Predictive Maintenance Solution for Automatic Lubrication Systems of Construction Machinery Utilizing Pressure Sensors and AI

Abstract

A method for providing a predictive maintenance solution for an automatic lubrication system of a machine device utilizing a pressure sensor and AI is disclosed. The method for providing a predictive maintenance solution for an automatic lubrication system of a machine device utilizing a pressure sensor and AI relates to a method for providing a predictive maintenance solution that utilizes a pressure sensor and artificial intelligence (AI) to analyze lubrication supply pressure data generated in the automatic lubrication system of a machine device in real time, and predicts and notifies the time when maintenance is required based on the results. Time-series data of pressure signals formed during the lubrication supply cycle are collected through a pressure sensor installed at the entrance of a distributor, and the minimum value, maximum value, average value, increase / decrease interval, and wave period of the pressure wave are extracted from the said data to calculate feature quantities. The calculated wave data is input into a pre-trained AI-based anomaly judgment model and classified into at least one maintenance state among normal, caution, warning, and service required. According to this method, moving away from the conventional simple supply determination method based on proximity sensors, quantitative maintenance diagnosis based on actual lubrication pressure fluctuations becomes possible. Furthermore, by providing only user notifications during the warning stage and performing control to stop the operation of the automatic lubrication device during the service requirement stage, equipment damage and failures caused by poor lubrication can be prevented in advance.

Description

Technology Field

[0001] The present invention relates to an automatic lubrication system for mechanical devices, and more specifically to a method for providing a predictive maintenance solution that predicts the lubrication status using an artificial intelligence model based on a lubricating oil supply pressure signal collected through a pressure sensor, and notifies the user of the time when maintenance is required according to the prediction result. Background Technology

[0002] Conventional automatic lubrication systems applied to machinery have primarily utilized a method of detecting the reciprocating motion of spool pins by installing proximity sensors inside the distributor. However, proximity sensors can only determine whether lubrication has been performed and have limitations in precisely identifying pressure changes during the actual lubrication supply process or abnormal signs occurring within the line.

[0003] In particular, the pressure waveform formed during lubrication supply exhibits a relatively constant pattern of increase and decrease under normal conditions; however, when abnormal situations occur—such as distributor blockage, lubrication point leakage, line obstruction, grease hardening, or viscosity changes—the waveform period, maximum and minimum pressures, and slope of increase and decrease become irregularly deformed. However, since conventional proximity sensor methods cannot measure or analyze these changes in pressure waveforms, there was a response delay issue where problems were only recognized after lubrication abnormalities had actually accumulated.

[0004] Furthermore, since the proximity sensor method relies on mechanical motion detection, it could not distinguish situations where the actual supply pressure inside the lubrication line was insufficient or where grease was not delivered to specific points, even when the distributor was performing reciprocating motion. As a result, lubrication failures could remain undetected for extended periods, leading to frequent issues with reduced operational reliability, such as bearing damage, friction surface deterioration, power loss, and unexpected equipment stoppages.

[0005] Therefore, conventional automatic lubrication systems lacked the technical foundation to precisely detect and judge the lubrication supply status itself, going beyond the level of simple operation verification, and had structural limitations in that signs of lubrication abnormalities could not be detected in advance.

[0006] In the utilization of automatic lubrication systems focused on user convenience, the aforementioned problems acted as a limiting factor in the growth and demand meeting of automatic lubrication systems, due to reasons such as the inability to provide precise predictive maintenance from the perspectives of user convenience and the management and maintenance of automatic machinery. The problem to be solved

[0007] The problem that an embodiment of the present invention aims to solve is to provide a method for providing a predictive maintenance solution for an automatic lubrication system of a machine device, which implements a predictive maintenance system capable of early prediction of abnormal conditions and advance notification to the user of the time when maintenance is required by quantitatively determining the lubrication quality and supply status based on the minimum pressure, maximum pressure, increase / decrease pattern, and periodic variation of pressure waves, and learning and applying the changes to an artificial intelligence model, in order to overcome the limitations of the proximity sensor-based lubrication detection method described above. means of solving the problem

[0008] According to one embodiment, a method for providing a predictive maintenance solution for an automatic lubrication system of a machine device utilizing a pressure sensor and AI comprises: a pressure data collection step of collecting time-series data of a pressure signal generated during lubricant supply per unit time from a pressure sensor included in an automatic lubrication device that supplies lubricant to a plurality of lubrication points through a distributor; a wave data extraction step of extracting information on the minimum value, maximum value, average value, increase / decrease interval, and wave period of a pressure wave formed during a lubricant supply cycle from the time-series data of the collected pressure signal; and a prediction step of inputting the wave data into a pre-trained AI-based anomaly judgment model to calculate at least one state value among a normal stage, a caution stage, a warning stage, and a service required stage. The present invention provides a predictive maintenance solution for an automatic lubrication system of a machine utilizing a pressure sensor and AI, comprising: a solution providing step in which, when the state value of the prediction step is calculated as a warning step or a failure step, a pre-inspection notification or a maintenance recommendation notification is generated and provided to the user, wherein in the warning step, the user is guided to the time when maintenance is required by a method that does not directly control the operation of the distributor, and in the failure step, the user is guided to the time when maintenance is required by a method that directly controls the operation of the distributor to stop.

[0009] In addition, the present invention processes time-series data by distinguishing between a standby section where lubrication is not supplied and an operating section where an actual pressure waveform is formed during the pressure wave analysis process. The wave data extraction step is controlled to collect the time-series data by distinguishing the temporal section of the pressure signal into a standby section where no wave appears and an operating section where a specific waveform appears. The prediction step detects irregular pressure changes when the minimum, maximum, or average value of the pressure wave in the operating section among the wave data deviates from a preset range, or when the length of the wave period or the pattern of the increase / decrease section does not match a pre-learned regularity. It can then determine to classify the irregular pressure change into one of a number of state values, such as a normal stage, a caution stage, a warning stage, or a service required stage.

[0010] In addition, the present invention can calculate the time when maintenance is required by subdividing the condition grade based on the range of change of the maximum value of the extracted pressure wave. The prediction step can be controlled such that if the maximum value of the pressure wave extracted from the wave data is 0 bar or more and less than 50 bar, it is determined to be a lubricating oil depletion or pump abnormal condition and set to the service required stage; if the maximum value of the pressure wave extracted from the wave data is 50 bar or more and less than 100 bar, it is set to a normal stage where no maintenance is required; if the maximum value of the pressure wave extracted from the wave data is greater than 100 bar and less than 150 bar, it is set to the caution stage where maintenance is expected to be required within a first time point; if the maximum value of the pressure wave extracted from the wave data is greater than 150 bar and less than 200 bar, it is set to the warning stage where maintenance is expected to be required within a second time point, which is closer than the first time point based on the current time point; and if the maximum value of the pressure wave extracted from the wave data exceeds 200 bar, it is set to the service required stage where immediate maintenance is required.

[0011] In addition, the present invention introduces a temperature-based correction factor to reflect the fact that pressure wave characteristics may vary depending on changes in lubricant viscosity and seasonal temperature differences. The prediction step is controlled to calculate the state value by applying an ambient temperature-specific correction factor to further reflect graph characteristics in which the amplitude and period of the wave change according to seasonal changes. The ambient temperature-specific correction values ​​include a first correction factor applied in a first temperature range corresponding to 15 ℃ to 25 ℃, a second correction factor applied in a second temperature range corresponding to 0 ℃ to 14 ℃, a third correction factor applied in a third temperature range corresponding to less than 0 ℃, and a fourth correction factor applied in a fourth temperature range corresponding to 26 ℃ or higher. The first to fourth correction factors can be adjusted to have values ​​between 0.8 and 1.2 by reflecting the temperature-specific viscosity according to the type of lubricant.

[0012] In addition, to control the prediction step to estimate the cause of the state value fluctuation rather than merely determining the state value, the present invention may further control the prediction step to predict the cause of the calculated state value. The prediction step may identify the cause by controlling it to determine that there is insufficient supply pressure or a decrease in lubricating oil flow rate if the minimum value of the pressure wave extracted from the wave data is lower than the normal range; determine that there is a long-term abnormal pressure accumulation state if the average value of the pressure wave extracted from the wave data deviates from the reference range; determine that there is a lubricating oil flow obstruction if the change in the slope of the increase / decrease section of the pressure wave extracted from the wave data differs from a pre-learned normal pattern; and determine that there is a stroke failure of the distributor or an increase in mechanical load if the wave period of the pressure wave extracted from the wave data becomes shorter or longer than the reference period. Effects of the invention

[0013] According to one embodiment, the present invention collects and analyzes pressure waves formed during the actual lubrication supply process through a pressure sensor as time-series data, thereby enabling precise diagnosis of detailed fluctuations in the lubrication state that could not be detected by conventional proximity sensor methods. In particular, as maintenance notifications and drive control methods are provided stepwise according to normal, caution, warning, and service required stages, the risk of equipment damage caused by the accumulation of lubrication abnormalities during operation can be prevented in advance, and a predictive maintenance effect can be secured to predict and manage maintenance timing without interrupting actual operation.

[0014] Furthermore, by analyzing pressure waveforms by dividing them into ambient and operating phases, and determining abnormalities based on minimum, maximum, and average values, slopes of increase and decrease, and the regularity of wave periods, internal defects such as lubrication point deviations, distributor stroke imbalances, line leakage, or blockages can be identified early. This allows for a significant improvement in diagnostic accuracy based on actual lubrication quality and delivery performance, moving beyond the conventional method of simply verifying supply status.

[0015] Furthermore, by classifying maximum pressure wave values ​​by range and tiering maintenance priorities, it provides maintenance decision-making criteria based on the severity of maintenance requirements, rather than simple warning notifications. In other words, since situations requiring caution, warning, or emergency action are clearly distinguished alongside normal conditions, the likelihood of failure propagation is reduced, and the timing of spare parts procurement and maintenance deployment can be managed efficiently.

[0016] In addition, by applying a correction factor that reflects the characteristics of ambient temperature and lubricant viscosity changes, fluctuations in pressure waveforms due to changes in seasons and environmental conditions are prevented from causing errors in judgment. Accordingly, stable lubrication condition diagnostic criteria can be maintained that are not affected by external factors such as lubricant hardening in extreme cold or viscosity reduction at high temperatures, and the reproducibility and reliability of diagnostic results can be ensured even if the equipment operating environment differs.

[0017] Furthermore, the present invention goes beyond classifying state values ​​to estimate the causes of state changes, thereby maximizing maintenance efficiency. Since it provides diagnostic functions for the causes of abnormal conditions—such as detecting supply pressure drops through a decrease in minimum values, identifying long-term abnormal accumulation based on deviations from average values, predicting flow failures through deviations in increase / decrease slopes, and determining signs of distributor stroke abnormalities through periodic variations—it is possible to move beyond simple warning-based maintenance and implement a high-resolution predictive maintenance system with clear grounds for maintenance decisions. Brief explanation of the drawing

[0018] FIG. 1 is an overall flowchart illustrating a method for providing a predictive maintenance solution for an automatic lubrication system of a machine using a pressure sensor and AI according to one embodiment. FIG. 2 is a flowchart illustrating a wave data extraction step according to an embodiment. FIG. 3 is a flowchart illustrating a method for predicting abnormalities according to one embodiment. FIG. 4 is a flowchart illustrating a method for classifying service stages based on the maximum value of wave data according to one embodiment. FIGS. 5 and 6 are drawings for explaining the components of an automatic lubrication device according to one embodiment. FIG. 7 is a drawing illustrating the attachment location of an automatic lubrication device and a distributor according to one embodiment. FIG. 8 is a drawing illustrating the pressure sensor placement structure of an automatic lubrication device according to another embodiment. Specific details for implementing the invention

[0019] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0020] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0021] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0022] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or coupled with that other component, or that there may be other components in between.

[0023] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0024] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0025] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0026] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. 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.

[0027] In the embodiments of the present invention, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in the embodiments of the present invention.

[0028] The shapes, sizes, ratios, angles, numbers, etc. disclosed in the drawings for explaining embodiments of the present invention are exemplary, and therefore the present invention is not limited to the depicted details. Furthermore, in describing the present invention, if it is determined that a detailed description of related known technology may unnecessarily obscure the essence of the present invention, such detailed description is omitted. Where terms such as "includes," "has," or "is made up" are used in this specification, other parts may be added unless "only" is used. Where a component is expressed in the singular, it includes cases where it includes the plural unless specifically stated otherwise.

[0029] In interpreting the components, they are interpreted to include a margin of error even in the absence of a separate explicit statement.

[0030] In the case of describing a positional relationship, for example, when the positional relationship between two parts is described using expressions such as 'on,' 'upper,' 'lower,' or 'next to,' one or more other parts may be located between the two parts unless 'immediately' or 'directly' is used.

[0031] When elements or layers are referred to as "on" another element or layer, this includes cases where another layer or element is placed directly on top of or in between. Throughout the specification, the same reference numerals refer to the same components.

[0032] The size and thickness of each component shown in the drawings are illustrated for convenience of explanation, and the present invention is not necessarily limited to the size and thickness of the illustrated components.

[0033] The features of each of the various embodiments of the present invention may be combined or combined with one another, either partially or wholly, and as will be fully understood by those skilled in the art, various technical interlocking and operation are possible, and each embodiment may be implemented independently of one another or together in an interlocking relationship.

[0034] FIG. 1 is an overall flowchart illustrating a method for providing a predictive maintenance solution for an automatic lubrication system of a machine using a pressure sensor and AI according to one embodiment.

[0035] Referring to FIG. 1, a method for providing a predictive maintenance solution for an automatic lubrication system of a machine using a pressure sensor and AI according to exemplary embodiments of the present invention may include a time series data collection step (S100), a wave data extraction step (S110), a prediction step (S120), and a solution provision step (S130).

[0036] The time series data collection step (S100) may be a step of collecting time series data of pressure signals generated during lubricant supply per unit time from a pressure sensor included in an automatic lubrication device that supplies lubricant to a plurality of lubrication points through a distributor.

[0037] The wave data extraction step (S110) may be a step of extracting information on the minimum value, maximum value, average value, increase / decrease interval, and wave period of the pressure wave formed during the lubricating oil supply cycle from the time series data of the collected pressure signal.

[0038] The prediction step (S120) may be a step of inputting the wave data into a pre-trained AI-based anomaly judgment model to calculate at least one state value among a normal stage, a caution stage, a warning stage, and a service required stage.

[0039] The above AI-based anomaly detection model may be a model trained using at least one of supervised learning, unsupervised learning, or semi-supervised learning methods, and in one embodiment, pressure waveform data in a normal lubrication state and pressure waveform data in various failure and abnormal states may be used as training data.

[0040] In one embodiment, the prediction step (S120) can be performed by detecting an irregular pressure change when the minimum, maximum, or average value of the pressure wave in the operating section among the wave data deviates from a preset range, or when the length of the wave period or the pattern of the increase / decrease section does not match a pre-learned regularity.

[0041] The above prediction step (S120) can be controlled to classify multiple maintenance states in stages based on the maximum value of the pressure wave, and the maximum value criterion applied thereto may be a value to which a correction factor reflecting the ambient temperature, lubricating oil viscosity, equipment usage environment, etc. is applied.

[0042] Additionally, the prediction step (S120) can be performed by determining to classify the irregular pressure change into one of the following: a normal state, a caution state, a warning state, and a service required state among a plurality of state values.

[0043] The solution provision step (S130) may be a step of generating and providing a pre-inspection notification or a maintenance recommendation notification to the user when the status value of the prediction step is calculated as a warning step or a failure step, wherein in the warning step, the user is guided to the time when maintenance is required in a manner that does not directly control the operation of the distributor, and in the failure step, the user is guided to the time when maintenance is required in a manner that directly controls the operation of the distributor to stop.

[0044] FIG. 2 is a flowchart illustrating a wave data extraction step according to an embodiment.

[0045] Referring to FIG. 2, the wave data extraction step (S110) according to exemplary embodiments of the present invention may be a step comprising a step of distinguishing between a waiting period and an operating period (S111), a step of determining whether it is an operating period (S112), a wave calculation step (S113) of calculating pressure wave characteristic values ​​such as a minimum value, a maximum value and an average value, or a step of extracting wave data (S114).

[0046] In the step (S111) of distinguishing between the standby section and the operating section, the standby section may be defined as a section where the pressure change is maintained below a reference threshold, and the operating section may be distinguished by comparing a simple threshold, detecting a slope change, or comparing with previous lubrication cycle data, as a section where a pressure waveform is repeatedly formed by pump driving or distributor stroke.

[0047] If the above wave data is confirmed to be an operating section, feature values ​​such as the minimum value, maximum value, average value, slope of the increase / decrease section, and wave period of the pressure wave can be calculated within the section, and the feature values ​​can be stored as structured wave data so that they can be used as input data in the subsequent prediction step (S120).

[0048] The calculation step (S113) can be performed by extracting the minimum, maximum, or average value of the pressure wave from the time series data in the operation interval.

[0049] The above wave data extraction step (S110) can be controlled to collect the time series data by dividing the time interval of the pressure signal into a waiting interval where no wave appears and an operating interval where a wave of a specific waveform appears.

[0050] FIG. 3 is a flowchart illustrating a method for predicting abnormalities according to one embodiment.

[0051] Referring to FIG. 3, the wave data input step (S121) according to exemplary embodiments of the present invention receives the wave data extracted in the wave data extraction step (S110), the abnormality determination step (S122) detects an irregular pressure change when the minimum value, maximum value, or average value of the pressure wave in the operating section among the wave data deviates from a preset range, or when the length of the wave period or the pattern of the increase / decrease section does not match a pre-learned regularity, and the state value calculation step (S123) can be controlled in a manner that determines to classify the irregular pressure change into one of a normal stage, a caution stage, a warning stage, and a service required stage among a plurality of state values.

[0052] FIG. 4 is a flowchart illustrating a method for classifying service stages based on the maximum value of wave data according to one embodiment.

[0053] Referring to FIG. 4, in one embodiment, the abnormality determination step (S122) may consist of a step of extracting the maximum value of the pressure wave in the operation section among the wave data (S1221), a step of reflecting a correction coefficient of the maximum value of the pressure wave (S1222), a step of determining a state value based on the maximum value reflecting the correction coefficient (S1223), and a service step classification step (S1224) of classifying an appropriate service provision step by determining the state value.

[0054] In one embodiment, the step (S1222) of reflecting the correction factor of the maximum value of the pressure wave is controlled to calculate the state value by applying an additional correction factor for each ambient temperature to further reflect the graph characteristics in which the amplitude and period of the wave change according to seasonal changes. In one embodiment, the correction value for each ambient temperature includes a first correction factor applied in a first temperature range corresponding to 15 ℃ to 25 ℃, a second correction factor applied in a second temperature range corresponding to 0 ℃ to 14 ℃, a third correction factor applied in a third temperature range corresponding to less than 0 ℃, and a fourth correction factor applied in a fourth temperature range corresponding to 26 ℃ or higher. The first to fourth correction factors can be adjusted to have values ​​between 0.8 and 1.2 by reflecting the viscosity for each temperature according to the type of lubricant.

[0055] The correction values ​​for each ambient temperature mentioned above are not absolutely limited to the given figures and may include a range that can be easily understood by a person with ordinary knowledge in the relevant technical field. Furthermore, to apply the correction values ​​to various mechanical devices, they may be applied as correction coefficients tailored to the unique driving pressure of the mechanical device.

[0056] In one embodiment, the step of determining the state value (S1223) can be controlled such that if the maximum value of the pressure wave extracted from the wave data is 0 bar or more and less than 50 bar, it is determined to be a lubricating oil depletion or pump abnormal state and set to the service required stage; if the maximum value of the pressure wave extracted from the wave data is 50 bar or more and less than 100 bar, it is set to a normal stage where no maintenance is required; if the maximum value of the pressure wave extracted from the wave data is greater than 100 bar and less than or equal to 150 bar, it is set to the caution stage where maintenance is expected to be required within a first time point; if the maximum value of the pressure wave extracted from the wave data is greater than 150 bar and less than or equal to 200 bar, it is set to the warning stage where maintenance is expected to be required within a second time point that is closer than the first time point based on the current time point; and if the maximum value of the pressure wave extracted from the wave data exceeds 200 bar, it is set to the service required stage where immediate maintenance is required.

[0057] The maximum value of the step (S1223) for determining the above state value may be calculated differently as specific numerical values ​​are reflected in the first to fourth correction coefficients, and is not absolutely limited to such numerical values, but may include values ​​within a range that can be easily understood and set by a person with ordinary knowledge in the relevant technical field.

[0058] The prediction step (S120) is controlled to further predict the cause of the calculated state value, and the prediction step may be controlled to determine that there is a supply pressure shortage or a decrease in lubricating oil flow rate if the minimum value of the pressure wave extracted from the wave data is lower than the normal range, determine that there is a long-term abnormal pressure accumulation state if the average value of the pressure wave extracted from the wave data deviates from the reference range, determine that there is a lubricating oil flow obstruction if the change in the slope of the increase / decrease section of the pressure wave extracted from the wave data differs from a pre-learned normal pattern, and determine that there is a stroke failure of the distributor or an increase in mechanical load if the wave period of the pressure wave extracted from the wave data becomes shorter or longer than the reference period.

[0059] In addition, if the wave period becomes abnormally short or long, the cause may be identified as a distributor stroke failure, valve stiction, or blockage of specific lubrication points, and the result of the cause prediction can be included in a user notification message.

[0060] FIGS. 5 and 6 are drawings for explaining the components of an automatic lubrication device according to one embodiment.

[0061] Referring to FIG. 5, in one embodiment, the automatic lubrication device (1) may consist of a pump (10), a lubricating oil storage unit (20), a lubricating oil supply line (30), a distributor (40), a lubricating oil discharge line (50), a drive control module (60), and a pressure sensor (70).

[0062] Referring to FIG. 6, in one embodiment, the automatic lubrication device (1) may be composed of a first unit including a pump (10), a lubricating oil storage unit (20), a lubricating oil supply line (30), and a drive control module (60), and a second unit including a lubricating oil supply line (30), a distributor (40), a lubricating oil discharge line (50), and a pressure sensor (70).

[0063] The first unit above can store lubricating oil and operate to distribute the lubricating oil to parts of the machine requiring lubrication through a pump and supply line according to a signal from the drive control module, and the second unit can supply the supplied lubricating oil to each lubrication discharge line through a distributor and measure the pressure of the supplied lubricating oil and transmit data.

[0064] FIG. 7 is a drawing illustrating the attachment location of an automatic lubrication device and a distributor according to one embodiment.

[0065] Referring to FIG. 7, in one embodiment, the second unit may be controlled to receive lubricating oil from the first unit and to properly discharge the lubricating oil through the distributor (40), and the second unit may be attached close to a location where lubrication of the machine is required. The attachment location may be a location within a range that can be easily understood and applied by a person skilled in the art.

[0066] FIG. 8 is a drawing illustrating the pressure sensor placement structure of an automatic lubrication device according to another embodiment.

[0067] Referring to FIG. 8, in another embodiment, the pressure sensor of the automatic lubrication device may be configured in either a method of measuring pressure by being connected to a lubricating oil discharge line (a) or a method of measuring pressure by being directly attached to a distributor (40) (b), and may also be configured to operate by being attached to the bonding portion between the lubricating oil supply line and the distributor. The attachment site of the pressure sensor is not limited to the above embodiments and may be a site within a range that can be easily utilized by a person skilled in the art.

[0068] Additionally, one or more pressure sensors (70) may be installed in a single distributor, and in another embodiment, they may be individually installed in each of the multiple lubricating oil discharge lines at the bottom of the distributor (40) to individually detect the pressure status of the lubricating oil discharge lines.

[0069] The predictive maintenance solution according to the present invention may be composed of an automatic lubrication device mounted on a mechanical device, a pressure sensor installed in a distributor of the device, a data processing control unit, a communication module, and a user terminal interface.

[0070] The automatic lubrication device may include a pump, a lubricating oil storage unit, a lubricating oil supply line, a distributor, a lubricating oil discharge line, a drive control module, and a pressure sensor. The pressure sensor may be installed at the distributor inlet, an internal port of the distributor, or the main line, and can measure pressure signals in real time during the progress of the lubrication cycle.

[0071] During lubrication supply, pressure signals are collected from the pressure sensor at unit time intervals, and the pressure signals are stored as time-series data after analog-to-digital conversion, and noise removal filtering, sensor drift correction, and reference level normalization can be performed.

[0072] The control unit divides the time axis of the lubrication cycle into a standby section and an operating section, designating only the actual section where the pressure waveform is formed as the analysis target. This prevents misjudgment caused by ambient noise and enables the clear extraction of wave characteristics.

[0073] In one embodiment, the control unit may be implemented inside the mechanical device ECU or as a separate independent control module, and the communication module is connected to a user terminal or remote server via CAN, RS485, LTE, Wi-Fi, etc., and notifications or status guidance may be provided through the equipment display or mobile terminal application.

[0074] Characteristics such as minimum, maximum, and average values, waveform period, rise and fall slopes, and amplitude fluctuation rates are calculated from the pressure waveform corresponding to the operating range. The maximum value reflects the instantaneous load condition within the lubrication line, and the minimum value indicates the stability of the lower limit of the supply pressure. The average value can be used to determine whether the system pressure is balanced over the long term.

[0075] In addition, the wave period is used to determine the distributor spool operating cycle and whether there is a point supply delay, and changes in the slope of the rising and falling sections can establish criteria for determining changes in lubricant viscosity, line friction resistance, and distributor stroke deviation.

[0076] The control unit can convert the above feature quantity into a data vector form and perform normalization and dimensionality reduction processing so that it can be input into an AI-based diagnostic model. At least one of a multilayer perceptron, CNN, LSTM, GRU, or a complex hybrid structure can be applied as an AI model for anomaly judgment. The model can receive a pressure wave feature vector as input and classify it into at least one of a normal stage, a caution stage, a warning stage, or a service required stage.

[0077] The status value is finally determined by including not only single-cycle results but also multi-cycle-based trend, mean deviation, and volatility analysis results. If necessary, hysteresis logic can be used to suppress false positives caused by transient spikes.

[0078] Pressure fluctuations that differ from the reference range or learned normal patterns are defined as irregular waves, and waveform distortion, periodic deformation, and amplitude deviation are detected. Such irregular pressure changes can be utilized for the early detection of internal structural abnormalities, such as lubrication point blockage, line leakage, distributor stroke failure, and lubricant viscosity hardening.

[0079] If the maximum pressure wave value falls within the normal range, it is set to the Normal stage; if the maximum value exceeds the critical range, it is subdivided into Caution, Warning, or Service Required stages. In the Service Required stage, operation stop control may be performed, whereas in the Warning stage, only notification services are provided while system operation is maintained. If the prediction result is in the Warning stage, a maintenance recommendation notification may be provided to the user; conversely, in the Service Required stage, control to stop the operation of the distributor may be performed. The notification delivery method can be configured based on multiple channels, such as equipment displays, remote control centers, and mobile applications. Additionally, it is possible to provide information for establishing maintenance plans, such as estimating remaining allowable operating time and linking with preventive maintenance scheduling.

[0080] In other words, the service requirement stage, which is a higher level than the warning stage, may include steps to protect users from situations where operating the system could be extremely dangerous by restricting or prohibiting system operations. In this stage, the necessity of system maintenance may be conveyed to users, etc., by providing notifications at a level higher than the warning stage, depending on system settings, etc.

[0081] Furthermore, the present invention applies a temperature correction factor to exclude the influence of ambient temperature changes on the pressure waveform. The correction factor is pre-stored as a value reflecting the viscosity characteristics of the lubricant for each temperature range, thereby stably maintaining normal diagnostic criteria even in sites with significant temperature fluctuations (such as winter or extreme heat). If the type of lubricant differs, the difference in the viscosity-temperature curve can be corrected by modifying the coefficient table.

[0082] In addition, to ensure universality applicable to various mechanical devices, the present invention may apply a correction factor tailored to the unique driving pressure according to the type or characteristics of the mechanical device to which the present invention is applied. Since driving pressure may vary from device to device and the normal driving pressure range in one mechanical device may differ from the normal driving pressure range in another machine, the accuracy of predictive maintenance of the mechanical device can be improved by correcting this value through the correction factor. For example, if the driving pressure considered as a standard is set to 50 bar or more and 100 bar or less, and the normal driving pressure of the first mechanical device is measured to be 50 bar or more and 100 bar or less, this is the driving pressure considered as a standard in this embodiment, so the application of a correction factor is not necessary. However, if the normal driving pressure of the second mechanical device is measured to be 55 bar or more and 110 bar or less, this differs from the driving pressure considered as a standard in this embodiment, so the correction factor can be applied to correct it to the driving pressure considered as a standard in this embodiment, which is 50 bar or more and 100 bar or less. As another example, if the normal driving pressure of the third machine is measured to be 40 bar or more and 95 bar or less, it differs from the driving pressure considered as a standard in this embodiment, and this can also be corrected to the driving pressure considered as a standard in this embodiment. Although the 'driving pressure considered as a standard' above is given as an example of 50 bar or more and 100 bar or less, it may also be a driving pressure standard within a range that can be easily utilized by a person with ordinary knowledge, and may not be limited to the numerical value of 50 bar or more and 100 bar or less.

[0083] After calculating the above status values, the system can predict the causes of abnormalities and provide them to the user. For example, a decrease in the minimum status value may indicate insufficient supply pressure or line leakage; a long-term deviation from the average value may indicate increased internal friction or blockage in the distributor; abnormal changes in the cycle may indicate spool stroke abnormalities or mechanical interference; and a decrease in the upward slope may indicate increased lubricant viscosity or flow path resistance. This enables the provision of maintenance guidance rather than simple abnormality detection, and accuracy can be enhanced through the continuous increase of training data.

[0084] As described above, pressure waveform data and maintenance history are accumulated on a server to continuously increase training data, thereby periodically improving the performance of the AI ​​model. The latest model can be applied to the equipment via OTA updates, and a comparative verification of false positive and false negative rates is performed before and after the update.

[0085] Pressure sensors can be installed at multiple locations and may include a fault location estimation function based on the comparison of pressure deviations across sections. Additionally, some computations can be performed on a cloud server, while only a lightweight judgment module may be installed on the equipment side. The wave analysis method also allows for further variations, such as frequency analysis, wavelet transform, and entropy analysis.

[0086] In one embodiment, the pressure sensor may be attached to the inlet side of the distributor, but the pressure sensor may also be installed at multiple points and may include a function to estimate the fault location through the comparison of pressure deviations by section. Additionally, some calculations may be performed on a cloud server, and only a lightweight judgment module may be installed on the equipment side. The wave analysis method may also be further modified and used, such as frequency analysis, wavelet transform, or entropy analysis. The possibility for such additional modifications may include the scope that a person with ordinary knowledge in the relevant technical field can generally understand and infer.

[0087] Additionally, in one embodiment, the pressure sensor may be installed at a specific point between the pump (10) and the distributor (40), and the point where it may be installed may be a point within the range that a person with ordinary knowledge can generally understand and infer.

[0088] The sampling period of pressure data can be variably set according to the lubrication cycle, pump driving time, distributor stroke speed, etc., and, for example, can be controlled to apply a relatively high sampling period while lubrication is being performed and a low sampling period during standby.

[0089] Although embodiments of the present invention have been described in more detail with reference to the attached drawings, the present invention is not necessarily limited to these embodiments and may be modified in various ways within the scope of the technical spirit of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. The scope of protection of the present invention shall be interpreted by the claims below, and all technical spirits within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention.

[0090] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below. Explanation of the symbols

[0091] 1 : Automatic lubrication device 10 : Pump 20: Lubricant reservoir 30: Lubricant supply line 40 : Distributor 50: Lubricant drain line 60: Drive control module 70: Pressure sensor S100: Pressure data collection step for collecting pressure signals generated during lubrication supply at unit time intervals from a pressure sensor coupled to the inlet of the distributor. S110: Wave data extraction step in which wave information including maximum value, minimum value, average value, period, and increase / decrease pattern is extracted from a collected pressure signal. S111: Standby / operation section distinction step that distinguishes between the lubrication standby state and the supply operation state S112: Step confirming that a wave is generated and that it is an operating section S113: Wave calculation step for calculating pressure wave characteristic values ​​such as minimum, maximum, and average values S114: Step for extracting wave data S120: Prediction step that classifies maintenance needs by inputting wave data into an AI model S121: Step of inputting wave data into the AI ​​model S122: A step for analyzing input wave data to determine whether there is an anomaly. S123: Step for calculating the state value S1221: Step to extract the maximum value from the input wave data S1222: Step of applying a correction factor to the maximum value S1223: Step of determining the state value based on the corrected maximum value S1224: A step for classifying service stages based on status values S130: Solution provision step that provides user inspection notifications, maintenance recommendations, or automatic drive stop commands based on classified status values.

Claims

Claim 1 A method for providing a predictive maintenance solution for an automatic lubrication system of a machine device utilizing a pressure sensor and AI, comprising: a pressure data collection step of collecting time-series data of pressure signals generated during lubricant supply per unit time from a pressure sensor included in an automatic lubrication device that supplies lubricant to a plurality of lubrication points through a distributor; a wave data extraction step of extracting information on the minimum value, maximum value, average value, increase / decrease interval, and wave period of pressure waves formed during a lubricant supply cycle from the time-series data of the collected pressure signals; and a prediction step of inputting the wave data into a pre-trained AI-based anomaly judgment model to calculate at least one state value among a normal stage, a caution stage, a warning stage, and a service required stage. A method for providing a predictive maintenance solution for an automatic lubrication system of a machine utilizing a pressure sensor and AI, comprising: a solution providing step in which, when the state value of the prediction step is calculated as a warning step or a failure step, a pre-inspection notification or a maintenance recommendation notification is generated and provided to the user, wherein in the warning step, the user is guided to the time when maintenance is required by a method that does not directly control the operation of the distributor, and in the failure step, the user is guided to the time when maintenance is required by a method that directly controls the operation of the distributor to stop; wherein the wave data extraction step is controlled to collect the time series data by dividing the temporal interval of the pressure signal into a standby section where no wave appears and an operating section where a wave of a specific waveform appears; and the prediction step detects as an irregular pressure change the case in which the minimum value, maximum value, or average value of the pressure wave in the operating section among the wave data deviates from a preset range, or the length of the wave period or the pattern of the increase / decrease section does not match a pre-learned regularity, and determines to classify the irregular pressure change into one of a normal stage, a caution stage, a warning stage, or a service required stage among a plurality of state values. Claim 2 delete Claim 3 A method for providing a predictive maintenance solution for an automatic lubrication system of a machine using a pressure sensor and AI according to claim 1, wherein the prediction step is controlled to determine a state of depletion of lubricating oil or pump malfunction and set to the service-required step when the maximum value of the pressure wave extracted from the wave data is 0 bar or more and less than 50 bar; to set to a normal step where no maintenance is required when the maximum value of the pressure wave extracted from the wave data is 50 bar or more and less than 100 bar; to set to the caution step where maintenance is expected to be required within a first time point when the maximum value of the pressure wave extracted from the wave data is 100 bar or more and less than 150 bar; to set to the warning step where maintenance is expected to be required within a second time point, which is closer than the first time point based on the current time point, when the maximum value of the pressure wave extracted from the wave data is 150 bar or more and less than 200 bar; and to set to the service-required step where immediate maintenance is required when the maximum value of the pressure wave extracted from the wave data exceeds 200 bar.

Images