor according to the real-time vibration energy dissipation efficiency, and construct a vibration suppression mechanism for the on-board graphics server that is both adaptive and robust, thereby effectively enhancing the anti-vibration stability and energy dissipation efficiency under complex road conditions.

[0124] Among them, the gain threshold combination refers to a gain control set formed by combining effective gain threshold items in frequency domain order for the servo compensation process, which can achieve precise amplification or attenuation of vibrations in different frequency bands. For example, the final determined gain threshold combination is: 25Hz gain 1.2, 35Hz gain 1.5, 45Hz gain 0.8. This combination can optimize the energy dissipation effect by adjusting the servo drive gain for the vibration energy of the corresponding frequency band.

[0125] As an embodiment of the present invention, the adaptive adjustment of the gain threshold combination of the real-time vibration spectrum in the servo compensation process based on the energy dissipation state includes: extracting energy consumption characteristic parameters corresponding to the energy dissipation state; generating a gain adjustment list corresponding to the energy consumption characteristic parameters; based on the gain adjustment list, calculating the corrected gain threshold corresponding to the real-time vibration spectrum item by item; performing frequency domain segmented screening on the corrected gain threshold to obtain an effective gain threshold item; based on the effective gain threshold item, adaptively adjusting the gain threshold combination of the real-time vibration spectrum in the servo compensation process.

[0126] Among them, the energy consumption characteristic parameters refer to key indicators extracted from the energy dissipation state that can reflect the vibration energy dissipation characteristics, such as energy attenuation rate, dissipated power, attenuation time constant, etc. For example, in the 50Hz vibration frequency band of a certain vehicle-mounted graphics server, after being controlled by the active memory controller, the vibration energy decays from 80dB to 60dB within 0.5 seconds, and the corresponding energy attenuation rate is 40dB / s, and the dissipated power is 2.3W. These data constitute the energy consumption characteristic parameters, which are used to characterize the energy dissipation efficiency; the gain adjustment list refers to a list of gain adjustment values ​​for different frequency points or frequency bands generated according to the energy consumption characteristic parameters, which clarifies the adjustment direction and amplitude of the gain of each frequency band. For example, based on the above-mentioned 50Hz frequency band energy consumption characteristics (attenuation rate 40dB / s), a gain adjustment list is generated: 20Hz frequency band gain +0.3, 35Hz frequency band gain -0.1, 50Hz frequency band gain +0.5, and the list Provide a quantitative basis for subsequent gain threshold correction; the modified gain threshold refers to the adjusted threshold calculated based on the original gain threshold and the gain adjustment list, which is used to adapt to the real-time energy dissipation state. For example, if the original gain threshold of a certain frequency band is 1.0 and needs to be increased by 0.5 according to the gain adjustment list, the modified gain threshold is 1.5; if another frequency band needs to be reduced by 0.2, it will be 0.8 after correction. Through this calculation, the gain threshold is more in line with the vibration energy dissipation requirements; the effective gain threshold item refers to the threshold item that can effectively improve the energy dissipation efficiency after the frequency domain segmentation screening of the modified gain threshold. For example, when screening the 20-60Hz modified gain threshold, the threshold items in the frequency band with energy dissipation rate below 30dB / s are eliminated, and only the modified threshold items of 25-45Hz (dissipation rate ≥40dB / s) are retained, such as 25Hz gain 1.2 and 40Hz gain 1.0. These are the effective gain threshold items.

[0127] Furthermore, the extraction of energy consumption characteristic parameters corresponding to the energy dissipation state can be achieved by a wavelet packet decomposition method, such as: using the db4 wavelet basis function to perform a 5-layer decomposition of the energy consumption signal and extracting the energy of each node, thereby obtaining the energy consumption characteristic parameters; the generation of the gain adjustment list corresponding to the energy consumption characteristic parameters can be achieved by a fuzzy reasoning method, such as: establishing a mapping rule library between energy consumption characteristics and gain adjustment amounts based on the Mamdani fuzzy system, thereby obtaining the gain adjustment list; the item-by-item calculation of the corrected gain threshold corresponding to the real-time vibration spectrum can be achieved by a gradient descent method, such as: using an adaptive learning rate algorithm to optimize the convergence process of the gain threshold of each frequency band, thereby obtaining the corrected gain threshold; the frequency domain segmentation screening of the corrected gain threshold can be achieved by a bandpass filtering method, such as: using a Butterworth filter group to extract the gain threshold component of the target frequency band, thereby obtaining the effective gain threshold item; the adaptive adjustment of the gain threshold combination of the real-time vibration spectrum in the servo compensation process can be achieved by a reinforcement learning method, such as: using a PPO algorithm to dynamically optimize the combination strategy of the gain threshold of each frequency band, thereby obtaining the gain threshold combination.

[0128] By calculating the seismic robustness index corresponding to the gain threshold combination, the present invention can quantitatively evaluate the anti-interference ability and stability margin of the servo compensation system in a wide-band vibration environment, provide an accurate numerical basis for the optimization of the seismic robustness strategy, and effectively improve the anti-resonance performance and long-term operation stability of the hardware under variable road conditions.

[0129] Among them, the seismic robustness index refers to a quantitative indicator (KR) of the system's seismic robustness under a combination of gain thresholds, calculated by the following formula, which combines gain parameters, energy dissipation parameters, etc. It is used to measure the system's ability to resist vibration interference and maintain stable operation. A higher value indicates stronger seismic robustness.

[0130] As an embodiment of the present invention, the calculating the seismic robustness index corresponding to the gain threshold combination includes:

[0131] The seismic robustness index corresponding to the gain threshold combination is calculated using the following formula:

[0132]

[0133] in, represents the seismic robustness index corresponding to the gain threshold combination, represents the number of gain parameters in the gain threshold combination, Indicates the number index of gain parameters, Indicates the The weight coefficient of the gain parameter, and They represent the start time and end time of the vibration time interval respectively. Indicates the The adjustment coefficient corresponding to the gain parameter is Indicates the Gain threshold over time The change function of represents the number of energy dissipation parameters in the gain threshold combination, represents the quantitative index of energy dissipation parameters, Indicates the The weight coefficient of the energy dissipation parameter, Indicates the The actual value of the energy dissipation parameter, Indicates the The reference value of the energy dissipation parameter.

[0134] In detail, the gain parameter refers to the parameter in the gain threshold combination used to adjust the degree of signal amplification or attenuation during the servo compensation process. It is the basic element for calculating the seismic robustness index. Different gain parameters correspond to compensation control in different frequency bands or scenarios. The weight coefficient refers to the quantitative assignment of the importance of the gain parameter or energy dissipation parameter in calculating the seismic robustness index, such as 、 , highlighting the impact of key parameters on the results, making the index more in line with the actual seismic characteristics; the vibration time interval refers to the vibration start time selected when analyzing vibration characteristics and calculating the seismic robustness index. To the end time The time period defines the scope of the vibration process to be studied, so that the calculation focuses on a specific vibration scenario; the adjustment coefficient is associated with the gain parameter, such as , used to adjust the gain parameter's effect on the servo compensation, reflecting the correction coefficient of the actual effectiveness of the gain parameter in anti-seismic adjustment; the change function refers to the image , describing the gain threshold over time The dynamically changing functional relationship reflects the time-varying characteristics of the gain parameter during the vibration process and provides a basis for accurate calculation. For example, for a continuous and stable vibration frequency band, a linear function can be used: (in is the gain adjustment rate, b is the initial gain value), for sudden transient vibration, a step function can be used: (in is the initial gain, is the adjusted gain, The energy dissipation parameter refers to the parameter related to vibration energy dissipation in the gain threshold combination, which reflects the system's ability to consume vibration energy and is the key to balancing the energy dimension when calculating the seismic robustness index; the actual value refers to the measured value of the energy dissipation parameter in the actual vibration scenario. , reflecting the actual energy dissipation performance of the system, and can be compared with the benchmark value to evaluate the actual seismic effect; the benchmark value refers to the reference standard for setting the energy dissipation parameters , as a basis for measuring whether the actual value is reasonable and whether the system has achieved the expected seismic performance, and is used to calculate the deviation term in the seismic robustness index.

[0135] For example, regarding the formula parameters 、 、 as well as The coefficient mapping table is as follows:

[0136]

[0137] The mechanism construction module 105 is used to generate the vibration characteristic pattern corresponding to the vehicle environment based on the seismic robustness index, extract the steady-state component, transient component and chaotic component in the vibration characteristic pattern, and construct a multi-cascade active seismic resistance mechanism corresponding to the graphics server in the vehicle environment based on the steady-state component, the transient component and the chaotic component.

[0138] Based on the seismic robustness index, the present invention generates vibration characteristic patterns corresponding to the vehicle environment, can accurately extract vibration patterns under complex road conditions, and build exclusive vibration patterns for the vehicle-mounted graphics server. This can help the system predict resonance risks under different working conditions, provide intuitive feature guidance for dynamic optimization of seismic resistance strategies, promote the upgrade of seismic resistance regulation from experience-based adaptation to data-driven, and continuously improve the stability and reliability of vehicle-mounted hardware in various vibration scenarios.

[0139] Among them, the vibration characteristic pattern refers to a characteristic map that reflects the vibration law of all working conditions based on the sequence vibration seed and combined with the dynamic generation of the vehicle environment, which contains information such as frequency, amplitude, and time series correlation. For example, for long-distance transportation scenarios, a characteristic map interwoven with "low-frequency continuous vibration patterns + high-frequency impact patterns" is generated. Different colors / shapes correspond to different vibration intensities and frequency distributions, which intuitively presents the vibration mode of the vehicle environment.

[0140] As an embodiment of the present invention, the generation of vibration characteristic patterns corresponding to the vehicle environment based on the seismic robustness index includes: analyzing a stable interval threshold corresponding to the seismic robustness index; extracting a vibration modal entropy value in the vehicle environment based on the stable interval threshold; constructing a vibration primitive sequence corresponding to the vehicle environment based on the vibration modal entropy value; iteratively reconstructing the vibration primitive sequence to obtain a sequence vibration seed; and generating the vibration characteristic patterns corresponding to the vehicle environment based on the sequence vibration seed.

[0141] Among them, the stability interval threshold refers to the parameter interval boundary value defined by the seismic robustness index (KR) and the system can stably resist vibration, reflecting the safety range for ensuring seismic performance. For example, when the KR of a certain vehicle-mounted system is [0.8, 1.2], the equipment is stable. If it is lower than 0.8, it is easy to resonate, and if it is higher than 1.2, the energy consumption exceeds the limit. This [0.8, 1.2] is the stability interval threshold, which is used to screen the effective vibration control range; the vibration modal entropy value refers to the quantitative value of the disorder and complexity of each vibration mode in the vehicle environment based on the stability interval threshold, reflecting the uniformity of the vibration energy distribution. For example, when a truck is driving, the 30Hz modal entropy value is 0.6 (energy concentration) and the 50Hz entropy value is 0.8 (energy dispersion). The higher the entropy value, the more complex the vibration mode, and the more refined the seismic resistance strategy is required; the vibration primitive sequence It refers to the sequence of the minimum characteristic units of vehicle vibration extracted according to the vibration modal entropy value, arranged in time / frequency. It is the basis for constructing characteristic patterns. For example, when analyzing the vibration of bumpy roads, the primitive sequence of "20Hz small vibration → 40Hz large impact → 20Hz small vibration" is extracted. Each primitive corresponds to a vibration segment of a specific frequency and amplitude. The sequence vibration seed refers to the core characteristic segment with self-similar expansion ability obtained after iterative reconstruction of the vibration primitive sequence. It can be used as a seed for generating complete patterns. For example, the above primitive sequence is simplified to the seed pattern of "[20Hz, weak] → [40Hz, strong] → [20Hz, weak]". This seed can be expanded into a complex pattern of "[20Hz, weak] × 3 → [40Hz, strong] × 2 → [20Hz, weak] × 3" according to the road conditions.

[0142] Furthermore, the analysis of the stability interval threshold corresponding to the seismic robustness index can be achieved through the quantile regression method, such as: using the Bootstrap resampling technique to calculate the 95% confidence interval of the exponential distribution, thereby obtaining the stability interval threshold; the extraction of the vibration modal entropy value in the vehicle environment can be achieved through the information entropy calculation method, such as: using the Shannon entropy formula to quantify the degree of uncertainty of each order modal energy, thereby obtaining the vibration modal entropy value; the construction of the vibration primitive sequence corresponding to the vehicle environment can be achieved through the symbolic aggregation approximation method, such as: using the SAX algorithm to convert the vibration time domain signal into a discrete symbol sequence, thereby obtaining a vibration primitive sequence; the iterative reconstruction of the vibration primitive sequence can be achieved through the LSTM neural network method, such as: constructing a sequence-to-sequence generation model based on the gated recurrent unit, thereby obtaining a sequence vibration seed; the generation of the vibration characteristic pattern corresponding to the vehicle environment can be achieved through the Gram angular field method, such as: using GAF to convert the one-dimensional vibration signal into a two-dimensional texture image, thereby obtaining a vibration characteristic pattern.

[0143] By extracting the steady-state component, transient component and chaotic component from the vibration characteristic pattern, the present invention can accurately disassemble the polymorphic characteristics of vehicle-mounted vibration, and provide a more fine-grained control basis for the anti-seismic strategy. The steady-state component can be used to build a basic anti-vibration model, the transient component adapts to sudden impact compensation, and the chaotic component responds to complex random vibrations. The three work together to allow the system to dynamically match different vibration forms, improve the accuracy and comprehensiveness of the anti-seismic performance of the vehicle-mounted graphics server, and build a solid defense line for stable hardware operation from multiple dimensions.

[0144] Among them, the steady-state component refers to the vibration component in the vibration characteristic pattern that is continuous, regular, and has small amplitude / frequency fluctuations, reflecting the long-term stable vibration mode of the vehicle environment. For example, when the vehicle system is traveling at a constant speed, the vibration caused by the continuous output of the engine is 100Hz and the amplitude is stable at 0.2g. Its frequency and intensity change very little over time, constituting a steady-state component and is the basic control object of seismic resistance; the transient component refers to the short-term, strongly changing vibration component in the vibration characteristic pattern caused by sudden working conditions (sudden braking, bumps), which is instantaneous and highly impactful. For example, when a vehicle brakes suddenly, a vibration pulse of 500 Hz and an amplitude of 1.5 g appears within 0.5 seconds. It has a short duration but concentrated energy. This type of short-term sudden vibration is a transient component and requires rapid response compensation; the chaotic component refers to the disordered and unpredictable vibration component in the vibration characteristic pattern caused by the coupling of multiple factors (road conditions + load + driving status), which manifests as random fluctuations in frequency / amplitude. For example, when a vehicle is driving at a low speed on a bumpy road, the vibration has irregular jumps between 20-200 Hz and random amplitude changes between 0.3-1.2 g. The chaotic characteristics are generated by the superposition of multiple vibration sources and require robust strategies to deal with. Optionally, the extraction of steady-state components, transient components and chaotic components in the vibration characteristic patterns can be achieved through empirical mode decomposition methods, such as using the EMD algorithm to adaptively decompose the vibration signal into an IMF component set, where the IMF component set covers steady-state components, transient components and chaotic components.

[0145] Furthermore, based on the steady-state component, the transient component and the chaotic component, the present invention constructs a multi-cascade active anti-vibration mechanism corresponding to the graphics server in the vehicle environment, which can accurately match different vibration characteristics through layered regulation - the steady-state component corresponds to the basic anti-vibration layer to achieve continuous energy dissipation, the transient component drives the fast response layer to suppress sudden impact, and the chaotic component activates the robust compensation layer to deal with random vibration. This mechanism can eliminate the impact of continuous vibration during uniform speed driving through multi-dimensional coordinated regulation, and fundamentally improve the vibration resistance and service life of the hardware in the full-scenario vehicle environment.

[0146] Among them, the multi-cascade active anti-vibration mechanism refers to a hierarchical collaborative control system built for different vibration characteristics (steady-state, transient, and chaotic components) in the vehicle environment. Through independent regulation and linkage of multiple functional levels, it can achieve precise suppression of complex vibrations. For example, the mechanism applied to the vehicle-mounted graphics server includes a three-layer architecture: the basic anti-vibration layer uses a metal spring with a stiffness of 10N / mm to continuously dissipate the steady-state vibration energy of 80Hz and an amplitude of 0.3g when driving at a constant speed; the fast response layer is equipped with a piezoelectric ceramic actuator, which can reduce the stiffness to 0.05 seconds when the vehicle passes over a speed bump and generates a transient impact of 300Hz and an amplitude of 1.8g. Increased to 20N / mm to offset the impact; the robust compensation layer uses a neural network algorithm to adjust the magnetorheological damper parameters in real time to deal with chaotic vibrations with random amplitude fluctuations (0.5-1.0g) between 20-150Hz in bumpy sections. The three levels work together according to the logic of "steady-state continuous control, transient rapid blocking, and chaotic intelligent adjustment" to ensure stable operation of the hardware under all road conditions. Optionally, the construction of a multi-cascade active anti-vibration mechanism corresponding to the graphics server in the vehicle environment can be achieved through a hierarchical control architecture method, such as: using a PID-H∞ hybrid control strategy to construct a cascade feedback control system, thereby obtaining a multi-cascade active anti-vibration mechanism.

[0147] Compared with the problems described in the background technology, the present invention can accurately construct a vibration energy spectrum to identify resonance modes by collecting real-time vibration data streams of the graphics server running in the vehicle environment, provide data support for dynamic damping control, and realize inertia compensation and nonlinear filtering under broadband vibration by analyzing vibration characteristics, and then generate servo drive instructions to optimize anti-seismic parameters, and finally construct a multi-cascade active anti-seismic mechanism to ensure the stable operation of the equipment under complex road conditions. Furthermore, the present invention can accurately locate the key frequency range with the highest vibration energy proportion by analyzing the dominant resonance mode corresponding to the damping control path, provide a core basis for the identification of critical oscillation nodes, lay an efficient modal analysis foundation for subsequent nonlinear filtering and servo drive instruction generation, and effectively enhance the anti-resonance ability of the vehicle-mounted graphics server in a broadband vibration environment. Furthermore, the present invention performs nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set, which can accurately adapt to the characteristics of the critical oscillation nodes, dynamically optimize the frequency domain filtering parameters, and make the anti-interference The scrambled band set effectively shields resonance energy, providing a pure spectrum foundation for subsequent servo drive command generation, and improving the system's anti-vibration stability under complex road conditions. Furthermore, the present invention loads the parameter allocation vector into a preset active memory controller and monitors the energy dissipation state in the active memory controller. This enables the active memory controller to accurately execute the anti-vibration strategy based on the dynamic stiffness parameter, ensuring that the servo drive system responds to vibration interference with the optimal parameter combination. This provides data support for the system's adaptive adjustment of the parameter allocation vector, ensuring the stable operation of the vehicle-mounted graphics server in complex vibration environments. Finally, based on the anti-vibration robustness index, the present invention generates a vibration characteristic pattern corresponding to the vehicle-mounted environment, accurately extracts the vibration pattern under complex road conditions, and constructs a dedicated vibration pattern for the vehicle-mounted graphics server. This can help the system predict the resonance risk under different working conditions, provide intuitive feature guidance for dynamically optimizing anti-vibration strategies, promote the upgrade of anti-vibration control from experience-based adaptation to data-driven, and continuously improve the stability and reliability of vehicle-mounted hardware in various vibration scenarios. Therefore, the embodiments of the present invention provide an active anti-vibration system and system for a vehicle-mounted graphics server in a high-vibration environment, which can ensure the anti-vibration stability of vehicle-mounted graphics computing.

[0148] Example 2:

[0149] like Figure 3 FIG. 1 is a flow chart of an active anti-vibration method for an in-vehicle graphics server in a high-vibration environment according to the present invention. In this embodiment, the active anti-vibration method for an in-vehicle graphics server in a high-vibration environment includes:

[0150] Collecting a real-time vibration data stream from a graphics server running in a vehicle environment, constructing a vibration energy spectrum corresponding to the real-time vibration data stream, analyzing the resonant mode clusters in the vibration energy spectrum, and generating a damping control path corresponding to the resonant mode clusters;

[0151] Analyzing a dominant resonance mode corresponding to the damping control path, identifying a critical oscillation node in the damping control path based on the dominant resonance mode, and calculating an inertia compensation coefficient corresponding to the critical oscillation node;

[0152] Based on the inertia compensation coefficient, nonlinear filtering is performed on the real-time vibration spectrum in the vehicle environment to obtain an anti-interference frequency band set, topological mapping is performed on the anti-interference frequency band set to generate a servo drive instruction set, dynamic stiffness parameters corresponding to the servo drive instruction set are queried, and parameter allocation vectors corresponding to the dynamic stiffness parameters are marked;

[0153] The parameter allocation vector is loaded into a preset active memory controller, and an energy dissipation state in the active memory controller is monitored. Based on the energy dissipation state, a gain threshold combination of the real-time vibration spectrum during a servo compensation process is adaptively adjusted, and a seismic robustness index corresponding to the gain threshold combination is calculated;

[0154] Based on the seismic robustness index, a vibration characteristic pattern corresponding to the vehicle environment is generated, and the steady-state component, transient component and chaotic component in the vibration characteristic pattern are extracted. Based on the steady-state component, the transient component and the chaotic component, a multi-cascade active seismic resistance mechanism corresponding to the graphics server in the vehicle environment is constructed.

[0155] In the several embodiments provided by the present invention, it should be understood that the provided systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and actual implementation may employ other division methods.

[0156] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0157] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An active anti-vibration system for an on-vehicle graphics server in a high-vibration environment, characterized in that: The system includes: a path generation module, a compensation coefficient calculation module, a vector marking module, an index calculation module and a mechanism construction module; The path generation module is configured to collect a real-time vibration data stream from a graphics server running in a vehicle-mounted environment, construct a vibration energy spectrum corresponding to the real-time vibration data stream, analyze the resonance mode clusters in the vibration energy spectrum, and generate a damping control path corresponding to the resonance mode clusters. The compensation coefficient calculation module is configured to analyze a dominant resonance mode corresponding to the damping control path, identify a critical oscillation node in the damping control path based on the dominant resonance mode, and calculate an inertia compensation coefficient corresponding to the critical oscillation node, wherein calculating the inertia compensation coefficient corresponding to the critical oscillation node includes: The inertia compensation coefficient corresponding to the critical oscillation node is calculated using the following formula: ； in, represents the inertia compensation coefficient corresponding to the critical oscillation node, Represent the lower and upper limits of the vibration frequency, Indicates frequency The corresponding vibration amplitude function, represents the number of different frequency components in the critical oscillation node, Indicates the number index of different frequency components, Indicates the The amplitude of the frequency component, Indicates the The frequency of the frequency component, represents the sampling period of the critical oscillation node, Indicates time Vibration velocity function when ; The vector marking module is configured to perform nonlinear filtering on the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set, perform topological mapping on the anti-interference frequency band set to generate a servo drive instruction set, query dynamic stiffness parameters corresponding to the servo drive instruction set, and mark parameter allocation vectors corresponding to the dynamic stiffness parameters; The index calculation module is configured to load the parameter allocation vector into a preset active memory controller, monitor the energy dissipation state in the active memory controller, adaptively adjust the gain threshold combination of the real-time vibration spectrum during the servo compensation process based on the energy dissipation state, and calculate the seismic robustness index corresponding to the gain threshold combination. Calculating the seismic robustness index corresponding to the gain threshold combination includes: The seismic robustness index corresponding to the gain threshold combination is calculated using the following formula: ； in, represents the seismic robustness index corresponding to the gain threshold combination, represents the number of gain parameters in the gain threshold combination, Indicates the number index of gain parameters, Indicates the The weight coefficient of the gain parameter, They represent the start time and end time of the vibration time interval respectively. Indicates the The adjustment coefficient corresponding to the gain parameter is Indicates the Gain threshold over time The change function of represents the number of energy dissipation parameters in the gain threshold combination, represents the quantitative index of energy dissipation parameters, Indicates the The weight coefficient of the energy dissipation parameter, Indicates the The actual value of the energy dissipation parameter, Indicates the Baseline values ​​of energy dissipation parameters; The mechanism construction module is used to generate the vibration characteristic pattern corresponding to the vehicle environment based on the seismic robustness index, extract the steady-state component, transient component and chaotic component in the vibration characteristic pattern, and construct a multi-cascade active seismic resistance mechanism corresponding to the graphics server in the vehicle environment based on the steady-state component, the transient component and the chaotic component.

2. The active anti-vibration system for an on-vehicle graphics server in a high-vibration environment according to claim 1, characterized in that: Generating a damping control path corresponding to the resonant mode cluster includes: Analyzing the modal frequency distribution in the resonant mode cluster; identifying energy convergence nodes in the modal frequency distribution; Mapping the phase offset interval corresponding to the energy convergence node; Analyzing the damping compensation amount corresponding to the energy convergence node based on the phase offset interval; Based on the damping compensation amount, a damping control path corresponding to the resonant mode cluster is generated.

3. The active anti-vibration system for a vehicle-mounted graphics server in a high-vibration environment according to claim 1, characterized in that: The analyzing the dominant resonance mode corresponding to the damping control path includes: Extracting path distribution characteristics in the damping control path; Based on the path distribution characteristics, dividing the resonance frequency band intervals corresponding to the damping control path; Marking the energy extreme value points in the resonance frequency band; Analyzing the modal coupling relationship corresponding to the energy extreme point; Based on the modal coupling relationship, a dominant resonance mode corresponding to the damping control path is determined.

4. The active anti-vibration system for a vehicle-mounted graphics server in a high-vibration environment according to claim 1, characterized in that: The step of identifying a critical oscillation node in the damping control path based on the dominant resonance mode includes: Analyzing the resonance distribution energy corresponding to the dominant resonance mode; Querying the energy diffusion period corresponding to the resonance distribution energy; extracting a peak oscillation segment within the energy diffusion period; generating a node oscillation label corresponding to the peak oscillation segment; Based on the node oscillation label, a critical oscillation node in the damping control path is identified.

5. The active anti-vibration system for a vehicle-mounted graphics server in a high-vibration environment according to claim 1, characterized in that: The nonlinear filtering of the real-time vibration spectrum in the vehicle environment based on the inertia compensation coefficient to obtain an anti-interference frequency band set includes: Analyzing the compensation response component corresponding to the inertia compensation coefficient; parsing spectrum response data corresponding to the real-time vibration spectrum in the vehicle environment based on the compensation response component; Querying a frequency band response weight set corresponding to the spectrum response data; screening an optimized frequency band group in the frequency band response weight set; Nonlinear filtering is performed on the parameters within the optimized frequency band group to obtain an anti-interference frequency band set.

6. The active anti-vibration system for an on-vehicle graphics server in a high-vibration environment according to claim 1, characterized in that: The parameter allocation vector corresponding to the dynamic stiffness parameter is marked, including: Parsing parameter identifiers corresponding to the dynamic stiffness parameters; Dividing the modal coupling domain corresponding to the dynamic stiffness parameter based on the parameter identifier; locating dominant stiffness components in the modal coupling domain; Analyzing the distribution offset value corresponding to the dominant stiffness component; The parameter allocation vector corresponding to the dynamic stiffness parameter is marked based on the allocation offset value.

7. The active anti-vibration system for a vehicle-mounted graphics server in a high-vibration environment according to claim 1, characterized in that: Adaptively adjusting the gain threshold combination of the real-time vibration spectrum during the servo compensation process based on the energy dissipation state includes: Extracting energy consumption characteristic parameters corresponding to the energy dissipation state; Generating a gain adjustment list corresponding to the energy consumption characteristic parameters; Calculating, item by item, a modified gain threshold corresponding to the real-time vibration spectrum based on the gain adjustment list; Performing frequency domain segmentation screening on the modified gain threshold to obtain an effective gain threshold item; Based on the effective gain threshold item, the gain threshold combination of the real-time vibration spectrum in the servo compensation process is adaptively adjusted.

8. The active anti-vibration system for a vehicle-mounted graphics server in a high-vibration environment as claimed in claim 1, characterized in that: Generating a vibration characteristic pattern corresponding to the vehicle environment based on the seismic robustness index includes: Analyzing a stability interval threshold corresponding to the seismic robustness index; extracting a vibration modal entropy value in the vehicle environment according to the stability interval threshold; constructing a vibration primitive sequence corresponding to the vehicle environment based on the vibration modal entropy value; Iteratively reconstructing the vibration primitive sequence to obtain a sequence vibration seed; Based on the sequence vibration seeds, a vibration characteristic pattern corresponding to the vehicle environment is generated.

Images