Search

KR-102961803-B1 - Real-time anomaly detection system for SOC facilities using skin sensors and artificial intelligence

KR102961803B1KR 102961803 B1KR102961803 B1KR 102961803B1KR-102961803-B1

Abstract

The present invention relates to a real-time anomaly detection system for SOC facilities that detects anomalies in facilities, such as railway tracks, tunnels, and buildings, by analyzing and predicting measurement data regarding deformation of the facilities through a machine learning algorithm before damage or destruction occurs. The system may include: a pressure sensor installed in close contact at a specific location of the facility, wherein a plurality of nodes having a certain area are arranged while maintaining spacing from each other, and the pressure sensor measures stress on the load acting on each node; a cloud database that collects and stores data measured by the pressure sensor; and an analysis device that performs time-series analysis on the stored data and predicts abnormal behavior of the facility based on the data learning results by artificial intelligence.

Inventors

  • 최정열
  • 이호현
  • 한재민
  • 안대희

Assignees

  • 주식회사 산강이엔씨

Dates

Publication Date
20260507
Application Date
20230331

Claims (6)

  1. A pressure sensor that is installed in close contact at a specific location of a facility, wherein a plurality of unit nodes having a certain area are arranged in rows and columns while maintaining spacing from each other, and measures the pressure caused by a vertical load acting on the area of each unit node to measure the pressure acting according to the deformation of the facility across the entire installed area; A cloud database that collects and stores data measured by the pressure sensor; and It includes an analysis device that performs time series analysis on the above-mentioned stored data and predicts abnormal behavior of facilities based on data learning results by artificial intelligence, The above analysis device is, An analysis section setting unit that, based on user input, sets a data analysis section in which data analysis is performed among the collected data, designates at least a portion of the analysis section as a learning section, and sets a prediction section after the analysis section; A management standard setting unit for setting management standard values for diagnosing facility defects; A prediction model generation unit that generates a prediction model by learning the data of the above-mentioned learning section; and It includes an abnormal behavior prediction unit that displays the trend of predicted data in a prediction interval using the above prediction model, and calculates and provides the time when the predicted data reaches a set management threshold. The above analysis device is, An SOC facility anomaly detection system characterized by further including a 3D visualization unit that identifies a unit node representing a maximum pressure value in the above data analysis section and visualizes the pressure values of all nodes in 3D at that point in time, thereby representing the pressure distribution over the entire installation area in a three-dimensional and continuous manner at a specific point in time.
  2. In paragraph 1, The above pressure sensor is installed between the sleeper vibration damping pad and the sleeper box to measure the track support stiffness of the sleeper floating track and the pressure change at the bottom of the sleeper, and is an SOC facility anomaly detection system that measures the pressure generated by the external environment at the bottom of the concrete sleeper.
  3. In paragraph 1, The above pressure sensor is a SOC facility anomaly detection system installed on the wall of a construction or civil engineering facility to measure vertical stress that causes deformation of the facility due to internal or external factors.
  4. In paragraph 1, The above analysis device is, An SOC facility anomaly detection system further comprising a time series analysis unit that generates and displays a scatter plot showing the pressure of each node over time through time series analysis in the above-determined data analysis section.
  5. In paragraph 4, The above analysis device is, An SOC facility anomaly detection system further comprising an algorithm selection unit for selecting a machine learning algorithm to learn data of the above-mentioned learning section.
  6. delete

Description

Real-time anomaly detection system for SOC facilities using skin sensors and artificial intelligence The present invention relates to a system and method for detecting abnormalities in Social Overhead Capital (SOC) facilities, and specifically, to a system and method capable of identifying abnormal behavior of SOC facilities in real time by utilizing a skin sensor and artificial intelligence. Currently, due to the population growth in cities in Korea, there is an increasing demand for social infrastructure facilities such as high-rise buildings, public transportation, and amenities. Furthermore, large-scale, deep excavation construction is being carried out adjacent to structures due to the overcrowding of urban areas. In urban areas, ensuring the safety of retaining walls and underground structures is critical for adjacent excavation. Therefore, automated measurement systems have been introduced to manage the safety of facilities in real time. However, the utilization of the results from these automated measurement systems is currently very low, and research on management standards and evaluation techniques for rational and reliable measurements is insufficient. Furthermore, research on behavioral analysis systems capable of responding in real-time to various environmental, structural, and material change factors occurring during construction is still insufficient. Unexpected external factors such as adjacent excavation, ground deterioration, and groundwater levels can cause sudden deformation and damage to underground structures and tracks, posing a significant threat to railway safety. To enable active real-time responses to various changes in conditions that may occur during construction and operation, research on AI-based algorithms for predicting structural safety regarding site conditions and various construction information (including material and structural characteristics) is currently required. FIG. 1 is a drawing showing a pressure sensor according to an embodiment of the present invention. FIG. 2 is a drawing showing an example of the installation of a pressure sensor according to the present invention. FIG. 3 is a drawing showing another installation example of a pressure sensor according to the present invention. FIG. 4 is a configuration diagram showing the main configuration of a facility abnormality detection system according to an embodiment of the present invention. FIG. 5 is a block diagram showing the detailed configuration of an analysis device according to an embodiment of the present invention. FIG. 6 is a drawing showing the display screen and UI of an analysis device according to an embodiment of the present invention. In order to provide a detailed explanation sufficient for a person skilled in the art to easily implement the technical concept of the present invention, the most preferred embodiment of the present invention will be described with reference to the accompanying drawings. First, it should be noted that when assigning reference numerals to the components of each drawing, the same components are assigned the same numeral whenever possible, even if they are shown on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of related known components or functions may obscure the essence of the invention, such detailed description is omitted. FIG. 1 is a drawing showing a pressure sensor according to an embodiment of the present invention. The pressure sensor (100) may use a skin sensor in which the magnitude of electrical resistance changes depending on force or pressure. By physical definition, pressure is the load divided by the cross-sectional area, and the magnitude of pressure changes depending on the area when the same load occurs. The skin sensor according to the present invention is made of a thin and lightweight material with a thickness of less than 0.5 mm and has strong resistance to impact. In the present invention, such a skin sensor is applied in the fields of railway (track), civil engineering, and construction. When an external load is applied to the skin sensor, conductive particles come close to the inside of the polymer, and the applied pressure can be measured through the fluctuating resistance values between the particles. Referring to FIG. 1, the pressure sensor (100) may be configured to include a plurality of unit nodes (101) having the same area. In the present invention, the node pressure (or stress) may be defined as the value obtained by dividing the load measured at each node by the node area. The main specifications of the skin sensor used in the embodiment of the present invention are as shown in the following [Table 1]. [Table 1] The number of nodes constituting the skin sensor can be determined based on the number of rows and columns, and in the present invention, it was composed of a total of 48*48 (2,304) nodes. The maximum detection pressure was determined by whether the ADC