Research On Human Factors Parameter Monitoring And Early Warning Equipment In Confined Space

The safety in industrial production and city construction has been threatened by frequent accidence happened in confined space,which is directly caused by the loss of consciousness,tiredness and incorrect operation of workers.At present,research on confined space mainly focuses on accident analysis,hazard identification,harmful gas detection,safety management,etc.The monitoring of human factors parameters of operators is still limited.Therefore,this study investigates the effects of poor working environment on the unsafe physiological state of human body by measuring the photoplethysmography(PPG)signal of human body,which has practical significance for ensuring safe production.Through the manned experiment,the human PPG signal in the confined space working environment such as high temperature,high humidity and low oxygen was obtained,and the human factor parameters analysis based on PPG signal was realized by relying on various signal processing methods.The specific signal processing methods include morphological feature parameter extraction(conduction period,rising phase rising rate,falling phase falling rate),complexity calculation(KC complexity,high-order KC complexity),based on Hilbert-Huang(HHT)transform signal Domain and frequency domain analysis and deep learning based signal recognition and classification.The proposed PPG signal processing methods provide theoretical basis and basis for subsequent experimental data processing and analysis.A 100-min manned experiment was designed for the situation of high humidity and low oxygen in a confined space environment caused by the spontaneous breathing of the operator.After 8 people entered the confined space of 15m3,within 19min,the environment became high humidity and low oxygen environment,the relative humidity reached 80%,the oxygen volume fraction decreased to 20.5%,the workers entered the physiological fatigue period,and the ambient humidity exceeded at the end of the experiment.90%,the oxygen volume fraction dropped to 20.2%.The characteristic parameters of the PPG signal during the fatigue period changed significantly,showing a significant increase in the conduction period(p<0.05),a decrease in the slope of the rising portion,a decrease in the amount of ejection of the heartbeat,a slowing of the ejection rate,an increase in the slope of the descending portion,and a characterization of the blood vessel.The resistance increases and the vascular compliance deteriorates.Aiming at the high temperature and high humidity in the confined space caused by long-term closure of the import and export and poor ventilation.The control experiment was designed.The high temperature and high humidity environment were set to 35? and the relative humidity was 80%.According to the Predicted Mean Vote(PMV),the temperature of the control group(ie normal comfort environment)was 25?,and the relative humidity was 30%.Through the establishment of the deep learning residual neural network model,the feature extraction and recognition of PPG signals in two cases are realized.The established network classification performance is excellent,and it has good generalization ability.It can enter the high temperature and high humidity environment for 0.5s.Early warning is issued,and the short warning time can ensure that the operator knows the current physiological state,improves vigilance and ensures his own safety.In the case of confined space operation,the inert gas squeezes the space oxygen,causing the ambient oxygen concentration to be too low.Design an extremely low-oxygen environment manned experiment to obtain the PPG signal of the human body when the space oxygen content is as low as 15.5%?16%.Through the empirical mode decomposition(EMD),it is determined that the Intrinsic mode function(IMF)in the human PPG signal is the component that characterizes the hemodynamic information of the human body.The frequency domain characteristic is instantaneous by Hilbert transform.The frequency is concentrated at 1.5-2.5 Hz.When the human body is in a low oxygen state,the amplitude of the Hilbert marginal spectrum of the frequency is significantly reduced.Further build a deep learning convolutional neural network to identify the difference of human PPG signal under normal oxygen concentration and extreme hypoxia.The network can complete state recognition within 4s,ensuring early warning of hypoxia injury ahead of human cognition.Finally,at the basis of summarizing the experimental research results,a watch device was developed and designed based on real-time measurement of PPG signals to realize human factor parameter monitoring and abnormal physiological state warning.The device circuit diagram is drawn by Altium Designer software,and the Printed Circuit Board(PCB)is fabricated according to the drawings.At the same time,the SolidWorks software is used to design the appearance of the device and to write a signal receiving program that can be applied in MATLAB software.This paper provides a scientific reference for the impact of unsafe working environment on human body who in confined spaces,and provides a reference for avoiding limited space accidents from the perspective of human factors.