Ai-based industrial safety management system

Abstract

The present invention relates to an AI-based industrial safety management system, characterized by comprising: a data collection unit provided within an industrial site and collecting industrial site data using wired and wireless communication; a risk determination unit that determines whether there is a risk to equipment and workers using the industrial site data collected by the data collection unit; an equipment control unit that controls the equipment by receiving a preset operation signal generated to be transmitted to equipment determined to be abnormal by the risk determination unit; and a monitoring unit that collects the industrial site data and the control signal transmitted by the equipment control unit and transmits them to a preset user terminal.

Description

Technology Field

[0001] The present invention relates to a safety management system for managing safety within an industrial site, and more specifically, to an AI-based industrial safety management system that performs safety status determination and management by conducting AI-based analysis of data collected from IoT and ICT devices installed within the site. Background Technology

[0003] An industrial site refers to a place where various production activities take place to manufacture or provide goods or services, and this can appear in various forms such as factories, construction sites, mines, and energy production facilities.

[0004] Industrial sites are regarded as one of the core sectors of economic activity both domestically and internationally, and they play a key role in productivity and economic growth in modern society.

[0005] These industrial sites play a crucial role in various aspects of modern society—including technology, economy, employment, and national development—through the production and manufacturing of diverse manufactured goods, job creation based on a vast workforce, technological advancement and innovation such as production technology automation, robotics, and information technology, a central role in the national economy, the circulation of the domestic economy through the efficient consumption of various resources and energy, and the development of urban and regional infrastructure and urbanization; thereby driving the overall development and stability of society.

[0006] Furthermore, since workers perform production and manufacturing in industrial settings amidst various facilities, industrial safety is considered from multiple perspectives with the aim of protecting both workers and the environment. Factors such as safety regulations and training, protective equipment and facilities, hazardous material management, accident prevention and response plans, health monitoring and provision of rest periods, the introduction of technological innovation and automation, and the improvement of cultural awareness are considered to enhance workplace safety. Industrial safety protects not only productivity but also the health and safety of workers, which can ultimately reduce social costs on a macro level.

[0007] However, industrial safety issues can manifest in various aspects. Representative examples include the labor environment of some companies, insufficient or inadequate compliance with safety regulations, a lack of safety training for on-site supervisors and workers, issues regarding exposure to and handling of hazardous materials in workplaces, risks of breakdowns and accidents due to insufficient maintenance and preventive measures for facilities and equipment, physical or mental fatigue of workers, and a lack of understanding of safety regulations and procedures resulting from poor communication between employers and workers. Depending on the scale and risk level of the industrial site, the severity of injuries sustained by workers following an accident can range from serious injury to fatality.

[0008] Therefore, research is required on an AI-based industrial safety management system that enables communication between various equipment and sensors within industrial sites and performs AI-based industrial safety management to improve productivity, predict and manage maintenance, strengthen safety management, automate logistics and warehouse management, make data-driven decisions, and predict demand, market trends, and production plans using advanced predictive analytics such as machine learning and deep learning, thereby enhancing competitiveness and reducing accident rates through safety management within industrial sites. Prior art literature

[0010] Korean Registered Patent No. 10-2005188 The problem to be solved

[0011] The present invention aims to prevent potential safety accidents in industrial sites by collecting multiple sensing data from equipment, site spaces, and workers within the industrial site and determining whether there are abnormalities. means of solving the problem

[0013] An AI-based industrial safety management system according to one embodiment of the present invention may include: a data collection unit provided within an industrial site and collecting industrial site data using wired and wireless communication; a risk determination unit that determines whether there is a risk to equipment and workers using industrial site data collected by the data collection unit; an equipment control unit that controls said equipment by receiving a preset operation signal generated to be transmitted to equipment determined to be abnormal by the risk determination unit; and a monitoring unit that collects said industrial site data and a control signal transmitted by said equipment control unit and transmits them to a preset user terminal.

[0014] Additionally, the data collection unit may include a video data collection unit that photographs equipment and workers within the industrial site using a preset video recording device, an equipment data collection unit that collects operation data of preset equipment provided within the industrial site, a sensor data collection unit that collects sensing data from an environmental measurement sensor that senses temperature, humidity, vibration, noise, fine dust, harmful gas concentration, pressure, light intensity, and magnetic field provided inside or outside the industrial site, and a worker data collection unit that collects health data including the body temperature, heart rate, blood pressure, blood sugar, and oxygen saturation of the worker from a preset terminal worn by the worker.

[0015] In addition, the risk judgment unit can perform worker behavior detection using video data collected by the data collection unit, and if it determines that there is an abnormality, transmit an alert to the worker's preset terminal and transmit a worker tracking request alert to the monitoring unit, and can perform part replacement timing prediction and fault diagnosis using equipment data collected from the equipment.

[0016] Additionally, the sensor data collection unit includes a sensor information collection unit that collects the supply voltage, consumption current, and output voltage supplied to the environment measurement sensor, a sensor error measurement unit that extracts a comparison target time, which is a time when the rate of change of the measurement value of the sensing data exceeds a preset ratio, and a data transmission unit that periodically transmits the comparison target time of the sensor error measurement unit to the monitoring unit, and if there is no time exceeding the preset ratio, the sensor error measurement unit can extract sensing data corresponding to n arbitrary times within the sensing data as the comparison target time.

[0017] In addition, the sensor data collection unit is equipped with an identical sensor within a preset radius from the environment measurement sensor as a comparison sensor, and the monitoring unit is,

[0018] Using the sensing data collected from the above environment measurement sensor and the comparison sensor and the comparison target time, the sensor failure probability (S) according to the following [Equation 1] TR It produces ),

[0019] [Mathematical Formula 1]

[0020]

[0021] (Here, S TR is the sensor failure probability, T ms1 is the measured value of the environmental measurement sensor at the first time point, T ms2 is the measured value of the environmental measurement sensor at the second time point, T ms3 is the measured value of the environmental measurement sensor at the third viewpoint, T msnis the measurement value of the environment measurement sensor at the nth time point, T cst1 is the measurement value of the comparison sensor at the first time point, T cst2 is the measurement value of the comparison sensor at the second time point, T cs3 is the measurement value of the comparison sensor at the third viewpoint, T csn is the measurement value of the comparison sensor at the nth time point, T msav is the average measurement value of the environmental measurement sensor, and SE represents the efficiency deviation of the environmental measurement sensor.

[0022] The efficiency deviation (SE) of the above-mentioned environmental measurement sensor is calculated according to the following [Equation 2], and

[0023] [Mathematical Formula 2]

[0024]

[0025] (Here, SE av is the average efficiency of the environmental measurement sensor, SE i represents the initial efficiency of the environmental measurement sensor)

[0026] The average efficiency of the above-mentioned environmental measurement sensor can be calculated according to the following [Equation 3].

[0027] [Mathematical Formula 3]

[0028]

[0029] (Here, W IN1 is the power supplied to the environmental measurement sensor at the first time point, W OUT1 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the first time point, W IN2 is the power supplied to the environmental measurement sensor at the second viewpoint, W OUT2 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the second viewpoint, W INn W is the power supplied to the environment measurement sensor at the n-th time point. OUTn represents the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the nth time point. Effects of the invention

[0031] According to the present invention, by collecting multiple sensing data from equipment, site spaces, and workers within an industrial site and determining whether there is an abnormality, it is possible to prevent potential safety accidents occurring within the industrial site in advance. Brief explanation of the drawing

[0033] FIG. 1 is a block diagram illustrating an AI-based industrial safety management system according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an intermediate block diagram of a data collection unit within an AI-based industrial safety management system according to an embodiment of the present invention. FIG. 3 is a diagram illustrating an intermediate block diagram of a sensor data collection unit within an AI-based industrial safety management system according to an embodiment of the present invention. Specific details for implementing the invention

[0034] Specific details regarding the problem to be solved, the means for solving the problem, and the effects of the invention as described above are included in the embodiments and drawings to be described below. The advantages and features of the present invention, and the methods for achieving them, will become clear by referring to the embodiments described below in detail together with the accompanying drawings.

[0035] The scope of the present invention is not limited to the embodiments described below, and various modifications can be made by those skilled in the art within the scope of the technical essence of the present invention.

[0036] Hereinafter, the AI-based industrial safety management system of the present invention will be described in detail with reference to the attached FIGS. 1 to 3.

[0037] First, FIG. 1 is a block diagram of an AI-based industrial safety management system according to an embodiment of the present invention, FIG. 2 is an intermediate block diagram of a data collection unit within an AI-based industrial safety management system according to an embodiment of the present invention, and FIG. 3 is an intermediate block diagram of a sensor data collection unit within an AI-based industrial safety management system according to an embodiment of the present invention.

[0038] Referring to FIG. 1, an AI-based industrial safety management system according to one embodiment of the present invention may include a data collection unit (110), a risk judgment unit (120), an equipment control unit (130), and a monitoring unit (140).

[0039] The above data collection unit (110) is provided within the industrial site and can collect industrial site data using wired and wireless communication.

[0040] More specifically, the data collection unit (110) is described in more detail with reference to FIG. 2.

[0041] Referring to FIG. 2, the data collection unit (110) may include an image data collection unit (111), an equipment data collection unit (112), a sensor data collection unit (113), and a worker data collection unit (114).

[0042] The above image data collection unit (111) can photograph equipment and workers within the industrial site using a preset image capturing device.

[0043] Here, the above-mentioned preset image capturing device may refer to an image capturing device such as a high-definition camera, a thermal imaging camera, or a motion detection camera.

[0044] The above equipment data collection unit (112) can collect operation data of a pre-configured equipment provided in the industrial site.

[0045] Here, the aforementioned pre-configured equipment refers to all equipment that operates by receiving power within an industrial site and may include conveyor belts, press equipment, automation equipment, etc.

[0046] The sensor data collection unit (113) can collect sensing data from environmental measurement sensors that sense temperature, humidity, vibration, noise, fine dust, harmful gas concentration, pressure, light intensity, and magnetic field, which are provided inside and outside the industrial site.

[0047] Here, the sensor data collection unit (113) may have the same sensor within a preset radius from the environment measurement sensor as a comparison sensor.

[0048] Here, the sensor data collection unit (113) is explained in more detail with reference to FIG. 3.

[0049] Referring to FIG. 3, the sensor data collection unit (113) may include a sensor information collection unit (113a), a sensor error measurement unit (113b), and a data transmission unit (113c).

[0050] The sensor information collection unit (113a) can collect the supply voltage, consumption current, and output voltage supplied to the environment measurement sensor.

[0051] More specifically, the sensor information collection unit (113a) can collect the supply voltage, consumption current, and output voltage supplied from the power source by further providing a voltage environment measurement sensor on the power supply line and output line of the environment measurement sensor.

[0052] At this time, the environment measurement sensor refers to the sensor mentioned in the sensor data collection unit (113), and an identical sensor (a sensor of the same model) within a preset radius from the environment measurement sensor may be provided as a comparison sensor.

[0053] Here, the preset radius is ideally located within a radius of 10 mm to 500 mm from the center position of the environment measurement sensor, and it can be determined that the closer it is to the environment measurement sensor, the higher the accuracy when compared.

[0054] The sensor error measurement unit (113b) can extract a comparison target rejection at a time when the rate of change of the measurement value of the sensing data exceeds a preset rate.

[0055] Here, the above preset ratio can be set to a ratio of 25% to 30%, and the setting can be changed from a minimum of 10% to a maximum of 90% depending on the type of environmental measurement sensor and the surrounding environment.

[0056] In addition, the method for calculating the rate of change of the measurement value of the sensing data is as follows: the measurement value T measured at time t2 within the time range of the sensing data collected by the sensor information collection unit (113a). t2 from t a The measured value T at the time interval t1 t1 T to the value after subtracting t1 The absolute value obtained by dividing the measured value and multiplying by 100 can be calculated as the rate of change of the measured value.

[0057] For example, if the time interval set by the user is 2 seconds, the total measurement time from the start of operation until termination of the temperature sensor, which is the current environment measurement sensor, is 2000 seconds, and the measured value at 1520 seconds is 21℃ and the measured value at 1518 seconds is 31.6℃, the rate of change in the measured value can be calculated as approximately 36.2% (100 * (21 - 28.6) / 21). In this case, if the ratio set by the user is 30%, 1518 seconds can be extracted as the time to be compared because the rate of change in the measured value exceeds 30%.

[0058] Here, the sensor error measuring unit (113b) can extract sensing data corresponding to n arbitrary times within the sensor data as the comparison target time when there is no time exceeding the preset ratio.

[0059] Here, the method for extracting n random times within the sensing data can be performed using a program that applies a programming language, and the programming language may include Python, Java, C, C++, JavaScript, Go, Ruby, Swift, Kotlin, PHP, C# (C Sharp), etc.

[0060] For example, to extract n random times from sensing data using Python, one can use the Random and Datetime modules to set the time range in which the sensor operates and then extract n random times, or use the NumPy module.

[0061] In addition, another method for randomly extracting n times from the sensing data is to apply the Fisher-Yates shuffle algorithm to the time range of the sensing data to randomly extract n times from the sensing data.

[0062] Meanwhile, among the n times extracted from the sensor error measurement unit (113b), n represents a natural number that may vary depending on the operating environment set by the user, and the meaning of the arbitrary time may be that it is a time extracted in a random format without regularity from the total time collected from the sensing data.

[0063] The data transmission unit (113c) can periodically transmit the comparison target time of the sensor error measurement unit to the monitoring unit (140).

[0064] Additionally, the data transmission unit (113c) can transmit data including the supply voltage, consumption current, and output voltage of the environment measurement sensor collected by the sensor information collection unit (113a) to the monitoring unit (140).

[0065] Through the above process, the sensor data collection unit (113) can transmit the sensing data, which is a set of measurement values ​​collected from multiple environmental measurement sensors, to a user terminal, and the user can classify the section in which the rate of change of the sensing data does not exceed a preset rate of change as a 'standard state section' corresponding to the time interval set by the user, and the section in which the rate of change of the sensing data exceeds the preset rate of change as an 'abnormal state section'.

[0066] Here, if the 'abnormal state interval' of the sensing data collected by the sensor data collection unit (113) using the input device including the user terminal exceeds the maximum time range of the 'abnormal state interval' set by the user, or if the number of times the 'abnormal state interval' occurs exceeds a certain number, the process stoppage, system temporary suspension, power cutoff, etc. of the process where the sensor is located can be set to be automatically performed. Through this, rapid response to the occurrence of an abnormal phenomenon can be performed.

[0067] The above worker data collection unit (114) can collect health data including the body temperature, heart rate, blood pressure, blood sugar, and oxygen saturation of the worker wearing the device from a preset terminal worn by the worker.

[0068] Meanwhile, the sensor data collection unit (113) transmits the current environmental state diagnosed using sensing data collected from multiple environmental measurement sensors provided in the AI-based industrial safety management system (100) to the monitoring unit (140) and a preset user terminal, thereby enabling the internal environment to be identified from the outside.

[0069] In addition, depending on the user settings, if the measured value deviates from the upper and lower limits of the relative humidity measurement set by the user, it may also be classified as an 'abnormal state range'.

[0070] For example, when the average value of relative humidity measured using a humidity environment measurement sensor located in the first facility within the industrial site is 90~95% and the user sets the sensing data change rate setting value to 20%, and the relative humidity measured from 0 seconds to 46,800 seconds is classified into a 'standard state section' or an 'abnormal state section', the section where the rate of change of relative humidity per second exceeds 20%, which is the same value as the sensing data change rate setting value, can be classified as an 'abnormal state section', and the section where the rate of change per second is less than 20% can be classified as a 'standard state section'.

[0071] Meanwhile, under the conditions described above, when the user sets the upper limit of relative humidity to 95% and the lower limit to 90%, and the relative humidity measured at 149 seconds is 72%, the relative humidity measured at 150 seconds is 91%, and the relative humidity measured at 151 seconds is 97%, the rate of change for the 149-150 second interval can be calculated as approximately 26.38% (100 * (91-72) / 72). At this time, since the rate of change per second in the 149-150 second interval exceeds 20%, it can be classified as an 'abnormal state interval'. In addition, the rate of change per second for the above 150-151 second interval is calculated to be approximately 6.59% (100 * (97-91) / 91), so the rate of change per second does not exceed 20%, but because the relative humidity measured at 151 seconds exceeds the upper limit of 95% set by the user, it can be classified as an 'abnormal state interval'.

[0072] Referring again to FIG. 1, the risk judgment unit (120) can determine whether there is a risk to equipment and workers using industrial site data collected by the data collection unit (110).

[0073] Here, the risk judgment unit (120) can perform worker behavior detection using the image data collected by the data collection unit (110) and, if it determines an abnormality, transmit a notification to the pre-configured terminal of the worker, transmit a worker tracking request notification to the monitoring unit (140), and perform part replacement time prediction and fault diagnosis using the equipment data collected from the equipment.

[0074] The above equipment control unit (130) can control the equipment by receiving a preset operation signal generated to be transmitted to the equipment determined to be abnormal by the above risk judgment unit (120).

[0075] The monitoring unit (140) can collect the industrial site data and the control signal transmitted from the equipment control unit (130) and transmit it to a preset user terminal.

[0076] Here, the pre-configured user terminal may refer to an electronic terminal capable of wired or wireless communication and input / output signal transmission, such as a PC, laptop, tablet PC, or smartphone.

[0077] In addition, the monitoring unit (140) uses the sensing data collected from the environment measurement sensor and the comparison sensor and the comparison target time to determine the sensor failure probability (S) according to the following [Equation 1]. TR ) can be produced.

[0078] [Mathematical Formula 1]

[0079]

[0080] (Here, S TR is the sensor failure probability, T ms1 is the measured value of the environmental measurement sensor at the first time point, T ms2 is the measured value of the environmental measurement sensor at the second time point, T ms3 is the measured value of the environmental measurement sensor at the third viewpoint, T msn is the measurement value of the environment measurement sensor at the nth time point, T cst1is the measurement value of the comparison sensor at the first time point, T cst2 is the measurement value of the comparison sensor at the second time point, T cs3 is the measurement value of the comparison sensor at the third viewpoint, T csn is the measurement value of the comparison sensor at the nth time point, T msav is the average measurement value of the environmental measurement sensor, and SE represents the efficiency deviation of the environmental measurement sensor.

[0081] At this time, the average efficiency (SE) of the above-mentioned environment measurement sensor av ) can be calculated according to the following [Mathematical Formula 2].

[0082] [Mathematical Formula 2]

[0083]

[0084] (Here, SE av is the average efficiency of the environmental measurement sensor, SE i represents the initial efficiency of the environmental measurement sensor)

[0085] The average efficiency of the above-mentioned environmental measurement sensor can be calculated according to the following [Equation 3].

[0086] [Mathematical Formula 3]

[0087]

[0088] (Here, W IN1 is the power supplied to the environmental measurement sensor at the first time point, W OUT1 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the first time point, W IN2 is the power supplied to the environmental measurement sensor at the second viewpoint, W OUT2 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the second viewpoint, W INn W is the power supplied to the environment measurement sensor at the n-th time point. OUTn represents the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the nth time point.

[0089] At this time, the n times mentioned in [Equation 1] to [Equation 3] above may refer to a number of times or randomly extracted times in which the rate of change of the measurement value of the sensing data extracted from the sensor error measurement unit (113b) exceeds a preset rate.

[0090] For example, in a state where a first temperature sensor provided as an environmental measurement sensor in the equipment and a second temperature sensor of the same model as the first temperature sensor provided as a comparison sensor within a 10mm radius are provided, the input / output voltage of the first temperature sensor and the current consumed inside the environmental measurement sensor can be displayed. At this time, if three times are extracted from the environmental measurement sensor, and the voltage supplied during the extracted first to third times is 4.9V, 5V, and 5V, and it is measured that the environmental measurement sensor consumes currents of 1mA, 0.9mA, and 1.1mA at the corresponding times and outputs 10mV, 9mV, and 11mV, then the input power (W at the first time) supplied to the environmental measurement sensor at the first time IN1 ) is 0.0049W(4.9V*0.001A), input power at the second time point (W IN2 ) is 0.0045W(5V*0.0009A), input power at the third time point (W IN3 ) can be calculated as 0.0055W (5V * 0.0011A). In addition, the output power (W at the first time point) output from the environment measurement sensor is out1 ) is 0.00001W(0.01V*0.001A), output power at the second time point (W OUT2- ) is 0.000081W(0.009V*0.0009A), output power at the third time point (W OUT3 ) can be calculated as 0.0000121W (0.011V * 0.0011A). Using this, the average efficiency (SE) of the above-mentioned environment measurement sensor av) can be calculated as 0.2(100 / 3*(0.00001 / 0.0049+0.000081 / 0.0045+0.0000121 / 0.0055)). At this time, the initial efficiency (SE provided by the manufacturer of the environmental measurement sensor being measured i If ) is 0.2%, the efficiency deviation (SE) of the environmental measurement sensor can be calculated as 0. Here, the average efficiency (SE) of the environmental measurement sensor av A lower output value indicates a more efficient sensor, and it can be determined that the higher the output value compared to the initial efficiency—which is calculated using measurements collected from the manufacturer or initially taken—the higher the probability of failure in the sensor's electrical domain. Meanwhile, the initial efficiency (SE) of the above-mentioned environment measurement sensor i If ) is not provided, the efficiency measured before being installed in the above equipment is the initial efficiency (SE i It can be set to ) and input into the monitoring unit (140).

[0091] Additionally, in an environment where the equipment is set to maintain a temperature of 60 degrees, sensing data resulting from the operation of the first temperature sensor can be collected from 0 seconds to 28,800 seconds. At this time, if the average measurement value collected from the first temperature sensor is 60.5℃ and three random times are extracted from the sensing data of the first temperature sensor using a sensing data processing program, the measurement values ​​at each time point may be extracted as 59℃ (485 seconds), 65℃ (1893 seconds), and 71℃ (2021 seconds). Here, when the measurement values ​​measured at the same time point and each time point from the second temperature sensor operating simultaneously with the first temperature sensor are 60.1℃ (485 seconds), 62.5℃ (1893 seconds), and 53℃ (2021 seconds), the failure probability (S) of the first temperature sensor TR ) can be calculated. That is, the failure probability (S) of the first temperature sensor above TR) can be calculated as approximately 3.19%(Min((|59-60.1|+|65-62.5|+|51-73|) / (3*60.5)*100+0), 100).

[0092] As another example, the monitoring unit (140) has a sensor failure probability (S) according to the following [Equation 1-2] which reflects a weighting factor based on the distance between the environment measurement sensor and the comparison sensor. TR ) can be produced.

[0093] [Mathematical Formula 1-2]

[0094]

[0095] (Here, D v )

[0096] At this time, the distance value weight (D-) of the environment measurement sensor and the comparison sensor is at this time. v ) is applied in correspondence with weights set by the user, and the user can freely change and apply the said weights.

[0097] For example, when the environment measurement sensor and the comparison sensor are provided in the position moving unit (123) as temperature environment measurement sensors of the same model, the distance weight (D) between the environment measurement sensor and the comparison sensor according to the distance (D) between the environment measurement sensor and the comparison sensor. v ) can be calculated according to [Table 1] below.

[0098] [Table 1]

[0099]

[0100] At this time, the distance weight (D v In the case of the same conditions as the above embodiment that do not reflect ), the distance weight (D v Failure probability of the first temperature sensor (S) reflecting ) TR ) can be calculated as approximately 3.19%(Min((|59-60.1|+|65-62.5|+|51-73|) / (3*60.5)*1*100+0), 100).

[0101] Through the above process, the monitoring unit (140) uses the sensing data collected from the sensor data collection unit (113) to determine the failure probability (S) of the sensor. TR ) can be produced.

[0102] In addition, the failure probability (S) calculated through the monitoring unit (140) TR If ) exceeds a preset ratio, a sensor failure notification, replacement notification, etc. can be transmitted to the preset user terminal.

[0103] Here, the above failure probability (S TR The preset ratio of ) can be set to 40% by default, and the user can reset the ratio to between 0 and 99% depending on the type and sensitivity of the sensor.

[0104] Through the above process, the AI-based industrial safety management system (100) can collect data from equipment, sensors, and wearable terminals worn by workers installed in the industrial site, use this data to predict equipment failures and calculate scheduled replacement dates for parts, and provide additional sensors of the same specifications to check and verify errors in the sensors themselves. Through this, even if a sensor fails, the failure can be determined, and through AI-based image data processing of the equipment temperature and worker movements via a video recording device, monitoring and notification measures can be performed when abnormal data or data exceeding a standard value is detected, thereby preventing safety accidents in advance. In addition, the probability of sensor failure can be calculated and quantified using a proprietary mathematical formula by utilizing the data collection unit (110) and the monitoring unit (140).

[0105] According to one embodiment of the present invention, by collecting multiple sensing data from equipment, site spaces, and workers within an industrial site and determining whether there is an abnormality, it is possible to prevent potential safety accidents occurring within the industrial site in advance.

[0106] As described above, although an embodiment of the present invention has been explained by limited embodiments and drawings, the embodiment of the present invention is not limited to the embodiments described above, and various modifications and variations are possible from this description by those skilled in the art to which the present invention pertains. Accordingly, an embodiment of the present invention should be understood only by the claims described below, and all equivalent or analogous variations thereof shall be considered to be within the scope of the inventive concept. Explanation of the symbols

[0108] 100 : AI-based industrial safety management system 110 : Data Collection Unit 111 : Image data collection unit 112 : Equipment Data Collection Unit 113: Sensor data collection unit 113a : Sensor information collection unit 113b : Sensor error measurement unit 113c : Data transmission section 114 : Worker Data Collection Department 120 : Risk Assessment Section 130 : Equipment Control Unit 140 : Monitoring section

Claims

Claim 1 A data collection unit provided within an industrial site and collecting industrial site data using wired and wireless communication; a risk determination unit that determines whether there is a risk to equipment and workers using industrial site data collected by the data collection unit; an equipment control unit that controls said equipment by receiving a preset operation signal generated to be transmitted to equipment judged as abnormal by the risk determination unit; and a monitoring unit that collects said industrial site data and control signals transmitted by said equipment control unit and transmits them to a preset user terminal; wherein the data collection unit comprises: a video data collection unit that photographs equipment and workers within said industrial site using a preset video recording device; an equipment data collection unit that collects operation data of preset equipment provided within said industrial site; and a sensor data collection unit that collects sensing data from environmental measurement sensors provided inside and outside said industrial site that sense temperature, humidity, vibration, noise, fine dust, hazardous gas concentration, pressure, light intensity, and magnetic field. An AI-based industrial safety management system comprising: a worker data collection unit that collects health data including the body temperature, heart rate, blood pressure, blood sugar, and oxygen saturation of a worker from a preset terminal worn by a worker; wherein the sensor data collection unit comprises: a sensor information collection unit that collects supply voltage, consumption current, and output voltage supplied to the environment measurement sensor; a sensor error measurement unit that extracts a comparison target time, which is a time when the rate of change of the measurement value of the sensing data exceeds a preset ratio; and a data transmission unit that periodically transmits the comparison target time of the sensor error measurement unit to the monitoring unit; wherein the sensor error measurement unit is characterized by extracting sensing data corresponding to n arbitrary times within the sensing data as a comparison target time when there is no time exceeding the preset ratio. Claim 2 delete Claim 3 An AI-based industrial safety management system according to claim 1, wherein the risk judgment unit performs worker behavior detection using image data collected by the data collection unit and, if an abnormality is determined, transmits a notification to a pre-configured terminal of the worker, transmits a worker tracking request notification to the monitoring unit, and performs part replacement timing prediction and fault diagnosis using equipment data collected from the equipment. Claim 4 delete Claim 5 In claim 1, the sensor data collection unit comprises an identical sensor within a preset radius from the environment measurement sensor as a comparison sensor, and the monitoring unit uses the sensing data collected from the environment measurement sensor and the comparison sensor and the comparison target time to determine the sensor failure probability (S) according to the following [Equation 1]. TR It produces ),[Mathematical Formula 1] (Here, S TR is the sensor failure probability, T ms1 is the measured value of the environmental measurement sensor at the first time point, T ms2 is the measured value of the environmental measurement sensor at the second time point, T ms3 is the measured value of the environmental measurement sensor at the third viewpoint, T msn is the measurement value of the environment measurement sensor at the nth time point, T cst1 is the measurement value of the comparison sensor at the first time point, T cst2 is the measurement value of the comparison sensor at the second time point, T cs3 is the measurement value of the comparison sensor at the third viewpoint, T csn is the measurement value of the comparison sensor at the nth time point, T msav is the average measurement value of the environmental measurement sensor, and SE represents the efficiency deviation of the environmental measurement sensor.) The efficiency deviation (SE) of the above environmental measurement sensor is calculated according to the following [Equation 2], and [Equation 2] (Here, SE av is the average efficiency of the environmental measurement sensor, SE i (where represents the initial efficiency of the environmental measurement sensor) An AI-based industrial safety management system characterized by calculating the average efficiency of the above environmental measurement sensor according to the following [Equation 3]. [Equation 3] (Here, W IN1 is the power supplied to the environmental measurement sensor at the first time point, W OUT1 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the first time point, W IN2 is the power supplied to the environmental measurement sensor at the second viewpoint, W OUT2 ε is the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the second viewpoint, W INn W is the power supplied to the environment measurement sensor at the n-th time point. OUTn represents the power measured when the environmental measurement sensor outputs the sensed value as an electrical signal at the nth time point.

Images