Enterprise safety supervision system and equipment based on data analysis

Abstract

The invention discloses an enterprise safety supervision system and equipment based on data analysis, and relates to the technical field of data analysis. Comprising an initial sensitivity monitoring module, a vehicle state data extraction module, an intelligent feature analysis module, a vehicle state evaluation and classification module, a low-dynamic disturbance monitoring module and a high-dynamic disturbance sensitivity adjustment module, the change of brake hydraulic pressure is captured; and the vehicle state data extraction module is used for acquiring various data information of vehicle driving in real time in the abnormal recognition process. Through the high-dynamic disturbance sensitivity adjustment module and deep learning evaluation, the system dynamically reduces the sensitivity of the pressure sensor in a high-frequency disturbance state, filters noise, and accurately monitors real abnormity; the driving state is divided through intelligent feature analysis, sensitivity self-adaptive adjustment is achieved, and the monitoring precision and the system stability under the complex working condition are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise safety supervision, and in particular to an enterprise safety supervision system and equipment based on data analysis. Background Art

[0002] Enterprise safety supervision based on data analysis refers to the use of modern data collection, processing and analysis technologies to comprehensively monitor and deeply mine various safety-related data generated during the operation of the enterprise, so as to achieve dynamic monitoring and intelligent early warning of safety hazards. In long-distance passenger transport enterprises, enterprises collect vehicle operation data (such as speed, location, braking data, fuel consumption, tire pressure, etc.) and driver behavior data (such as fatigue driving, speeding, and sudden braking) in real time, combined with historical accident data and external environmental factors such as weather and road conditions, and use big data analysis and machine learning models to identify possible risk points. The analysis results can provide a scientific basis for decisions such as vehicle scheduling, driver management, and route planning, timely adjust operation strategies, reduce the accident rate, and comprehensively improve the company's driving safety and compliance management level.

[0003] Pressure sensors play a key role in identifying vehicle braking safety. They monitor the hydraulic pressure changes of the brake fluid in real time and capture the hydraulic signal generated when the driver steps on the brake pedal to evaluate the working status of the brake system. When the brake fluid pressure is lower than the normal range (for example, due to leakage, bubbles or wear) or the pressure changes abnormally (such as delayed response or unstable fluctuations), the pressure sensor can detect these problems and trigger an alarm signal. In addition, the pressure sensor can also help evaluate the degree of match between the brake pedal force and the hydraulic response, and identify potential safety hazards caused by driver misoperation or brake system failure. Through linkage with other vehicle data systems, this information can provide precise support for real-time monitoring, early warning and maintenance of brake safety to ensure driving safety.

[0004] The prior art has the following deficiencies:

[0005] During the brake hydraulic pressure identification process, the pressure sensor is usually set to a higher sensitivity to accurately capture the slight changes in the brake fluid hydraulic pressure and promptly identify potential abnormalities or hidden dangers in the brake system. However, when the vehicle is traveling on extremely uneven roads (such as construction sites or gravel roads) and is subjected to continuous high-frequency vibrations, excessive sensitivity may lead to serious consequences. High-frequency vibrations can cause small and rapid fluctuations in the brake fluid pressure signal, and high-sensitivity sensors may misjudge these fluctuations as abnormal pressure and frequently trigger system warnings. This will not only interfere with the driver's normal operation, but may also cause driver fatigue or ignore real warning signals due to too many false alarms, thereby weakening trust in the brake monitoring system and ultimately threatening vehicle driving safety.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide an enterprise safety supervision system and equipment based on data analysis. Through the high dynamic disturbance sensitivity adjustment module and deep learning model evaluation, the system dynamically reduces the sensitivity of the pressure sensor under high-frequency dynamic disturbance state, effectively filters the noise signal caused by high-frequency vibration, avoids false alarms and accurately monitors real hydraulic anomalies. At the same time, the intelligent feature analysis module generates a dynamic disturbance index, divides the vehicle state into "high-frequency dynamic disturbance" and "low dynamic disturbance", and realizes intelligent adjustment of sensitivity. Under low dynamic disturbance, the initial sensitivity is maintained to accurately capture braking anomalies, comprehensively improving the adaptability, stability and anomaly detection capabilities of the system under complex working conditions, so as to solve the problems in the above-mentioned background technology.

[0008] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: an enterprise safety supervision system and equipment based on data analysis, including an initial sensitivity monitoring module, a vehicle status data extraction module, an intelligent feature analysis module, a vehicle status assessment classification module, a low dynamic disturbance monitoring module and a high dynamic disturbance sensitivity adjustment module:

[0009] Initial sensitivity monitoring module: During vehicle driving, the pressure sensor captures the change of brake fluid pressure with initial sensitivity, i.e. warning threshold;

[0010] The vehicle status data extraction module obtains various data information of vehicle driving in real time during the abnormality identification process, and extracts key features reflecting that the vehicle is in a high-frequency vibration state from the acquired vehicle driving data information;

[0011] The intelligent feature analysis module performs detailed analysis on the extracted key features under the detection window, inputs the analyzed features into the pre-trained deep learning model, and uses the deep learning model to perform intelligent evaluation on the vehicle driving status;

[0012] The vehicle status assessment classification module divides the vehicle driving status into two categories: "high-frequency dynamic disturbance" and "low dynamic disturbance" based on the assessment results of the deep learning model;

[0013] Low dynamic disturbance monitoring module, when the vehicle is in "low dynamic disturbance", continues to capture changes in brake fluid pressure with initial sensitivity to identify potential abnormalities in the brake system;

[0014] The high dynamic disturbance sensitivity adjustment module dynamically reduces the initial sensitivity based on the evaluation results when the vehicle is in "high-frequency dynamic disturbance", reduces noise signal interference and false alarms caused by high-frequency vibration, and focuses on monitoring the actual hydraulic pressure changes.

[0015] Preferably, key features reflecting that the vehicle is in a high-frequency vibration state are extracted from the acquired vehicle driving data information, and the extracted features include the impact response characteristics of the vehicle during driving and the degree of excitation of the vehicle vibration by the road surface roughness. Under the detection window, after analyzing the impact response characteristics of the vehicle during driving and the degree of excitation of the vehicle vibration by the road surface roughness, an impact response spectrum index and a road surface roughness excitation index are generated respectively. The impact response spectrum index quantifies the vibration response characteristics of the vehicle caused by external impacts during driving, and the road surface roughness excitation index quantifies the overall excitation degree of the vehicle vibration by the road surface roughness.

[0016] Preferably, the specific steps of analyzing the impulse response characteristics of the vehicle during driving in the detection window to generate the impulse response spectrum index are as follows:

[0017] First, the three-dimensional acceleration signal of the vehicle is collected, including the vertical, lateral and longitudinal directions, and the three-dimensional acceleration signal is converted into the frequency domain to capture the vibration characteristics at different frequencies. The frequency domain signal is expressed as follows:

[0018] A=[A x (f), A y (f), A z (f)],f∈[f min , f max ]

[0019] , where A is the frequency domain three-dimensional acceleration signal matrix, A x (f) A y (f) and A z (f) represents the acceleration signal amplitude of the vehicle in the lateral x, longitudinal y and vertical z directions at frequency f, f min and f max are the lowest and highest frequencies of the band, respectively;

[0020] The three-dimensional acceleration signal is further processed to calculate the impact response characteristics of the vibration signal in each direction. The calculation expression is as follows:

[0021]

[0022] In the formula, A i (f) is the amplitude of the vibration signal of the vehicle in direction i at frequency f, W(f) is the weight function, R i(f) is the impact response characteristic function, which represents the impact response characteristics of the vehicle in direction i under different frequencies f;

[0023] Based on the impulse response characteristic function, the three-dimensional directional signals are integrated to calculate the spectrum energy density matrix, which is used to quantify the total energy distribution in different frequency bands. The calculation formula of the impulse spectrum energy density matrix is ​​as follows:

[0024]

[0025] Where E(f) is the energy density matrix of the impulse spectrum, φ i is the directional weight coefficient;

[0026] The impulse response spectrum index is calculated through the impulse spectrum energy density matrix E(f). The calculation expression is as follows:

[0027]

[0028] Where γ is the amplification factor, q is the nonlinear adjustment coefficient, and SRSI is the shock response spectrum index.

[0029] Preferably, the specific steps of analyzing the degree of excitation of vehicle vibration by road roughness in the detection window to generate a road roughness excitation index are as follows:

[0030] First, the vertical acceleration signal of the vehicle is collected and converted into the frequency domain. At the same time, it is analyzed in combination with the power spectrum density of the road excitation to extract the original vibration energy. The extraction formula is as follows:

[0031]

[0032] In the formula, R raw is the original vibration energy, a(f) is the acceleration amplitude in the frequency domain, P(f) is the power spectrum density of the road surface excitation, and f min and f max are the lowest and highest frequencies of the band, respectively;

[0033] Further refinement of the original vibration energy through dynamic modulation raw The frequency distribution characteristics of the frequency range are divided into several frequency bands, and different weights are assigned to each frequency band. The vibration energy after modulation is calculated. The calculation expression is as follows:

[0034]

[0035] In the formula, R mod is the modulated vibration energy, R raw，k is the original vibration energy of the kth frequency band, G(k) is the dynamic weight function, and N is the number of frequency bands;

[0036] Through feature enhancement and nonlinear mapping, the modulated vibration energy R mod Converted into sensitive enhanced vibration energy, the formula is as follows:

[0037]

[0038] In the formula, R enh is the enhanced vibration energy, α is the nonlinear mapping gain coefficient, β is the nonlinear mapping threshold adjustment coefficient, θ is the frequency change rate weight coefficient, is the frequency change rate characteristic, is a nonlinear function;

[0039] Will enhance the vibration energy R enh Converted into the final road roughness incentive index, the conversion formula is as follows:

[0040]

[0041] In the formula, R index is the road roughness excitation index, δ is the normalization factor, η is the sensitivity coefficient, R th is the reference threshold.

[0042] Preferably, the impact response spectrum index and road roughness excitation index generated after analysis are input into a pre-learned deep learning model, a dynamic disturbance index is generated by the deep learning model, and the vehicle driving state is intelligently evaluated by the dynamic disturbance index.

[0043] Preferably, the dynamic disturbance index generated when the vehicle driving state is intelligently evaluated by the pre-learned deep learning model is compared and analyzed with a pre-set dynamic disturbance index reference threshold, and the vehicle driving state is divided, and the division steps are as follows:

[0044] If the dynamic disturbance index is greater than a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "high frequency dynamic disturbance";

[0045] If the dynamic disturbance index is less than or equal to a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "low dynamic disturbance".

[0046] Preferably, when the vehicle is in a "high frequency dynamic disturbance", the specific steps of dynamically reducing the initial sensitivity based on the evaluation results are as follows:

[0047] After determining that the vehicle is in a high-frequency dynamic disturbance state, the dynamic disturbance index DPI is used to x The initial sensitivity is dynamically adjusted to calculate the current adjusted sensitivity. The adjustment formula is as follows:

[0048]

[0049] In the formula, S init is the initial sensitivity, DPI x is the dynamic disturbance index, DPI ref is the reference threshold of the dynamic disturbance index, max(DPI x , DPI ref ) is the normalization factor, S adj is the sensitivity after dynamic adjustment;

[0050] After adjusting the sensitivity, the key signals that truly reflect the abnormal hydraulic pressure are monitored, and the small fluctuations caused by high-frequency vibrations are ignored. The monitoring of hydraulic pressure changes is based on the effective pressure difference calculated based on dynamic sensitivity. The calculation expression of the effective pressure difference is as follows:

[0051] ΔP eff =max(|PP avg |-S adj ,0)

[0052] Where ΔP eff is the effective pressure difference, P is the current brake fluid pressure signal value, P avg It is the average value of the brake fluid pressure.

[0053] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0054] The present invention uses a high dynamic disturbance sensitivity adjustment module to dynamically reduce the initial sensitivity of the pressure sensor based on the evaluation results of the deep learning model when the vehicle is in a "high-frequency dynamic disturbance" state. This adjustment effectively filters the noise signal caused by high-frequency vibrations, avoids misjudging normal vibrations as abnormal hydraulic pressure, and reduces invalid warnings triggered by false alarms. This mechanism significantly improves the stability and reliability under complex working conditions and solves the problem of frequent false alarms caused by the high sensitivity of pressure sensors in the prior art. At the same time, the accuracy and adaptability are improved, and even in complex scenarios such as extremely uneven roads (such as construction sites or gravel sections), the real abnormalities of the brake system can be accurately monitored to ensure the efficiency and accuracy of monitoring.

[0055] The present invention uses an intelligent feature analysis module and a vehicle state assessment classification module, and utilizes the dynamic disturbance index generated by a deep learning model to accurately divide the vehicle driving state into two categories: "high-frequency dynamic disturbance" and "low dynamic disturbance". This dynamic disturbance index quantitative assessment mechanism based on the impact response characteristics and the degree of road roughness excitation not only solves the problem of inaccurate driving state judgment under complex road conditions, but also realizes the intelligent dynamic adjustment of sensitivity. In a low dynamic disturbance environment, the system maintains the initial sensitivity to ensure that subtle changes in the brake hydraulic pressure are captured in a timely manner, thereby realizing accurate identification of potential abnormalities in the brake system. In a high-frequency dynamic disturbance environment, the sensitivity adjustment can effectively and centrally monitor the real abnormal changes in hydraulic pressure to avoid masking the real fault signal due to the high-frequency vibration environment. This mechanism enhances the system's adaptability to various working conditions and improves the comprehensiveness and reliability of abnormal detection of the brake system. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0057] Figure 1 This is a module schematic diagram of an enterprise safety supervision system and equipment based on data analysis of the present invention. DETAILED DESCRIPTION

[0058] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0059] The present invention provides Figure 1 The enterprise safety supervision system based on data analysis shown in the figure includes an initial sensitivity monitoring module, a vehicle status data extraction module, an intelligent feature analysis module, a vehicle status assessment classification module, a low dynamic disturbance monitoring module and a high dynamic disturbance sensitivity adjustment module:

[0060] Initial sensitivity monitoring module: During vehicle driving, the pressure sensor captures the change of brake fluid pressure with initial sensitivity, i.e. warning threshold;

[0061] The main purpose of this step is to establish a basic monitoring level to ensure that the sensor can accurately detect abnormalities in the brake system under normal driving conditions. The initial sensitivity setting needs to be calibrated according to the normal operating range and historical data of the brake system to ensure that the sensor can effectively identify potential pressure anomalies without generating false alarms. For example, the warning threshold should be set within the safe range of brake fluid pressure, neither too high to cause missed alarms nor too low to cause false alarms.

[0062] The vehicle status data extraction module obtains various data information of vehicle driving in real time during the abnormality identification process, and extracts key features reflecting that the vehicle is in a high-frequency vibration state from the acquired vehicle driving data information;

[0063] These data include but are not limited to the vehicle's speed, acceleration, steering angle, road condition information, and vibration intensity. By deploying multiple sensors at key locations on the vehicle, the system can fully capture the vehicle's operating status under different operating conditions. The key to this step is to ensure the timeliness and accuracy of data collection so that subsequent feature extraction and analysis can be performed based on the latest vehicle status, thereby improving the accuracy and response speed of abnormal identification.

[0064] The intelligent feature analysis module performs detailed analysis on the extracted key features under the detection window, inputs the analyzed features into the pre-trained deep learning model, and uses the deep learning model to perform intelligent evaluation on the vehicle driving status;

[0065] From the acquired vehicle driving data information, key features reflecting that the vehicle is in a high-frequency vibration state are extracted. The extracted features include the impact response characteristics during vehicle driving and the degree of excitation of vehicle vibration by road roughness. Under the detection window, after analyzing the impact response characteristics during vehicle driving and the degree of excitation of vehicle vibration by road roughness, an impact response spectrum index and a road roughness excitation index are generated respectively. The impact response spectrum index quantifies the vibration response characteristics caused by external impacts (such as potholes, protrusions or obstacles) during vehicle driving, especially the energy distribution in different frequency ranges. The road roughness excitation index quantifies the overall excitation degree of vehicle vibration by road roughness, including the tire-road contact pressure fluctuation caused by the rough surface, the dynamic response amplitude of the suspension system, and the vehicle chassis vibration characteristics.

[0066] When the impact response characteristics of the vehicle change frequently and significantly during driving, it indicates that the vehicle is in a high-frequency vibration state, because this change directly reflects the dynamic characteristics of the vehicle being subjected to continuous external impacts in a short period of time and responding quickly. Specifically, a core feature of the high-frequency vibration state is that the proportion of high-frequency components in the vibration signal of the vehicle increases significantly, and frequent impact responses are the main source of this high-frequency characteristic. When the vehicle is driving on a complex road surface (such as potholes, speed bumps or gravel sections), the irregularities of the road surface will cause continuous and violent impacts, which cause the suspension system and the body to vibrate at high frequencies and large amplitudes in the vertical direction (Z axis). The increase in the frequency of the impact response and the shortening of the time interval further enhance the cumulative effect of the vibration, causing the overall vibration characteristics of the vehicle to change from low-frequency stability to high-frequency dynamic disturbances. In addition, high-frequency impacts will also be reflected in the vehicle's impact response spectrum as the concentration and increase of high-frequency energy, which is a typical manifestation of the high-frequency vibration state. Therefore, the frequent and significant changes in the impact response characteristics can clearly reflect that the vehicle is currently in a high-frequency vibration environment.

[0067] The specific steps for analyzing the impact response characteristics of the vehicle during driving in the detection window to generate the impact response spectrum index are as follows:

[0068] First, the three-dimensional acceleration signal of the vehicle is collected, including the vertical, lateral and longitudinal directions, and the three-dimensional acceleration signal is converted into the frequency domain to capture the vibration characteristics at different frequencies. The frequency domain signal is expressed as follows:

[0069] A=[A x (f), A y (f), A z (f)],f∈[f min , f max ]

[0070] , where A is the frequency domain three-dimensional acceleration signal matrix, A x (f) A y (f) and A z (f) represents the acceleration signal amplitude of the vehicle in the lateral x, longitudinal y and vertical z directions at frequency f, f min and f max are the lowest and highest frequencies of the band, respectively;

[0071] The frequency components of the vibration signal in different directions are captured through frequency domain conversion, providing a basis for subsequent spectrum feature extraction.

[0072] The three-dimensional acceleration signal is further processed to calculate the impact response characteristics of the vibration signal in each direction. The impact response characteristic function quantifies the concentration of high-frequency components in the signal by weighted integration of the frequency. The calculation expression is as follows:

[0073]

[0074] , where A i (f) is the amplitude of the vibration signal of the vehicle in direction i at frequency f, W(f) is the weight function, the weighting factor for frequency f, which is used to enhance or weaken the influence of a specific frequency band on the result, R i (f) is the shock response characteristic function, which represents the shock response characteristics of the vehicle at different frequencies f in direction i (lateral x, longitudinal y and vertical z), and quantifies the vibration energy intensity of the vehicle after being impacted in a specific direction;

[0075] The impact response characteristic function highlights the impact characteristics of the high-frequency band by weighting the high-frequency signal, reflecting the impact intensity of the vehicle in all directions.

[0076] Based on the impulse response characteristic function, the three-dimensional directional signals are integrated to calculate the spectrum energy density matrix, which is used to quantify the total energy distribution in different frequency bands. The calculation formula of the impulse spectrum energy density matrix is ​​as follows:

[0077]

[0078] Where E(f) is the shock spectrum energy density matrix, which represents the comprehensive energy density of the vehicle vibration signal at frequency f and is used to quantify the shock intensity in the current frequency band. i is the direction weight coefficient, which indicates the contribution ratio of each direction i to the total impulse response and is used to adjust the weight of each direction to the total energy;

[0079] The impact spectrum energy density matrix integrates the vibration signals in three-dimensional directions and clearly displays the impact energy distribution of the vehicle in each frequency band through a unified energy representation.

[0080] The impulse response spectrum index is calculated through the impulse spectrum energy density matrix E(f). The calculation expression is as follows:

[0081]

[0082] Where γ is the amplification factor used to enhance the impact of low-energy signals, q is the nonlinear adjustment coefficient used to limit the excessive impact of extreme high-energy signals on the overall results, and SRSI is the shock response spectrum index.

[0083] Through nonlinear processing of spectral energy, the shock response spectrum index SRSI provides a comprehensive quantitative result that can accurately indicate whether the vehicle is in a high-frequency vibration state.

[0084] The larger the performance value of the impact response spectrum index generated after analyzing the impact response characteristics of the vehicle during driving under the detection window, the higher the performance value, indicating that the vehicle is in a high-frequency vibration state during driving; conversely, if the index value is smaller, it indicates that the vehicle is not in a high-frequency vibration state. This is because the impact response spectrum index quantifies the intensity and distribution of the high-frequency energy in the vehicle vibration signal. When the vehicle is subjected to frequent short-term high-amplitude impacts (such as dense potholes or gravel roads), the vibration energy in the high-frequency band increases significantly, which directly leads to an increase in the performance value of the impact response spectrum index, reflecting the characteristics of high-frequency vibration. Under stable driving or low dynamic disturbance conditions, the vehicle vibration is mainly concentrated in the low-frequency band, and the energy in the high-frequency band is small, so the performance value of the impact response spectrum index is also low.

[0085] When the excitation degree of vehicle vibration due to road roughness is significantly enhanced, it does indicate that the vehicle is in a high-frequency vibration state during driving, because the increase in road roughness directly leads to a significant increase in the proportion of high-frequency components in the vehicle vibration signal. Rough roads (such as gravel roads and unpaved roads) will cause drastic fluctuations in the contact pressure between the tire and the ground, requiring the vehicle suspension system to frequently adjust to adapt to the changing road conditions, resulting in continuous fluctuations in the vibration amplitude and high-frequency dynamic response. Specifically, the high-frequency vibration in the vertical direction (Z axis) of the vehicle acceleration signal is intensified, the dynamic friction sliding phenomenon is more obvious, and the vibration density (the frequency of vibration events) is significantly increased. In addition, vibration noise may occur in the vehicle chassis, body and internal structural parts, further reflecting the high-frequency excitation characteristics. The key feature of the high-frequency vibration state is that the vibration intensity is concentrated in a short period and the energy of the high-frequency component is significantly increased, and the excitation of the road roughness is the main reason for triggering these characteristics. Therefore, the significant increase in the degree of road roughness excitation is one of the important signs that the vehicle enters a high-frequency vibration state.

[0086] The specific steps for analyzing the degree of excitation of vehicle vibration by road roughness in the detection window to generate the road roughness excitation index are as follows:

[0087] First, the vertical acceleration signal of the vehicle is collected and converted into the frequency domain. At the same time, it is analyzed in combination with the power spectrum density of the road excitation to extract the original vibration energy. The extraction formula is as follows:

[0088]

[0089] In the formula, R raw is the original vibration energy, which represents the total energy of the vibration signal within the frequency range. a(f) is the frequency domain acceleration amplitude, which is the amplitude of the vehicle acceleration signal in the frequency domain after Fourier transform, reflecting the vibration intensity of the vehicle at different frequencies. P(f) is the road excitation power spectrum density, which reflects the frequency response characteristics of road roughness to vehicle vibration and describes the energy distribution at different frequencies. minand f max are the lowest and highest frequencies of the band, respectively;

[0090] This step extracts vibration energy by combining acceleration signals and road excitation characteristics, removes temporal noise, and constructs basic indicators of road roughness excitation.

[0091] Further refinement of the original vibration energy through dynamic modulation raw The frequency distribution characteristics of the frequency range are divided into several frequency bands, and different weights are assigned to each frequency band. The vibration energy after modulation is calculated. The calculation expression is as follows:

[0092]

[0093] In the formula, R mod is the modulated vibration energy, which is the total amount of vibration energy after dynamic weight adjustment, reflecting the comprehensive impact of different frequency bands on high-frequency vibration. raw，k is the original vibration energy of the kth frequency band, G(k) is the dynamic weight function, which is the weight set according to the characteristics of different frequency bands and is used to adjust the contribution of each frequency band to the total vibration energy, and N is the number of frequency bands, which refines the frequency range to improve the resolution;

[0094] This step highlights the relative contribution of different frequency bands to vibration through dynamic weight adjustment, making the modulated result closer to the actual excitation characteristics of high-frequency vibration.

[0095] Through feature enhancement and nonlinear mapping, the modulated vibration energy R mod Converted into sensitive enhanced vibration energy, the formula is as follows:

[0096]

[0097] In the formula, R enh is the enhanced vibration energy, which indicates the vibration energy value after feature enhancement, and is used to further quantify the impact of road excitation on vehicle vibration. α is the nonlinear mapping gain coefficient, which is used to control the degree of enhancement of vibration energy by the nonlinear function. β is the nonlinear mapping threshold adjustment coefficient, which is used to control the threshold smoothness of the nonlinear enhancement function and affect the distribution range of vibration energy enhancement. θ is the frequency change rate weight coefficient, which is used to control the weight of the frequency change rate feature in enhancing vibration energy. is the frequency change rate characteristic, indicating the vibration energy R raw，k The sum of the frequency change rates in different frequency bands k is used to quantify the severity of the frequency distribution. is a nonlinear function used to amplify larger R mod value, while suppressing smaller values;

[0098] This step amplifies high-value features through nonlinear mapping and identifies irregular excitations through the frequency change rate, further enhancing the ability to identify vibration features.

[0099] Will enhance the vibration energy R enh Converted into the final road roughness incentive index, the conversion formula is as follows:

[0100]

[0101] In the formula, R index is the road roughness incentive index, δ is the normalization factor, ensuring that the generated road roughness incentive index R index The range is within the set interval, η is the sensitivity coefficient, which controls the response degree of the road roughness excitation index to the input change, R th It is the reference threshold, indicating the upper limit of the normal vibration range, and is used to distinguish high-frequency vibration states.

[0102] This step enhances the exponential mapping of vibration energy to intuitively quantify the high-frequency excitation level of the vehicle under the current road conditions and generate a clear and discriminative index R index .

[0103] The larger the performance value of the road roughness excitation index generated by analyzing the degree of excitation of vehicle vibration by road roughness under the detection window, the more the vibration excitation intensity of the vehicle per unit time increases, the higher the proportion of high-frequency components, and the frequent response of the vehicle suspension system and body structure, which is usually a typical feature of high-frequency vibration state. In this case, the fluctuation amplitude and frequency of the acceleration signal are significantly enhanced, and the dynamic friction sliding phenomenon is more intense, reflecting that the vehicle is experiencing continuous excitation of complex and rough road surface; when the performance value of the road roughness excitation index is small, it indicates that the road surface is relatively flat, the vehicle vibration excitation is low, the vibration frequency and intensity are within the normal range, and the vehicle is not in a high-frequency vibration state.

[0104] The impact response spectrum index and road roughness excitation index generated after analysis are input into the pre-learned deep learning model, and the dynamic disturbance index is generated by the deep learning model. The dynamic disturbance index is used to perform an intelligent evaluation of the vehicle's driving status.

[0105] The pre-learned deep learning model refers to a deep learning algorithm model that has been trained offline and optimized in the intelligent vehicle driving state evaluation system. The model is based on a large amount of historical data (including vehicle impact response characteristics, road roughness excitation degree, and the corresponding real driving state label). Through repeated training, it learns to extract effective information from input features and establish a nonlinear mapping relationship between input features and vehicle driving state. Its function is to convert complex and multi-dimensional input data (such as impact response spectrum indicators and road roughness excitation indicators) into dynamic disturbance indices that are easy to quantify and understand, so as to accurately evaluate the vehicle's driving state. The pre-trained model usually contains multiple hidden layers of deep neural networks (such as LSTM, CNN or Transformer), which can capture the hidden dynamic characteristics and timing relationships during vehicle driving, thereby improving the accuracy and robustness of the evaluation.

[0106] In the offline training phase, the pre-learned deep learning model uses a large amount of labeled data to adjust and optimize parameters to achieve the best fit between the input data and the target output. The training process includes feature extraction, weight update, hyperparameter tuning, etc., to ensure that the model has good generalization ability for unseen data. In the online evaluation phase, the model performs intelligent calculations based on the extracted real-time input data (such as impact response spectrum indicators and road roughness excitation indicators) to generate a dynamic disturbance index. In this process, the deep learning model uses its learned feature extraction capabilities to map the input features into the evaluation results of the vehicle's driving status. This "pre-learning-online application" architecture enables the model to efficiently process real-time data while having the characteristics of high precision and low latency, and is the core technical foundation of the intelligent monitoring system.

[0107] The deep learning model is not limited here, and can realize the impact response spectrum index SRSI and road roughness excitation index R index Perform comprehensive analysis to generate the Dynamic Perturbation Index (DPI) x In order to realize the technical solution of the present invention, the present invention provides a specific generation method;

[0108] Dynamic Perturbation Index DPI x The generation formula is as follows:

[0109]

[0110] Where v1 and v2 are the shock response spectrum index SRSI and the road roughness excitation index R respectively. index The proportional coefficient is preset, and both v1 and v2 are greater than 0.

[0111] It can be seen from the dynamic disturbance index that the larger the performance value of the impact response spectrum index generated after analyzing the impact response characteristics of the vehicle during driving under the detection window, the larger the performance value of the road roughness excitation index generated after analyzing the excitation degree of road roughness on vehicle vibration under the detection window, then the larger the performance value of the dynamic disturbance index generated when the vehicle driving state is intelligently evaluated by the pre-learned deep learning model, indicating that the vehicle is in a high-frequency vibration state, otherwise it indicates that the vehicle is not in a high-frequency vibration state.

[0112] The preset proportionality coefficient is used to calculate the dynamic disturbance index for different indicators (such as the shock response spectrum index SRSI and the road roughness excitation index R index ) parameters for assigning weights. The setting of these coefficients reflects the importance or contribution of each indicator in the comprehensive analysis. Specifically, the preset proportional coefficients can be set based on historical data, vehicle test results or actual working conditions, through experience or optimization methods (such as regression analysis, machine learning parameter adjustment), to ensure that the influence of different indicators on the final result in the dynamic disturbance index calculation is reasonable and practical. For example, if the impact of the impact response on the vehicle state is more significant, v1>v2 can be set, and if the road roughness excitation contributes more to the vehicle vibration, v2>v1. The optimization of these proportional coefficients directly affects the accuracy and evaluation effect of the dynamic disturbance index.

[0113] The vehicle status assessment classification module divides the vehicle driving status into two categories: "high-frequency dynamic disturbance" and "low dynamic disturbance" based on the assessment results of the deep learning model;

[0114] The dynamic disturbance index generated by the pre-learned deep learning model when intelligently evaluating the vehicle driving state is compared with the pre-set dynamic disturbance index reference threshold to divide the vehicle driving state. The division steps are as follows:

[0115] If the dynamic disturbance index is greater than a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "high frequency dynamic disturbance";

[0116] If the dynamic disturbance index is less than or equal to a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "low dynamic disturbance".

[0117] High-frequency dynamic disturbance refers to the high-frequency, large-amplitude vibration state caused by the complexity of the external environment or road conditions during the vehicle's driving. This state usually occurs when the vehicle is driving on an extremely uneven road surface (such as a gravel road, a construction site, or a road with dense potholes) or is subjected to continuous impact (such as continuous speed bumps or sudden obstacles). Low dynamic disturbance refers to the vehicle's stable, low-vibration operating state during driving, which usually occurs when the vehicle is driving on a flat road (such as an asphalt road or a highway).

[0118] Low dynamic disturbance monitoring module, when the vehicle is in "low dynamic disturbance", continues to capture changes in brake fluid pressure with initial sensitivity to identify potential abnormalities in the brake system;

[0119] When the vehicle is in "low dynamic disturbance", the initial sensitivity continues to capture changes in brake fluid pressure. The purpose is to fully utilize the high sensitivity of the sensor when the vehicle is in a relatively stable driving state, accurately monitor small fluctuations in brake fluid pressure, and promptly identify potential abnormalities in the brake system. This monitoring method can effectively detect problems such as brake fluid leakage, pipeline bubbles, brake valve jams, or brake pad wear. These problems may not have obvious external manifestations in the early stages, but high-sensitivity monitoring can capture signs of abnormal pressure changes. In a low-dynamic disturbance state, vibration interference is small and the monitoring environment is relatively ideal. The sensor can take advantage of its high sensitivity to maximize the safety of the brake system, prevent potential problems from developing into serious failures, and ensure driving safety.

[0120] High dynamic disturbance sensitivity adjustment module, when the vehicle is in "high-frequency dynamic disturbance", dynamically reduces the initial sensitivity based on the evaluation results, reduces noise signal interference and false alarms caused by high-frequency vibration, and focuses on monitoring the real hydraulic pressure changes;

[0121] When the vehicle is in a "high frequency dynamic disturbance", the specific steps for dynamically reducing the initial sensitivity based on the evaluation results are as follows:

[0122] After determining that the vehicle is in a high-frequency dynamic disturbance state, the dynamic disturbance index DPI is used to x The initial sensitivity is dynamically adjusted to calculate the current adjusted sensitivity. The adjustment formula is as follows:

[0123]

[0124] In the formula, S init is the initial sensitivity, which is the sensitivity set by the pressure sensor in the default state, DPI x is the dynamic disturbance index, DPI ref is the reference threshold of the dynamic disturbance index, max(DPI x , DPI ref) is the normalization factor to ensure that the sensitivity adjustment range is controlled, S adj is the sensitivity after dynamic adjustment;

[0125] Through the above steps, the sensitivity adjustment is inversely proportional to the intensity of the dynamic disturbance. The stronger the disturbance, the more the sensitivity decreases, thereby reducing the noise signal interference caused by high-frequency vibration.

[0126] After adjusting the sensitivity, the key signals that truly reflect the abnormal hydraulic pressure are monitored, and the small fluctuations caused by high-frequency vibrations are ignored. The monitoring of hydraulic pressure changes is based on the effective pressure difference calculated based on dynamic sensitivity. The calculation expression of the effective pressure difference is as follows:

[0127] ΔP eff =max(|PP avg |-S adj ,0)

[0128] In the formula, ΔP eff is the effective pressure difference, filtering out small fluctuations caused by high-frequency vibrations and retaining only abnormal changes that exceed the sensitivity threshold. P is the current brake fluid pressure signal value, and P avg It is the average value of the brake fluid pressure.

[0129] Through this step, we can effectively focus on significant hydraulic pressure anomalies, ensuring that potential problems such as brake fluid leakage and pipe blockage can still be accurately identified in a high-frequency vibration environment, while reducing false alarms caused by noise and improving monitoring accuracy.

[0130] When the vehicle is in "high-frequency dynamic disturbance", the initial sensitivity is dynamically reduced based on the evaluation results. Its main function is to adjust the monitoring strategy to reduce the noise signal interference and false alarms caused by high-frequency vibration, so as to accurately capture the truly abnormal hydraulic pressure changes. In the high-frequency vibration state, the dynamic disturbance index of the vehicle increases, and the violent fluctuations in vibration frequency and amplitude will be transmitted to the pressure sensor through the hydraulic system, so that it detects a large number of tiny but meaningless pressure changes. If the high sensitivity continues to be maintained, the sensor may misjudge these fluctuations as abnormal signals and frequently trigger warnings. This will not only interfere with the driver's judgment and increase the system's false alarm rate, but also cause the real hydraulic abnormalities to be covered up. Therefore, dynamically reducing sensitivity can effectively filter out high-frequency noise caused by vibration, while retaining the ability to detect significant pressure abnormalities. By adjusting the sensitivity, the system will focus on monitoring abnormal changes with large amplitude and persistence in the hydraulic signal, avoiding the driver ignoring the real warning signal due to false alarms, and at the same time improving the reliability and accuracy of the brake monitoring system under complex working conditions. This dynamic adjustment mechanism maximizes driving safety while ensuring the efficient operation of the system.

[0131] The present invention uses a high dynamic disturbance sensitivity adjustment module to dynamically reduce the initial sensitivity of the pressure sensor based on the evaluation results of the deep learning model when the vehicle is in a "high-frequency dynamic disturbance" state. This adjustment effectively filters the noise signal caused by high-frequency vibrations, avoids misjudging normal vibrations as abnormal hydraulic pressure, and reduces invalid warnings triggered by false alarms. This mechanism significantly improves the stability and reliability under complex working conditions and solves the problem of frequent false alarms caused by the high sensitivity of pressure sensors in the prior art. At the same time, the accuracy and adaptability are improved, and even in complex scenarios such as extremely uneven roads (such as construction sites or gravel sections), the real abnormalities of the brake system can be accurately monitored to ensure the efficiency and accuracy of monitoring.

[0132] The present invention uses an intelligent feature analysis module and a vehicle state assessment classification module, and utilizes the dynamic disturbance index generated by a deep learning model to accurately divide the vehicle driving state into two categories: "high-frequency dynamic disturbance" and "low dynamic disturbance". This dynamic disturbance index quantitative assessment mechanism based on the impact response characteristics and the degree of road roughness excitation not only solves the problem of inaccurate driving state judgment under complex road conditions, but also realizes the intelligent dynamic adjustment of sensitivity. In a low dynamic disturbance environment, the system maintains the initial sensitivity to ensure that subtle changes in the brake hydraulic pressure are captured in a timely manner, thereby realizing accurate identification of potential abnormalities in the brake system. In a high-frequency dynamic disturbance environment, the sensitivity adjustment can effectively and centrally monitor the real abnormal changes in hydraulic pressure to avoid masking the real fault signal due to the high-frequency vibration environment. This mechanism enhances the system's adaptability to various working conditions and improves the comprehensiveness and reliability of abnormal detection of the brake system.

[0133] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0134] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0135] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0136] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0137] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0138] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0139] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0141] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0142] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An enterprise safety supervision system based on data analysis, characterized in that: It includes initial sensitivity monitoring module, vehicle status data extraction module, intelligent feature analysis module, vehicle status assessment and classification module, low dynamic disturbance monitoring module and high dynamic disturbance sensitivity adjustment module: Initial sensitivity monitoring module: During vehicle driving, the pressure sensor captures the change of brake fluid pressure with initial sensitivity, i.e. warning threshold; The vehicle status data extraction module obtains various data information of vehicle driving in real time during the abnormality identification process, and extracts key features reflecting that the vehicle is in a high-frequency vibration state from the acquired vehicle driving data information; The intelligent feature analysis module performs detailed analysis on the extracted key features under the detection window, inputs the analyzed features into the pre-trained deep learning model, and uses the deep learning model to perform intelligent evaluation on the vehicle driving status; The vehicle status assessment classification module divides the vehicle driving status into two categories: "high-frequency dynamic disturbance" and "low dynamic disturbance" based on the assessment results of the deep learning model; Low dynamic disturbance monitoring module: When the vehicle is in "low dynamic disturbance", it continues to capture changes in brake fluid pressure with initial sensitivity to identify potential abnormalities in the brake system; High dynamic disturbance sensitivity adjustment module, when the vehicle is in "high-frequency dynamic disturbance", dynamically reduces the initial sensitivity based on the evaluation results, reduces noise signal interference and false alarms caused by high-frequency vibration, and focuses on monitoring the real hydraulic pressure changes.

2. According to the data analysis-based enterprise safety supervision system of claim 1, it is characterized in that: From the acquired vehicle driving data information, key features reflecting that the vehicle is in a high-frequency vibration state are extracted. The extracted features include the impact response characteristics of the vehicle during driving and the degree of excitation of the vehicle vibration by the road roughness. Under the detection window, after analyzing the impact response characteristics of the vehicle during driving and the degree of excitation of the vehicle vibration by the road roughness, an impact response spectrum index and a road roughness excitation index are generated respectively. The impact response spectrum index quantifies the vibration response characteristics of the vehicle caused by external impact during driving, and the road roughness excitation index quantifies the overall excitation degree of the vehicle vibration by the road roughness.

3. The enterprise safety supervision system based on data analysis according to claim 2 is characterized in that: The specific steps for analyzing the impact response characteristics of the vehicle during driving in the detection window to generate the impact response spectrum index are as follows: First, the three-dimensional acceleration signal of the vehicle is collected, including the vertical, lateral and longitudinal directions, and the three-dimensional acceleration signal is converted into the frequency domain to capture the vibration characteristics at different frequencies. The frequency domain signal is expressed as follows: A=[A x (f),A y (f),A z (f)],f∈[f min ,f max ] Where A is the frequency domain three-dimensional acceleration signal matrix, A x (f) A y (f) and A z (f) represents the acceleration signal amplitude of the vehicle in the lateral x, longitudinal y and vertical z directions at frequency f, f min and f max are the lowest and highest frequencies of the band, respectively; The three-dimensional acceleration signal is further processed to calculate the impact response characteristics of the vibration signal in each direction. The calculation expression is as follows: In the formula, A i (f) is the amplitude of the vibration signal of the vehicle in direction i at frequency f, W(f) is the weight function, R i (f) is the impact response characteristic function, which represents the impact response characteristics of the vehicle in direction i under different frequencies f; Based on the impulse response characteristic function, the three-dimensional directional signals are integrated to calculate the spectrum energy density matrix, which is used to quantify the total energy distribution in different frequency bands. The calculation formula of the impulse spectrum energy density matrix is ​​as follows: Where E(f) is the energy density matrix of the impulse spectrum, φ i is the directional weight coefficient; The impulse response spectrum index is calculated through the impulse spectrum energy density matrix E(f). The calculation expression is as follows: Where γ is the amplification factor, q is the nonlinear adjustment coefficient, and SRSI is the shock response spectrum index.

4. The enterprise safety supervision system based on data analysis according to claim 2 is characterized in that: The specific steps for analyzing the degree of excitation of vehicle vibration by road roughness in the detection window to generate the road roughness excitation index are as follows: First, the vertical acceleration signal of the vehicle is collected and converted into the frequency domain. At the same time, it is analyzed in combination with the power spectrum density of the road excitation to extract the original vibration energy. The extraction formula is as follows: In the formula, R raw is the original vibration energy, a(f) is the acceleration amplitude in the frequency domain, P(f) is the power spectrum density of the road surface excitation, and f min and f max are the lowest and highest frequencies of the band, respectively; Further refinement of the original vibration energy through dynamic modulation raw The frequency distribution characteristics of the frequency range are divided into several frequency bands, and different weights are assigned to each frequency band. The vibration energy after modulation is calculated. The calculation expression is as follows: In the formula, R mod is the modulated vibration energy, R raw , k is the original vibration energy of the kth frequency band, G(k) is the dynamic weight function, and N is the number of frequency bands; Through feature enhancement and nonlinear mapping, the modulated vibration energy R mod Converted into sensitive enhanced vibration energy, the formula is as follows: In the formula, R enh is the enhanced vibration energy, α is the nonlinear mapping gain coefficient, β is the nonlinear mapping threshold adjustment coefficient, θ is the frequency change rate weight coefficient, is the frequency change rate characteristic, is a nonlinear function; Will enhance the vibration energy R enh Converted into the final road roughness incentive index, the conversion formula is as follows: In the formula, R index is the road roughness excitation index, δ is the normalization factor, η is the sensitivity coefficient, R th is the reference threshold.

5. The enterprise safety supervision system based on data analysis according to claim 2 is characterized in that: The impact response spectrum index and road roughness excitation index generated after analysis are input into the pre-learned deep learning model, and the dynamic disturbance index is generated by the deep learning model. The dynamic disturbance index is used to perform an intelligent evaluation of the vehicle's driving status.

6. The enterprise safety supervision system based on data analysis according to claim 5 is characterized in that: The dynamic disturbance index generated by the pre-learned deep learning model when intelligently evaluating the vehicle driving state is compared with the pre-set dynamic disturbance index reference threshold to divide the vehicle driving state. The division steps are as follows: If the dynamic disturbance index is greater than a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "high frequency dynamic disturbance"; If the dynamic disturbance index is less than or equal to a preset dynamic disturbance index reference threshold, the vehicle driving state is classified as "low dynamic disturbance".

7. The enterprise safety supervision system based on data analysis according to claim 6 is characterized in that: When the vehicle is in "high frequency dynamic disturbance", the specific steps for dynamically reducing the initial sensitivity based on the evaluation results are as follows: After determining that the vehicle is in a high-frequency dynamic disturbance state, the dynamic disturbance index DPI is used to x The initial sensitivity is dynamically adjusted to calculate the current adjusted sensitivity. The adjustment formula is as follows: In the formula, S init is the initial sensitivity, DPI x is the dynamic disturbance index, DPI ref is the reference threshold of the dynamic disturbance index, max(DPI x , DPI ref ) is the normalization factor, S adj is the sensitivity after dynamic adjustment; After adjusting the sensitivity, the key signals that truly reflect the abnormal hydraulic pressure are monitored, and the small fluctuations caused by high-frequency vibrations are ignored. The monitoring of hydraulic pressure changes is based on the effective pressure difference calculated based on dynamic sensitivity. The calculation expression of the effective pressure difference is as follows: ΔP eff =max(|P-P avg |-S adj ,0) In the formula, ΔP eff is the effective pressure difference, P is the current brake fluid pressure signal value, P avg It is the average value of the brake fluid pressure.

8. An enterprise safety supervision device based on data analysis, comprising the enterprise safety supervision system based on data analysis described in any one of claims 1-7.

Images