Method for displaying physical health state of personnel before equipment operation based on intelligent monitoring platform

Abstract

The invention provides a method for displaying the physical health state of a person before equipment operation based on an intelligent monitoring platform, and the method comprises the steps: firstly carrying out the cleaning, standardization and fusion processing of a collected multi-dimensional data source, generating multi-modal health data, and selecting the input variables of the health features of an operator through the Pearson correlation analysis; a convolutional neural network is combined with a gating circulation unit to extract spatial-temporal features of the monitoring feature quantity, and an attention mechanism is used to distribute corresponding weights for the gating circulation unit; repeatedly training the monitored health data by using exponential weighted moving average, and obtaining an alarm threshold value of a human body health state in combination with a model evaluation index root-mean-square error; based on a membership function combining a semi-trapezoid and a semi-ridge shape, determining a health state grade of the operator; and finally, the intelligent monitoring platform judges the body health state of the operator and whether the operator is allowed to operate the equipment or not before operating the equipment by the operator, and monitors and forcibly intercepts the operation authority of the person in an unhealthy state in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of human health monitoring, and in particular to a method for displaying the physical health status of a person before equipment operation based on an intelligent monitoring platform. Background Art

[0002] In the context of energy interconnection in today's society, the safe and stable operation of hydropower plant power generation equipment plays an increasingly important role in the interconnection of power systems and energy supply. The switching operation of the equipment is one of the tasks performed by the operators of hydropower plants. When traditional equipment operators perform equipment switching operations, they need to report to the person in charge on duty after arriving at the site before operating the equipment. The operator's physical health will affect the normal equipment operation process to a certain extent. However, during the traditional equipment operation process, the person in charge on duty can only understand the health status of the equipment operator by phone. This method cannot fully, real-timely, and understand the operator's physical health status throughout the entire operation process. There is a risk of poor health of the operator and subsequent misoperation, which affects the operation of the equipment. Summary of the Invention

[0003] The present invention aims to provide a method for displaying the physical health status of personnel before equipment operation based on a smart monitoring platform. By utilizing multimodal data fusion processing, multi-task deep learning models, adaptive learning algorithms and other technical means, the smart monitoring platform is used to judge the physical health status of the operator before operating the equipment, determine whether the operator is allowed to operate the equipment, monitor in real time and forcibly intercept the operating permissions of personnel in unhealthy conditions, thereby improving the safety level of power station operation and maintenance and greatly promoting the development of smart power stations.

[0004] In order to achieve the above-mentioned technical features, the purpose of the present invention is achieved as follows: a method for displaying the physical health status of personnel before equipment operation based on an intelligent monitoring platform. This method first cleans, standardizes and fuses the collected multi-dimensional data sources by utilizing multimodal data fusion processing, multi-task deep learning models, and adaptive learning algorithm technologies to generate multimodal health data. Then, a deep learning model is used for feature extraction and analysis, and an adaptive learning algorithm is used to perform health status assessment. Finally, based on the assessment results, the intelligent monitoring platform automatically determines whether the operator is allowed to operate the equipment, monitors in real time and forcibly intercepts the operating permissions of unhealthy personnel, and improves the safety level of power station operation and maintenance.

[0005] Preferably, the method specifically includes the following steps: Step 1: Pre-process the various indicator data of normal operators and ambient temperature collected by the smart monitoring platform; Step 2: Select the input variables of operator health characteristics through Pearson correlation analysis; Step 3: Use a convolutional neural network combined with a gated recurrent unit to extract the spatiotemporal features of the monitoring feature quantity, and use an attention mechanism to assign corresponding weights to the gated recurrent unit; Step 4: Use the exponentially weighted moving average to repeatedly train the monitored health data, and combine it with the model evaluation indicator root mean square error to obtain the alarm threshold of the human health status; Step 5: Determine the health status level of the operator based on the membership function combining the semi-trapezoidal and semi-ridged forms; Step 6: When the operator receives an instruction to operate the device, the smart monitoring platform installed on the device connects with the smart bracelet worn by the operator to collect the operator's health status data and perform data preprocessing operations; Step 7: Use the monitoring data from step 6 as the feature input of the model trained in step 4, and compare the output root mean square error evaluation index with the alarm threshold of the human health status obtained in step 4; Step 8: If the root mean square error of the operator's health status evaluation index is less than the alarm threshold, the smart monitoring platform issues a command to operate the equipment. Otherwise, the smart monitoring platform issues an alarm and locks the equipment operation process, terminating the operation.

[0006] Preferably, the process of collecting various indicators of normal operators by the smart monitoring platform in step 1 is as follows: when an operator in normal health approaches the device to be operated, the smart monitoring platform installed in front of the device to be operated automatically detects the operator's approach and connects with the smart bracelet worn by the operator in real time to collect the operator's health status data; The health status data of the equipment operator collected by the smart monitoring platform include: body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, stress monitoring data, and emotion monitoring data. Among them, body temperature data, blood oxygen concentration data, and respiratory rate data are directly measured by the infrared technology installed in the smart monitoring platform, and heart rate data, blood pressure data, maximum oxygen uptake data, electrocardiogram data, stress monitoring data, and emotion monitoring data are measured by the smart bracelet worn by the operator and transmitted to the smart monitoring platform synchronously.

[0007] Preferably, among the various indicator data of the normal operator in step 1: the normal range of body temperature data is 36°C to 37.2°C, the normal range of heart rate data is 60-100 times / minute, the normal range of blood pressure data is systolic pressure: 90-139 mmHg, diastolic pressure: 60-89 mmHg, the normal range of respiratory rate data is 12-20 times / minute, the normal data of maximum oxygen uptake is 2500-3500 ml / minute, the normal data of electrocardiogram is 60-100 times / minute, the normal range of blood oxygen concentration data is 94%-100%, the normal range of pressure monitoring data is 50-70, and the normal range of emotion monitoring data is 0-100. The higher the score, the greater the emotional fluctuation. It is uniformly generated by the smart bracelet based on the various indicator data of the operator in the past day.

[0008] Preferably, in step 1, the various indicator data of normal operators collected by the smart monitoring platform are normalized to the maximum and minimum values ​​and mapped to the range [0, 1], wherein the normalization formula is: ; Where: is the normalized data; 、 are the minimum and maximum values ​​of the sample data set respectively; is the original sample data.

[0009] Preferably, the Pearson correlation coefficient in step 2 r It is used to measure the similarity between two indicators. The larger the correlation coefficient, the stronger the correlation between the two variables; conversely, the weaker the correlation. The calculation formula is: ; in: for X The average value of for Y The average value of n To monitor the number of samples.

[0010] Preferably, the convolutional neural network in step 3 consists of an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer; The gated recurrent unit in step 3 is the input gate and the forget gate in the long short-term memory network model combined into the update gate in the gated recurrent unit model. Z t It determines how much information from the previous moment is retained at the current moment. The larger the value, the more information from the previous moment is retained, and vice versa. r t Controls ignoring the information of the previous moment. The larger the value, the less information of the previous moment is retained. The input of the gated recurrent unit model is x t , combined with z t and r t Get the output h t The formula is as follows: ; ; ; ; Where, Represents input x t With the previous output h t-1 The combination of W (z) 、 U (z) 、 W (r) 、 U (r) 、 U and W represents the training parameter matrix, Represents the element composite relationship, σ and tanh are the hyperbolic tangent functions of Sigmoid and tanh respectively; The attention mechanism in step 3 is used to strengthen the influence on the output variable, and its calculation formula is shown as follows: ; ; ; Where: For the GRU layer t Output at any moment, b is the bias vector, is the weight coefficient, C For the attention layer t Output at any moment.

[0011] Preferably, in step 3, a convolutional neural network is combined with a gated recurrent unit to extract the spatiotemporal features of the feature quantity, and an attention mechanism is used to assign corresponding weights to the gated recurrent unit. The structure of the combined network model consists of an input layer, a CNN layer, a GRU layer, an attention layer, and an output layer. First, the CNN extracts the spatial features of the original data and uses them as the input of the GRU network. Second, the GRU extracts the temporal features and inputs the results into the attention layer. Finally, the attention layer calculates the weights based on the input data. The input layer data is the collected health status data of normal operators: body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, stress monitoring data, and emotion monitoring data; the parameters of the CNN layer are set to a kernel length of 1 and a number of 128; the GRU model is constructed with two layers, with 64 neurons in the first layer and 128 neurons in the second layer. The model optimizer uses Adam and the activation function is the ReLU function, implemented based on the Python programming language; the input of the attention layer is the output of the GRU model, and the output of the CNN-GRU-Attention combination model is the operator's health status level.

[0012] Preferably, the root mean square error formula of the evaluation indicator in step 4 is as follows: ; Where: n Indicates the sample size of the test set; X act ( i )and X pred ( i ), i =1, 2,… n , respectively i The true value and predicted value of the moment detection value; If the input data is the operator's healthy state data and can be adapted to the model, the root mean square error of the prediction is small. If the input data is the operator's unhealthy state data and cannot be adapted to the model, the root mean square error of the prediction will increase. Based on the predicted values ​​obtained by the CNN-GRU-Attention combination model, the exponentially weighted moving average is used to set a threshold to observe the changing trend of the root mean square error. The operator's health status is judged based on the changing trend. The formula for calculating the exponentially weighted moving average is shown below: ; Where: is the weight of historical monitoring data; is the control line statistic of the exponentially weighted moving average; is the root mean square error; The threshold for monitoring the operator's health status is the upper limit of the exponentially weighted moving average, and its calculation formula is as follows: ; Where, is the mean of the root mean square error, is the standard deviation of the root mean square error; X is a constant related to the threshold position.

[0013] Preferably, the health status level of the operator is determined in step 5 and represented by a membership function combining a semi-trapezoidal and a semi-ridged form, which is defined as follows: ; Where: g 1( x m ), g 2( x m ), g 3( x m ), g 4( x m ) represent the four levels of health status of operators: healthy, good, caution, and serious. Warning prompts are given for caution level, and alarms and locking of equipment operation permissions are given for serious level. g i The four intervals are 0~0.3, 0.1~0.6, 0.4~0.9, 0.7~1, b i The six cutoff values ​​are 0.1, 0.3, 0.4, 0.6, 0.7, and 0.9.

[0014] The present invention has the following beneficial effects: The patent of this invention proposes a method and system for displaying the physical health status of personnel before equipment operation based on a smart monitoring platform. By utilizing multimodal data fusion processing, multi-task deep learning models, adaptive learning algorithms and other technical means, the smart monitoring platform is used to judge the physical health status of the operator before operating the equipment, whether the operator is allowed to operate the equipment, and real-time monitoring and forced interception of the operating permissions of unhealthy personnel, thereby improving the safety level of power station operation and maintenance and greatly promoting the development of smart power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings and examples.

[0016] Figure 1 This is a method flow of the present invention for displaying the physical health status of personnel before equipment operation based on an intelligent monitoring platform.

[0017] Figure 2 This is the convolutional neural network structure of the present invention.

[0018] Figure 3 This is the LSTM network model of the present invention.

[0019] Figure 4 This is the GRU network model of the present invention.

[0020] Figure 5 This is a structural diagram of the CNN-GRU-Attention combined model based on the present invention.

[0021] Figure 6 This is the membership function distribution diagram of the present invention.

[0022] Figure 7 The operator of the present invention wears a smart bracelet.

[0023] Figure 8 This is the overall layout diagram of the operating equipment of the present invention.

[0024] Figure 9 It is an intelligent monitoring platform installed on the operating equipment of the present invention.

[0025] Figure 10 This is a correlation analysis diagram between the monitoring feature variables of the present invention.

[0026] Figure 11 This is a prediction model for the health status of the person operating the device of the present invention.

[0027] Figure 12 The health status of the person operating the device of the present invention. DETAILED DESCRIPTION

[0028] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0029] Example 1: See also Figure 1-12 A method for displaying the physical health status of personnel before equipment operation based on a smart monitoring platform. This method uses multimodal data fusion processing, multi-task deep learning models, and adaptive learning algorithm technologies to first clean, standardize, and fuse the collected multidimensional data sources to generate multimodal health data. It then uses a deep learning model to extract and analyze features, and combines it with an adaptive learning algorithm to perform health status assessment. Finally, based on the assessment results, the smart monitoring platform automatically determines whether the operator is allowed to operate the equipment, monitors in real time, and forcibly blocks the operating permissions of personnel in unhealthy conditions, thereby improving the safety level of power plant operation and maintenance.

[0030] A method based on a smart monitoring platform to display the health status of personnel before equipment operation is as shown in the attached Figure 1 The specific steps are as follows: Step 1: Pre-process the various indicator data of normal operators and ambient temperature collected by the smart monitoring platform; The process of collecting various indicators of normal operators by the smart monitoring platform in step 1 is as follows: when an operator in normal physical health approaches the equipment to be operated, the smart monitoring platform installed on the front of the equipment to be operated automatically detects the operator's approach and connects with the smart bracelet worn by the operator in real time to collect the operator's health status data. The health status data of the equipment operator collected by the smart monitoring platform specifically includes: body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, stress monitoring data, emotion monitoring data, etc. Among them, body temperature data, blood oxygen concentration data, and respiratory rate data are directly measured by infrared technology installed in the smart monitoring platform, and heart rate data, blood pressure data, maximum oxygen uptake data, electrocardiogram data, stress monitoring data, and emotion monitoring data are measured by the smart bracelet worn by the operator and synchronously transmitted to the smart monitoring platform.

[0031] Among the various indicator data of the normal operator in step 1: the normal range of body temperature data is 36°C to 37.2°C, the normal range of heart rate data is 60-100 times / minute, the normal range of blood pressure data is systolic pressure (high pressure): 90-139 mmHg, diastolic pressure (low pressure): 60-89 mmHg, the normal range of respiratory rate data is 12-20 times / minute, the normal data of maximum oxygen uptake is 2500-3500 ml / minute, the normal data of electrocardiogram is 60-100 times / minute, the normal range of blood oxygen concentration data is 94%-100%, the normal range of stress monitoring data is 50-70, and the normal range of emotion monitoring data is 0-100. The higher the score, the greater the emotion fluctuation. The score is uniformly generated by the smart bracelet based on the various indicator data of the operator in the past day.

[0032] In step 1, the various indicator data of normal operators collected by the smart monitoring platform are normalized to the maximum and minimum values ​​and mapped to the range [0, 1]. The normalization formula is: ; Where: is the normalized data; 、 are the minimum and maximum values ​​of the sample data set respectively; is the original sample data.

[0033] Step 2: Select the input variables of operator health characteristics through Pearson correlation analysis; The Pearson correlation coefficient in step 2 r It is used to measure the similarity between two indicators. The larger the correlation coefficient, the stronger the correlation between the two variables; conversely, the weaker the correlation. The calculation formula is: ; in: for X The average value of for Y The average value of n To monitor the number of samples.

[0034] Step 3: Use a convolutional neural network combined with a gated recurrent unit to extract the spatiotemporal features of the monitoring feature quantity, and use an attention mechanism to assign corresponding weights to the gated recurrent unit; The convolutional neural network (CNN) in step 3 is a feedforward neural network that uses local connections and shared weights to efficiently process data, which can effectively reduce the complexity and overfitting defects in the data processing process. Its structure consists of an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. Figure 2 CNN has the characteristics of a small number of free parameters, multi-dimensional network learning data features, and good robustness. By alternating convolutional layers and pooling layers, it can reduce the dimension of feature data and improve the quality of feature data.

[0035] The Gated Recurrent Unit (GRU) in step 3 is an upgrade based on the Long Short-Term Memory (LSTM) model, which solves the problems of complex LSTM parameter training and long training time. Its internal structure is shown in Figures (3-4). The input gate and forget gate in the LSTM model are merged into the update gate in the GRU model. Z t It determines how much information from the previous moment is retained at the current moment. The larger the value, the more information from the previous moment is retained, and vice versa. r t Controls ignoring the information of the previous moment. The larger the value, the less information of the previous moment is retained.

[0036] The input of the GRU model is x t , combined with z t and r t Get the output h t The formula is as follows: ; ; ; ; Where, Represents input x t With the previous output ht-1 The combination of W (z) 、 U (z) 、 W (r) 、 U (r) 、 U and W represents the training parameter matrix, Represents the element composite relationship, σ and tanh are the hyperbolic tangent functions of Sigmoid and tanh respectively; The attention mechanism in step 3 is used to strengthen the influence on the output variable, and its calculation formula is shown as follows: ; ; ; Where: For the GRU layer t Output at any moment, b is the bias vector, is the weight coefficient, C For the attention layer t Output at any moment.

[0037] In step 3, the convolutional neural network is combined with the gated recurrent unit to extract the spatiotemporal features of the feature quantity, and the attention mechanism is used to assign corresponding weights to the gated recurrent unit. The structure of the combined network model is as follows: Figure 5 As shown in the figure, the model consists of an input layer, a CNN layer, a GRU layer, an attention layer, and an output layer. First, the CNN extracts spatial features from the raw data and uses them as input to the GRU network. Second, the GRU extracts temporal features and feeds the results into the attention layer. Finally, the attention layer calculates weights based on the input data. The input layer data consists of collected health data of normal operators: body temperature, heart rate, blood pressure, respiratory rate, maximum oxygen uptake, electrocardiogram (ECG), blood oxygen concentration (BOS), stress monitoring data, and mood monitoring data. The CNN layer parameters are set to a kernel length of 1 and a number of 128 neurons. The GRU model is constructed with two layers, with 64 neurons in the first layer and 128 neurons in the second layer. The model optimizer uses Adam and the activation function is the ReLU function, implemented in the Python programming language. The input to the attention layer is the output of the GRU model, and the output of the combined CNN-GRU-Attention model is the operator's health status level.

[0038] Step 4: Use the exponentially weighted moving average to repeatedly train the monitored health data, and combine it with the model evaluation indicator root mean square error to obtain the alarm threshold of the human health status; The root mean square error formula of the evaluation index in step 4 is as follows: ; Where: n Indicates the sample size of the test set; X act ( i )and X pred ( i ), i =1, 2,… n , respectively i The true value and predicted value of the moment detection value; If the input data is the operator's healthy state data and can be adapted to the model, the root mean square error of the prediction is small. If the input data is the operator's unhealthy state data and cannot be adapted to the model, the root mean square error of the prediction will increase. Based on the predicted values ​​obtained by the CNN-GRU-Attention combination model, the exponentially weighted moving average is used to set a threshold to observe the changing trend of the root mean square error. The operator's health status is judged based on the changing trend. The formula for calculating the exponentially weighted moving average is shown below: ; Where: is the weight of historical monitoring data; in this embodiment, it is 0.3; is the control line statistic of the exponentially weighted moving average; is the root mean square error; The threshold for monitoring the operator's health status is the upper limit of the exponentially weighted moving average, and its calculation formula is as follows: ; Where, is the mean of the root mean square error, is the standard deviation of the root mean square error; X is a constant related to the threshold position. X =3.

[0039] Step 5: Determine the health status level of the operator based on the membership function combining the semi-trapezoidal and semi-ridged forms; In step 5, the health status level of the operator is determined and represented by a membership function combining a semi-trapezoidal and a semi-ridged form, which is defined as follows: ; Where: g1( x m ), g 2( x m ), g 3( x m ), g 4( x m ) represent the four levels of health status of operators: healthy, good, caution, and serious. Warning prompts are given for caution level, and alarms and locking of equipment operation permissions are given for serious level. g i The four intervals are 0~0.3, 0.1~0.6, 0.4~0.9, 0.7~1, b i The six cutoff values ​​are 0.1, 0.3, 0.4, 0.6, 0.7, and 0.9. The membership function combining the semi-trapezoidal and semi-ridge shapes is shown in the attached figure. Figure 6 shown.

[0040] Step 6: When the operator receives an instruction to operate the device, the smart monitoring platform installed on the device connects with the smart bracelet worn by the operator to collect the operator's health status data and perform data preprocessing operations; The smart bracelet worn by the operator in step 6 is as shown in the attached Figure 7 As shown, the overall layout of the operating equipment is as follows Figure 8 As shown in the figure, the intelligent monitoring platform installed on the operating equipment is as follows Figure 9 shown.

[0041] Step 7: Use the monitoring data from step 6 as the feature input of the model trained in step 4, and compare the output root mean square error evaluation index with the alarm threshold of the human health status obtained in step 4; Step 8: If the root mean square error of the operator's health status evaluation index is less than the alarm threshold, the smart monitoring platform issues a command to operate the equipment. Otherwise, the smart monitoring platform issues an alarm and locks the equipment operation process, terminating the operation.

[0042] Example 2: This embodiment selects the data information collected from 400 equipment operators as test samples for analysis. In order to facilitate the analysis of the effect of the model, the physical conditions of the first 65% of the equipment operators are normal, and the physical conditions of the last 35% of the equipment operators are abnormal. A method for displaying the physical health status of personnel before equipment operation based on an intelligent monitoring platform and the data information monitored by the system include the operator's body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, pressure monitoring data, and pressure monitoring data. However, too many variables will cause data redundancy and affect the accuracy of the prediction model. The target variable of the patent of the present invention is the health status of the equipment operator, and the input variables include: the operator's body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, pressure monitoring data, and emotion monitoring data. Through Pearson correlation analysis, variable features with high correlation with the operator's health status are selected as input variables, and finally variables with correlation coefficients greater than 0.8 are selected as input variables. The correlation of each input variable is as follows: Figure 10 The selection results are shown in Table 1. In summary: After Pearson correlation analysis, the operator's body temperature data, heart rate data, blood pressure data, respiratory rate data, blood oxygen concentration data, stress monitoring data, and emotion monitoring data were finally selected as the feature input variables of the CNN-GRU-Attention combination model.

[0043] Table 1 Model input variables selected by Pearson correlation analysis

[0044] The health status of the equipment operator is determined by calculating the root mean square error, and the threshold is set using the exponentially weighted moving average. The monitoring feature variables selected by Pearson correlation analysis are used as the input of CNN-GRU-Attention. The prediction model of the equipment operator's health status is shown in the attached figure. Figure 11 As shown in the figure, it can be seen that in the prediction results of the first 260 feature sequences, the root mean square error is less than the threshold set by the exponential weighted moving average, indicating that the health status of the equipment operator is normal. After the 260th feature, the root mean square error is greater than the threshold set by the exponential weighted moving average, indicating that the health status of the equipment operator is abnormal, which is consistent with the actual feature input. It can be judged that the prediction performance of the model proposed in the present invention is effective. The health status of the equipment operator is shown in the attached figure. Figure 12 As shown in the figure, the first 260 feature quantities are judged to be healthy and good, which are normal, and the smart monitoring platform issues instructions to operate the equipment; the next 140 feature quantities are judged to be cautionary and serious, which are abnormal, and the smart monitoring platform issues an alarm and locks the operation process of the equipment operator, terminating the operation.

[0045] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for displaying the health status of personnel before equipment operation based on a smart monitoring platform, characterized in that: This method utilizes multimodal data fusion processing, multi-task deep learning models, and adaptive learning algorithm technologies to first clean, standardize, and fuse the collected multidimensional data sources to generate multimodal health data. It then uses a deep learning model to extract and analyze features, and combines it with an adaptive learning algorithm to assess health status. Finally, based on the assessment results, the smart monitoring platform automatically determines whether the operator is allowed to operate the equipment, monitors in real time, and forcibly blocks the operating permissions of unhealthy personnel, thereby improving the safety level of power plant operation and maintenance.

2. According to claim 1, a method for displaying the health status of personnel before equipment operation based on a smart monitoring platform is characterized in that: The specific steps include: Step 1: Pre-process the various indicator data of normal operators and ambient temperature collected by the smart monitoring platform; Step 2: Select the input variables of operator health characteristics through Pearson correlation analysis; Step 3: Use a convolutional neural network combined with a gated recurrent unit to extract the spatiotemporal features of the monitoring feature quantity, and use an attention mechanism to assign corresponding weights to the gated recurrent unit; Step 4: Use the exponentially weighted moving average to repeatedly train the monitored health data, and combine it with the model evaluation indicator root mean square error to obtain the alarm threshold of the human health status; Step 5: Determine the health status level of the operator based on the membership function combining the semi-trapezoidal and semi-ridged forms; Step 6: When the operator receives an instruction to operate the device, the smart monitoring platform installed on the device connects with the smart bracelet worn by the operator to collect the operator's health status data and perform data preprocessing operations; Step 7: Use the monitoring data from step 6 as the feature input of the model trained in step 4, and compare the output root mean square error evaluation index with the alarm threshold of the human health status obtained in step 4; Step 8: If the root mean square error of the operator's health status evaluation index is less than the alarm threshold, the smart monitoring platform issues a command to operate the equipment. Otherwise, the smart monitoring platform issues an alarm and locks the equipment operation process, terminating the operation.

3. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 2, characterized in that: The process of collecting the various indicators of the normal operator by the smart monitoring platform in step 1 is as follows: when an operator in good health approaches the equipment to be operated, the smart monitoring platform installed in front of the equipment to be operated automatically detects the operator's approach and connects with the smart bracelet worn by the operator in real time to collect the operator's health status data; The health status data of the equipment operator collected by the smart monitoring platform include: body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, stress monitoring data, and emotion monitoring data. Among them, body temperature data, blood oxygen concentration data, and respiratory rate data are directly measured by the infrared technology installed in the smart monitoring platform, and heart rate data, blood pressure data, maximum oxygen uptake data, electrocardiogram data, stress monitoring data, and emotion monitoring data are measured by the smart bracelet worn by the operator and transmitted to the smart monitoring platform synchronously.

4. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 3, characterized in that: Among the various indicator data of the normal operator in step 1: the normal range of body temperature data is 36°C to 37.2°C, the normal range of heart rate data is 60-100 times / minute, the normal range of blood pressure data is systolic pressure: 90-139 mmHg, diastolic pressure: 60-89 mmHg, the normal range of respiratory rate data is 12-20 times / minute, the normal data of maximum oxygen uptake is 2500-3500 ml / minute, the normal data of electrocardiogram is 60-100 times / minute, the normal range of blood oxygen concentration data is 94%-100%, the normal range of stress monitoring data is 50-70, and the normal range of emotion monitoring data is 0-100. The higher the score, the greater the emotion fluctuation. The score is uniformly generated by the smart bracelet based on the various indicator data of the operator in the past day.

5. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 4, characterized in that: In step 1, the various indicator data of normal operators collected by the smart monitoring platform are normalized to the maximum and minimum values ​​and mapped to the range [0, 1]. The normalization formula is: ； Where: is the normalized data; 、 are the minimum and maximum values ​​of the sample data set respectively; is the original sample data.

6. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 2, characterized in that: The Pearson correlation coefficient in step 2 r It is used to measure the similarity between two indicators. The larger the correlation coefficient, the stronger the correlation between the two variables; conversely, the weaker the correlation. The calculation formula is: ； in: for X The average value of for Y The average value of n To monitor the number of samples.

7. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 6, characterized in that: The convolutional neural network in step 3 consists of an input layer, a convolutional layer, a pooling layer, a fully connected layer and an output layer; The gated recurrent unit in step 3 is the input gate and the forget gate in the long short-term memory network model combined into the update gate in the gated recurrent unit model. Z t It determines how much information from the previous moment is retained at the current moment. The larger the value, the more information from the previous moment is retained, and vice versa. r t Controls ignoring the information of the previous moment. The larger the value, the less information of the previous moment is retained. The input of the gated recurrent unit model is x t , combined with z t and r t Get the output h t The formula is as follows: ； ； ； ； Where, Represents input x t With the previous output h t-1 The combination of W (z) 、 U (z) 、 W (r) 、 U (r) 、 U and W represents the training parameter matrix, Represents the element composite relationship, σ and tanh are the hyperbolic tangent functions of Sigmoid and tanh respectively; The attention mechanism in step 3 is used to strengthen the influence on the output variable, and its calculation formula is shown as follows: ； ； ； Where: For the GRU layer t Output at any moment, b is the bias vector, is the weight coefficient, C For the attention layer t Output at any moment.

8. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 7, characterized in that: In step 3, a convolutional neural network is combined with a gated recurrent unit to extract the spatiotemporal features of the feature quantity, and an attention mechanism is used to assign corresponding weights to the gated recurrent unit. The structure of the combined network model consists of an input layer, a CNN layer, a GRU layer, an attention layer, and an output layer. First, the CNN extracts the spatial features of the original data and uses them as the input of the GRU network; secondly, the GRU extracts the temporal features and inputs the results into the attention layer; finally, the attention layer calculates the weights based on the input data. The input layer data is the collected health status data of normal operators: body temperature data, heart rate data, blood pressure data, respiratory rate data, maximum oxygen uptake data, electrocardiogram data, blood oxygen concentration data, stress monitoring data, and emotion monitoring data; the parameters of the CNN layer are set to a kernel length of 1 and a number of 128; the GRU model is constructed with two layers, with 64 neurons in the first layer and 128 neurons in the second layer. The model optimizer uses Adam and the activation function is the ReLU function, implemented based on the Python programming language; the input of the attention layer is the output of the GRU model, and the output of the CNN-GRU-Attention combination model is the operator's health status level.

9. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 8, characterized in that: The root mean square error formula of the evaluation index in step 4 is as follows: ； Where: n Indicates the sample size of the test set; X act ( i )and X pred ( i ), i =1, 2,… n , respectively i The true value and predicted value of the moment detection value; If the input data is the operator's healthy state data and can be adapted to the model, the root mean square error of the prediction is small. If the input data is the operator's unhealthy state data and cannot be adapted to the model, the root mean square error of the prediction will increase. Based on the predicted values ​​obtained by the CNN-GRU-Attention combination model, the exponentially weighted moving average is used to set a threshold to observe the changing trend of the root mean square error. The operator's health status is judged based on the changing trend. The formula for calculating the exponentially weighted moving average is shown below: ； Where: is the weight of historical monitoring data; is the control line statistic of the exponentially weighted moving average; is the root mean square error; The threshold for monitoring the operator's health status is the upper limit of the exponentially weighted moving average, and its calculation formula is as follows: ； Where, is the mean of the root mean square error, is the standard deviation of the root mean square error; X is a constant related to the threshold position.

10. The method for displaying the health status of personnel before equipment operation based on a smart monitoring platform according to claim 9, characterized in that: In step 5, the health status level of the operator is determined and represented by a membership function combining a semi-trapezoidal and a semi-ridged form, which is defined as follows: ； Where: g 1( x m ), g 2( x m ), g 3( x m ), g 4( x m ) represent the four levels of health status of operators: healthy, good, caution, and serious. Warning prompts are given for caution level, and alarms and locking of equipment operation permissions are given for serious level. g i The four intervals are 0~0.3, 0.1~0.6, 0.4~0.9, 0.7~1, b i The six cutoff values ​​are 0.1, 0.3, 0.4, 0.6, 0.7, and 0.9.

Images