Monitoring method of escape respirator storage device and related equipment

Abstract

The invention discloses a monitoring method of an escape respirator storage device and related equipment, and belongs to the field of intelligent safety equipment.The method comprises the following steps that a corresponding first dynamic threshold value interval is calculated based on collected basic data, whether the real-time statistical value of the basic data of the position node of the escape respirator storage device is located in the first dynamic threshold value interval or not is judged; analyzing the escape respirator storage device with the abnormal state through a pre-trained abnormity diagnosis model, and determining abnormal factors, abnormal reasons and abnormal time-space information; and when the abnormal reason is data acquisition abnormality, recording and log comparison are performed on the equipment environment information of the current position node to obtain the basic data of the truth-returning node, a second dynamic threshold interval is calculated according to the basic data of the truth-returning node, and whether the real-time statistical value is located outside the second dynamic threshold interval is judged, so that an alarm is given. The problem that fault detection cannot be performed on the escape respirator storage device in real time in the prior art can be solved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent safety devices, and particularly relates to a monitoring method for an escape breathing apparatus storage device and related equipment. Background Art

[0002] In the construction field of underground confined spaces such as tunnels and shafts, the escape breathing apparatus is a key emergency device for coping with sudden accidents and ensuring the life safety of construction workers. The performance of its storage device has a crucial impact on the efficiency of emergency rescue. However, the current escape breathing apparatus storage solutions commonly adopted in the industry have a series of significant technical defects, seriously restricting the improvement of construction safety guarantee levels.

[0003] The storage devices of traditional escape breathing apparatuses mostly adopt simple open cabinet structures. These devices are exposed to high-dust, humid, and corrosive environments for a long time, resulting in an accelerated aging rate of components such as the cylinder sealing rings of the breathing apparatus, and the valves are prone to blockage, thus reducing the reliability and service life of the equipment. In addition, the traditional storage solution relies entirely on manual inspections to monitor the device status, which is not only inefficient but also prone to misreading or omission due to human factors. Especially for subtle changes such as progressive air pressure decay, manual inspections are often difficult to detect. Moreover, the traditional storage devices lack environmental perception capabilities and do not integrate sensors for monitoring key parameters such as harmful gas concentrations (such as CO / CO2), particulate matter (such as PM2.5), and temperature and humidity. This leads to the inability to timely warn of potential risks such as harmful gas penetration or filter element saturation, thereby increasing the possibility of mask failure. In actual working scenarios, multiple operations such as opening the cabinet, self-checking, and connecting the mask need to be manually performed, which takes a long time and far exceeds the golden escape window period, seriously affecting the timeliness of emergency response. At the same time, due to the messy construction site environment, the storage of escape breathing apparatuses often cannot be uniformly managed, which not only increases the difficulty of daily management but also reduces the efficiency of emergency response. In case of abnormal situations or emergencies, construction workers need to spend a lot of time checking and preparing the equipment, thus delaying the escape opportunity. When problems occur with the escape breathing apparatus storage device or the cylinder, such as the device cannot be opened, the cylinder pressure is insufficient, or the cylinder air outlet is blocked, it will seriously threaten the life safety of construction workers and have a significant impact on the progress and efficiency of the overall construction project. Summary of the Invention

[0004] The purpose of the present invention is to provide a monitoring method for an escape breathing apparatus storage device and related equipment to solve the problem that the prior art cannot perform real-time fault detection on the escape breathing apparatus storage device.

[0005] To achieve the above object, the present invention adopts the following technical solutions: In a first aspect, a monitoring method for an escape breathing apparatus storage device includes the following steps: Collect the basic data of the escape breathing apparatus storage devices in local tunnels and shafts; Based on the basic data, calculate the corresponding first dynamic threshold range, and determine whether the real-time statistical value of the basic data of the position node of the escape breathing apparatus storage device is within the first dynamic threshold range. If the real-time statistical value of the current position node is outside the first dynamic threshold range, the status of the corresponding escape breathing apparatus storage device is abnormal; otherwise, the status is normal; Analyze the escape breathing apparatus storage devices with abnormal status through a pre-trained anomaly diagnosis model to determine the abnormal factors, abnormal causes, and abnormal spatio-temporal information; When the abnormal cause is abnormal data collection, record the device environment information of the current position node and compare it with the log to obtain the true node basic data. Calculate the second dynamic threshold range based on the true node basic data, and determine whether the real-time statistical value is outside the second dynamic threshold range. If the real-time statistical value is outside the second dynamic threshold range, alarm and output the abnormal factors, abnormal causes, and abnormal spatio-temporal information; otherwise, do not alarm.

[0006] In some embodiments, the basic data includes the spatial data, cylinder pressure data, temperature and humidity data, CO data, CO2 data, PM2.5 data, and illuminance data in local tunnels and shafts.

[0007] In some embodiments, the anomaly diagnosis model is trained through the following steps: Obtain the marked release curve of the cylinder manufacturer, and use the basic data to label the abnormal factors of the escape breathing apparatus storage device as the original training sample set; Based on the real-time statistical value, perform abnormal factor training on the original training sample set to form a real sample set; Input the real sample set into a convolutional neural network for feature extraction, establish the relationship between the abnormal cause and the abnormal factor, and capture the abnormal spatio-temporal information of the escape breathing apparatus storage device within the corresponding time to establish an anomaly diagnosis model, and repeatedly train the anomaly diagnosis model through the real-time statistical value.

[0008] In some embodiments, the convolutional neural network is an A3D network or an As3D network, and both convolutional neural networks include multiple cascaded convolutional layers; The size of the convolutional kernel in the convolutional layer of the A3D network is 3×3×3, and the convolutional kernel of the first convolutional layer in the As3D network is 3×7×7, and the other convolutional layers are 3×3×3; There are residual connections between the convolutional layers in the As3D network.

[0009] In some embodiments, the reasons for anomalies include: abnormal data acquisition, abnormal battery, abnormal pressure, abnormal temperature, abnormal humidity, abnormal CO content, abnormal CO2, abnormal dust, and abnormal light; The abnormal factors for the abnormal data acquisition include: terminal offline and missing terminal data; The abnormal factors for the abnormal battery include: too low supply battery voltage, too high supply battery voltage, unbalanced supply battery current, over-discharged supply battery, and over-charged supply battery; The abnormal factors for the abnormal pressure include: too high gas cylinder pressure, too low gas cylinder pressure, fluctuating gas cylinder pressure, continuous pressure relief of the gas cylinder, and loss of gas cylinder pressure; The abnormal factors for the abnormal temperature include: too high equipment temperature environment and too low equipment ambient temperature; The abnormal factors for the abnormal humidity include: abnormal humidity in the equipment environment; Abnormal CO content includes: abnormal CO content in the equipment environment; Abnormal CO2 includes: abnormal CO2 content in the equipment environment; Abnormal dust: abnormal dust in the equipment environment; Abnormal light: abnormal illuminance in the equipment environment.

[0010] In some embodiments, the number of connections inside the abnormal diagnosis model is where L is the number of layers of the convolutional neural network; The abnormal diagnosis model further includes a constraint condition, which is established through the following steps: Assume that at the running time t of the escape breathing apparatus storage device, the control factor of the node (x, y) is G(x, y, t). When the node moves to (x + dx, y + dy) at time t + dt, a formula for the control factor G(x + dx, y + dy, t + dt) to be equal to the control factor G(x, y, t) is obtained; Through Taylor series expansion, after eliminating the quadratic terms, the velocities u and v of the node (x, y) in the x-direction and y-direction are introduced respectively. Finally, G(x, y, t) is eliminated and the high-order terms with an order greater than or equal to 2 are ignored to obtain the constraint condition.

[0011] In a second aspect, a monitoring system for an escape breathing apparatus storage device includes: A data acquisition module for collecting basic data of the escape breathing apparatus storage device in the local tunnel and shaft; An abnormal state detection module, configured to calculate a corresponding first dynamic threshold range based on the basic data, and determine whether the real-time statistical value of the basic data of the position node of the escape breathing apparatus storage device is within the first dynamic threshold range. If the real-time statistical value of the current position node is outside the first dynamic threshold range, the state of the corresponding escape breathing apparatus storage device is abnormal; otherwise, the state is normal. An abnormality diagnosis module, configured to analyze the escape breathing apparatus storage device with abnormal state through a pre-trained abnormality diagnosis model to determine abnormal factors, abnormal reasons, and abnormal spatio-temporal information. An abnormal result verification module, configured to, when the abnormal reason is data acquisition abnormality, record the device environment information of the current position node and perform log comparison to obtain the basic data of the verification node, calculate a second dynamic threshold range based on the basic data of the verification node, and determine whether the real-time statistical value is outside the second dynamic threshold range. If the real-time statistical value is outside the second dynamic threshold range, an alarm is issued and the abnormal factors, abnormal reasons, and abnormal spatio-temporal information are output; otherwise, no alarm is issued.

[0012] In a third aspect, an electronic device includes a memory, a processor, and a computer program stored in the memory and executable in the processor. When the processor executes the computer program, the steps of the monitoring method for an escape breathing apparatus storage device are implemented.

[0013] In a fourth aspect, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the monitoring method for an escape breathing apparatus storage device are implemented.

[0014] In a fifth aspect, a computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the monitoring method for an escape breathing apparatus storage device are implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: By collecting the basic data of the escape breathing apparatus storage device in real time and calculating the dynamic threshold range, the present invention can monitor the status of the device in real time. Once it is found that the real-time statistical value exceeds the threshold range, it can be determined that the device status is abnormal, and an early warning can be issued in a timely manner, which helps to quickly respond to and handle potential problems; by using a pre-trained anomaly diagnosis model to analyze the escape breathing apparatus storage device with abnormal status, the abnormal factors, abnormal causes, and abnormal spatio-temporal information can be accurately determined. This helps to deeply understand the root cause of the anomaly and provides strong support for subsequent repairs and maintenance; when the abnormal cause is abnormal data collection, by recording and comparing the device environment information, the basic data of the true node is obtained, and the second dynamic threshold range is calculated for verification. This step helps to eliminate false alarms caused by incorrect data collection and improve the accuracy and reliability of monitoring.

[0016] Furthermore, the present invention considers various possible abnormal causes and abnormal factors to ensure a comprehensive detection of the status of the escape breathing apparatus storage device. The real sample set is input into the convolutional neural network for feature extraction, the relationship between the abnormal causes and abnormal factors is established, and the abnormal spatio-temporal information of the escape breathing apparatus storage device within the corresponding time is captured. An anomaly diagnosis model is established, and the anomaly diagnosis model is repeatedly trained through the real-time statistical value to realize the intelligent monitoring and diagnosis of the status of the escape breathing apparatus storage device, and continuously improve the accuracy and adaptability of the model. Brief Description of the Drawings

[0017] Figure 1 It is a schematic diagram of multi-dimensional convolution provided in this embodiment; Figure 2 It is a schematic diagram of the neural network characteristics inside the anomaly diagnosis model provided in this embodiment; Figure 3 It is a flowchart of the monitoring method of the escape breathing apparatus storage device provided in this embodiment; Figure 4 It is a structural diagram of the monitoring system of the escape breathing apparatus storage device provided in this embodiment; Figure 5 It is a flowchart of anomaly diagnosis provided in this embodiment; Figure 6 It is a flowchart of truth-preserving processing provided in this embodiment. Detailed Embodiment

[0018] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solution of the present invention will be further described in detail below with reference to the drawings. The content is an explanation of the present invention rather than a limitation.

[0019] It should be noted that the terms "including" and "having" and any variations thereof in the description and claims of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, systems, products, or devices.

[0020] As Figure 3 shown, this embodiment provides a monitoring method for an escape breathing apparatus storage device, including the following steps: S101. Collect the basic data of the escape breathing apparatus storage devices in the local tunnel and shaft, and establish the basic data of the escape breathing apparatus storage devices in the local tunnel and shaft.

[0021] Among them, the basic data of the escape breathing apparatus storage device includes spatial data in the tunnel and shaft, gas cylinder pressure data, temperature and humidity data, CO data, CO2 data, PM2.5 data, illuminance data, etc. The gas cylinder pressure data is the air pressure value of the gas cylinder in the escape breathing apparatus storage device, which will have consumption and loss along with the parameter characteristics of the gas cylinder equipment, and it is necessary to define a reasonable threshold for consumption and loss.

[0022] Specific to each item of data, a basic database can be established through the front-end data acquisition module. Gas cylinder pressure data (including the factory production parameters), battery-powered operation data, basic information of the temperature and humidity in the tunnel or underground, and CO, CO2 data, etc.; the collected calculation model data includes the air pressure gauge, current load, voltage parameters, battery status, operating conditions, alarm information, etc. of the escape breathing apparatus storage device; the cloud model data includes the information of the equipment of the escape breathing apparatus storage device in the tunnel or shaft, on-site environmental data (temperature, humidity, air content, etc.), production cycle data; gas cylinder off-cabinet data, self-check data, etc.

[0023] S102. Calculate the data thresholds required for the cloud data model using the terminal basic data, and use the continuously learning remote mathematical model to determine whether the maintenance value of the node escape breathing apparatus storage device is abnormal. Specifically, calculate the corresponding first dynamic threshold interval based on the basic data, and determine whether the real-time statistical value of the basic data of the position node of the escape breathing apparatus storage device is within the first dynamic threshold interval. If the real-time statistical value of the current position node is outside the first dynamic threshold interval, the corresponding escape breathing apparatus storage device is in an abnormal state; otherwise, the state is normal.

[0024] The statistical value of the escape breathing apparatus device is the difference between the total number of escape gas cylinders in the storage cabinet collected and calculated in real time and the number of construction workers. For example, a construction team is generally equipped with 6-8 escape gas cylinders. As the number of team members changes, it is necessary to maintain the difference in the personnel allocation ratio. The air pressure data of the gas cylinders is collected by differential input. The differential value needs to be adjusted according to different manufacturers of the gas cylinders. The statistical value of air pressure leakage and damage is the difference obtained by periodically collecting the escape breathing apparatus storage device and based on the cloud data model. By calculating the difference between the on-site data of the gas cylinder and the air pressure data provided by the gas cylinder brand using the data collected by the escape breathing device collection module, the threshold calculated by the cloud large model can be used to judge whether the on-site node is abnormal. When judging whether the gas cylinders of the escape breathing apparatus storage device are abnormal, the air pressure threshold mathematical model cultivated by the cloud can also be used to judge different types of gas cylinders.

[0025] The air pressure threshold mathematical model is established by the autonomous learning of the cloud big data model based on the on-site collection data of the escape breathing apparatus storage device in the tunnel or shaft node. When the on-line monitoring data value of the air pressure of the gas cylinder at the on-site node is within the normal value range of the air pressure model, the node device works normally and the monitoring and maintenance normal log is generated. If the on-line monitoring statistical value of the gas cylinder is outside the interval determined by the air pressure model threshold, it is determined that the air pressure monitoring statistical value of the escape breathing apparatus storage device is abnormal. Among them, for the statistical value of the air pressure monitoring data of the gas cylinders in the escape breathing apparatus storage device and the specific origin of the air pressure threshold model, the determination method of the data model threshold collected by the sensor and established can be referred to.

[0026] After determining whether the air pressure threshold statistical value is abnormal, it is possible to determine whether to perform an abnormal diagnosis according to the judgment result. Specifically, if the judgment result is yes, the steps of step S103 are executed; if the judgment result is no, no operation is required.

[0027] S103. Mark the escape breathing apparatus storage device at the tunnel or shaft node where the air pressure threshold mathematical model determines an abnormality, retrieve the log in the background for data aggregation, and determine the abnormal factors, abnormal causes, and abnormal environment information that cause the abnormal escape breathing device at the node through a pre-trained abnormal diagnosis model.

[0028] Among them, the abnormal environment information includes the time information and geographical space information of the tunnel or shaft where the alarm or abnormality occurs (such as the location of the abnormal device, the geographical location of the construction tunnel or shaft, etc.). Specifically, when a node terminal alarm or a cloud model abnormality occurs, the abnormal environment information is the time, location, temperature, humidity, and equipment information of the escape breathing apparatus storage device at the fault alarm node.

[0029] In the embodiments of the present invention, since data collection of on-site nodes is required to complete the sample library, an abnormal diagnosis model of an escape breathing apparatus storage device can be preset. Among them, the abnormal diagnosis model of the gas cylinder pressure is a model obtained after training with the parameter curve of the gas cylinder manufacturer as the training sample. This model can realize the abnormal diagnosis of the escape breathing apparatus storage device through the decision tree algorithm. Since the types of abnormal reasons of the device are limited, continuous model training is required to make the process of abnormal diagnosis accurate and fast.

[0030] The training process of the real sample set of the abnormal diagnosis model of the escape breathing apparatus storage device includes: S1031. Obtain the marked discharge curve of the gas cylinder manufacturer as the original training sample set; S1032. According to the abnormal factor judgment rule, for each threshold marked in the sample model, combine the data collected in real time by the on-site node device to perform abnormal factor training and train the real sample set; S1033. Perform abnormal factor statistics on the real sample set, determine the abnormal factors and abnormal reasons, and record them in the real sample set.

[0031] Specifically, first create a mathematical model of the escape breathing apparatus holding device, and then perform model training. Model training is to determine the characteristic attributes (i.e., abnormal factors) and obtain training samples, and further complete the analysis of abnormal factors of the breathing apparatus holding device, especially the gas cylinder device, and establish an association relationship between the abnormal factors and the abnormal reasons based on the analysis results.

[0032] Determining the cause of the failure can be completed manually, that is, the abnormal reasons in the real training sample set can be manually marked. For example, it can be divided into 9 categories and 19 features such as data acquisition abnormality, battery abnormality, pressure abnormality, temperature abnormality, humidity abnormality, and light abnormality as shown in Table 1.

[0033] Table 1 Abnormal reasons and abnormal factors

[0034] Obtaining the training samples is realized by the node device in the tunnel or shaft. The basic data comes from the theoretical values of normal temperature, normal pressure and the gas cylinder parameter curve. The judgment threshold data is set according to the theoretical data in the initial stage of the abnormal sample, including the classification of abnormal reasons. Through the abnormal factor judgment rule, the abnormal phenomenon characteristics (i.e., abnormal factors) of the on-site escape breathing apparatus holding device are automatically analyzed and marked, that is, the original training sample set is obtained. Among them, the versions and serial numbers in the original training sample set and the real training sample set are only used to distinguish the different information marked in the samples in the sample set, and there are no other meanings such as sequence and order. Because the training sample set is updated in real time, the cloud big data model will collect on-site data in real time and train a new model.

[0035] The main work in the model training stage is to continuously collect on-site data, calculate the occurrence frequency of each abnormal influencing factor of the device in the real training samples and the conditional probability estimation of each abnormal factor division for each abnormal cause, and record the results. Its input is the real training samples, and after the training is completed, an abnormal diagnosis model of the escape breathing apparatus storage device can be obtained.

[0036] Since there are more than 5 abnormal factors, the theoretical basis adopts the convolutional network and neural network learning theory. Multidimensional convolution is to stack multiple consecutive abnormal factors to form a three-dimensional factor cube, and then operate on the factor cube with a three-dimensional convolution kernel. Through three-dimensional convolution, feature extraction can be performed on a continuous and long-term normal or abnormal factor sequence. Each feature after the convolutional layer is connected to multiple adjacent factor states in the previous layer, so the operation information of the escape breathing apparatus holding device within a period of time can be captured.

[0037] The schematic diagram of multidimensional convolution is as Figure 1 shown. The formula is established: The value of the true position (x, y, z) of the nth judgment factor in the mth layer can be expressed as

[0038] In the formula: is the output at (x, y, z) in the feature map of the nth factor in the mth layer; tanh is the activation function; is the shared bias of the factor feature map; PQR are the height, width and time length of the multidimensional convolution kernel; is the weight between the (p, q, r) in the feature map of the nth factor in the mth layer and the feature map of the oth in the (m - 1)th layer; u is the input from the (m - 1)th layer to the mth layer. Based on the A3D and As3D network structures constructed by multidimensional convolution, the following matrix:

[0039] In the above matrix, the number of abnormal factors, that is, the number of channels, is omitted. The multidimensional convolution kernel is represented as a cube of L×H×W, where L, H, and W are the time length, height, and width of the multidimensional convolution kernel respectively. In addition, the feature factor sampling layer is omitted in the table. The network has a total of 8 convolution operations, and the various convolutional layers are cascaded. Among them, the size of the convolution kernel is 3×3×3. As shown in the above calculation formula, after the data model network passes through two fully connected layers, an abnormal diagnosis model of the escape breathing apparatus storage device that can judge the abnormal state is finally obtained.

[0040] The network is equivalent to the second version of the network. Except Except that the convolution kernel size is 3×7×7, the rest are all 3×3×3. The network uses three-dimensional convolution in the residual network to extract various eigen data collected in real time at the terminal. In the convolutional layer , , , residual connections are used. The residual network is easier to optimize and can solve the problem of gradient disappearance caused by the increase in network depth. Thus, the accuracy of the cloud data model's autonomous learning and abnormal factor judgment is improved.

[0041] The abnormal diagnosis model of the escape breathing apparatus storage device needs to have the ability of autonomous learning and updating, so it also needs to have the characteristics of a neural network inside, such as Figure 2 shown. For a traditional network with L layers (i.e., a hierarchical progressive network), there are a total of L connections; for the neural network of the escape breathing apparatus storage device mathematical model, there are connections. If the transformation function of the th layer is denoted as , and the output is , then the change of each layer of the mathematical model neural network can be described by a simple formula, that is In the formula, means that the factor feature map is connected and stable.

[0042] Table 2 Estimation of Abnormal Causes by Abnormal Factors

[0043] Based on the estimation of each influencing factor for the current situation of abnormal causes from the real-time collected data samples of the tunnel or shaft escape breathing apparatus storage device nodes, the establishment of the data model involved is as follows: The control factor at the data collection point in the tunnel or shaft escape breathing apparatus storage device node is . At the device operation time , the data collected at this point reaches , so the abnormal factor judgment value at can be denoted as , assuming it is equal to (1) Expand the left side of equation (1) using the Taylor series, and after calculation and eliminating the quadratic term, we get: (2) (3) (4) where represents the high-order term with an order greater than or equal to 2, and eliminating and ignore it is necessary to collect the constraint equations of the data model: (5) Based on the above formulas (1)-(5) and the statistical results of the training sample set, an abnormal diagnosis model of the escape breathing apparatus storage device based on the convolutional network and big data model algorithm can be constructed, and it can learn autonomously and be used for diagnosing the reasons for the abnormalities of the escape breathing apparatus storage device. When (a certain abnormal state feature is repeatedly established and learned by the data model), an abnormal factor will be established belonging to the data model of the network type.

[0044] The process of abnormal diagnosis by the abnormal diagnosis model of the escape breathing apparatus storage device is the abnormal diagnosis of the parameter states of the node gas cylinders in the construction tunnel or shaft. As Figure 5 shown, it is a schematic diagram of the model training process provided by the embodiment of the present invention. Specifically, the training and diagnosis process includes: S1034. Analyze the node escape breathing apparatus storage device according to the abnormal factor judgment rule to determine the target abnormal factor and abnormal space-time information; S1035. Use the correlation relationship between the abnormal factor and the internal members of the device to establish a mathematical model of various factor abnormalities; S1036. Use the real-time online collected data to improve the mathematical model of factor abnormalities.

[0045] Among them, assuming that the abnormal factors affecting the escape breathing apparatus storage device are abnormal gas cylinder pressure, and the reasons for the device abnormality include battery abnormality, data acquisition abnormality, pressure abnormality, temperature abnormality, CO abnormality, CO2 abnormality, etc., then according to the multi-dimensional convolutional network equation, it can be obtained:

[0046] From the mathematical model established from the data of this network equation, the abnormal pressure factor has been established in the model cube, and then its various abnormal reasons are monitored and collected in real time and compared with the mathematical model: the abnormal pressure factor will be accurately marked in the matrix network sequence. When determining what kind of target abnormality the escape breathing apparatus storage device is, the matrix abnormal reasons can be compared with the feedback to determine the type of target abnormality. Apply the method provided by the embodiments of the present invention to monitor the oxygen pressure value of the gas cylinder of the escape breathing apparatus storage device in real time, display it on the screen in real time, and transmit it to the cloud through the Internet of Things module. Use the cloud big data model to establish an online mathematical model of the escape breathing apparatus storage device at each construction tunnel and underground node online, so as to realize the model management and non-time-sharing maintenance of the escape breathing apparatus storage device. When the operation data of the escape breathing apparatus storage device at a certain node is abnormal, the cloud big data model will report the abnormality, and at the same time start the abnormal diagnosis of the cloud mathematical model to determine the fault point of the escape breathing apparatus storage device at the node, and use the threshold value of the mathematical model to judge whether the collected value is abnormal; if so, input the collected data into the abnormal diagnosis mathematical model of the escape breathing apparatus storage device for abnormal cause analysis to determine the abnormal factors, abnormal causes and abnormal time and space information that cause the device to be abnormal; among them, the abnormal diagnosis model is a mathematical model obtained by combining theoretical training models with real data learning.

[0047] It should be noted that based on the above embodiments, the embodiments of the present invention also provide corresponding improvement solutions. In the preferred / improved embodiments, the same steps or corresponding steps as those in the above embodiments can be referred to each other, and the corresponding beneficial effects can also be referred to each other, and will not be repeated one by one in the preferred / improved embodiments of this article.

[0048] Preferably, considering the abnormal collection caused by the complex on-site environment, there are many interference factors in the establishment of the mathematical model. To improve the accuracy of the abnormal diagnosis mathematical model of the escape breathing apparatus storage device and reduce unnecessary abnormal alarms, it is currently still at the level of improving the accuracy and authenticity of front-end collection, lacking targeted means to deal with complex external environments. The pertinence of the establishment of the abnormal factor model is not strong, and it cannot effectively solve the substantial problems, and the improvement of reliability is limited.

[0049] Based on the abnormal diagnosis method of the escape breathing apparatus storage device provided by the embodiments of the present invention, the real cause can be diagnosed through the manifested phenomena, which will effectively reduce the layout and maintenance costs of the escape breathing apparatus holding device during tunnel and shaft construction; at this time, the mathematical model can be targeted for truth-seeking processing to reduce the ratio of false alarms of on-site abnormalities of the device (such as gas cylinder pressure loss), which will greatly improve the work efficiency of tunnel and shaft construction personnel and management personnel. Therefore, step S104 is added after step S103, as Figure 6 shown, which is a schematic diagram of the truth-seeking processing flow provided by the embodiments of the present invention, that is: S104, truth-seeking processing: S1041, when the abnormal cause is data collection abnormality, record the device environment information of the current position node and compare it with the log to obtain the basic truth-seeking node data; S1042. Establish a mathematical model using the basic data of the truth nodes, and use the threshold of the mathematical model to calculate the second dynamic threshold interval based on the basic data of the truth nodes, and determine whether the real-time statistical value is outside the second dynamic threshold interval, so as to determine whether the real-time statistical value is abnormal; S1043. If the real-time statistical value is outside the second dynamic threshold interval, an alarm is issued and the abnormal factors, abnormal reasons and abnormal spatio-temporal information are output. Otherwise, no alarm is issued and a prompt message for checking data acquisition anomalies is output.

[0050] The monitoring method provided in this embodiment has the following advantages: (1) Applying the method provided in the embodiment of the present invention, the oxygen pressure value of the gas cylinder of the escape breathing apparatus storage device is monitored in real time, displayed on the real-time screen, and transmitted to the cloud through the IOT module (Internet of Things). The mathematical model of the escape breathing apparatus storage device for each construction tunnel and underground node is established online using the cloud big data model, realizing the model management and non-time-sharing maintenance of the escape breathing apparatus storage device. When the operation data of the escape breathing apparatus storage device at a certain node is abnormal, the cloud big data model will report the anomaly, and at the same time start remote anomaly diagnosis to determine the fault point of the escape breathing apparatus storage device at the node. Combining the emergency release ERC (Emergency Release Connector), TAG (Tagging) positioning, digital relay protection, and multi-information interaction, the remote maintenance of the escape breathing apparatus storage device is completed.

[0051] (2) When it is determined that an alarm or failure occurs in the escape breathing apparatus storage device at a certain tunnel or shaft node, the cloud model records the fault information and enters it into the model system. The cloud client program independently trains the corresponding mathematical model for the gas cylinder brand, gas cylinder capacity, gas cylinder factory date, and gas cylinder pressure module. This mathematical model can be used as a model reference for the monitoring, maintenance, parameter characteristics, and consumable selection of the escape breathing device gas cylinder. When the abnormal reporting of the gas cylinder pressure monitoring data occurs, the cloud big data records and issues an alarm, and the local escape breathing storage device liquid crystal screen alarms, and the gas cylinder self-check and maintenance program is started, automatically locking the anomalies caused by pressure, leakage, etc. The embodiment of the invention can shorten the time for tunnel and shaft construction personnel to check the gas cylinder pressure to within 2 minutes. Reduce the time for tunnel and shaft construction personnel to take out the escape breathing device by 1 minute and 30 seconds, and optimize the time for construction personnel to wear the breathing escape mask by 2 minutes. When emergencies occur in the tunnel and shaft, the escape breathing apparatus holding device will win precious escape time for construction personnel and ensure the safety of personal and property.

[0052] This embodiment also provides an intelligent storage cabinet for an escape breathing apparatus, which is composed of modules such as a storage vehicle body, a liquid crystal display screen, a pressure acquisition sensor, a MEMS sensor, a carbon monoxide sensor, a carbon dioxide sensor, a PM2.5 sensor, and a central processing unit. The electrical characteristics and mechanical structure design are combined with each other to standardize the daily storage of the breathing apparatus, manage the intelligent escape breathing apparatus, and facilitate the inspection and maintenance of the escape breathing apparatus. It realizes the visual overall planning and digital management of the escape breathing apparatus during the construction process.

[0053] As Figure 4 shown, this embodiment provides a monitoring system for an escape breathing apparatus storage device, including: A data acquisition module for acquiring the basic data of the escape breathing apparatus storage device in the local tunnel and shaft; An abnormal state detection module for calculating a corresponding first dynamic threshold interval based on the basic data, and determining whether the real-time statistical value of the basic data of the position node of the escape breathing apparatus storage device is within the first dynamic threshold interval. If the real-time statistical value of the current position node is outside the first dynamic threshold interval, the state of the corresponding escape breathing apparatus storage device is abnormal, otherwise the state is normal; An abnormal diagnosis module for analyzing the escape breathing apparatus storage device with an abnormal state through a pre-trained abnormal diagnosis model to determine the abnormal factors, abnormal reasons, and abnormal spatio-temporal information; An abnormal result verification module for, when the abnormal reason is data acquisition abnormality, recording the device environment information of the current position node and comparing it with the log to obtain the verified node basic data, calculating a second dynamic threshold interval based on the verified node basic data, and determining whether the real-time statistical value is outside the second dynamic threshold interval. If the real-time statistical value is outside the second dynamic threshold interval, an alarm is given and the abnormal factors, abnormal reasons, and abnormal spatio-temporal information are output, otherwise no alarm is given.

[0054] The division of modules in the embodiments of the present invention is illustrative, only a logical function division. In actual implementation, there may be other division methods. In addition, in each embodiment of the present invention, the functional modules can be integrated in one processor, or can exist independently physically, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0055] In this embodiment, a computer device is further provided. The computer device includes a processor and a memory. The memory is used to store a computer program (in this embodiment, the computer program includes a calculation component and an iteration component, and can perform model calculation and model update). The computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function. The processor described in the embodiment of the present invention can be used for the operation of a monitoring method of an escape breathing apparatus storage device.

[0056] This embodiment also provides a storage medium, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in a computer device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and, of course, the extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And in this storage space, one or more instructions suitable for being loaded and executed by the processor are also stored. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the monitoring method of an escape breathing apparatus storage device in the above embodiment.

[0057] This embodiment also provides a computer program product. The computer program product includes a computer program. When the computer program is executed by the processor, the corresponding steps of the monitoring method of an escape breathing apparatus storage device in the above embodiment are implemented.

[0058] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0059] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0060] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that realize the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0062] Finally, 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 them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: still modifications or equivalent substitutions can be made to the specific embodiments of the present invention, and any modification or equivalent substitution that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A monitoring method for an escape breathing apparatus storage device, characterized in that: The following steps are involved: Collect basic data on escape breathing apparatus storage devices in local tunnels and shafts; Calculate the corresponding first dynamic threshold interval based on the basic data, determine whether the real-time statistical value of the basic data of the location node of the escape respirator storage device is within the first dynamic threshold interval, if the real-time statistical value of the current location node is outside the first dynamic threshold interval, then the state of the corresponding escape respirator storage device is abnormal, otherwise the state is normal; The abnormal state of the escape respirator storage device is analyzed through the pre-trained abnormal diagnosis model to determine the abnormal factors, abnormal causes and abnormal spatiotemporal information; When the cause of the abnormality is data collection abnormality, the device environment information of the current location node is recorded and log-compared to obtain the return-to-true node basic data, the second dynamic threshold interval is calculated based on the return-to-true node basic data, and it is determined whether the real-time statistical value is outside the second dynamic threshold interval. If the real-time statistical value is outside the second dynamic threshold interval, an alarm is issued and the abnormal factors, abnormal causes and abnormal time and space information are output, otherwise no alarm is issued.

2. The monitoring method of the escape breathing apparatus storage device according to claim 1 is characterized in that: The basic data includes local tunnel and shaft space data, gas cylinder pressure data, temperature and humidity data, CO data, CO2 data, PM2.5 data and illumination data.

3. The monitoring method of the escape breathing apparatus storage device according to claim 1 is characterized in that: The abnormal diagnosis model is trained by the following steps: Obtain the discharge curve marked by the gas cylinder manufacturer, and annotate the abnormal factors of the escape breathing apparatus storage device based on the basic data as the original training sample set; Based on the real-time statistical value, abnormal factor training is performed on the original training sample set to form a real sample set; The real sample set is input into a convolutional neural network for feature extraction, the relationship between the abnormal cause and the abnormal factor is established, and the abnormal spatiotemporal information of the escape respirator storage device within the corresponding time is captured to establish an abnormal diagnosis model, which is repeatedly trained through real-time statistical values.

4. The monitoring method of the escape breathing apparatus storage device according to claim 3 is characterized in that: The convolutional neural network is an A3D network or an As3D network, and the convolutional neural network includes a plurality of convolutional layers connected in series; The convolution kernel size in the convolution layer of the A3D network is 3×3×3. In the As3D network, except for the first convolution layer, the convolution kernel size is 3×7×7, and the other convolution layers are all 3×3×3. Residual connections are formed between the convolutional layers in the As3D network.

5. The monitoring method of the escape breathing apparatus storage device according to claim 1 is characterized in that: Abnormal causes include: data collection abnormality, battery abnormality, pressure abnormality, temperature abnormality, humidity abnormality, CO content abnormality, CO2 abnormality, dust abnormality and light abnormality; The abnormal factors of data collection abnormality include: terminal offline and terminal data missing; The abnormal factors of the battery abnormality include: the power supply battery voltage is too low, the power supply battery voltage is too high, the power supply battery current is unbalanced, the power supply battery is over-discharged, and the power supply battery is over-charged; The abnormal factors of the abnormal pressure include: excessively high cylinder pressure, excessively low cylinder pressure, cylinder pressure fluctuation, continuous cylinder pressure release and cylinder pressure loss; The abnormal factors of the temperature abnormality include: the device temperature environment is too high and the device environment temperature is too low; The abnormal factors of the abnormal humidity include: abnormal humidity of the equipment environment; Abnormal CO content includes: abnormal CO content in the equipment environment; CO2 abnormalities include: abnormal CO2 content in the equipment environment; Abnormal dust: Abnormal dust in the equipment environment; Abnormal lighting: The lighting in the device environment is abnormal.

6. The monitoring method of the escape breathing apparatus storage device according to claim 3 is characterized in that: The number of connections within the anomaly diagnosis model is ,in L is the number of layers of the convolutional neural network; The abnormality diagnosis model also includes constraints, which are established by the following steps: Assuming that the escape respirator storage device is running at time t, the control factor of the node (x, y) is G(x, y, t), and at time t+dt, when the node moves to (x+dx, y+dy), the formula that the control factor G(x+dx, y+dy, t+dt) is equal to the control factor G(x, y, t) is obtained; After Taylor series expansion and elimination of quadratic terms, the velocities u and v of the node (x, y) in the x and y directions respectively are introduced. Finally, G(x, y, t) is eliminated and high-order terms with an order greater than or equal to 2 are ignored to obtain the constraint conditions.

7. A monitoring system for an escape breathing apparatus storage device, characterized in that: include: Data collection module, used to collect basic data of escape breathing apparatus storage devices in local tunnels and shafts; an abnormal state detection module, used to calculate the corresponding first dynamic threshold interval based on the basic data, and determine whether the real-time statistical value of the basic data of the location node of the escape respirator storage device is within the first dynamic threshold interval; if the real-time statistical value of the current location node is outside the first dynamic threshold interval, the state of the corresponding escape respirator storage device is abnormal, otherwise the state is normal; An abnormality diagnosis module is used to analyze the escape respirator storage device in an abnormal state through a pre-trained abnormality diagnosis model to determine the abnormal factors, abnormal causes and abnormal spatiotemporal information; The abnormal result return module is used to record and compare the equipment environment information of the current location node with the log to obtain the return node basic data when the abnormal cause is data collection abnormality, calculate the second dynamic threshold interval according to the return node basic data, and determine whether the real-time statistical value is outside the second dynamic threshold interval. If the real-time statistical value is outside the second dynamic threshold interval, an alarm is issued and the abnormal factor, abnormal cause and abnormal time and space information are output; otherwise, no alarm is issued.

8. An electronic device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of a monitoring method for an escape breathing apparatus storage device as claimed in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the monitoring method of an escape breathing apparatus storage device according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, the steps of the monitoring method of an escape breathing apparatus storage device according to any one of claims 1 to 6 are implemented.

Images