Artificial intelligence intelligent safety management method using three-stage index-based algorithm

Abstract

The present invention relates to an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm, comprising: a step of refining sensing data from a plurality of intelligent sensors each including at least one of a displacement gauge, an inclinometer, a settlement gauge, a crack gauge, and an accelerometer, and transmitting the refined data; a step of receiving and transmitting the refined data transmitted through the plurality of intelligent sensors at a plurality of smart gateways; a step of determining whether there is a precursor to collapse by analyzing displacement, displacement velocity, and displacement acceleration for each sensor using the refined data received through the plurality of smart gateways at a cloud server; a step of calculating a Predictive Behavioral Risk Index (PBRI) for each sensor at the cloud server based on a state value, the displacement velocity, the displacement acceleration, pattern anomaly, external influence, and a prediction value; and a step of calculating a Gateway-level Risk Index (GWRI) for each gateway by analyzing the direction of increase / decrease, the rate of increase / decrease, and the correlation with the refined data and the Predictive Behavioral Risk Index (PBRI) at the cloud server. A step of generating a set of coefficients of a non-linear reference function for applying dynamic thresholds for each sensor using the Predictive Risk Behavior Index (PBRI) at the cloud server and transmitting it to the smart gateway; a step of receiving the set of coefficients of the non-linear reference function at the plurality of smart gateways and transmitting them to each of the plurality of intelligent sensors; a step of receiving the set of coefficients of the non-linear reference function for each sensor at the plurality of intelligent sensors, reconstructing the non-linear reference function, and then calculating the sensor data to determine whether a collapse warning is required; and a step of calculating a Calculated Priority Risk Index (CPRI) using the refined data, the Gateway-specific Zone Risk Index (GWRI), and the Predictive Risk Behavior Index (PBRI) at the cloud server to determine the overall site risk.By including this, not only can structural safety management be performed effectively, but false alarms can be drastically reduced, precursors to structural collapse can be detected early, and effective safety management can be carried out through the latest analysis information stored and managed even when external communication is cut off.

Description

Technology Field

[0001] The present invention relates to an intelligent safety management method using a three-stage artificial intelligence index-based algorithm, which is based on a closed-loop structure in which a plurality of intelligent sensors, a plurality of smart gateways, and a cloud server are interconnected in three layers, and provides a three-measurement circulation hardware structure in which data measured and refined by a plurality of intelligent sensors is transmitted up to the cloud server through the smart gateway, non-linear reference function coefficients generated by the cloud server for each sensor are transmitted down individually to each sensor through the smart gateway, and the next measurement data of each sensor to which the updated polynomial equation is applied is transmitted up again to the cloud server, thereby enabling effective safety management of a structure, significantly reducing false alarms, detecting early signs of structural collapse, and performing effective safety management through the latest analysis information stored and managed even when external communication is cut off. Background Technology

[0002] As is well known, structural collapse and ground subsidence in underground excavations, slopes, tunnels, and bridges at construction sites are serious accidents that cause massive loss of life and property damage, and safety monitoring systems are widely operated to prevent this.

[0003] Conventional safety measurement monitoring systems have the following fundamental limitations.

[0004] First, as a limitation of the simple data collection structure, conventional technology is provided with a unidirectional open-loop structure in which the sensor transmits measurement data to a server and compares it with a pre-set fixed threshold. Since the server's analysis results do not automatically follow changes in the sensor's alarm criteria, manual intervention by an administrator is essential, and there is a problem in that it is difficult to respond in real time to changes in field conditions.

[0005] Second, regarding the specific limitations of conventional technology, Korean Registered Patent No. 10-1431237 (System for Detecting Abnormal Behavior of Structures and Evaluating Safety) discloses a technology for automatically updating management thresholds based on probability distributions; however, this maintains a single constant threshold value through ex post updates based on cumulative history and does not include a gateway layer or an individual polynomial reference function structure for each sensor. Additionally, Korean Registered Patent No. 10-2073323 (AI-based Structural Health Management System) discloses time-series prediction using AI, but it does not include a closed-loop feedback structure that converts the prediction results into a polynomial reference function and automatically transmits them to the sensors.

[0006] Accordingly, there is a need for the development of intelligent safety management technology based on a closed-loop structure in which multiple intelligent sensors, smart gateways, and cloud servers are interconnected in three layers. Prior art literature

[0007] 1. Korean Registered Patent No. 10-1431237 (Registered Aug. 11, 2014) 2. Korean Registered Patent No. 10-2073323 (Registered Jan. 23, 2020) The problem to be solved

[0008] The present invention aims to provide an intelligent safety management method using a 3-stage artificial intelligence index-based algorithm, which is based on a closed-loop structure in which a plurality of intelligent sensors, a plurality of smart gateways, and a cloud server are interconnected in three layers, and provides a 3-measurement cyclic hardware structure in which data measured and refined by a plurality of intelligent sensors is transmitted up to the cloud server through the smart gateway, non-linear reference function coefficients generated by the cloud server for each sensor are transmitted down individually to each sensor through the smart gateway, and the next measurement data of each sensor to which the updated polynomial equation is applied is transmitted up again to the cloud server. This allows for effective safety management of a structure, significantly reduces false alarms, detects early signs of structural collapse, and enables effective safety management through the latest analysis information stored and managed even when external communication is cut off.

[0009] The purposes of the embodiments of the present invention are not limited to those mentioned above, and other unmentioned purposes will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0010] According to an embodiment of the present invention, the method comprises: a step of refining sensing data from a plurality of intelligent sensors each including at least one of a displacement gauge, an inclinometer, a settlement gauge, a crack gauge, and an accelerometer, and transmitting the refined data; a step of receiving and transmitting the refined data transmitted through the plurality of intelligent sensors at a plurality of smart gateways; a step of determining whether there is a precursor to collapse by analyzing displacement, displacement velocity, and displacement acceleration for each sensor using the refined data received through the plurality of smart gateways at a cloud server; a step of calculating a Predictive Behavioral Risk Index (PBRI) for each sensor at the cloud server based on a state value, the displacement velocity, the displacement acceleration, pattern anomaly, external influence, and a prediction value; and a step of calculating a Gateway-level Risk Index (GWRI) by analyzing the direction of increase / decrease, the rate of increase / decrease, and the correlation with the refined data and the Predictive Behavioral Risk Index (PBRI) at the cloud server. An intelligent safety management method using an artificial intelligence 3-stage index-based algorithm may be provided, comprising: a step of generating a set of coefficients of a non-linear reference function for applying dynamic thresholds for each sensor using the Predictive Risk Behavior Index (PBRI) at the cloud server and transmitting it to the smart gateway; a step of receiving the set of coefficients of the non-linear reference function at the plurality of smart gateways and transmitting them to each of the plurality of intelligent sensors; a step of receiving the set of coefficients of the non-linear reference function for each sensor at the plurality of intelligent sensors, reconstructing the non-linear reference function, and then calculating the sensor data to determine whether a collapse warning is required; and a step of calculating a Calculated Priority Risk Index (CPRI) using the refined data, the Gateway-specific Zone Risk Index (GWRI), and the Predictive Risk Behavior Index (PBRI) at the cloud server to determine the overall risk of the site.

[0011] In addition, according to an embodiment of the present invention, the step of transmitting the refined data may be provided as an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm that obtains the refined data by multi-sampling the sensing data from the plurality of intelligent sensors, and then performing outlier removal, noise filtering, and representative value calculation.

[0012] In addition, according to an embodiment of the present invention, the step of determining whether there is a precursor to collapse may be provided by an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm, wherein the step of determining whether there is a precursor to collapse is performed by performing a 3-stage measurement analysis of the displacement, displacement velocity, and displacement acceleration for each sensor on the cloud server, and determining that there is a precursor to collapse if the displacement acceleration exceeds a preset acceleration threshold.

[0013] In addition, according to an embodiment of the present invention, the step of determining whether there is a precursor to collapse may be provided by an intelligent safety management method using an artificial intelligence three-stage index-based algorithm, wherein the step of determining the precursor to collapse involves comparing and analyzing the absolute value of the displacement and a preset management standard value through a first-stage displacement analysis on the cloud server, analyzing the rate of change and increasing trend per unit time through a second-stage displacement velocity analysis, and comparing and analyzing the displacement acceleration and a preset acceleration threshold value through a third-stage displacement acceleration analysis.

[0014] In addition, according to an embodiment of the present invention, the step of calculating the Predictive Risk Behavior Index (PBRI) may be calculated by quantifying the Predictive Risk Behavior Index (PBRI) as 0-100 on the cloud server, and an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm may be provided in which an exponential non-linear weight is applied to the Predictive Risk Behavior Index (PBRI) when the displacement acceleration exceeds the preset acceleration threshold alone.

[0015] In addition, according to an embodiment of the present invention, the step of calculating the gateway-specific zone risk index (GWRI) may be provided as an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm that calculates the gateway-specific zone risk index (GWRI) in the range of 0-100 based on the predictive risk behavior index (PBRI) for each smart gateway, the maximum value, the average value, the increasing trend, the ratio of sensors exceeding a threshold value, and the correlation-based concentration in the cloud server.

[0016] In addition, according to an embodiment of the present invention, the step of determining whether a collapse warning is given may be provided by an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm, wherein the step of determining whether a collapse warning is given is to receive a set of coefficients of the non-linear reference function from the plurality of intelligent sensors, reconstruct the non-linear reference function, and then apply the next sensing data provided from the plurality of intelligent sensors to determine whether a collapse warning is given.

[0017] In addition, according to an embodiment of the present invention, the step of determining the total site risk may be provided by an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm that calculates the Comprehensive Risk Prediction Index (CPRI) by analyzing the correlation between the Predictive Risk Behavior Index (PBRI) and the Zone Risk Index (GWRI) for each smart gateway and the direction of increase / decrease, the speed of increase / decrease, and the area of ​​increase / decrease in the cloud server. Effects of the invention

[0018] The present invention is based on a closed-loop structure in which a plurality of intelligent sensors, a plurality of smart gateways, and a cloud server are interconnected in three layers, and provides a three-measurement cyclic hardware structure in which data measured and refined by a plurality of intelligent sensors is transmitted up to the cloud server through the smart gateway, non-linear reference function coefficients for each sensor generated by the cloud server are transmitted down individually to each sensor through the smart gateway, and the next measurement data of each sensor to which the updated polynomial equation is applied is transmitted up again to the cloud server. By providing this structure, not only can safety management of the structure be effectively performed, but false alarms can be drastically reduced, signs of structural collapse can be detected early, and effective safety management can be performed through the latest analysis information stored and managed even when external communication is cut off. Brief explanation of the drawing

[0019] FIG. 1 is a flowchart illustrating an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention, and FIG. 2 is a block diagram of an intelligent safety management system that performs an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention, and FIGS. 3 to 7 are drawings for explaining the detailed technology of an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention. Specific details for implementing the invention

[0020] The advantages and features of the embodiments 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. However, the present invention is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Throughout the specification, the same reference numerals refer to the same components.

[0021] In describing the embodiments of the present invention, specific descriptions of known functions or configurations will be omitted if it is determined that such detailed descriptions could unnecessarily obscure the essence of the invention. Furthermore, the terms described below are defined in consideration of their functions in the embodiments of the present invention, and these definitions may vary depending on the intentions or practices of the user or operator. Therefore, such definitions should be based on the content throughout this specification.

[0022] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0023] FIG. 1 is a flowchart illustrating an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention, FIG. 2 is a block diagram of an intelligent safety management system performing an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention, and FIG. 3 to 7 are drawings for explaining the detailed technology of an intelligent safety management process using an artificial intelligence 3-stage index-based algorithm according to an embodiment of the present invention.

[0024] Referring to FIGS. 1 to 7, sensing data can be refined and refined data transmitted from a plurality of intelligent sensors (100) each including at least one of a displacement sensor, an inclinometer, a settlement sensor, a crack sensor, and an accelerometer (step 10).

[0025] In the step (10) of transmitting the refined data, the refined data can be obtained by multi-sampling the sensing data from a plurality of intelligent sensors (100), and then performing outlier removal, noise filtering, and representative value calculation.

[0026] For example, a plurality of intelligent sensors (100) may include at least one of a displacement sensor, an inclinometer, a settlement sensor, a crack sensor, and an accelerometer, and may be installed in multiple units in underground excavation, slopes, tunnels, bridges, etc. at construction sites as a device that refines sensing data and transmits refined data, and may be responsible for the Edge AI Layer, and each may include a first communication module (110), a measurement unit (120), a first processor unit (130), a storage unit (140), etc.

[0027] Here, the first communication module (110) is a component that includes a wireless communication module and connects to a wireless communication network (400), and performs wireless communication between a plurality of intelligent sensors (100) and a plurality of smart gateways (200). For example, it may include a Bluetooth wireless communication module, a BLE wireless communication module, a Wi-Fi wireless communication module, a LoRa wireless communication module, a Zigbee wireless communication module, etc., to perform wireless communication between a plurality of intelligent sensors (100) and a plurality of smart gateways (200).

[0028] This first communication module (110) can transmit refined data for each sensor to a corresponding smart gateway among a plurality of smart gateways (200) via a wireless communication network (400) under the control of the first processor (130).

[0029] The measuring unit (120) is a component that provides sensing data detected according to a preset frequency using at least one of a displacement gauge, an inclinometer, a settlement gauge, a crack gauge, and an accelerometer. At least one of a displacement gauge, an inclinometer, a settlement gauge, a crack gauge, and an accelerometer is installed in multiple numbers (e.g., individually or in parallel) in an underground excavation, slope, tunnel, bridge, etc., at a construction site, and the corresponding physical quantity (e.g., displacement value, inclination value, settlement value, crack value, acceleration value, etc.) at each installation location is sensed according to a preset frequency and the sensing data can be transmitted to the first processor (130).

[0030] Here, the preset frequency can be set to multiple times within a unit time, preferably set to 10 times or more per second, and more preferably set to 50 times or more per second.

[0031] The first processor unit (130) is a component that controls the first communication module unit (110) to transmit the acquired refined data to a smart gateway (200), and performs sensing control of the measurement unit (120), multi-samples the sensing data, performs outlier removal, noise filtering, and calculates a representative value to obtain refined data, and transmits the obtained refined data to a smart gateway (200). It can control the first communication module unit (110) to multi-sample the sensing data transmitted from the measurement unit (120) according to a preset frequency, and to transmit the refined data to a plurality of smart gateways (200) by performing outlier removal, noise filtering, and calculation of a representative value on the multi-sampled data.

[0032] For example, the first processor unit (130) can continuously collect multiple sensing data according to a preset frequency for the same physical quantity, then apply statistical techniques based on average values, standard deviations, etc. to remove outliers (i.e., abnormally large or small values) or outliers that exceed preset physical limits, perform noise filtering using a moving average filter, a low-pass filter, a median filter, etc., and calculate a representative value by giving greater weight to recent data or highly reliable data in the average value or median value, and obtain refined data with the representative value.

[0033] The storage unit (140) stores sensing data, refined data, etc., and can store various data and information obtained from each of the multiple intelligent sensors (100) by including non-volatile memory (e.g., flash memory, ROM, EEPROM, FRAM, MRAM, etc.).

[0034] Meanwhile, the wireless communication network (400) may include, for example, a short-range wireless communication network, a broadband wireless communication network, etc. In the case of a short-range wireless communication network, it may include wireless communication methods such as Bluetooth, BLE (Bluetooth Low Energy), Wi-Fi, LoRa, and ZigBee, and in the case of a broadband wireless communication network, it may include synchronous and asynchronous methods, and may include CDMA, GSM, LTE and all future mobile communication methods, and may include a network capable of performing wireless communication using an IoT-based distributed communication method utilizing a lightweight message-based publish and subscribe structure.

[0035] Through this wireless communication network (400), wireless communication between a plurality of intelligent sensors (100), a plurality of smart gateways (200), and a cloud server (300) can be supported.

[0036] And, the refined data transmitted through each of the multiple intelligent sensors (100) at the multiple smart gateways (200) can be received and transmitted (step 20)

[0037] For example, a plurality of smart gateways (200) are devices that receive and transmit refined data transmitted through a plurality of intelligent sensors (100), and each performs the gateway layer, and may include a second communication module (210), a second processor (220), a temporary storage unit (230), etc.

[0038] Here, the second communication module (210) is a component that includes a wireless communication module and is connected to a wireless communication network (400), and performs wireless communication between a plurality of intelligent sensors (100) and a plurality of smart gateways (200), and wireless communication between a plurality of smart gateways (200) and a cloud server (300). For example, it can perform wireless communication between a plurality of intelligent sensors (100) and a plurality of smart gateways (200) by including a Bluetooth wireless communication module, a BLE wireless communication module, a Wi-Fi wireless communication module, a LoRa wireless communication module, a Zigbee wireless communication module, etc., and can perform wireless communication between a plurality of smart gateways (200) and a cloud server (300) by including a CDMA wireless communication module, a GSM wireless communication module, an LTE wireless communication module, etc.

[0039] This second communication module (210) can receive refined data for each sensor transmitted from a plurality of intelligent sensors (100) via a wireless communication network (400) under the control of the second processor (220) and transmit it to the second processor (220), and can transmit the refined data for each sensor to a cloud server (300) via a wireless communication network (400) under the control of the second processor (220).

[0040] The second processor unit (220) can control the second communication module unit (210) to collect refined data received from a plurality of intelligent sensors (100) and transmit it to a cloud server (300).

[0041] The temporary storage unit (230) is a component that temporarily stores refined data, etc., and can temporarily store various data and information obtained from each of the multiple smart gateways (200) by including volatile memory (e.g., RAM, cache memory, register, etc.).

[0042] Next, the displacement, displacement velocity, and displacement acceleration for each sensor can be analyzed using the refined data received from each smart gateway (200) in the cloud server (300) to determine whether there is a precursor to collapse (step 30).

[0043] In the step (30) of determining whether there is a precursor to collapse, a three-stage measurement analysis of displacement, displacement velocity, and displacement acceleration is performed for each sensor on the cloud server (300), and if the displacement acceleration exceeds a preset acceleration threshold, it can be determined that there is a precursor to collapse.

[0044] In addition, in the step (30) of determining whether there is a precursor to collapse, the cloud server (300) can compare and analyze the absolute value of the displacement and the pre-set management standard value through the first stage displacement analysis, analyze the rate of change and increasing trend per unit time through the second stage displacement velocity analysis, and compare and analyze the displacement acceleration and the pre-set acceleration threshold value through the third stage displacement acceleration analysis.

[0045] For example, the cloud server (300) is responsible for the Cloud AI Layer and may include a third communication module (310), a collapse judgment unit (320), a PBRI calculation unit (330), a GWRI calculation unit (340), a reference function generation unit (350), a CPRI calculation unit (360), etc.

[0046] Here, the third communication module (310) is a component that includes a wireless communication module and connects to a wireless communication network (400), and performs wireless communication between a plurality of smart gateways (200) and a cloud server (300). For example, it may include a CDMA wireless communication module, a GSM wireless communication module, an LTE wireless communication module, etc., to perform wireless communication between a plurality of smart gateways (200) and a cloud server (300).

[0047] This third communication module (310) can receive refined data for each sensor transmitted via a wireless communication network (400) from a plurality of smart gateways (200) and transmit it to a collapse determination unit (320).

[0048] And, the collapse judgment unit (320) is a component that performs a three-stage measurement analysis of displacement, displacement velocity, and displacement acceleration for each sensor using refined data transmitted through the third communication module (310), and determines that it is a precursor to collapse if the displacement acceleration exceeds a preset acceleration threshold. It can compare and analyze the absolute value of the displacement with a preset management standard value through the first stage displacement analysis, analyze the rate of change and increasing trend per unit time through the second stage displacement velocity analysis, and compare and analyze the displacement acceleration with a preset acceleration threshold through the third stage displacement acceleration analysis.

[0049] For example, in the first stage displacement analysis, the currently measured absolute displacement value and the pre-set management standard value can be compared and analyzed, in the second stage displacement velocity analysis, the displacement velocity can be calculated to analyze the rate of change and increasing trend per unit time, and in the third stage displacement acceleration analysis, the displacement acceleration can be calculated to compare and analyze the displacement acceleration with the pre-set acceleration threshold value.

[0050] This collapse judgment unit (320) can detect early signs of collapse through a three-stage hierarchical analysis of displacement (absolute value), displacement velocity, and displacement acceleration, and if the displacement acceleration exceeds a preset acceleration threshold through a three-stage displacement acceleration analysis, it can determine that it is a sign of collapse regardless of the displacement (absolute value) and control the transmission of the warning signal to multiple smart gateways (200) through the second communication module (310) so as to immediately issue a collapse warning. Of course, the collapse judgment unit (320) may analyze displacement, velocity, and acceleration separately, or analyze displacement and velocity, displacement and acceleration, and displacement, velocity, and acceleration separately.

[0051] Accordingly, the second processor unit (220) provided in the plurality of smart gateways (200) can receive a warning signal for a pre-collapse warning transmitted from the cloud server (300) via the wireless communication network (400) through the second communication module unit (210) and control it to transmit to the plurality of intelligent sensors (100), and can control the smart gateway to issue a pre-collapse warning visually and audibly through a separate warning device (e.g., speaker, warning light, etc.).

[0052] In addition, the first processor unit (130) provided in the plurality of intelligent sensors (100) can receive a warning signal for a pre-collapse warning transmitted from the plurality of smart gateways (200) through the wireless communication network (400) via the first communication module unit (110) and control the corresponding intelligent sensors (e.g., a plurality of intelligent sensors installed in the area managed by the corresponding smart gateway) to issue a pre-collapse warning visually and audibly through a separate warning device (e.g., a speaker, a warning light, etc.).

[0053] As described above, by using displacement acceleration as an independent alarm trigger, monitoring and management can be performed to issue an early warning for a collapse precursor in the collapse precursor stage, where the absolute value of the displacement has not yet exceeded a preset management threshold value.

[0054] Next, a Predictive Behavioral Risk Index (PBRI) can be calculated based on the state value, displacement velocity, displacement acceleration, pattern anomaly, external influence, and prediction value for each sensor in the cloud server (300) (step 40).

[0055] In the step (40) of calculating the above predictive risk behavior index (PBRI), the predictive risk behavior index (PBRI) is quantified from 0 to 100 in the cloud server (300), and an exponential non-linear weight may be applied to the predictive risk behavior index (PBRI) when the displacement acceleration exceeds a preset acceleration threshold alone.

[0056] For example, the PBRI calculation unit (330) provided in the cloud server (300) is a component that calculates the predictive risk behavior index (PBRI) on a scale of 0 to 100 based on the state value, displacement velocity, displacement acceleration, pattern anomaly, external influence, and prediction value for each sensor, and can assign an exponential non-linear weight to the predictive risk behavior index (PBRI) when the displacement acceleration exceeds a preset acceleration threshold value alone.

[0057] This PBRI calculation unit (330) can quantify the Predictive Risk Behavior Index (PBRI) from 0 to 100 based on the state value (S), displacement velocity (V), displacement acceleration (A), pattern anomaly (P), external influence (E), and prediction value (F) for each intelligent sensor. The state value (S) represents a normalized score of 0 to 100 based on the current absolute displacement value, the displacement velocity (V) represents the rate of change in displacement per unit time, and can assign weights in a non-linear increasing trend, the displacement acceleration (A) represents the second derivative value, and can assign exponential non-linear weights to the Predictive Risk Behavior Index (PBRI) when it is exceeded alone, the pattern anomaly (P) represents the degree of deviation from the normal range of the time series pattern, allowing for early detection of abnormal patterns, and the external influence (E) can reflect process variables such as excavation depth, bracing stage, rainfall amount, and temperature. The predicted value (F) can be used for preemptive standard updates through future state prediction by time series AI.

[0058] Here, the Predictive Risk Behavior Index (PBRI) has a range of 0 to 100, and a higher value indicates a dangerous state. When the displacement acceleration (A) exceeds a preset acceleration threshold, an exponential non-linear weight is applied to the Predictive Risk Behavior Index (PBRI) to manage the risk of collapse in advance.

[0059] The PBRI calculation unit (330) described above can analyze the predictive risk behavior index (PBRI) based on the degree of dispersion of values ​​(e.g., standard deviation and variance) in a time concept by utilizing the characteristics of change and change value according to time intervals, and such PBRI calculation can be performed through standard deviation and variance based on time using (change value + speed + acceleration).

[0060] Meanwhile, the cloud server (300) can calculate the Gateway-level Risk Index (GWRI) by analyzing the refined data and the predictive risk behavior index (PBRI), the direction of increase / decrease, the speed of increase / decrease, and the correlation (step 50).

[0061] In the step (50) of calculating the zone risk index (GWRI) for each gateway, the zone risk index (GWRI) for each gateway can be calculated in the range of 0-100 based on the predictive risk behavior index (PBRI) for each smart gateway, the maximum value, the average value, the increasing trend, the ratio of sensors exceeding a threshold, and the correlation-based concentration at the cloud server (300).

[0062] For example, the GWRI calculation unit (340) provided in the cloud server (300) is a component that calculates the zone risk index (GWRI) for each gateway by analyzing refined data, the predictive risk behavior index (PBRI), the direction of increase / decrease, the rate of increase / decrease, and the correlation based on dynamic thresholds, and can calculate the zone risk index (GWRI) for each gateway in the range of 0-100 based on the predictive risk behavior index (PBRI) for each smart gateway, the maximum value, the average value, the increasing trend, the ratio of sensors exceeding the threshold, and the correlation-based concentration.

[0063] The GWRI calculation unit (340) can calculate the gateway-specific zone risk index (GWRI) by combining refined data and predictive risk behavior index (PBRI) of multiple intelligent sensors (100) within the zone managed by each smart gateway (200). It can also quantify the overall risk level of the zone by analyzing the direction of increase / decrease, the rate of increase / decrease, and the correlation between the predictive risk behavior index (PBRI) and refined data corresponding to multiple intelligent sensors (100) within the zone.

[0064] The Gateway Area Risk Index (GWRI) described above can be calculated differently for each gateway based on various data obtained from multiple intelligent sensors (100) included in the area under its charge, and can also be calculated differently depending on the installation location of multiple intelligent sensors (100) installed in underground excavation, slopes, tunnels, bridges, etc. at the construction site.

[0065] Next, a set of coefficients of a non-linear reference function for applying dynamic thresholds to each sensor using a predictive risk behavior index (PBRI) can be generated in the cloud server (300) and transmitted to the smart gateway (200) (step 60).

[0066] For example, the reference function generation unit (350) provided in the cloud server (300) is a component that generates a set of coefficients of a non-linear reference function having displacement, displacement velocity, and displacement acceleration as independent variables for each sensor using a predictive risk behavior index (PBRI) and transmits them to a third communication module unit (310) to transmit them to each of the multiple smart gateways (200), and can generate a non-linear reference function having displacement (S), velocity (V), and acceleration (A) as independent variables to apply a dynamic threshold value as shown in Equation 1 below.

[0067]

[0068] Here, S_norm represents the normalized sensor measurement value (displacement (S)), V_norm represents the rate of change (velocity (V), slope), and A_norm represents acceleration (A), which is the change in the rate of change; it can be adjusted per site as w1+w2+w3=1, and the coefficients a0-a4 can be calculated and updated in real time based on the Predictive Risk Behavior Index (PBRI) and its prediction result (i.e., prediction value (F)), and each sensor has a different set of coefficients This can be generated, and the set of coefficients of the generated non-linear reference function can be transmitted to the third communication module (310) to be transmitted to each of the multiple smart gateways (200).

[0069] And, a set of coefficients of a non-linear reference function can be received from a plurality of smart gateways (200) and transmitted to a plurality of intelligent sensors (100) respectively (step 70).

[0070] For example, a second communication module (210) provided in a plurality of smart gateways (200) can receive a set of coefficients of a non-linear reference function for each sensor transmitted from a cloud server (300) via a wireless communication network (400) under the control of a second processor (220) and transmit it to a plurality of intelligent sensors (100) via a wireless communication network (400).

[0071] Here, the second processor unit (220) can receive a warning signal for a pre-collapse warning transmitted from the cloud server (300) via the wireless communication network (400) through the second communication module unit (210) and control it to transmit to a plurality of intelligent sensors (100), and can control the pre-collapse warning to be issued visually and audibly through a separate warning device (e.g., speaker, warning light, etc.) at the smart gateway.

[0072] Additionally, the temporary storage unit (230) can temporarily store a set of coefficients of a non-linear reference function.

[0073] As described above, multiple smart gateways (200) can be operated to preserve and store the set of coefficients of the last received non-linear reference function for each gateway when communication with the cloud server (300) is interrupted, and to continuously maintain zone-unit alarms of multiple intelligent sensors (100) included in the area in charge, thereby ensuring a high fail-safe level even when communication with the cloud server (300) is interrupted.

[0074] Next, a set of coefficients of a non-linear reference function is received for each sensor from a plurality of intelligent sensors (100), and after reconstructing the non-linear reference function, the collapse warning can be determined by calculating on the sensing data (step 80).

[0075] In the step (80) of determining whether a collapse warning is given, a set of coefficients of a non-linear reference function is received from a plurality of intelligent sensors (100) and the non-linear reference function is reconstructed, and then the next sensing data provided from the measurement unit (120) is applied to determine whether a collapse warning is given.

[0076] For example, a first communication module (110) provided in a plurality of intelligent sensors (100) can receive a set of coefficients of a non-linear reference function for each sensor transmitted via a wireless communication network (400) from a smart gateway among a plurality of smart gateways (200) and transmit it to a first processor (130).

[0077] And, the first processor unit (130) can receive a set of coefficients of a nonlinear reference function received through the first communication module unit (110), reconstruct the nonlinear reference function, and then apply the next sensing data provided from the measurement unit (120) to determine whether a collapse warning is issued. For a nonlinear reference function that models the normal behavior of a structure, the set of coefficients of a nonlinear reference function transmitted from a plurality of smart gateways (200) and received through the first communication module unit (110) can be received and applied to the nonlinear reference function to reconstruct it.

[0078] Next, the first processor unit (130) can calculate a predicted value by substituting the next sensing data transmitted from the measurement unit (120) according to a preset frequency into a reconstructed non-linear reference function, and can determine the risk of collapse by comparing the deviation corresponding to the difference between the measured value and the predicted value with a preset threshold value. It can determine that there is a risk of collapse if any one of the following cases is present: the risk of collapse persists for a preset time, the deviation increases rapidly, the multiple sensing data exceeds a preset threshold value, or the deviation from the reference function increases statistically significantly. Accordingly, it can control the output of a warning signal regarding the risk of collapse visually and audibly through a separate output device (e.g., a speaker, a warning light, etc.).

[0079] Of course, the first processor (130) can control the first communication module (110) to transmit various data and information calculated and acquired for collapse risk judgment to a plurality of smart gateways (200), and the storage unit (140) can store the non-linear reference function and coefficient set and the judgment result regarding whether a collapse warning is issued.

[0080] A plurality of intelligent sensors (100) as described above can be continuously operated to autonomously determine the risk of collapse by storing the set of coefficients of the last received non-linear reference function for each sensor when communication with a plurality of smart gateways (200) is interrupted, and by inputting sensing data into the non-linear reference function in real time even when displacement, velocity, and acceleration change rapidly. Through this, a fundamentally high fail-safe completeness can be secured compared to a simple constant maintenance method.

[0081] Meanwhile, the total risk of the site can be determined by calculating the Calculated Priority Risk Index (CPRI) using refined data, the Gateway-specific Area Risk Index (GWRI), and the Predictive Risk Behavior Index (PBRI) from the cloud server (300) (step 90).

[0082] In the step (90) of determining the total risk of the site, the Predictive Risk Behavior Index (PBRI) and the Zone Risk Index (GWRI) for each smart gateway can be analyzed in the cloud server (300) to calculate the Comprehensive Risk Prediction Index (CPRI) by analyzing the correlation between the direction of increase / decrease, the speed of increase / decrease, and the zone of increase / decrease.

[0083] For example, the CPRI calculation unit (360) provided in the cloud server (300) can calculate the Comprehensive Risk Prediction Index (CPRI) representing the overall risk level of the site by comprehensively analyzing the increase and decrease of the Gateway-specific Zone Risk Index (GWRI) of all multiple smart gateways (200) and the Predictive Risk Behavior Index (PBRI) of each sensor of multiple intelligent sensors (100) through an artificial intelligence-based anomaly detection model, and this Comprehensive Risk Prediction Index (CPRI) can determine the overall risk of the site by combining the trends of adjacent areas even if only the Gateway-specific Zone Risk Index (GWRI) of a specific area increases rapidly.

[0084] Here, regarding the AI-based anomaly detection model, it can be applied primarily based on management thresholds and statistics, secondarily based on Isolation Forest or LOF (Local Outlier Factor), thirdly based on time-series anomaly detection using Autoencoder and LSTM, and fourthly, a rule engine for cause analysis after an alarm is issued. By applying it in multiple layers in this way, false alarms can be prevented.

[0085] For example, artificial intelligence-based anomaly detection models can be applied in various ways, as shown in Table 1 below.

[0086]

[0087] When a dangerous state is determined according to the comprehensive risk prediction index (CPRI) as described above, an alarm issuance signal corresponding to the dangerous state can be transmitted to a smart gateway adjacent to the smart gateway among the multiple smart gateways (200) through the third communication module (310). At the smart gateway adjacent to the smart gateway, an alarm can be issued visually and audibly through a separate alarm device (e.g., speaker, warning light, etc.), and at the same time, an alarm issuance signal can be transmitted to multiple intelligent sensors (100) in charge of the smart gateway and multiple intelligent sensors (100) in charge of the smart gateway adjacent to the smart gateway. Similarly, at the multiple intelligent sensors (100) in the area, an alarm can be issued visually and audibly through a separate alarm device (e.g., speaker, warning light, etc.), thereby allowing the dangerous state to be immediately transmitted to the site in multiple places, and thereby preventing casualties from the risk of collapse.

[0088] Meanwhile, the intelligent safety management method using a three-stage index-based algorithm according to the embodiment of the present invention as described above consists of a triple structure of multi-sampling and data cleaning, analysis of change, change rate and change acceleration, and generation of dynamic criteria, and provides a closed-loop structure through feedback between the sensor and the server so that the dynamic criteria are transmitted to the sensor, thereby enabling the sensor itself to effectively determine whether a collapse has occurred.

[0089] Accordingly, according to an embodiment of the present invention, a closed-loop structure is provided in which a plurality of intelligent sensors, a plurality of smart gateways, and a cloud server are interconnected in three layers, and a three-measurement cyclic hardware structure is provided in which data measured and refined by a plurality of intelligent sensors is transmitted up to the cloud server through the smart gateway, the coefficients of a sensor-specific non-linear reference function generated by the cloud server are transmitted down individually to each sensor through the smart gateway, and the next measurement data of each sensor to which the updated polynomial equation is applied is transmitted up again to the cloud server. This allows for effective safety management of a structure, significantly reduces false alarms, detects early signs of structural collapse, and enables effective safety management through the latest analysis information stored and managed even when external communication is cut off.

[0090] Although various embodiments of the present invention have been presented and described in the above description, the present invention is not necessarily limited thereto, and those skilled in the art will readily understand that various substitutions, modifications, and changes are possible within the scope of the technical concept of the present invention. Explanation of the symbols

[0091] 100: Multiple intelligent sensors 200: Multiple smart gateways 300 : Cloud Server 400 : Wireless communication network

Claims

Claim 1 A step of refining sensing data from a plurality of intelligent sensors, each comprising at least one of a displacement gauge, an inclinometer, a settlement gauge, a crack gauge, and an accelerometer, and transmitting the refined data; a step of receiving and transmitting the refined data transmitted through the plurality of intelligent sensors at a plurality of smart gateways; a step of determining whether there is a precursor to collapse by analyzing displacement, displacement velocity, and displacement acceleration for each sensor using the refined data received through the plurality of smart gateways at a cloud server; a step of calculating a Predictive Behavioral Risk Index (PBRI) for each sensor at the cloud server based on a state value, the displacement velocity, the displacement acceleration, pattern anomaly, external influence, and a prediction value; a step of calculating a Gateway-level Risk Index (GWRI) for each gateway by analyzing the direction of increase / decrease, the rate of increase / decrease, and the correlation with the refined data and the Predictive Behavioral Risk Index (PBRI) at the cloud server; and a step of using the Predictive Behavioral Risk Index (PBRI) at the cloud server for each The method comprises: a step of generating a set of coefficients of a nonlinear reference function for applying dynamic thresholds to each sensor and transmitting it to the smart gateway; a step of receiving the set of coefficients of the nonlinear reference function at the plurality of smart gateways and transmitting them to each of the plurality of intelligent sensors; a step of receiving the set of coefficients of the nonlinear reference function for each sensor at the plurality of intelligent sensors, reconstructing the nonlinear reference function, and then calculating the sensor data to determine whether a collapse warning is issued; and a step of determining the overall site risk by calculating a Calculated Priority Risk Index (CPRI) using the refined data, the Gateway-specific Zone Risk Index (GWRI), and the Predictive Risk Behavior Index (PBRI) at the cloud server; wherein the step of determining whether there is a collapse precursor involves performing a three-stage measurement analysis of the displacement, displacement velocity, and displacement acceleration for each sensor at the cloud server.If the above displacement acceleration exceeds a preset acceleration threshold, it is determined to be a precursor to collapse; the step of determining whether there is a precursor to collapse comprises comparing and analyzing the absolute displacement value and the preset management standard value through a first-stage displacement analysis on the cloud server, analyzing the rate of change and increasing trend per unit time through a second-stage displacement velocity analysis, and comparing and analyzing the above displacement acceleration and the preset acceleration threshold through a third-stage displacement acceleration analysis; the step of calculating the above Predictive Risk Behavior Index (PBRI) comprises calculating the above Predictive Risk Behavior Index (PBRI) by quantifying it from 0 to 100 on the cloud server, wherein if the above displacement acceleration exceeds the preset acceleration threshold alone, an exponential non-linear weight is assigned to the above Predictive Risk Behavior Index (PBRI); and the step of calculating the above Predictive Risk Behavior Index (PBRI) comprises, for each of the plurality of intelligent sensors, the above state value, displacement velocity, displacement acceleration, Quantify the above Predictive Risk Behavior Index (PBRI) based on pattern anomalies, external influences, and predicted values, wherein the state value represents a normalized score of 0-100 based on the current absolute displacement value, the displacement velocity represents the rate of change of displacement per unit time, assigning weights in a non-linear increasing trend, and the displacement acceleration is the second derivative value; when exceeded alone, an exponential non-linear weight is assigned to the above Predictive Risk Behavior Index (PBRI), the above pattern anomaly represents the degree of deviation of the time-series pattern from the normal range to detect abnormal patterns early, the above external influences reflect at least one process variable among excavation depth, bracing stage, rainfall, and temperature, the above predicted value is utilized for preemptive standard updates through future state prediction by time-series AI, and analyze the above Predictive Risk Behavior Index (PBRI) based on the degree of dispersion of values ​​in a time concept using the characteristics of changes and changed values ​​according to time intervals, and the above predictive The Risk Behavior Index (PBRI) is an intelligent safety management method utilizing a 3-stage artificial intelligence index-based algorithm calculated through standard deviation and variance based on time using (change value + velocity + acceleration). Claim 2 An intelligent safety management method using an artificial intelligence 3-stage index-based algorithm, wherein the step of transmitting the refined data comprises multi-sampling the sensing data from the plurality of intelligent sensors, and then performing outlier removal, noise filtering, and representative value calculation to obtain the refined data. Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 In claim 2, the step of calculating the Gateway-specific Area Risk Index (GWRI) is an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm that calculates the Gateway-specific Area Risk Index (GWRI) in the range of 0-100 based on the Predictive Risk Behavior Index (PBRI) for each smart gateway, maximum value, average value, increasing trend, ratio of sensors exceeding a threshold, and correlation-based concentration at the cloud server. Claim 7 An intelligent safety management method using an artificial intelligence 3-stage index-based algorithm, wherein the step of determining whether a collapse warning is issued is to receive a set of coefficients of the non-linear reference function from the plurality of intelligent sensors, reconstruct the non-linear reference function, and then apply the next sensing data provided from the plurality of intelligent sensors to determine whether a collapse warning is issued. Claim 8 In claim 7, the step of determining the total site risk comprises an intelligent safety management method using an artificial intelligence 3-stage index-based algorithm that calculates the Comprehensive Risk Prediction Index (CPRI) by analyzing the correlation between the Predictive Risk Behavior Index (PBRI) and the Zone Risk Index (GWRI) for each smart gateway and the direction of increase / decrease, the speed of increase / decrease, and the zone of increase / decrease in the cloud server.

Images