Mechanical system vibration detection device

Abstract

A mechanical system vibration detection device is disclosed. The mechanical system vibration detection device according to an embodiment of the present invention comprises: a mounting part having a bracket structure that is mounted on one side surface of a monitoring object and has a vibration detection part mounted at a predetermined distance from one side surface of the monitoring object, and is bent multiple times; a vibration detection part mounted on the mounting part in a structure that is detachably mounted by a fastening member and detects the vibration of the monitoring object in real time and transmits it to a data transmission part; a data transmission part that wirelessly transmits the data transmitted from the vibration detection part to a data analysis corresponding part using a preset wireless communication means; and a data analysis corresponding part that collects the data transmitted from the data transmission part, analyzes the state of the monitoring object based on the data collected for a preset time, and outputs the analysis result to an operator in real time.

Description

Technology Field

[0001] The present invention relates to a mechanical system vibration detection device, and more specifically, to a mechanical system vibration detection device capable of ensuring the safety and reliability of a mechanical system by measuring vibration and detecting the wear state of a rotating bearing. Background Technology

[0002] Bearings applied to rotary drive units in conventional mechanical systems are consumable parts that accumulate wear, deterioration, and reduced lubrication as they are exposed to high-speed rotation and repetitive load conditions for extended periods. In particular, for equipment such as dry vacuum pumps used in semiconductor processes, vacuum facilities, and industrial pump systems, even minute bearing abnormalities can lead to a decrease in the overall performance of the equipment or a halt in the process; therefore, precise management of bearing conditions is required. However, in conventional technology, management primarily relies on periodic inspections or operator experience rather than a system that manages bearings by reflecting their actual operating status in real time.

[0003] Conventional bearing condition management methods often remain limited to periodic disassembly inspections or visual and auditory assessments of noise and heat. Such management approaches have drawbacks, as they rely on equipment shutdown or only detect abnormal signs after they have progressed significantly. In particular, early diagnosis is difficult during the initial wear stage because external abnormalities are rarely visible, and problems are frequently discovered only after an actual failure has occurred.

[0004] Furthermore, while existing technologies partially utilize methods for measuring bearing temperature, simple temperature-based diagnosis can only indirectly assess the degree of friction increase and has limitations in precisely analyzing complex mechanical degradation conditions such as eccentric rotation, vibration amplification, and localized wear. Since temperature rise is heavily influenced by the surrounding environment, it suffers from the problem of failing to accurately reflect the physical wear state of the bearing itself.

[0005] While some technologies utilizing vibration analysis have been proposed, conventional vibration measurement techniques mostly rely on temporary measurements using portable diagnostic devices or expensive analytical equipment. Such equipment presents challenges, including the difficulty of establishing a continuous monitoring system, reliance on the interpretation capabilities of skilled professionals, and excessive installation and operating costs. Consequently, these structures make practical implementation difficult in small and medium-sized industrial sites.

[0006] In particular, conventional technologies often remain at the level of simply recording vibration data, leading to a problem where measured data fails to translate into practical maintenance decisions. Even when data is accumulated, the lack of a systematic analysis environment prevents the automatic determination of correlations between vibration patterns and signs of failure, resulting in a structure that relies on the subjective judgment of managers.

[0007] Another problem is the inefficiency of the sensor installation structure. Conventional vibration sensors are often attached directly to the surface of the bearing housing or temporarily without a separate fixing structure, leading to inaccurate vibration transmission or the sensor itself falling off. This structural instability significantly reduces the reliability of vibration measurements.

[0008] Furthermore, existing systems suffer from a lack of connectivity between measurement devices and the central control system, which prevents data collection and analysis from being organized into a single integrated system. Vibration data measured at the individual equipment level is managed independently, and the absence of a network-based integrated monitoring system limits the simultaneous management of multiple devices.

[0009] In terms of communication methods, conventional technologies face difficulties in real-time data transmission due to reliance on wired connections or unstable wireless environments. Particularly in industrial settings, stable communication is often not ensured due to factors such as electromagnetic interference, distance issues, and separation between equipment, making the reliable collection and transmission of vibration data challenging.

[0010] Furthermore, the absence of an intelligent judgment system to automatically detect bearing abnormalities results in a problem where measured data remains at the level of simple alarm signals. A simple alarm structure based on whether a certain threshold is exceeded causes issues such as generating unnecessary alarms even in normal conditions or, conversely, missing gradually developing high-risk states.

[0011] As such, conventional technology has limitations in various elements, such as real-time monitoring of bearing conditions, automation of data analysis, stable installation structures, communication reliability, and intelligent judgment systems. Consequently, preventive maintenance and proactive response to failures in mechanical systems are currently being carried out inefficiently. Therefore, the present invention aims to solve the problems of conventional technology. Prior art literature

[0012] Korean Patent Publication No. 10-2097944 (Registration Date: April 1, 2020) The problem to be solved

[0013] The objective of the present invention is to provide a mechanical system vibration detection device capable of preventing mechanical system failures in advance by detecting vibrations generated in a rotating bearing in real time and analyzing them to identify the bearing wear condition at an early stage. Furthermore, the invention aims to provide a device that enables the stable collection and wireless transmission of vibration data, and allows for the quantification and automation of maintenance decisions by facility managers by providing analysis results in real time. means of solving the problem

[0014] A mechanical system vibration detection device according to one aspect of the present invention for achieving such an objective may comprise: a mounting part having a bracket structure that is mounted on one side surface of a monitoring object and has a vibration detection part mounted at a predetermined distance from one side surface of the monitoring object, and is bent multiple times; a vibration detection part that is mounted on the mounting part in a structure that is detachably mounted by a fastening member and detects the vibration of the monitoring object in real time and transmits it to a data transmission part; a data transmission part that wirelessly transmits the data transmitted from the vibration detection part to a data analysis corresponding part using a preset wireless communication means; and a data analysis corresponding part that collects the data transmitted from the data transmission part, analyzes the state of the monitoring object based on the data collected for a preset time, and outputs the analysis result to an operator in real time.

[0015] In one embodiment of the present invention, the mounting portion may comprise: a detachable portion having a structure that is bolted to one side surface of a monitoring object or attached by magnetism; a vibration transmission portion having an integral structure with the detachable portion and extending by a predetermined height from the detachable portion; and a sensor mounting portion having a bracket structure that is mounted integrally on the upper end of the vibration transmission portion and extends laterally by a predetermined width to mount the vibration detection portion in a manner that surrounds the outer surface of the vibration detection portion.

[0016] In one embodiment of the present invention, the vibration detection unit may be configured to include: a housing unit having a cylindrical structure that is detachably mounted on the sensor mounting unit, a temperature detection unit mounted on the upper surface of the cylindrical structure, and a data cable connected to a data transmission unit mounted on the lower surface; a temperature detection unit mounted on the upper surface of the housing unit that detects the temperature of one side surface of a monitoring object in real time and transmits the detected temperature data to a data transmission unit through a data cable; and a vibration detection unit mounted inside the housing unit that detects the vibration of a monitoring object in real time and transmits the detected vibration data to a data transmission unit through a data cable.

[0017] In one embodiment of the present invention, the data transmission unit may be configured to include: a data receiving module having a plurality of connection ports capable of independently receiving a plurality of data cables connected to a vibration detection unit; a data preprocessing module that labels and temporarily stores data transmitted from the data receiving module according to each monitoring target, detection location, detection time, and detection purpose; and a wireless transmission module that transmits the data processed and temporarily stored by the data preprocessing module to a data analysis corresponding unit in real time using a preset wireless communication protocol.

[0018] In one embodiment of the present invention, the data analysis response unit may comprise: an AI judgment module that classifies and determines the danger state of a monitored object according to preset danger levels based on data transmitted from the wireless transmission module; a status warning module that outputs a preset warning signal based on the result determined by the AI ​​judgment module and stores data related to the history of the warning signal; and a safety measure module that, based on the result determined by the AI ​​judgment module, cuts off power supplied to the monitored object, urgently activates emergency safety equipment, and reports to a safety manager and emergency dispatch personnel when rapid emergency measures are required. Effects of the invention

[0019] According to the mechanical system vibration detection device of the present invention, vibrations of a rotating bearing are measured in real time, and the measured data is transmitted wirelessly for analysis, thereby enabling early determination of the bearing's wear condition. Consequently, abnormal conditions can be detected before equipment failure occurs, significantly improving the stability of the mechanical system. Furthermore, unlike conventional methods that rely on simple temperature or noise assessment, the bearing condition can be diagnosed more precisely through vibration-based analysis. In addition, by integrating and analyzing data from multiple sensors, the overall condition of the equipment can be comprehensively assessed, and data-driven maintenance becomes possible, reducing unnecessary inspections and lowering maintenance costs. Moreover, the reliability of system operation can be maximized through the provision of real-time warnings and the linkage of safety measures. Brief explanation of the drawing

[0020] FIG. 1 is a perspective view showing a mechanical system vibration detection device according to one embodiment of the present invention. FIG. 2 is a photograph showing the state in which a mechanical system vibration detection device according to one embodiment of the present invention is installed. FIG. 3 is a schematic diagram showing two embodiments of a detachable part of a mounting part according to one embodiment of the present invention, respectively, showing a bolt fastening method and a magnetic member usage method. FIG. 4 is a schematic diagram showing the mounting position of a vibration detection unit according to one embodiment of the present invention. FIG. 5 is a block diagram showing a mechanical system vibration detection device according to one embodiment of the present invention. FIG. 6 is a schematic diagram showing a communication method of a data transmission unit according to an embodiment of the present invention. FIG. 7 is a schematic diagram showing a communication method of a data transmission unit according to an embodiment of the present invention, and is a schematic diagram showing Read analog data of a Command and Read analog data of a Response transmitted from a Host System to an AMP Box. FIG. 8 is a schematic diagram showing a communication method of a data transmission unit according to an embodiment of the present invention, and is a schematic diagram showing Read analog data of a Command and Read analog data of a Response transmitted from a Host System to an AMP Box. Figure 9 is a schematic diagram showing a method for calculating SUM1 by converting the lower two digits of the hexadecimal value resulting from the sum of character codes from STX to ETX into ASCII for data containing STX and ETX. Figure 10 is a schematic diagram showing a method for calculating SUM2 by converting the lower two digits of the hexadecimal value resulting from the addition of character codes into ASCII for data that does not include STX and ETX. Specific details for implementing the invention

[0021] Preferred embodiments of the present invention will be described in detail below with reference to the drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the present invention.

[0022] Throughout this specification, when it is stated that one component is located "on" another component, this includes not only cases where one component is in contact with another component, but also cases where another component exists between the two components. Throughout this specification, when it is stated that a part "includes" a component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0023] FIG. 1 shows a perspective view of a mechanical system vibration detection device according to one embodiment of the present invention, and FIG. 2 shows a photograph showing the state in which a mechanical system vibration detection device according to one embodiment of the present invention is installed.

[0024] Referring to these drawings, the machine system vibration detection device (100) according to the present embodiment is equipped with a mounting part (110) that performs a specific role, a vibration detection part (120), a data transmission part (130), and a data analysis response part (140), thereby measuring and analyzing vibration signals generated from a monitoring target object to evaluate the condition of a bearing, and thus providing a machine system vibration detection device capable of improving the safety and reliability of the machine system.

[0025] Hereinafter, with reference to FIG. 1 and FIG. 10, each component constituting the mechanical system vibration detection device (100) according to the present embodiment will be described in detail.

[0026] Detailed description of the mounting part (110)

[0027] The mounting part (110) is a mechanical interface structure designed to be stably fixed to one side surface of the object being monitored, while allowing the vibration detection part (120) to receive the actual vibration characteristics of the object without distortion. It is not a simple fixing member, but is formed as a composite structure that simultaneously satisfies vibration transmission efficiency and sensor protection functions, and is configured to maintain structural stability even in a long-term operating environment.

[0028] The mounting portion (110) has a bracket structure that is bent multiple times, and this bent structure serves to improve rigidity against external forces compared to a single planar structure. In particular, it is designed to minimize the problem of the mounting angle being deformed due to micro-vibrations, repetitive loads, thermal expansion, etc., that occur during the operation of rotating equipment, thereby preventing positional displacement of the vibration detection portion (120).

[0029] The mounting part (110) has a structure capable of accommodating various fixing types so as to enable stable contact regardless of the surface condition of the object being monitored. It is designed to be compatible with various coupling structures, such as bolt fastening and magnetic attachment methods, thereby ensuring flexibility regarding the installation environment, and can be applied without separate processing even when the equipment structure is changed.

[0030] The mounting part (110) eliminates unnecessary elastic elements to increase the vibration transmission efficiency and adopts a rigid structure mainly made of metal material. This allows for minimizing vibration loss in the high-frequency range and reliably transmitting even minute vibration values ​​to the vibration detection part (120).

[0031] In addition, the mounting part (110) is designed to maintain a structurally stable state even in the event of external shock or equipment vibration amplification. The reliability of the mechanical fastening is enhanced to prevent sensor detachment or positional error, and it can be operated stably for a long period of time even in environments where vibration and shock are repeated, such as industrial sites.

[0032] Detailed description of the detachable part (111)

[0033] The detachable part (111) is an interface area where the mounting part (110) is coupled to the object to be monitored, and is configured to simultaneously satisfy mechanical fixation reliability and installation convenience. It is a structure that plays an important role not only in vibration transmission but also in maintenance.

[0034] The detachable part (111) may be configured to include one or more of a bolt-fastening structure or a magnetic attachment structure, and the fixing method is selected according to the facility environment. In the case of a bolt-fastening structure, it has excellent durability against vibration and can be applied to high-speed rotating bodies, while a magnetic attachment structure has the advantage of allowing a worker to quickly install or detach it.

[0035] Since the detachable part (111) functions as a direct transmission path for vibration in addition to its simple fixing purpose, the rigidity of the contact surface is very important. Accordingly, a structure that minimizes surface irregularities and a tight-fitting design are applied, thereby reducing the possibility of resonance with the mounting surface and minimizing transmission errors.

[0036] In addition, the detachable part (111) is structurally designed so that the fastening force does not decrease even with repeated detachment. A material and structure are applied that take into account the accumulation of fatigue in the fastening part, weakening of magnetism, and wear of the fastening surface, so that the fixing force is stably maintained even during long-term operation.

[0037] The detachable part (111) is designed with a modular structure to flexibly respond to changes in the installation environment. It can be quickly reinstalled without separate equipment processing when replacing equipment, moving locations, or replacing sensors, thus providing excellent maintenance efficiency.

[0038] Detailed description of the vibration transmission part (112)

[0039] The vibration transmission unit (112) is structured to transmit the vibration of the monitoring object transmitted to the detachable unit (111) to the vibration detection unit (120), and is a core functional unit responsible for vibration transmission within the mounting unit (110). It is a major structural element that determines the accuracy of vibration sensitivity.

[0040] The vibration transmission part (112) has an extension structure that extends by a predetermined height from the detachable part (111), and this extension structure is designed to transmit vibration signals without loss while separating the sensor from the surface of the object. It provides a stable vibration measurement environment while avoiding external heat dissipation structures or wiring interference.

[0041] The vibration transmission section (112) is formed of a high-rigidity material so that the vibration signal is not attenuated, and the cross-sectional shape and structural rigidity are precisely designed to minimize resonance or signal distortion that may occur in the intermediate insert structure. Accordingly, signal responsiveness is improved and both low-frequency and high-frequency vibrations are accurately reflected.

[0042] In addition, the vibration transmission unit (112) also performs a structural role of protecting the sensor from external environmental factors. By separating the vibration detection unit (120) by a certain distance from temperature changes, dust, oil droplets, etc. inside the facility, it has the effect of extending the sensor's lifespan.

[0043] The vibration transmission part (112) is designed so that torsion or bending deformation does not occur, as it is structured to mechanically connect the mounting part (110) and the sensor mounting part (113). This minimizes distortion in the direction of vibration and enables precise vibration measurement of the actual rotating body behavior.

[0044] Detailed description of the sensor mounting part (113)

[0045] The sensor mounting part (113) is a mechanical interface structure that allows the vibration detection part (120) to stably receive the vibration transmitted through the vibration transmission part (112). It is not a simple fixed part, but performs the function of precisely maintaining the installation position and direction of the sensor so that the vibration signal is not distorted.

[0046] The sensor mounting portion (113) is formed in the shape of a bracket and has a structure that surrounds the outer surface of the vibration detection portion (120), and this structure serves to protect the sensor from movement, falling, external impact, etc. A compression force distribution structure is applied so that the detection portion does not become displaced or the binding force does not decrease even under industrial equipment conditions with severe vibration environments.

[0047] The sensor mounting part (113) is formed with a binding structure that allows the vibration detection part (120) to be detachably attached, making maintenance and replacement easy. As a result, in the event of a sensor failure, only the vibration detection part can be quickly replaced without disassembling the entire equipment, thereby reducing maintenance costs.

[0048] The sensor mounting portion (113) also serves as a mechanical filter to minimize external vibration interference. The structure is optimized to suppress unnecessary resonance, thereby reducing the inflow of disturbance signals other than the target vibration.

[0049] In addition, the sensor mounting part (113) is formed to be compatible with vibration detection parts (120) having various outer diameters, and the system has excellent expandability because the mounting part can be utilized as is even when replacing sensors through a standardized structure.

[0050] Detailed description of the vibration detection unit (120)

[0051] The vibration detection unit (120) is a core sensor module that converts vibrations generated from a monitored object into electrical signals. It is formed with a high-sensitivity structure to detect even minute vibrations caused by bearing wear or abnormal rotation.

[0052] The vibration detection unit (120) is detachably mounted on the sensor mounting unit (113) so that it can be quickly replaced during maintenance. The fixed form is designed with a tight-fitting structure to maximize vibration transmission efficiency, and fastening stability is ensured so that no structural gaps or play occur.

[0053] The vibration detection unit (120) is not limited to simple vibration detection but can capture minute changes according to real-time signal generation characteristics and can obtain data at a level capable of distinguishing bearing damage, eccentric rotation, external impact, etc.

[0054] In addition, the vibration detection unit (120) is formed with a structure that enhances environmental resistance, so it operates stably even in high temperature, high humidity, and dusty environments. A sealed structure and a protective housing are applied so that reliable measurement data can be provided even under harsh conditions in industrial sites.

[0055] The vibration detection unit (120) is connected to the data transmission unit (130) via a wire to stably transmit vibration signals, and the data cable connection structure is optimized to minimize errors caused by signal loss or external electromagnetic interference.

[0056] Detailed description of the housing unit (121)

[0057] The housing unit (121) is a protective case structure that forms the outer shape of the vibration detection unit (120) and serves to protect the sensor's internal circuit and the detection unit from the external environment. It is not a simple external case, but a functional structure designed with durability and heat dissipation in mind.

[0058] The housing unit (121) is formed in a cylindrical structure and is stably connected to the sensor mounting part (113), preventing rotation or detachment due to external force. Anti-slip protrusions or fixing grooves may be formed on the outer wall of the cylinder.

[0059] The housing unit (121) is equipped with a temperature detection unit (122) at the top and has an opening formed at the bottom for connecting a data cable. This structure allows for stable temperature measurement at the top and efficient signal transmission at the bottom.

[0060] Additionally, the housing unit (121) is designed so that its internal space protects the vibration detection unit (123) and the circuit, and may include a cushioning structure to prevent external shocks from being directly transmitted to the internal components. This reduces the possibility of sensor malfunction and damage.

[0061] The housing unit (121) is formed of a metal or reinforced resin with consideration for corrosion resistance and heat resistance, and various materials may be applied depending on the installation environment. Accordingly, stable performance is maintained without external damage or structural deterioration even during long-term operation.

[0062] Detailed description of the temperature detection unit (122)

[0063] The temperature detection unit (122) is an auxiliary sensing component for indirectly determining the bearing operating state by detecting the surface temperature of the object being monitored in real time. It is not a simple temperature sensor, but a monitoring element designed to analyze the operating state from various angles by combining it with vibration data. It performs the role of detecting abnormal heat generation phenomena caused by bearing wear, poor lubrication, misalignment, etc., at an early stage.

[0064] The temperature detection unit (122) is mounted on the upper surface of the vibration detection unit (120) and is designed to measure temperature while in direct contact with the outside or in close proximity. This arrangement structure is designed to minimize interference with the vibration detection unit (123) while ensuring heat transfer efficiency. Through this, the temperature measurement signal and the vibration measurement signal are structurally separated, allowing for the independent acquisition of highly reliable data.

[0065] The temperature detection unit (122) is designed to detect not only rapid temperature changes but also gradual temperature increases, and continuously provides information to the data transmission unit (130) so that the reference temperature pattern in a long-term operating environment and real-time temperature data can be compared and analyzed. Since bearing failure does not occur suddenly but undergoes a deterioration process that accumulates over a certain period of time, a structure that secures such accumulated temperature data plays a very important role in failure prediction.

[0066] In conventional technology, the detection method for exceeding a simple temperature limit is predominant, which has the problem of frequently causing unnecessary alarms due to environmental factors or temporary load increases. However, since the temperature detection unit (122) of the present invention is interpreted in conjunction with vibration information, it can distinguish between simple thermal changes and heat generation caused by mechanical abnormalities. Accordingly, false alarms are reduced, and the accuracy of judgment regarding the actual possibility of failure is improved.

[0067] In addition, the temperature detection unit (122) may utilize a sensor element having heat resistance characteristics to operate stably even in high-temperature environments, and is designed to provide reliable data for a long period even in harsh industrial environments such as vacuum equipment and high-speed rotating bodies. Through this, equipment reliability and management efficiency are significantly improved compared to conventional technology.

[0068] Detailed description of the vibration detection unit (123)

[0069] The vibration detection unit (123) is a key component that detects vibrations occurring in the bearing and rotating part and converts them into electrical signals. Since vibration is a physical quantity that directly reflects various abnormal conditions such as bearing wear, eccentricity, lack of lubrication, and assembly defects, the technical core of the present invention is focused on the vibration detection unit (123).

[0070] The vibration detection unit (123) is mounted inside the housing unit (121) and is protected from external shocks, dust, moisture, etc., and is positioned to directly receive the signal transmitted from the vibration transmission unit (112). Through this structure, pure mechanical vibration data can be stably acquired without external environmental influence.

[0071] The vibration detection unit (123) is designed to detect even very minute vibration changes occurring in the early stages of bearing wear, and can identify frequency components, amplitude changes, and repeating patterns of the rotating body vibration. Through this, it is possible to obtain data that can estimate the wear progression stage, going beyond the level of simple fault detection.

[0072] In conventional technology, vibration was mostly measured temporarily, and the structure required analysis by an expert. In contrast, the vibration detection unit (123) can form an automatic analysis structure by collecting data continuously and transmitting it to the data analysis response unit (140). As a result, dependence on skilled personnel is reduced, and the equipment management system is automated.

[0073] In addition, the vibration detection unit (123) is structurally designed to minimize the influence of disturbance signals, and data characteristics are secured to distinguish between vibrations occurring in adjacent equipment or shock signals from the surrounding environment and the natural vibration of the bearing. Accordingly, measurement reliability is improved and diagnostic accuracy is enhanced.

[0074] Detailed description of the data transmission unit (130)

[0075] The data transmission unit (130) is a relay module that receives data collected from the temperature detection unit (122) and the vibration detection unit (123) and stably transmits it to the data analysis response unit (140) in a wireless environment. It performs the role of a core communication hub that ensures data consistency and reliability, rather than being a simple communication device.

[0076] The data transmission unit (130) has a structure capable of independently receiving each signal so that multiple sensor data are not mixed, and performs data format matching and temporary storage functions. A flow control function is built in to prevent data loss, duplication, or errors from occurring during this process.

[0077] In conventional technology, sensors and analysis devices are connected via wires, resulting in significant installation constraints and complex wiring management. However, the data transmission unit (130) of the present invention is based on a wireless communication environment, allowing for stable data transmission even in environments where equipment is spaced apart. Accordingly, the degree of freedom in installation is improved, and system scalability is ensured.

[0078] The data transmission unit (130) is capable of immediate reporting when abnormal data occurs based on a periodic communication environment with the Host System. Unlike conventional technology, the equipment does not autonomously transmit signals, but operates in a response structure according to the command of the Host System, thereby ensuring consistent management of the data flow.

[0079] In addition, the data transmission unit (130) ensures data reliability by including an error detection and verification mechanism during transmission, and enables long-term status analysis of the equipment through an accumulated data management structure. As a result, base data is accumulated that can analyze long-term deterioration trends beyond short-term diagnosis.

[0080] Detailed description of the data acceptance module (131)

[0081] The data receiving module (131) serves as an input interface for receiving multiple sensor data transmitted from the vibration detection unit (120) and the temperature detection unit (122). It is not a simple set of input ports, but a control module that includes a structure for independently managing the data flow for each sensor and preventing interference between signals.

[0082] The data receiving module (131) is equipped with multiple connection ports so that multiple vibration detection units (120) can be connected independently. Accordingly, multiple bearings or multiple monitoring points can be managed simultaneously, and it can be flexibly applied even in large-scale facilities or complex equipment environments.

[0083] In conventional technology, signal collision or data loss problems frequently occurred in structures where multiple sensors were connected in parallel. However, the data receiving module (131) of the present invention structurally blocks such data interference problems by providing an independent receiving structure for each sensor.

[0084] In addition, the data receiving module (131) includes a function to verify the basic format and validity of the collected data. This function serves to filter out abnormal data in advance, such as sensor output errors or disconnections.

[0085] The data receiving module (131) is designed with a modular structure to facilitate the connection of additional sensors when expanding the device. Even if additional monitoring targets are added due to changes in the field environment, it is possible to respond by expanding the data receiving module (131) without the need to replace the entire system.

[0086] Detailed description of the data preprocessing module (132)

[0087] The data preprocessing module (132) is a step-by-step processing module that refines raw data received through the data receiving module (131) before analysis. It is a key element that determines the overall system reliability by performing the function of correcting data quality before analysis.

[0088] The data preprocessing module (132) automatically labels the data based on sensor identification information, measurement location, time information, etc. Through this, the data is processed into a structure suitable for analysis rather than a simple list of numerical values.

[0089] In conventional technology, data sorting and classification tasks were performed manually, which placed a heavy burden on management. However, the data preprocessing module (132) of the present invention automates this process to minimize operator intervention. Accordingly, management efficiency is improved and human error is reduced.

[0090] The data preprocessing module (132) improves the accuracy of analysis by filtering out outliers, missing values, noise signals, etc., from the measurement data. This prevents temporary vibrations or disturbance signals caused by environmental factors from being reflected in the AI ​​analysis.

[0091] The data preprocessing module (132) includes a temporary storage function to enable time-based data accumulation, and this accumulated information is used for bearing degradation trend analysis. This serves as a basis for enabling long-term condition prediction rather than short-term diagnosis.

[0092] Detailed description of the wireless transmission module (133)

[0093] The wireless transmission module (133) is a communication interface that wirelessly transmits data processed by the data preprocessing module (132) to the data analysis corresponding unit (140). It includes a data transmission control structure that ensures stability and consistency, rather than a simple communication means.

[0094] The wireless transmission module (133) communicates with the Host System using a protocol-based response method. Since the device does not arbitrarily transmit data but responds according to the Host System's request, data flow management is systematically performed.

[0095] In conventional technology, problems such as data loss, communication collisions, and connection instability frequently occurred in wireless environments, but the wireless transmission module (133) of the present invention structurally improves these problems through a communication structure including a verification code.

[0096] The wireless transmission module (133) includes state control logic that manages communication delay, error occurrence, retransmission, etc. Through this, communication reliability is ensured and a real-time monitoring environment is realized.

[0097] The wireless transmission module (133) is designed to enable many-to-one communication with the Host System even when multiple equipment is in operation, thereby enabling the implementation of a centralized monitoring system. Accordingly, the scope of equipment management is expanded, and an integrated monitoring environment can be established.

[0098] Detailed explanation of the communication method of the wireless transmission module (133)

[0099] The communication method of the wireless transmission module (133) is a protocol that operates based on a request-response structure between the Host System (PMS) and the AMP Box. The key feature of this communication method is that the AMP Box is designed to not arbitrarily transmit data, but to return a response only after a command is received from the Host System. Through this, even if multiple AMP Boxes exist on the network, irregular data collisions or interference do not occur, and the Host System, which is a central control system, can completely control the entire communication flow. Therefore, even in environments with high communication volume, traffic is managed consistently, and abnormal data surges or unpredictable patterns are minimized.

[0100] The wireless transmission module (133) is operated under the premise that the minimum interval between commands transmitted from the Host System to each AMP Box is set to 1.0 second. This time interval constraint takes into account the internal processing time of the AMP Box, the data collection cycle, and the analog data sampling and preprocessing time, thereby preventing problems such as the device's internal buffer being overloaded or response delays accumulating due to the transmission of commands at excessively short intervals. In addition, since a Command-Response structure of a certain period is maintained, the Host System can reliably perform timestamp-based data alignment and equipment status history management. This method is an important operational design element for simultaneously ensuring communication stability and data consistency.

[0101] The Command and Response data frames processed by the wireless transmission module (133) have formats separated by function. For example, the Read analog data Command transmitted from the Host System to the AMP Box is a command to request analog data and is structured to include a set ASCII character range and SUM1 data for checking. In response, the Read analog data Response transmitted from the AMP Box to the Host System is transmitted in a structured format that includes analog data values ​​for each channel and SUM2 data for verifying them. This structure goes beyond simple value transmission and performs the role of minimizing transmission errors that may occur in the wireless section by embedding its own error verification means within the data frame.

[0102] In addition, the wireless transmission module (133) supports not only analog data but also a Read AMP Status data command and a Response format for transmitting equipment status information. The Read AMP Status Command transmitted from the Host System has a relatively simple structure including SUM1, but the Read AMP Status Response returned from the AMP Box has a more extended structure including Warning data and Reserve data. The first byte of the Warning data is designed to include, in bit units, whether a warning has occurred for channels 1 to 4, and the second byte is designed to include whether a warning has occurred for channels 5 to 8. Through this structure, the abnormal status of multiple channels can be compressed and expressed in just two bytes, and the Reserve data area can be utilized as an extension field to store future function expansion or additional status information.

[0103] In the communication method of the wireless transmission module (133), different checksums (SUM1, SUM2) are applied depending on whether STX and ETX are included. In the case of a data frame containing STX and ETX, SUM1 is defined to be calculated by converting the lower two hexadecimal digits of the result of adding all character codes from STX to ETX into ASCII. On the other hand, in the case of data not containing STX and ETX, SUM2 is calculated by converting the lower two hexadecimal digits of the result of adding all character codes into ASCII. This dual-structure verification method is intended to secure an error detection capability optimized according to the data frame format, and enables the receiving side to quickly detect abnormal situations such as single-bit errors, partial data loss, or partial damage that may occur during the wireless transmission process. As a result, the wireless transmission module (133) performs the role of supporting high-reliability data communication required in industrial sites, going beyond simple transmission and reception functions.

[0104] Detailed explanation of the data analysis response unit (140)

[0105] The data analysis response unit (140) is a central analysis and response module that receives vibration data and temperature data wirelessly transmitted from the data transmission unit (130), quantitatively analyzes the condition of the monitored object, and comprehensively performs warning output, safety measures, history management, etc., according to the results. The data analysis response unit (140) is not a simple monitoring device, but serves as a core brain that enables the establishment of preventive maintenance strategies for the machine system and pre-failure response. In particular, it is designed to comprehensively determine complex condition changes such as bearing wear, abnormal vibration, and increased heat generation, and to perform multi-stage response logic according to preset risk levels.

[0106] The data analysis response unit (140) is configured to perform time-axis-based trend analysis on the received data. For example, vibration RMS values, peak values, vibration magnitude in a specific frequency band, average temperature values, and fluctuation ranges for each bearing location are stored at regular time intervals, and long-term trend changes are identified based on this. This structure goes beyond simply determining whether "current values ​​are high or low" and enables quantitative tracking of signs of wear that gradually increase over time, abnormal patterns that appear repeatedly, and abnormal vibrations that occur only under specific operating conditions. For example, if a pattern is detected in which the vibration magnitude in a specific frequency band gradually increases over a month in the bearing of a Dry Vacuum Pump, the data analysis response unit (140) can classify this as a sign of potential bearing wear progression and set it to a preliminary warning state.

[0107] The data analysis response unit (140) can also operate by setting risk levels in multiple stages for each monitored object. For example, stages such as Normal, Attention, Warning, and Emergency can be classified, and a standard value for the combination of vibration and temperature corresponding to each stage can be pre-set or dynamically adjusted according to the learning results of the AI ​​judgment module (141). As a specific embodiment, for a specific bearing, if the vibration RMS value rises to 1.5 times the normal average and the temperature increases slightly, it can be classified as the "Attention" stage; if the vibration peak value exceeds a certain threshold or an abnormal pattern persists in a specific high-frequency band, it can be classified as the "Warning" stage; and if an abnormal rapid rise in both vibration and temperature occurs simultaneously, it can be classified as the "Emergency" stage. This stage-based classification is used as a direct criterion for determining maintenance timing, changing operating conditions, and determining whether to stop immediately.

[0108] The data analysis response unit (140) can provide the analysis results to the operator in the form of graphs, numerical lists, warning messages, etc. For example, real-time vibration and temperature status, risk level, recent trend graphs, and warning history for each bearing can be visually displayed on a web-based monitoring screen or a local HMI screen. In one embodiment, in a semiconductor process line where multiple Dry Vacuum Pumps are installed, the data analysis response unit (140) can display the bearing status of each piece of equipment in colors (green / yellow / orange / red) so that the operator can identify the dangerous equipment at a glance. In addition, the analysis results can be stored in the form of CSV, log files, or databases and utilized later for cause of failure analysis (RCA) or equipment improvement design.

[0109] The data analysis response unit (140) operates in conjunction with the AI ​​judgment module (141), the status warning module (142), and the safety measure module (143). Based on the risk score or judgment label calculated by the AI ​​judgment module (141), the status warning module (142) can generate visual and auditory warnings to the operator, and the safety measure module (143) can be configured to execute specific control commands, such as power cutoff, emergency stop, and activation of auxiliary equipment, when necessary. For example, if the AI ​​judgment module (141) determines the dangerous state of a specific bearing as "urgent," the data analysis response unit (140) can be operated in a structure that immediately transmits the judgment result to the safety measure module (143), outputs a power cutoff signal for the equipment, and simultaneously transmits an emergency dispatch alarm to the safety manager terminal or control server.

[0110] Detailed explanation of the AI ​​judgment module (141)

[0111] The AI ​​judgment module (141) is an artificial intelligence-based analysis engine that operates inside or outside the data analysis response unit (140) and performs the role of automatically classifying and judging the condition of a monitored object using vibration and temperature data. Unlike a simple threshold comparison method, it is an advanced analysis module designed to reflect complex characteristics such as patterns over time, changes in frequency composition, and correlations between multi-sensors. Through this, it is possible to diagnose the condition and predict failures with higher precision than existing maintenance methods that relied on human experience or intuition.

[0112] The AI ​​judgment module (141) extracts various features from vibration data and utilizes them for learning and inference. In one embodiment, time domain features such as RMS, peak-to-peak value, variance, kurtosis, and skewness may be used, and frequency domain features such as energy by band centered on a specific bearing defect frequency, spectral peak position, and spectral energy distribution may be used. Additionally, temperature data may be utilized as features that reflect the progression of deterioration, such as average value, slope of rise, and maximum value over a certain period of time. These features are organized in a vector form and provided as input to the AI ​​model, and based on this, the model can determine whether the condition is normal / abnormal, early / intermediate / severe wear stage, and whether an emergency stop is required.

[0113] The AI ​​judgment module (141) can apply various forms of machine learning or deep learning models. For example, in a relatively simple embodiment, supervised learning models such as Random Forest or Support Vector Machine (SVM) can be used to classify whether it is normal or abnormal. In a more advanced embodiment, a 1D Convolutional Neural Network (CNN) can be used to directly input the time-frequency pattern of a vibration signal to classify the type of defect, or a Long Short-Term Memory (LSTM) based model can be used to learn the deterioration trend over time to predict the time of failure. In addition, if initial data is insufficient, unsupervised learning-based anomaly detection (Autoencoder, Isolation Forest, etc.) can be utilized to determine a signal that deviates from the normal pattern as an anomaly.

[0114] The AI ​​judgment module (141) has various embodiments depending on the method of configuring the learning dataset. In one embodiment, past equipment operation data and actual failure history data are collected, and a supervised learning model can be trained using label information such as the bearing replacement time, the time of failure, and the time of warning. In another embodiment, only normal operation data is sufficiently collected initially, and then an Autoencoder model that has learned normal patterns is built. Subsequently, the extent to which input data deviates from the normal pattern is evaluated as a reconstruction error, and if a certain standard is exceeded, it is determined to be an abnormal state. In this way, the AI ​​judgment module (141) can be flexibly configured according to the construction environment and the level of data acquisition.

[0115] The judgment result of the AI ​​judgment module (141) may be output simply as a binary value (normal / abnormal), or as a risk score or a probability value for a risk level. For example, a risk score from 0 to 1 may be assigned to each bearing, and the range from 0.0 to 0.3 may be mapped as normal, the range from 0.3 to 0.6 as caution, the range from 0.6 to 0.8 as warning, and 0.8 or higher as emergency, so that the data analysis response unit (140) can utilize this in the step-by-step response logic. In one embodiment, if the AI ​​judgment module (141) analyzes the bearing condition of a Dry Vacuum Pump and a risk level of 0.78 is calculated for a specific bearing, the system may classify this as a "warning level" and automatically register the priority inspection of the equipment in the maintenance plan.

[0116] Additionally, the AI ​​judgment module (141) may include a feedback learning structure to maintain and improve model performance over time. Since data from the time when a bearing is replaced or a failure occurs in the actual field can be utilized as a very important learning resource, the data analysis response unit (140) can be configured to provide this event information back to the AI ​​judgment module (141) and periodically retrain the model or update the reference value. Through such a structure, the AI ​​judgment module (141) functions as an adaptive diagnostic engine that is continuously corrected by field data, rather than a fixed model at the time of initial construction, and can effectively overcome the limitations of the conventional fixed threshold-based diagnostic method.

[0117] Detailed description of the status warning module (142)

[0118] The status warning module (142) is a module that performs the role of notifying the operator in real time of the danger status of the monitored object based on the analysis results transmitted from the AI ​​judgment module (141). It is not a simple alarm output device, but an integrated warning management module that systematically controls the information display method, warning intensity, and transmission path so that the level of risk can be intuitively recognized.

[0119] The status warning module (142) is configured to output different types of warning signals depending on the danger level. In the normal stage, only a simple status display is provided; in the caution stage, guidance focused on visual display is provided; and in the warning and emergency stages, multiple warning methods such as audible alarms, message outputs, and pop-up notifications are used in conjunction with visual signals. Through this differentiated output structure for each stage, the operator can intuitively understand the current status of the equipment.

[0120] The status warning module (142) includes a function for systematically storing warning history. For each warning, the time of occurrence, identification information of the target equipment, the warning level, and data on the basis for judgment are recorded together, and this information is used for post-analysis or maintenance history management. Through this, the history of equipment deterioration progression can be systematically managed beyond simple alarm occurrence history.

[0121] The status warning module (142) can be linked so that warning information is transmitted to a specific operator terminal or a central control server. For example, in an environment where multiple pieces of equipment are managed, such as a semiconductor process line, all warning information can be aggregated and displayed on a central control room server. This enables integrated management at the equipment group level rather than at the individual equipment level.

[0122] The status warning module (142) is configured to allow setting the warning output method and criteria according to the operating environment. Flexible management according to the operating policy is possible by setting the warning sensitivity high for specific equipment and the allowable range wide for other equipment. As a result, the problem of false alarms caused by the application of uniform thresholds as in the past can be effectively reduced.

[0123] Detailed description of the safety measure module (143)

[0124] The safety measure module (143) is an execution module that performs actual equipment control operations according to the risk judgment result transmitted from the AI ​​judgment module (141). It is a key response means that goes beyond simple warnings and prevents the spread of failures and damage to personnel and equipment by controlling the physical operating state of the machine system.

[0125] The safety measure module (143) is configured to apply different control scenarios depending on the risk level. For example, in the caution level, soft control such as easing operating conditions and reducing load may be performed, and in the warning level, some function restrictions and interlock operation induction may be performed. If determined to be an emergency level, strong control such as forced stop, power cutoff, and generation of emergency output signals is immediately performed.

[0126] The safety measure module (143) can be directly linked with the equipment control unit and can provide various types of control signals, such as relay output, digital signal output, and network-based control commands. Through this, an automatic safety control system can be established by linking the system of the present invention without modifying existing equipment.

[0127] In addition, the safety measure module (143) includes a function to record the history of emergency measures, thereby systematically managing when and what measures were executed for which equipment. This history is used as an important reference for analyzing the cause of failure, post-reporting, and establishing maintenance plans.

[0128] The safety measure module (143) may include a function to automatically notify safety managers, maintenance personnel, and emergency response personnel of emergency situations. For example, emergency notification may be made via SMS, mobile application notifications, messenger integration, email, etc., thereby preventing delays in human response. Unlike conventional technology, this configuration automates a structure that relied on manual notification methods, thereby having the effect of significantly improving the overall speed of safety response.

[0129] As described above, the present invention is fundamentally distinguished from the prior art in that it is not a structure in which the mounting unit (110), vibration detection unit (120), data transmission unit (130), and data analysis response unit (140) are merely functionally arranged, but rather forms an integrated structure in which the state of the mechanical system is measured, transmitted, analyzed, and responded to as a single signal flow. When a mechanical abnormality occurs in the mechanical system, the effect first manifests in the form of vibration, and this vibration is transmitted without deformation through the structure of the mounting unit. Subsequently, the vibration detection unit converts this vibration into an electrical signal, and the data transmission unit processes the corresponding data into a form optimized for the communication environment and stably transmits it to the analysis system. Finally, the data analysis response unit classifies the state through statistical analysis and AI judgment, and forms a closed loop by automatically linking to the warning and control stages.

[0130] The mounting part (110) serves as the starting point of this flow and acts as a mechanical reference point to minimize distortion of the physical signal known as vibration. Problems of mismeasurement caused by poor attachment, unstable contact, and deviation in installation position, which commonly occurred in conventional technology, are resolved through the structural design of the mounting part itself. The mounting part, designed to maintain fastening stability even in a repetitive vibration environment and to prevent structural deterioration during long-term operation, is a key element that determines the reliability of measurement data. Since accurate diagnosis is impossible without such a structure even if advanced analysis technology is applied later, the present invention is distinguished in that it defines the sensor position accuracy itself as a technical element.

[0131] In this structurally secured stable vibration transmission environment, the vibration detection unit (120) can detect vibration changes down to the micro level. Abnormalities such as eccentricity, cracking, lubrication deterioration, and local wear of the rotating body do not manifest merely as simple amplitude changes, but as changes in frequency components, distortion of signal patterns, and energy concentration phenomena in specific bands. The vibration detection unit functions as a data generation device capable of reflecting these complex changes, going beyond a single numerical measuring instrument. This has an essential technical difference in that, unlike the conventional method which used high values ​​as simple alarm criteria, it is a structure that secures precise data suitable for condition analysis.

[0132] The data transmission unit (130) goes beyond simply transmitting such sensor data and acts as a gateway to manage the information flow of the entire system. The communication structure, which responds only to requests from the central system, prevents unnecessary data overload and ensures data consistency by controlling the order of communication. This structure structurally suppresses data loss issues that occur due to the nature of a wireless environment and is designed to control communication more efficiently as the number of devices increases. Since a data integrity verification mechanism is built-in, the data flowing into the analysis system is limited to pre-verified data. This is an example of structurally implementing data quality assurance at the communication stage.

[0133] The data analysis response unit (140) functions not as a simple receiving device, but as an intelligent control system that interprets the meaning of sensor data and performs a response. It quantifies the equipment status based on the change trend of the vibration signal, the repeating pattern, and the correlation with temperature to classify the risk level, and this result leads to the output of a warning and the execution of a control command. In particular, the AI-based discrimination structure determines the state by learning the pattern and speed of change and long-term patterns, rather than whether a specific value is exceeded, so initial abnormalities and temporary noise are distinguished. Accordingly, the problems of false alarms and non-detection, which were the biggest problems of conventional technology, are simultaneously improved.

[0134] Consequently, the present invention provides a structure in which the entire process, from measurement to response, is automated. In the past, structures where each stage was separated and required human judgment are integrated into a single system in this invention. When an equipment malfunction occurred, the workflow in which a manager would hold a vibrator to take measurements, interpret records, and decide on actions is replaced by a data flow. This goes beyond a simple machine monitoring device and technically redefines the method of facility management itself.

[0135] Furthermore, this integrated structure can be expanded from the management of individual equipment to a comprehensive facility management system. Since vibration data collected from multiple pieces of equipment flows into a single analysis system, it enables status comparison at the equipment group level, prioritization of hazardous equipment, and the automatic establishment of predictive maintenance policies. This represents a technological advancement that overcomes the limitations of conventional technology, which managed each piece of equipment individually, by integrating the entire process into a single management unit.

[0136] In summary, the present invention is not merely a simple set of functions for individual elements, but an intelligent maintenance system in which physical signal transmission structures, data flow structures, judgment structures, and response structures are combined into a single framework. Each component is designed to complement the others, and the failure of even one component degrades overall functionality, thereby enabling the technological effect to manifest at the system level. Consequently, equipment operational stability, prediction accuracy, and maintenance efficiency are simultaneously improved, and the equipment management paradigm shifts from a reactive, post-event approach to a proactive, prevention-oriented approach.

[0137] The above detailed description of the present invention describes only specific embodiments thereof. However, it should be understood that the present invention is not limited to the specific forms mentioned in the detailed description, but rather should be understood to include all variations, equivalents, and substitutions within the spirit and scope of the invention as defined by the appended claims.

[0138] In other words, the present invention is not limited to the specific embodiments and descriptions described above, and any person skilled in the art to which the present invention pertains can make various modifications without departing from the essence of the invention as claimed in the claims, and such modifications fall within the scope of protection of the present invention. Explanation of the symbols

[0139] 100: Mechanical system vibration detection device 110: Mounting part 111: Detachable part 111a: Magnetic component 112: Vibration transmission section 113: Sensor mounting section 120: Vibration detection unit 121: Housing Unit 122: Temperature detection unit 123: Vibration detection unit 130: Data transmission section 131: Data reception module 132: Data Preprocessing Module 133: Wireless transmission module 140: Data Analysis Response Department 141: AI Judgment Module 142: Status Warning Module 143: Safety Measure Module 10: Monitoring Target

Claims

Claim 1 A mounting part (110) having a bracket structure that is bent multiple times and is mounted on one side surface of a monitoring object and has a vibration detection part (120) mounted at a predetermined distance from one side surface of the monitoring object; a vibration detection part (120) that is mounted on the mounting part (110) in a structure that is detachably mounted by a fastening member and detects the vibration of the monitoring object in real time and transmits it to a data transmission part (130); and a data transmission part (130) that wirelessly transmits the data transmitted from the vibration detection part (120) to a data analysis corresponding part (140) using a preset wireless communication means. The device includes a data analysis response unit (140) that collects data transmitted from the data transmission unit (130), analyzes the state of the monitored object based on the collected data for a preset time, and outputs the analysis results to the operator in real time; and the mounting unit (110) comprises: a detachable part (111) having a structure that is bolted to one side surface of the monitored object or attached by magnetism; and a vibration transmission part (112) having an integrated structure with the detachable part (111) and extending from the detachable part (111) by a predetermined height. and a sensor mounting part (113) of a bracket structure that is mounted in an integrated structure on the upper part of the vibration transmission part (112) and extends laterally by a predetermined width to surround the outer surface of the vibration detection part (120); wherein the vibration detection part (120) is a cylindrical structure that is mounted in a detachable structure to the sensor mounting part (113), and a housing unit (121) having a temperature detection unit (122) mounted on the upper surface of the cylindrical structure and a data cable connected to a data transmission part (130) mounted on the lower surface;A temperature detection unit (122) is mounted on the upper surface of the housing unit (121) at a predetermined distance from the vibration detection unit (123) to detect the temperature of one side surface of the object to be monitored in real time, to capture abnormal heat generation caused by preset causes including bearing wear, poor lubrication, and misalignment, and to detect not only rapid temperature changes but also gradually progressing temperature rises, and is designed to continuously transmit information to a data transmission unit (130) via a data cable so as to compare and analyze real-time temperature data with a reference temperature pattern in the operating environment during a preset period; The device includes a vibration detection unit (123) mounted inside the housing unit (121) to detect vibrations of a monitoring target in real time and to transmit the detected vibration data to a data transmission unit (130) via a data cable; wherein the data transmission unit (130) comprises: a data receiving module (131) having a plurality of connection ports capable of independently receiving a plurality of data cables connected to the vibration detection unit (120); a data preprocessing module (132) that labels and temporarily stores the data transmitted from the data receiving module (131) according to each monitoring target, detection location, detection time, and detection purpose; and a wireless transmission module (133) that transmits the data processed and temporarily stored by the data preprocessing module (132) to a data analysis corresponding unit (140) in real time using a preset wireless communication protocol.The data analysis response unit (140) includes vibration and temperature data transmitted from the wireless transmission module (133) to classify and determine the dangerous state of the monitored object according to preset danger levels, reflects complex features including patterns over time, changes in frequency configuration, and correlations between multi-sensors, learns the deterioration trend over time using an LSTM (Long Short-Term Memory) based model to predict the failure time, and if initial data is insufficient, uses an unsupervised learning-based anomaly detection (Autoencoder, Isolation Forest) to determine a signal deviating from the normal pattern as an anomaly, and initially collects only normal operation data sufficiently, then builds an Autoencoder model that learns the normal pattern, evaluates how much the subsequently input data deviates from the normal pattern as a reconstruction error, and determines an abnormal state if it exceeds a certain standard; the AI ​​judgment module (141) outputs a preset warning signal based on the result determined by the AI ​​judgment module (141) and stores data related to the history of the warning signal. A mechanical system vibration detection device characterized by including: a safety measure module (143) that, based on the result determined by the AI ​​judgment module (141), cuts off the power supplied to the monitored object, urgently activates emergency safety equipment, and reports to the safety manager and emergency dispatch personnel when rapid emergency measures are required. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete

Images