Rail transit construction risk prediction method and system

Abstract

The invention relates to the technical field of engineering construction risk monitoring, in particular to a rail transit construction risk prediction method and system, which synchronously captures environment variables such as noise spectrum and support deformation through a precision sensor network, and acquires physiological indexes such as electroencephalogram rhythm, eye movement characteristics and heart rate variability in a linkage manner. Therefore, the system can identify the composite risk state of the sudden change of environmental pressure and the decline of physiological function. Functional fusion is carried out on a regional dynamic pressure coefficient and a real-time fatigue index of a worker, a risk prediction model with a cross-dimensional conduction characteristic is constructed, risk prevention and control are upgraded from passive response to man-machine-environment cooperative defense, the limitation of a traditional monitoring system on separation analysis of environment parameters and personnel states is broken through, and the risk prediction accuracy is improved. Correlation monitoring of the construction environment and the personnel physiological pressure state is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering construction risk monitoring, in particular to a rail transit construction risk prediction method and system. BACKGROUND

[0002] Rail transit construction is a huge and precise system engineering, not only with many types of work, but also with extremely complex construction environment. The risks in which work not only come from the use of mechanical equipment, the special environment of underground operation, but also closely related to the psychological state of construction personnel and the implementation of management system. And the existing rail transit construction risk prevention and control mainly relies on discrete monitoring means, and the split analysis of environmental parameters and personnel state leads to early warning lag.

[0003] Traditional risk monitoring focuses on geological deformation, equipment operation and other physical indicators, although noise, support deformation and other data can be obtained through a sensor network, but it is not dynamically associated with the physiological state of personnel. Personnel monitoring mostly uses periodic physical examination or subjective questionnaire, which cannot capture sudden fatigue and stress response in real time. Especially in high-risk scenes such as shield tunneling and deep foundation pit, the coupling effect of environmental pressure change and personnel physiological decline often causes chain accidents. Therefore, there is an urgent need for a dynamic risk prediction method that can associate construction environment and personnel physiological state. SUMMARY

[0004] The technical problem solved by the present application is to provide a rail transit construction risk prediction method and system to solve the technical problem that traditional risk monitoring cannot monitor the coupling effect of environmental pressure change and personnel physiological decline.

[0005] The basic scheme provided by the present application is a rail transit construction risk prediction method, comprising the steps of: S1: Real-time acquisition of environmental pressure parameters of each construction area, the environmental pressure parameters including real-time noise level and support structure deformation rate; S2: According to the environmental pressure parameters, the dynamic pressure value K is calculated through the regional pressure coefficient model:

[0006] In the formula, is the real-time noise level, unit: dB; is the support structure deformation rate, unit: mm / h; , is the weight coefficient and satisfies ; , are respectively the preset threshold values of the real-time noise level and the support structure deformation rate; S3: Obtain the worker's vital sign data stream and the worker's positioning information through a wearable device, and extract the physiological fatigue index F; S4: constructing a risk coupling prediction function to predict the work risk value R of the target worker according to the dynamic pressure value of the construction area where the target worker is located and the physiological fatigue index of the target worker, the risk coupling prediction function being:

[0007] wherein a and b are coupling coefficients of environment and physiology; c is an electromyographic attenuation correction factor; sgn is an electromyographic signal slope sign function; S5: performing a hierarchical intervention on the target worker according to the risk value.

[0008] Further, the real-time noise level in S1 is collected by a capacitive microphone deployed in the target area; and the support structure deformation rate is collected by a MEMS tilt sensor deployed in the foundation pit support structure.

[0009] Further, the weight coefficients in S2 are dynamically updated by an entropy weight method.

[0010] Further, the sign data stream in S3 includes the power ratio of brain waves and waves, the eye movement closed eye length proportion, and the heart rate.

[0011] Further, the calculation formula of the physiological fatigue index F in S3 is:

[0012] wherein, is the power ratio of brain waves and waves; is the eye movement closed eye length proportion; is the root mean square difference in heart rate variability, is the individual resting heart rate variability baseline; is the normalized weight.

[0013] Further, the coupling coefficients a and b of environment and physiology in S4 are optimized by transfer learning:

[0014] wherein W is a weight matrix trained from historical engineering accident cases, and are historical environmental pressure data sets and historical fatigue state data sets, respectively.

[0015] Further, the hierarchical intervention in S5 includes: ​​​​​​​​​When R < 0.6, the worker state is normal, and no intervention is needed; When 0.6 <= R < 0.8, the target worker is reminded in real time through a wearable device to make autonomous state adjustment; When 0.8 <= R < 1.0, the target worker is forced to evacuate.

[0016] A rail transit construction risk prediction system applies any of the above rail transit construction risk prediction methods.

[0017] The principle and advantages of the present application are that the present application realizes the advanced pre-control of man-machine collaborative risk by establishing a double-mode coupling mechanism of the construction environment and the physiological state of personnel. The present application synchronously captures environmental variables such as noise spectrum and support deformation through a precision sensing network, and simultaneously collects physiological indexes such as electroencephalogram rhythm, eye movement characteristics and heart rate variability, so that the system can identify the composite risk state of the dramatic change of environmental pressure and the decline of physiological function. The regional dynamic pressure coefficient and the real-time fatigue index of the worker are functionally fused to construct a risk prediction model with cross-dimensional conduction characteristics, which breaks through the limitations of traditional monitoring systems on the fragmented analysis of environmental parameters and personnel state, and realizes the associated monitoring of the construction environment and the physiological pressure state of personnel. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is a step flowchart of the present application, a rail transit construction risk prediction method embodiment one. DETAILED DESCRIPTION

[0019] The following will be further described in detail through specific embodiments: The specific implementation process is as follows: Embodiment one is shown in the accompanying drawings Figure 1 A rail transit construction risk prediction method, including steps: S1: Real-time collection of environmental pressure parameters of each construction area, the environmental pressure parameters including real-time noise level and support structure deformation rate.

[0020] S2: According to the environmental pressure parameters, the dynamic pressure value K is calculated through the regional pressure coefficient model.

[0021] S3: Obtain the worker's vital sign data stream and the worker's positioning information through a wearable device, and extract the physiological fatigue index F; S4: Construct a risk coupling prediction function, and predict the work risk value R of the target worker according to the dynamic pressure value of the construction area where the target worker is located and the physiological fatigue index thereof. S5: According to the risk value, execute the hierarchical intervention on the target worker.

[0022] The environmental variables such as noise spectrum and support deformation are synchronously captured through a precise sensor network, and physiological indexes such as electroencephalogram rhythm, eye movement characteristics and heart rate variability are collected in linkage, so that the system can identify the compound risk state of environmental pressure change and physiological function decline. The regional dynamic pressure coefficient and the real-time fatigue index of workers are functionally fused to construct a risk prediction model with cross-dimensional conduction characteristics, realizing the correlation monitoring of construction environment and physiological pressure state of workers.

[0023] Specifically, the environmental pressure parameters of each construction area are first collected, including real-time noise level and support structure deformation rate. The real-time noise level in this embodiment is collected by a capacitive microphone deployed in the target area, and an IEC 61672 Class 1 sound level meter is used in this embodiment to calculate the equivalent continuous sound level every 1s. The support structure deformation rate is collected by a MEMS tilt sensor array deployed in the foundation pit support structure, and the MEMS tilt sensor array is arranged in the steel support node in this embodiment to calculate the angle change rate every 10min. In this embodiment, time synchronization of all sensors is realized through PTP protocol, and the worker position is mapped to the construction area pressure field by the UWB positioning system. The problem of time and space misalignment in traditional monitoring is solved, and aligned basic data is provided for coupled analysis.

[0024] Then, the dynamic pressure value K is calculated according to the environmental pressure parameters through the regional pressure coefficient model:

[0025] In the formula, is the real-time noise level, unit: dB; is the support structure deformation rate, unit: mm / h. , is the weight coefficient and satisfies In this embodiment, the weight coefficient , is dynamically updated based on the parameter volatility through entropy weight method. , are the preset threshold values of the real-time noise level and the support structure deformation rate respectively. Specifically, in this embodiment, the upper limit of noise tolerance is 110dB; the critical deformation rate of support is 3mm / h. This scheme analyzes the regional pressure coefficient through discrete sensor data, and quantifies the real-time risk intensity of each construction area.

[0026] The worker's vital sign data stream includes electroencephalogram wave and wave power ratio, eye movement closed eye duration ratio, and heart rate. The wearable device in this embodiment adopts a multi-channel dry electrode electroencephalogram cap integrated in an industrial safety cap (sampling rate 256 Hz, impedance <10 kΩ) and an embedded tool electrocardiogram patch. The front brim of the industrial safety cap collects the worker's brain waves through the built-in multi-channel dry electrode wave power ratio wave power ratio; and it is deployed with an infrared miniature camera, 60 fps, resolution 640x480, for collecting the worker's eye movement closed eye duration data. The embedded tool electrocardiogram patch is used to collect the worker's heart rate, sampling rate 250 Hz.

[0027] According to the worker's vital sign data stream, the physiological fatigue index F is extracted:

[0028] In the formula, is the brain wave wave power ratio wave power ratio is the eye movement closed eye duration ratio; is the root mean square difference in heart rate variability, mainly reflecting the instantaneous fluctuation of consecutive heart interval, the stronger the parasympathetic nerve activity, the greater the fluctuation of consecutive heart interval, and the higher the RMSSD value; is the individual resting heart rate variability baseline, which is the RMSSD standard value of the worker in the unloading resting state, representing the autonomic nerve regulation ability; , , is the normalized weight. This scheme adopts wavelet packet decomposition to eliminate motion artifacts, combined with ICA to separate eye interference. F>0.7 indicates that the worker has been in a serious fatigue state.

[0029] Then a risk coupling prediction function is constructed to predict the work risk value R of the target worker according to the dynamic pressure value of the construction area where the target worker is located and the physiological fatigue index of the target worker, and the risk coupling prediction function is:

[0030] In the formula, a and b are the coupling coefficients of environment and physiology; c is the electromyographic attenuation correction factor, which is determined by the sign of the median frequency slope of sEMG; sgn is the slope sign function of electromyographic signal. Among them, the coupling coefficients a and b of environment and physiology are optimized by transfer learning:

[0031] In the formula, W is the weight matrix trained from historical engineering accident cases, and The historical environmental stress data set and the historical fatigue state data set are respectively, the process quantifies the nonlinear superposition effect of environmental stress and physiological fatigue, and the muscle fatigue sudden disability response is enhanced by the electromyography attenuation sign function.

[0032] Finally, a hierarchical intervention is performed on the target worker according to the risk value, wherein the hierarchical intervention includes: When R < 0.6, the worker state is normal, and no intervention is performed; When 0.6≤R<0.8, the target worker is reminded to adjust the state autonomously in real time through the wearable device; When 0.8≤R<1.0, the target worker is forced to evacuate.

[0033] In summary, the scheme synchronously captures environmental variables such as noise spectrum and support deformation through a precision sensing network, and simultaneously collects physiological indicators such as electroencephalogram rhythm, eye movement characteristics, and heart rate variability, so that the system can identify the combined risk state of sudden environmental stress change and physiological function decline. The regional dynamic stress coefficient and the real-time fatigue index of the worker are functionally fused to construct a risk prediction model with cross-dimensional conduction characteristics, upgrading the risk prevention and control from "passive response" to "human-machine-environment collaborative defense", and providing a universal intelligent safety framework for rail transit construction.

[0034] Embodiment Two Embodiment Two differs from Embodiment One only in that Embodiment Two is a rail transit construction risk prediction system applying the above-mentioned rail transit construction risk prediction method.

[0035] The above is only an embodiment of the present application, and common knowledge of specific structures and characteristics in the scheme is not described in detail here. Ordinary skilled personnel in the art know all ordinary technical knowledge in the field of the application before the filing date or the priority date, can know all existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary skilled personnel in the art can improve and implement the present scheme based on their own abilities under the guidance of this application, and some typical known structures or known methods should not be an obstacle for the implementation of the present application. It should be noted that, for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which should also be considered within the scope of protection of the present application, and these will not affect the effectiveness and practicality of the present application. The scope of protection claimed in this application should be subject to the content of its claims, and the specific implementation in the specification can be used to explain the content of the claims.

Claims

1. A method for predicting construction risks in rail transit, characterized in that, Including the following steps: S1: Real-time acquisition of environmental pressure parameters for each construction area, including real-time noise level and deformation rate of support structure; S2: Calculate the dynamic pressure value K based on environmental pressure parameters using a regional pressure coefficient model. In the formula, Real-time noise level, unit: dB; The deformation rate of the support structure is expressed in mm / h. , The weighting coefficients are satisfied. ; , These are preset thresholds for real-time noise level and deformation rate of support structure, respectively. S3: Obtain worker vital sign data streams and worker location information through wearable devices, and extract the physiological fatigue index F; S4: Construct a risk coupling prediction function to predict the work risk value R of the target worker based on the dynamic pressure value and physiological fatigue index of the construction area where the target worker is located. The risk coupling prediction function is as follows: In the formula, a and b are the coupling coefficients between the environment and physiology; c is the electromyography attenuation correction factor; and sgn is the electromyography signal slope sign function. S5: Implement tiered interventions for target workers based on risk values.

2. The method for predicting construction risks in rail transit according to claim 1, characterized in that: The real-time noise level in S1 is acquired by a condenser microphone deployed in the target area; the deformation rate of the support structure is acquired by a MEMS tilt sensor deployed in the foundation pit support structure.

3. The method for predicting construction risks in rail transit according to claim 2, characterized in that: The weighting coefficient in S2 , Dynamically updated using the entropy weight method.

4. The method for predicting construction risks in rail transit according to claim 3, characterized in that: The vital signs data stream in S3 includes electroencephalogram (EEG) data. Waves and Wave power ratio, eye movement duration with eyes closed as a percentage, and heart rate.

5. The method for predicting construction risks in rail transit according to claim 4, characterized in that: The formula for calculating the physiological fatigue index F in S3 is as follows: In the formula, EEG Waves and Wave power ratio; The percentage of time spent with eyes closed during eye movements; The root mean square difference in heart rate variability. Baseline for individual resting heart rate variability; , , For normalized weights.

6. The method for predicting construction risks in rail transit according to claim 5, characterized in that: The coupling coefficients a and b between environment and physiology in S4 are optimized through transfer learning: In the formula, W is a weight matrix trained from historical engineering accident cases. and These are historical environmental stress datasets and historical fatigue state datasets, respectively.

7. The method for predicting construction risks in rail transit according to claim 6, characterized in that, The graded intervention in S5 includes: When R < 0.6, the worker is in a normal condition and no intervention is needed; When 0.6 ≤ R < 0.8, wearable devices will remind the target worker to adjust their autonomous state in real time. When 0.8 ≤ R < 1.0, the target workers are forcibly evacuated.

8. A rail transit construction risk prediction system, characterized in that: The method for predicting construction risks in rail transit, as described in any one of claims 1-7, was applied.

Images