Failure sign detection system, failure sign detection method, and failure sign detection program

Abstract

The present disclosure provides a technology capable of automatically detecting deviations from the normal state of equipment without requiring a user to set a monitoring threshold value in advance.　The present technology provides a failure sign detection system comprising: a degree-of-abnormality calculation unit that calculates the degree of abnormality of equipment on the basis of a vibration acceleration frequency spectrum included in vibration acceleration data of the equipment; a state change detection unit that detects a state change of the equipment on the basis of the root mean square (RMS) of a vibration acceleration included in the vibration acceleration data; and a failure prediction unit that detects a sign of a failure of the equipment on the basis of the degree of abnormality and the state change.

Description

Failure sign detection system, failure sign detection method, and failure sign detection program

[0001] The technology disclosed herein (hereinafter also referred to as "the technology") relates to a failure sign detection system, a failure sign detection method, and a failure sign detection program.

[0002] Conventionally, monitoring is performed based on data such as vibration, current, and temperature sensed from the equipment, and deviations from these indicators are detected as signs of failure based on monitoring thresholds set by the user.

[0003] For example, Patent Document 1 discloses a technology relating to "a pump abnormality detection system characterized by comprising: a model storage means for storing a model of normal operation data when no abnormality has occurred in a pump, which indicates the mutual correlation between the output values ​​of multiple vibration sensor means that measure vibrations of a pump that delivers fluid, installed within a power plant; and a pump abnormality detection means for comparing the correlation determined by the model recorded in the model storage means with the correlation between measurement data for an arbitrary period obtained from the multiple vibration sensor means, monitoring the occurrence of abnormal values ​​based on the amount of breakdown in the correlation, and extracting a predetermined amount of breakdown in the relationship as an alarm event."

[0004] JP 2016-4298 A

[0005] In conventional technology, the accuracy of failure sign detection is greatly affected by the appropriate setting of monitoring thresholds. Setting these thresholds requires sufficient knowledge about the normal operation of the equipment, but it is difficult to accumulate such knowledge when equipment malfunctions rarely.

[0006] Therefore, the main purpose of this technology is to provide a technology that can automatically detect signs of equipment failure without the user having to set monitoring thresholds in advance.

[0007] The present technology provides a failure sign detection system including: an abnormality degree calculation unit that calculates an abnormality degree of an equipment based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the equipment; and a failure sign detection unit that detects a state change of the equipment based on a vibration acceleration root mean square (RMS) included in the vibration acceleration data and detects a sign of a failure of the equipment based on the abnormality degree and the state change. The abnormality degree calculation unit may learn a vibration pattern of the equipment during normal operation based on time series data of the vibration acceleration frequency spectrum, and calculate an abnormality degree according to a degree of deviation from normal operation. The abnormality degree calculation unit may learn the vibration pattern of the equipment during normal operation using statistical analysis. The abnormality degree calculation unit may learn the vibration pattern of the equipment during normal operation using a machine learning algorithm. The sign detection unit may learn the vibration pattern of the equipment during normal operation based on time series data of the vibration acceleration RMS, and detect a state change corresponding to a deviation from normal operation. The failure sign detection unit may use statistical analysis to learn the vibration pattern of the equipment during normal operation. The failure sign detection unit may use a machine learning algorithm to learn the vibration pattern of the equipment during normal operation. The system may further include a preprocessing unit that preprocesses the vibration acceleration data, and if there is a period during which the value of the vibration acceleration data does not satisfy a reference value, the preprocessing unit may exclude the vibration acceleration data for that period. The failure sign detection system may further include a sensor information acquisition unit that senses and acquires the vibration acceleration data. The failure sign detection system may further include an operation information acquisition unit that acquires operating condition information of the equipment, and the failure sign detection unit may further detect a failure sign of the equipment based on the operating condition information. When the equipment is a rotating equipment, the operating condition information may include information regarding the rotation of the rotating equipment. The failure sign detection system may further include a visualization display unit that visualizes and displays the detection result. The failure sign detection system may further include a notification unit that notifies a user of the detection result. The equipment may be a rotating equipment.The present technology also provides a failure sign detection method including: calculating an abnormality level of an equipment based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the equipment; detecting a state change of the equipment based on the vibration acceleration RMS included in the vibration acceleration data, and detecting a sign of failure of the equipment based on the abnormality level and the state change. The present technology also provides a failure sign detection program that causes a computer to execute: calculating an abnormality level of the equipment based on the vibration acceleration frequency spectrum included in vibration acceleration data acquired from the equipment; detecting a state change of the equipment based on the vibration acceleration RMS included in the vibration acceleration data, and detecting a sign of failure of the equipment based on the abnormality level and the state change.

[0008] According to the present technology, it is possible to provide a technology that can automatically detect signs of equipment failure without the need for a user to set a monitoring threshold in advance. Note that the effects described herein are not necessarily limited to those described herein and may be any of the effects described in this disclosure.

[0009] 1 is a graph showing an example of time series data of vibration acceleration RMS. FIG. 1 is a graph showing an example of time series data of vibration acceleration RMS. FIG. 2 is a graph showing an example of time series data of vibration acceleration RMS. FIG. 3 is a graph showing an example of time series data of vibration acceleration RMS. FIG. 4 is a block diagram showing an example configuration of a failure sign detection system 100 according to an embodiment of the present technology. FIG. 5 is a flowchart showing an example of a processing flow of an abnormality degree calculation unit 111 according to an embodiment of the present technology. FIG. 6 is a graph showing an example of time series data of vibration acceleration frequency spectrum. FIG. 7 is a graph showing an example of time series data of an abnormality degree. FIG. 8 is a flowchart showing an example of a processing flow of a sign detection unit 112 according to an embodiment of the present technology. FIG. 9 is a graph showing an example of time series data of vibration acceleration RMS. FIG. 10 is a graph showing an example of time series data of an abnormality degree. FIG. 11 is a block diagram showing an example configuration of a failure sign detection system 100 according to an embodiment of the present technology. FIG. 12 is a graph showing an example of time series data of vibration acceleration RMS. 1 is a graph showing an example of a screen displayed by a visualization display unit according to an embodiment of the present technology. 2 is a graph showing an example of a screen displayed by a visualization display unit according to an embodiment of the present technology. 3 is a graph showing an example of a screen displayed by a visualization display unit according to an embodiment of the present technology. 4 is a block diagram showing an example of a configuration of a failure sign detection system 100 according to an embodiment of the present technology. 5 is a flowchart showing an example of a processing flow of a report generation unit 16 according to an embodiment of the present technology. 6 is a block diagram showing an example of a configuration of a failure sign detection system 100 according to an embodiment of the present technology. 7 is a flowchart showing an example of a flow of a failure sign detection method according to an embodiment of the present technology.

[0010] Hereinafter, preferred embodiments for implementing the present technology will be described with reference to the drawings. Note that the embodiment described below shows an example of a typical embodiment of the present technology, and does not limit the scope of the present technology. In addition, the present technology can be combined with any of the following examples and their modifications.

[0011] In the following description of the embodiments, configurations may be described using terms including "approximately," such as "approximately parallel" and "approximately perpendicular." For example, "approximately parallel" does not only mean completely parallel, but also means substantially parallel, i.e., including a state where the orientation is deviated from the completely parallel state by, for example, a few percent. The same applies to other terms including "approximately." Furthermore, each figure is a schematic diagram and is not necessarily an accurate depiction. The scale of the drawings is exaggerated to make the features of the technology easier to understand. Therefore, it should be noted that the scale of the drawings and the scale of the actual device are not necessarily the same.

[0012] Unless otherwise specified, in the drawings, "top" means the top or upper side in the drawing, "bottom" means the bottom or lower side in the drawing, "left" means the left or left side in the drawing, and "right" means the right or right side in the drawing. Furthermore, in the drawings, the same or equivalent elements or members are given the same reference numerals, and redundant explanations will be omitted.

[0013] The description will be given in the following order: 1. First embodiment of the present technology (example 1 of failure sign detection system) (1) Overview (2) System configuration (3) Abnormality degree calculation unit (4) Sign detection unit 2. Second embodiment of the present technology (example 2 of failure sign detection system) 3. Third embodiment of the present technology (example 3 of failure sign detection system) 4. Fourth embodiment of the present technology (example 4 of failure sign detection system) 5. Fifth embodiment of the present technology (example 5 of failure sign detection system) 6. Sixth embodiment of the present technology (example 6 of failure sign detection system) 7. Seventh embodiment of the present technology (example of failure sign detection method) 8. Eighth embodiment of the present technology (example of failure sign detection program)

[0014] [1. First Embodiment of the Present Technology (Example 1 of a Failure Sign Detection System)] [(1) Overview] A conventional equipment monitoring process consists of four main steps. In the first step, sensors for measuring vibration, current, temperature, etc. are installed in the equipment to be monitored. At this stage, data transmission from the sensors and connection to the failure sign detection system are configured. In the second step, monitoring thresholds for determining failure signs are configured in the failure sign detection system. Based on the configured thresholds, the system periodically collects sensing data from the monitored equipment and monitors fluctuations in the data. In the third step, if the monitored data exceeds the configured thresholds, the system detects a failure sign and issues an alert. In the final step, upon detection of a failure sign, the user performs the necessary maintenance work on the target equipment. This work includes repairs and part replacement. Through this process, it is possible to take preventive measures before a failure occurs.

[0015] However, this process faces challenges. In particular, setting monitoring thresholds is a significant issue due to its difficulty. Conventional technologies employ a method in which a warning is issued when a threshold set through monitoring of sensing data is exceeded. This threshold is set based on knowledge of abnormalities and failures in the monitored equipment. However, for equipment in which abnormalities and failures are rare, such knowledge is scarce, making it difficult to set appropriate thresholds.

[0016] This issue will be explained with reference to FIG. 1, which is a graph showing an example of time series data of vibration acceleration root mean square (RMS). This graph plots how the vibration acceleration RMS value changes over time, with the horizontal axis representing time and the vertical axis representing the vibration acceleration RMS value. The data points are displayed in a scatter plot format, and two thresholds, i.e., a first threshold T1 and a second threshold T2, are indicated by horizontal lines.

[0017] When the vibration acceleration RMS exceeds the first threshold T1, it indicates that attention needs to be paid to the condition of the equipment (a sign of failure), and when it exceeds the second threshold T2, it suggests that there is a high possibility that maintenance or repair is required.

[0018] However, vibration acceleration RMS data can vary significantly depending on the individual device, and a uniformly set monitoring threshold may not be appropriate for all devices. As a result, there is a risk of over-maintenance due to excessive detection even when there is a margin of error before failure. Over-maintenance leads to unnecessary increases in maintenance costs and excessive downtime of devices, resulting in reduced productivity.

[0019] Furthermore, if monitoring thresholds are not set appropriately, there is a risk that signs of failure will be overlooked. If monitoring thresholds are set too high, the system may not detect signs of failure even when maintenance is actually required. This can lead to sudden equipment failures and unexpected downtime, resulting in significant production losses and safety risks. Therefore, setting appropriate thresholds tailored to the characteristics of each piece of equipment is the key to an effective monitoring and maintenance strategy.

[0020] This will be explained with reference to Fig. 2, which is a graph showing an example of time-series data of vibration acceleration RMS. Fig. 2 shows the change over time in vibration acceleration RMS sensed by a sensor attached to a device, with the vibration acceleration RMS plotted on the vertical axis and time on the horizontal axis.

[0021] FIG. 2A shows changes in vibration acceleration RMS sensed by a sensor attached to a pump (hereinafter, pump A). ​​Based on a vibration acceleration RMS value T3 at which failures frequently occurred in the past, a specific vibration acceleration RMS value T4 is set as a threshold value with a high failure rate.

[0022] On the other hand, Figure 2B shows the vibration acceleration RMS obtained from a sensor attached to another pump (hereinafter referred to as pump B), and indicates that a failure may occur at a specific vibration acceleration RMS value T5. In reality, there is no knowledge of a failure in pump B, and it is difficult to know the value T5. If the monitoring threshold value T4 of pump A is used to monitor pump B, a false detection problem occurs in which an early warning is issued at time t1 even though the possibility of failure is low.

[0023] As described above, conventional technologies have problems with false detection and inaccurate failure prediction due to setting a uniform threshold without fully considering the differences in usage and environmental conditions for each device. Furthermore, it is difficult to apply the same monitoring threshold to different devices, even if they are the same type, and setting an appropriate monitoring threshold for some devices is an issue.

[0024] Furthermore, when monitoring rotating equipment using inverter control, one challenge is distinguishing between changes in vibration acceleration data caused by changes in rotational speed and changes in vibration acceleration data due to abnormalities or failures. Conventional technology has difficulty accurately distinguishing between changes in vibration acceleration data due to these different causes, which can result in incorrectly identifying them as signs of failure. This issue is a common occurrence in rotating equipment whose rotational speed changes, making monitoring tasks more difficult. To address this issue, a new approach is needed to effectively distinguish between changes in rotational speed and changes in vibration due to failures.

[0025] This will be explained with reference to Fig. 3. Fig. 3 is a graph showing an example of time-series data of vibration acceleration RMS. In this graph, the horizontal axis represents time and the vertical axis represents vibration acceleration RMS.

[0026] During time period L when the equipment is operating at low speed, it can be observed that the equipment rotation speed is low and, accordingly, the vibration acceleration RMS is also low. At time t2, the equipment switches to high-speed operation. During time period H when the equipment is operating at high speed, it is shown that the vibration acceleration RMS also increases as the rotation speed increases. Then, at time t3, the vibration acceleration RMS increases further due to signs of equipment failure.

[0027] With conventional technology, it is difficult to distinguish whether such a change in vibration acceleration RMS is due to a change in the operating conditions of the equipment (a change from low speed to high speed) or due to an equipment failure. Because changes in operating conditions and changes in vibration acceleration RMS due to a failure are measured using the same index, there is a risk that the change in operating conditions will be mistakenly recognized as a sign of a failure. In this example graph, there is a risk that the change in vibration acceleration RMS at time t2 will be mistakenly detected as being caused by a sign of an equipment failure.

[0028] To address this issue, a method can be considered in which additional operating condition information (such as rotation speed or inverter frequency) is acquired from the monitored equipment and analyzed to select appropriate operating conditions. Combining the operating condition information with the vibration acceleration RMS makes it possible to detect signs of failure. However, this method requires labor for analysis and the construction of a system for acquiring operating condition information, and there are still challenges in easily detecting signs of failure. To address this issue, the development of a simple detection method is an important challenge, especially in fields where automation of analysis processes and efficient information acquisition are required.

[0029] This technology is a new method for detecting signs of equipment failure, and unlike conventional failure sign detection technologies, it does not require users to set monitoring thresholds in advance. Conventional technologies require users to set monitoring thresholds, and when a value exceeds those thresholds, it is determined to be a sign of a failure. However, because failures occur infrequently, there is an issue in that it is difficult to accumulate the experience and knowledge needed to set appropriate monitoring thresholds.

[0030] To solve this problem, our technology employs a method in which the monitoring model itself learns the normal state of the equipment and automatically determines deviations from that normal state, without the user having to set monitoring thresholds in advance. This approach makes it possible to effectively detect equipment anomalies and signs of failure even when detailed knowledge about the characteristics and failures of the monitored equipment is lacking. This technology offers significant advantages, especially when monitoring equipment where failures occur rarely and signs of failure are difficult to identify.

[0031] [(2) System Configuration] The failure sign detection system disclosed in this specification is a system that can detect a critical change in the state of an equipment based on the vibration acceleration RMS included in the vibration acceleration data acquired from the equipment, calculate the degree of abnormality of the equipment using the vibration acceleration frequency spectrum included in the vibration acceleration data, detect a change in the state of the equipment based on the degree of abnormality, and detect a sign of failure. This system can efficiently and accurately grasp the abnormal state of the equipment from the vibration acceleration data, and provides important information for predictive maintenance.

[0032] Specifically, the present technology provides a failure sign detection system including an abnormality degree calculation unit that calculates the abnormality degree of the equipment based on the vibration acceleration frequency spectrum included in vibration acceleration data acquired from the equipment, and a sign detection unit that detects a change in the state of the equipment based on the vibration acceleration RMS included in the vibration acceleration data, and detects signs of failure of the equipment based on the abnormality degree and the state change.

[0033] Vibration RMS acceleration is an index used to evaluate the vibration state of equipment. This index is the time average of the acceleration fluctuations of the vibration signal, and quantitatively indicates the energy content of the vibration. Vibration RMS acceleration is calculated by squaring the instantaneous value of the vibration acceleration, averaging it over a certain period, and then taking the square root of the average value.

[0034] Vibration acceleration RMS plays an especially important role in the fields of condition monitoring and predictive maintenance of equipment and components. Detecting abnormal vibrations is often a sign of mechanical failure, wear, or damage. Therefore, by regularly monitoring vibration acceleration RMS, it is possible to perform preventative maintenance before a failure occurs, thereby extending the life of equipment and components.

[0035] For example, an acceleration sensor is used to measure vibration. This sensor continuously captures the acceleration of vibrations caused by the operation of equipment and converts it into an electrical signal. The RMS value can be calculated from the obtained acceleration data, making it possible to quantitatively evaluate the vibration condition of the equipment. By monitoring the vibration acceleration RMS, the condition of the equipment can be checked periodically and the appropriate timing for maintenance can be determined.

[0036] The time-series data of vibration acceleration RMS is measured at specific time intervals, and the date and time of the measurement are recorded. This data structure provides a basis for accurately tracking and analyzing the dynamic behavior of equipment and its changes. By recording vibration acceleration RMS periodically, it is possible to maintain the health of equipment and contribute to fault prediction.

[0037] Vibration acceleration RMS data is an important indicator of the health of equipment. The database structure in this technology manages time-series data of this vibration acceleration RMS. The database structure will be described with reference to FIG. 4. FIG. 4 shows a portion of a database for managing time-series data of vibration acceleration RMS. As shown in FIG. 4, in this example, the database arranges dates (YY.MM.DD) and times (HH:MM:SS) in columns, and records corresponding vibration acceleration RMS values ​​in rows.

[0038] Specific embodiments for implementing the present technology will be described below. A configuration example of an embodiment of a failure sign detection system based on the present technology will be described with reference to Fig. 5. Fig. 5 is a block diagram showing a configuration example of a failure sign detection system 100 according to an embodiment of the present technology. The system configuration of this failure sign detection system 100 is made up of multiple components required to collect and analyze vibration acceleration data from equipment and detect failure signs.

[0039] As shown in FIG. 5, the failure sign detection system 100 includes a sensor information acquisition unit 3, a communication interface 2, and a failure sign detection device 1.

[0040] The failure sign detection device 1 includes a calculation unit 11, a memory 12, a power supply unit 13, a user interface 14, and the like.

[0041] The sensor information acquisition unit 3 senses and acquires vibration acceleration data. The data acquired by the sensor information acquisition unit 3 is transmitted to the system in real time or at regular intervals and used for analysis.

[0042] Specifically, the sensor information acquisition unit 3 receives information from a sensor 4 that precisely captures the vibration state of the monitored equipment. This sensor 4 has a sensitivity that allows the vibration state of the equipment to be reflected in the data, and the acquired data reflects the operating state of the equipment in detail, such as the magnitude, frequency, and duration of vibration. This data forms the basis for the failure sign detection system to distinguish between normal and abnormal operating patterns of the equipment and to detect signs of failure at an early stage.

[0043] The sensor 4 is attached directly to the equipment 5 to collect vibration acceleration data from the equipment 5. The sensor 4 measures the vibration acceleration RMS and vibration acceleration frequency spectrum of the equipment and transmits the data to the system.

[0044] The sensor 4 that senses vibration acceleration data preferably includes a piezoelectric element. Piezoelectric elements generate electric charge when subjected to physical force and have the ability to accurately convert changes in that charge into an electrical signal. This makes them suitable for detecting changes in force, such as vibration and acceleration, with high sensitivity, and they are particularly suited to vibration measurement. Piezoelectric sensors are capable of precisely measuring a wide range of vibration levels, from minute vibrations to large impacts, and are recognized as being effective in applications requiring high sensitivity, such as monitoring industrial machinery and structures. Furthermore, sensors using piezoelectric elements are highly durable and reliable, making them suitable for long-term use in harsh operating environments. Therefore, when continuous sensing of vibration acceleration data with high precision is required, a sensor with a piezoelectric element is preferred. However, the sensor is not limited to using a piezoelectric element.

[0045] The piezoelectric element may be, for example, Pb(Zr,Ti)O3 [PZT], PbTiO3, Pb(Mg 1 / 3Nb 2 / 3 )O3-PbTiO3[PMN-PT], Pb(Zn 1 / 3 Nb 2 / 3 )O3-PbTiO3[PZN-PT], BaTiO3[BT], (K,Na)NbO3[KNN], KNbO3, NaNbO3, ( K,Na,Li)NbO3, (K,Na,Li)(Nb,Ta,Sb)O3, (Sr,Ba)Nb2O6, (Sr,Ca)NaNb5O 15 , (Na,K)Ba2NbO 15 , BiFeO3, Bi4Ti3O 12 , (Bi 1 / 2 K 1 / 2 ) TiO3, (Bi 1 / 2 Na 1 / 2 )TiO3, BaTiO3-(Bi 1 / 2 K 1 / 2 )TiO3, BaTiO3-(Bi 1 / 2 Na 1 / 2 ) TiO3, AlN, LiNbO3, LiTaO3, alpha-SiO2, GaPO4, LiB4O7, La3Ga5SiO 14 , La3Ta 0.5 Ga 5.5 O 14 The piezoelectric material is made of at least one material selected from the group consisting of MgSiO3, ZnO, polyvinylidene fluoride [PVDF], halide perovskite piezoelectric material, polylactic acid [PLLA], cellulose, and polypeptide.

[0046] The sensor information acquisition unit 3 can receive and temporarily store data from the sensor 4. The sensor information acquisition unit 3 has the function of converting the data from the sensor 4 into a digital format and performing initial data preprocessing.

[0047] The communication interface 2 exchanges data between the sensor information acquisition unit 3 and the calculation unit 11. The communication interface 2 supports wired (e.g., Ethernet) or wireless (e.g., Wi-Fi, Bluetooth (registered trademark), etc.) communication.

[0048] The calculation unit 11 is the heart of the system and executes main processes such as data analysis. For example, a central processing unit (CPU) or a graphics processing unit (GPU) is used as the calculation unit 11. The calculation unit 11 includes an anomaly degree calculation unit 111 and a sign detection unit 112. The anomaly degree calculation unit 111 and the sign detection unit 112 will be described in detail later.

[0049] The memory 12 temporarily stores data being processed. This system uses a RAM (random access memory) that allows for high-speed access. Storage (such as an HDD or SSD) may also be used for long-term data retention and historical analysis.

[0050] The power supply unit 13 is a power source for operating the entire system. A stable power supply is required, and in some cases a backup power supply or a battery is used.

[0051] The user interface 14 receives input from a user and includes, for example, a keyboard and a mouse.

[0052] These components may be selected and integrated according to the requirements of the failure sign detection system 100. The specific components may differ depending on the type of device, the usage environment, and even the needs of the user.

[0053] [(3) Abnormality Degree Calculation Unit] The abnormality degree calculation unit 111 has a function of calculating the abnormality degree of the equipment based on the vibration acceleration frequency spectrum included in the vibration acceleration data acquired from the equipment. The abnormality degree calculation unit 111 learns the vibration pattern of the equipment during normal operation based on the time-series data of the vibration acceleration frequency spectrum, and calculates the abnormality degree according to the degree of deviation from normal operation.

[0054] Specifically, the anomaly degree calculation unit 111 collects and learns vibration acceleration data during normal operation and uses this as reference data. It then compares vibration acceleration data acquired during actual operation with this reference data to evaluate the degree to which the data at any given time deviates from the normal pattern. The degree of abnormality indicated by this deviation is quantified and calculated as the anomaly degree. This anomaly degree calculation process makes it possible to accurately grasp changes in the equipment's condition and detect signs of failure early.

[0055] The period that defines normal operation can be determined by human judgment or automatically by a computer based on past statistical data. This reference period could be, for example, one week. For example, if a device shuts down on weekends and operates in roughly the same pattern every week, one week can be set as the reference period. Furthermore, if the device tends to operate at full capacity at the end of each month, one month can be selected as the reference period. Setting a reference period in this way makes it possible to collect data during normal operation that is tailored to the device's operating schedule and characteristics, thereby improving the accuracy of fault detection.

[0056] The abnormality degree calculation unit 111 can have a function of learning the vibration pattern during normal operation of the equipment using statistical analysis. Through this function, the abnormality degree calculation unit can learn the basic characteristics and patterns of vibration in a normal state based on the vibration acceleration data acquired from the equipment. The abnormality degree calculation unit 111 calculates statistical quantities such as the mean, variance, and standard deviation from the vibration acceleration data during normal operation using statistical analysis, and defines a normal range for the vibration acceleration data based on these values.

[0057] After learning the vibration pattern during normal operation, the anomaly degree calculation unit 111 evaluates the degree to which newly acquired vibration acceleration data deviates from the previously identified normal pattern. This deviation is quantified and calculated as an anomaly degree, which can contribute to equipment status monitoring and failure sign detection. The statistical analysis methods used include calculation of the average value and standard deviation, but the selection of these methods can vary depending on the type of equipment to be monitored, the operating environment, the required sensitivity, etc.

[0058] The anomaly degree calculation unit 111 can also have a function of learning vibration patterns during normal operation of the equipment using a machine learning algorithm. By using a machine learning algorithm, it becomes possible to automatically learn normal vibration patterns from a large amount of vibration acceleration data and evaluate the degree to which abnormal vibration patterns deviate from the normal vibration patterns. Various machine learning methods, such as supervised learning, unsupervised learning, and semi-supervised learning, can be applied to this process, but the present failure sign detection system uses unsupervised learning.

[0059] Unsupervised learning automatically identifies patterns and clusters in data and detects anomalies.

[0060] By using a machine learning algorithm, the anomaly degree calculation unit 111 extracts complex patterns and relationships present in the vibration acceleration data of the equipment and calculates the anomaly degree based on this. This method provides more advanced anomaly detection capabilities than conventional methods, and significantly improves the accuracy of equipment status monitoring and failure sign detection.

[0061] The anomaly degree calculation unit 111 can utilize various representative algorithms and models widely used in the fields of statistical analysis and machine learning. Examples of statistical methods include principal component analysis (PCA) and clustering. Examples of machine learning algorithms include one-class support vector machines (OCSVM), k-nearest neighbor methods, and autoencoders based on deep learning. It is recommended that these models and algorithms be selected taking into consideration criteria such as measurement data characteristics, prediction accuracy, computational efficiency, and ease of implementation. This optimizes the analysis of vibration acceleration data obtained from equipment, contributing to more accurate anomaly detection and early detection of failure signs.

[0062] As described above, the method for calculating the abnormality level is not particularly limited, but an example of the calculation method is shown below. It goes without saying that the calculation method is not limited to this.

[0063] An example of the processing flow of the abnormality degree calculation unit 111 will be described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of the processing flow of the abnormality degree calculation unit 111 according to an embodiment of the present technology.

[0064] 6, first, in step S11, the abnormality degree calculation unit 111 collects vibration acceleration data. The collected vibration acceleration data includes, for example, vibration acceleration RMS and vibration acceleration frequency spectrum.

[0065] Next, in step S12, the abnormality degree calculation unit 111 converts the collected vibration acceleration data into a frequency spectrum using a mathematical method such as Fourier transform or fast Fourier transform (FFT). Through this conversion, information is reconstructed from time domain data into frequency domain data.

[0066] Vibration acceleration frequency spectrum analysis is a method for analyzing the frequency components of vibration signals generated by equipment, and plays an important role in equipment health monitoring and fault diagnosis. This analysis converts vibration signals captured in the time domain into the frequency domain, allowing the individual frequency components contained in the signal and their amplitudes to be identified.

[0067] Different machine elements and failure modes vibrate at their own unique frequencies. Therefore, analysis of the vibration acceleration frequency spectrum can distinguish between normal operating patterns and abnormal vibration patterns. For example, certain faults, such as bearing failure, shaft imbalance, and gear damage, show increased vibration in specific frequency ranges. Utilizing these characteristics and observing the vibration spectrum enables early detection of abnormalities and planning of predictive maintenance, contributing to improved system reliability and performance.

[0068] In addition, analysis of vibration acceleration frequency spectrum is non-invasive and allows for real-time monitoring without shutting down equipment, which allows for continuous monitoring of equipment status and allows for planned maintenance and repairs, contributing to operational efficiency and cost reduction.

[0069] The data structure of the time series data of vibration acceleration frequency spectra will be described with reference to FIG. 7. FIG. 7 shows a part of a database for managing the time series data of vibration acceleration frequency spectra. As shown in FIG. 7, the time series data structure of vibration acceleration frequency spectra is made up of vibration accelerations measured at regular time intervals and their corresponding frequency spectra. Specifically, each row of the database contains the measurement date and time (YY.MM.DD HH:MM:SS) as well as acceleration values ​​(a1, a2, ..., a) representing the top n points of the vibration acceleration frequency spectrum. n ) and their corresponding frequencies (f1, f2, ..., f n ) is described.

[0070] This database aims to enable detailed analysis of vibration acceleration data obtained from monitored equipment and to track vibration characteristics over time. The frequency spectrum data here provides important information for detecting the vibration behavior of equipment and evaluating the health of the equipment through its fluctuations.

[0071] Returning to the explanation of Fig. 6, next, in step S13, the anomaly degree calculation unit 111 learns the vibration pattern observed during normal operation of the equipment based on the time-series data of the converted vibration acceleration frequency spectrum. In this process, the characteristics of frequency components that indicate the sound state of the equipment and the vibration amplitude distribution are analyzed, and a reference pattern for normal operation is defined.

[0072] The anomaly degree calculation unit 111 acquires time-series data of the vibration acceleration frequency spectrum over a certain period of time from the start of operation of the equipment. Specifically, for example, 100 data points are acquired from the start of operation and defined as reference data during normal operation. The acceleration and frequency during normal operation fluctuate to some extent, but remain within a certain range.

[0073] Next, in step S14, the anomaly degree calculation unit 111 calculates the degree of deviation from a predefined normal operating pattern for data from the 101st data point onwards as the degree of anomaly. In this calculation process, the degree of deviation, frequency, and range of impact are comprehensively taken into consideration to quantify the current state of the equipment. For example, Mahalanobis distance or k-nearest neighbor method is used as a measure of the degree of anomaly. This makes it possible to quantitatively express the degree of anomaly.

[0074] The calculation of the degree of abnormality will be described with reference to Figs. 8 and 9. Fig. 8 is a graph showing an example of time-series data of a vibration acceleration frequency spectrum. As shown in Fig. 8, the time-series data of the vibration acceleration frequency spectrum has time on the horizontal axis and the vibration frequency spectrum on the vertical axis, and the intensity of the vibration acceleration is displayed using different colors. This graph makes it possible to grasp in detail the changes in frequency components and their intensity at a specific time.

[0075] Next, Fig. 9 is a graph showing, in the form of a scatter plot, the degree of abnormality calculated based on the vibration acceleration frequency spectrum data shown in Fig. 8. In Fig. 9, the horizontal axis represents time and the vertical axis represents the degree of abnormality, and the change in the degree of abnormality over time is plotted. In this diagram, it is particularly noteworthy that the degree of abnormality rises sharply toward the right end, indicating that some change has occurred over time and that this is manifested as an increase in the degree of abnormality.

[0076] The positions where the degree of abnormality has changed significantly are indicated by multiple vertical lines. The abnormality degree calculation unit 111 calculates the degree of deviation from normal operation of the equipment as the degree of abnormality. The states indicated by the vertical lines indicate that the equipment is in a dangerous state.

[0077] Returning to the description of Fig. 6, next, in step S15, the abnormality degree calculation unit 111 notifies the sign detection unit 112 of the calculated abnormality degree.

[0078] [(4) Predictor Detector] The predictor detector 112 has a function of detecting a change in the state of the equipment based on the vibration acceleration RMS included in the vibration acceleration data, and detecting a predictor of equipment failure based on the degree of abnormality calculated from the vibration acceleration frequency spectrum and the state change. The predictor detector 112 learns the vibration pattern of the equipment during normal operation based on the time-series data of the degree of abnormality calculated from the vibration acceleration frequency spectrum and the time-series data of the vibration acceleration RMS, and detects a state change corresponding to a deviation from normal operation.

[0079] Specifically, the predictor detection unit 112 continuously monitors vibration acceleration RMS data obtained from the equipment and builds a reference model of the equipment's vibration pattern from this data. This reference model represents the vibration behavior of the equipment when it is operating normally. The predictor detection unit 112 then compares the actual vibration pattern with this reference model to detect abnormal changes or deviations in the vibration pattern. If such deviations are observed, they are interpreted as a sign that could lead to equipment failure.

[0080] The symptom detection unit 112 may have a function of learning the vibration pattern of the equipment during normal operation using statistical analysis. When the symptom detection unit 112 uses statistical analysis, the vibration pattern of the equipment during normal operation is identified using a statistical method. In this approach, statistics such as the mean value, variance, and standard deviation are calculated from a data set in a normal state, and a normal vibration pattern is defined based on these.

[0081] For example, the predictor detection unit 112 analyzes the statistical characteristics of vibration acceleration data acquired from the equipment. This analysis identifies the basic characteristics of the vibration pattern during normal operation of the equipment. The calculated statistics serve as a reference point indicating normal vibration behavior and serve as a basis for detecting abnormal vibration patterns by later comparison with vibration acceleration data.

[0082] If this method detects deviations from vibration acceleration data during normal operation, it indicates that some change may have occurred in the equipment's condition. For example, if the standard deviation increases suddenly, this could be a sign that vibrations are becoming more irregular and that equipment failure is imminent.

[0083] The advantage of using statistical analysis is that it can objectively define and monitor normal operating patterns of equipment even from complex vibration acceleration data. This allows the predictive detection unit to detect signs of equipment failure early and provide valuable information for appropriate maintenance. This technology improves equipment reliability while also contributing to minimizing damage caused by unexpected failures.

[0084] The sign detection unit 112 may have a function of learning vibration patterns during normal operation of the equipment using a machine learning algorithm. When the sign detection unit 112 uses a machine learning algorithm, it automatically learns more complex data patterns and vibration behaviors, and learns vibration patterns during normal operation. The machine learning model learns normal patterns from a large amount of vibration acceleration data, and effectively detects signs of failure by evaluating the degree to which new data deviates from the normal patterns.

[0085] This method applies a variety of machine learning algorithms, including supervised learning and unsupervised learning, to analyze vibration acceleration data. These algorithms are capable of detecting unknown failure modes and subtle vibration changes by learning normal vibration patterns. Training a machine learning model requires a large amount of data collected from equipment in various states. The model uses this data to learn normal conditions and evaluates the extent to which new data deviates from those normal conditions.

[0086] The advantage of using machine learning is that it can automatically extract features from data and perform advanced pattern recognition. This makes it possible to detect subtle changes in vibration and previously unknown failure patterns that were difficult to detect using conventional statistical analysis. Furthermore, machine learning models can continuously learn and improve their performance based on newly acquired data, making it possible to maintain high detection accuracy even when the equipment's operating environment changes.

[0087] The processing flow of the sign detection unit 112 systematically shows a series of steps for detecting a change in the state of the equipment based on the vibration acceleration RMS included in the vibration acceleration data, and further detecting a sign of equipment failure based on the abnormality degree calculated by the abnormality degree calculation unit 111. An example of the processing flow of the sign detection unit 112 will be described with reference to Fig. 10. Fig. 10 is a flowchart showing an example of the processing flow of the sign detection unit 112 according to an embodiment of the present technology.

[0088] 10, first, in step S21, the sign detection unit 112 receives the degree of abnormality from the abnormality degree calculation unit 111. This data indicates the degree to which the current operating state of the equipment deviates from normal.

[0089] Next, in step S22, the sign detection unit 112 receives the vibration acceleration RMS included in the vibration acceleration data. This data indicates the overall strength of the vibration of the device.

[0090] Next, in step S23, the sign detection unit 112 applies statistical analysis or a machine learning algorithm to the time-series data of the vibration acceleration RMS and compares it with data obtained during normal operation of the equipment to detect a change in the state.The sign detection unit 112 also applies statistical analysis or a machine learning algorithm to the time-series data of the abnormality degree and compares it with data obtained during normal operation of the equipment to detect a change in the state.

[0091] Next, in step S24, the sign detection unit 112 detects signs of equipment failure based on the abnormality level and state change. At this stage, the relationship between fluctuations in the abnormality level and fluctuations in the vibration acceleration RMS is analyzed.

[0092] This analysis will be described with reference to Fig. 11 and Fig. 12. Fig. 11 is a graph showing an example of time-series data of vibration acceleration RMS. Fig. 12 is a graph showing an example of time-series data of the degree of abnormality.

[0093] The vertical lines L1 located near the center of the graph in Fig. 11 and the vertical lines L2 located on the right side indicate positions where large changes in the vibration acceleration RMS value were observed. The vertical lines L3 located on the right side of the graph in Fig. 12 indicate positions where large changes in the degree of anomaly value were observed.

[0094] When combined with the changes indicated by vertical line L3, it becomes clear that vertical line L2 on the right indicates a "danger" state. This "danger" state vertical line L2 warns that the equipment is approaching failure or that there is a high possibility of failure, and makes it possible to distinguish it from the "caution" state indicated by vertical line L1 on the left. The "caution" state here refers to a sign of equipment failure, and at this stage it functions as a signal to strengthen equipment monitoring and encourage prompt action.

[0095] In this way, the failure sign detection unit 112 plays an important role in maintaining the reliability of the equipment and preventing losses due to sudden failures by detecting failure signs early and carrying out maintenance at the appropriate time based on the time series data of the vibration acceleration RMS and the vibration acceleration frequency spectrum.

[0096] This process flow enables the symptom detection unit 112 to effectively use data collected from devices to detect symptoms of device failure at an early stage and to identify in advance when device maintenance or repair will be required. Note that the specific implementation of each step may vary depending on the type of device, the usage environment, and available technology.

[0097] This technology also solves the difficulty of monitoring inverter-controlled equipment. In inverter-controlled equipment, changes in rotation speed due to operating conditions cause changes in sensing data, but similar changes occur due to abnormalities or failures. With conventional technology, it has been difficult to effectively identify these changes.

[0098] In a failure sign detection system based on this technology, the monitoring model learns changes in sensing data caused by operating conditions, making it possible to distinguish between changes in operating conditions and changes caused by abnormalities or failures. This makes it possible to accurately detect failure signs in inverter-controlled equipment without being influenced by operating conditions.

[0099] The equipment to which this technology is applied may be a rotating equipment. Rotating equipment is a general term for machines that primarily use rotational motion to perform their functions, and these machines have the characteristic of generating vibrations during operation. Rotating equipment is widely used throughout industry and has a wide variety of applications. Specific examples of rotating equipment include pumps, fans, and compressors.

[0100] The equipment to which this technology can be applied is not limited to rotating equipment, but can also be applied to various production facilities equipped with drive systems.

[0101] This technology demonstrates its functionality regardless of the type or characteristics of the equipment. This is because it is possible to detect signs of failure through the analysis of vibration acceleration data, regardless of whether the equipment is rotating equipment or other types of production equipment with drive systems. Therefore, this technology is effective for the maintenance management of various machines and facilities in a wide range of industrial fields, minimizing damage caused by unexpected equipment shutdowns and failures and contributing to improved operational efficiency.

[0102] The above description of the failure sign detection system according to the first embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0103] 2. Second Embodiment of the Present Technology (Example 2 of Failure Sign Detection System) In one embodiment of the present technology, data from periods when monitored equipment is stopped is excluded from analysis of failure sign detection, thereby enabling more accurate detection of failure signs. During periods when equipment is stopped, the likelihood of equipment deterioration progressing is low, and therefore data acquired during those periods is considered to be less relevant in the analysis of failure signs.

[0104] Therefore, if the vibration acceleration RMS does not meet a preset reference value, the equipment is determined to be in a non-operating state, and data from that period is excluded from analysis. This reference value can be set to an appropriate value depending on the characteristics of the monitored equipment.

[0105] Therefore, it is desirable that the failure sign detection system according to an embodiment of the present technology further include a preprocessing unit that preprocesses the vibration acceleration data. This will be described with reference to Fig. 13. Fig. 13 is a block diagram showing an example configuration of the failure sign detection system 100 according to an embodiment of the present technology.

[0106] As shown in FIG. 13 , the failure sign detection system 100 further includes a preprocessing unit 113 that preprocesses the vibration acceleration data. If there is a period in which the vibration acceleration data value does not satisfy a reference value, the preprocessing unit 113 excludes the vibration acceleration data for that period. A period in which the vibration acceleration data value does not satisfy a reference value may be, for example, a period in which the equipment is stopped. This preprocessing includes filtering, excluding data based on a reference value, and the like. This improves the accuracy of the analysis.

[0107] The vibration acceleration RMS data collected from monitored equipment reflects the operating state of the equipment, and excluding data from periods when the equipment is stopped contributes to more accurate detection of signs of failure. This will be explained with reference to FIG. 14. FIG. 14 shows sample data of vibration acceleration RMS obtained from monitored equipment in a time series, with the horizontal axis representing the sample number and the vertical axis representing the vibration acceleration RMS value. The data points on the graph indicate the magnitude of vibration in each sample.

[0108] If the vibration acceleration RMS does not meet a predetermined reference value, i.e., if the vibration is extremely small or nonexistent, the data D is considered to be data from when the equipment is stopped, and the preprocessing unit 113 excludes this data D from the analysis. This reference value is set according to the characteristics of the equipment to be monitored and is used to prevent irrelevant data from erroneously affecting the analysis results. This approach ensures that only data from the period when the equipment is in operation is used for reliable analysis, achieving accurate condition monitoring.

[0109] This technology enables accurate calculation of the degree of abnormality in operating conditions, enabling more reliable determination of the need for maintenance that reflects the actual progression of equipment deterioration. This is expected to avoid excessive maintenance and optimize the operational efficiency and lifespan of equipment.

[0110] The above description of the failure sign detection system according to the second embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0111] 3. Third Embodiment of the Present Technology (Example 3 of Failure Sign Detection System) A failure sign detection system according to an embodiment of the present technology may further include an operation information acquisition unit that acquires operating condition information of the equipment. This will be described with reference to FIG. 15 . FIG. 15 is a block diagram showing an example configuration of a failure sign detection system 100 according to an embodiment of the present technology. As shown in FIG. 15 , the failure sign detection system 100 further includes an operation information acquisition unit 114 that acquires operating condition information of the equipment.

[0112] The operation information acquisition unit 114 acquires information related to the operation status of the equipment. This information is extremely important in determining whether the equipment is functioning normally or is in an abnormal operating state that could lead to a malfunction. For example, if the equipment is being operated at a speed outside the set operating range, this is likely to be a sign of a malfunction.

[0113] When the equipment is a rotating equipment, the operating condition information includes information about the rotation of the rotating equipment. This information about the rotation includes the number of rotations and the rotation speed. Based on this information, it is possible to determine the operating state of the equipment, such as whether it is operating at a low speed or a high speed.

[0114] The sign detection unit 112 detects signs of equipment failure based on the operating condition information provided by the operating information acquisition unit 114. The operating state of the equipment is estimated through analysis based on the operating condition information, and an evaluation is made as to whether the state is within the normal operating range of the equipment or indicates an abnormality. This evaluation is performed in combination with vibration acceleration data and information obtained from other sensors, thereby achieving more accurate failure sign detection.

[0115] The above description of the failure sign detection system according to the third embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0116] 4. Fourth Embodiment of the Present Technology (Fourth Example of Failure Sign Detection System) A failure sign detection system according to an embodiment of the present technology may further include a visualization display unit that visualizes and displays a detection result. This will be described with reference to FIG. 16 . FIG. 16 is a block diagram showing an example configuration of a failure sign detection system 100 according to an embodiment of the present technology. As shown in FIG. 16 , the failure sign detection system 100 further includes a visualization display unit 15 that visualizes and displays a detection result. The visualization display unit 15 visualizes and displays the detection result, providing information in a form that can be easily understood by a user. The visualization display unit 15 may be, for example, a display monitor. This display monitor may have a touch screen function, allowing for intuitive operation.

[0117] The failure sign detection system of the present technology processes time-series data of vibration acceleration RMS and vibration acceleration frequency spectrum acquired from the monitored equipment and calculates the abnormality level of the equipment based on this data. The calculated abnormality level is compared with data from a learning period immediately after monitoring begins, which reflects normal operation of the equipment. If a change in the equipment status is detected through this comparison, the detection timing is preferably displayed on the visualization display unit 15. The displayed information preferably includes the timing of the initial and final abnormality of the equipment, allowing the user to intuitively understand the status of the equipment.

[0118] Therefore, this system visually displays the results of detection of changes in the device's status, helping users to respond quickly.

[0119] An example of visualizing and displaying the detection result will be described with reference to Fig. 17 and Fig. 18. Fig. 17 and Fig. 18 are graphs showing an example of a screen displayed by a visualization display unit according to an embodiment of the present technology.

[0120] 17 shows time-series data of vibration acceleration RMS. This graph shows a learning period S during which the vibration acceleration RMS value remains at a substantially constant low level over time. This learning period represents the period during which the system learns data from normal conditions. After this, a significant increase in the vibration acceleration RMS value is observed, and multiple vertical lines L1 indicating "caution" timing and multiple vertical lines L2 indicating "danger" timing are shown.

[0121] This "Caution" is an early sign that the vibration acceleration RMS value is gradually increasing and may exceed the normal operating range of the equipment. At this stage, it has not yet reached "Danger," but it does alert users to a change in the equipment's condition. Equipment operators can view this sign as a warning and use it as an important indicator for planning early maintenance and repairs. By detecting and responding to abnormal signs at the appropriate time, it is possible to prevent equipment failures and ensure long-term operational stability.

[0122] Figure 18 shows time-series data of the vibration acceleration frequency spectrum. Multiple vertical lines are shown at times when the vibration intensity changes. This indicates that a specific change has occurred in the vibration frequency component, suggesting that a problem may be occurring in a specific part of the equipment.

[0123] This technology monitors vibration acceleration RMS and detects abnormalities that exceed thresholds corresponding to caution and danger. Furthermore, it monitors the vibration acceleration frequency spectrum and distinguishes between caution and danger. Specifically, when only the vibration acceleration RMS changes, it is judged to be at the caution level, and when both the vibration acceleration RMS and the degree of abnormality calculated from the vibration acceleration frequency spectrum change, it is judged to be at the danger level.

[0124] The failure sign detection system according to the present technology is also compatible with monitoring of inverter-controlled equipment. Figures 19 and 20 are graphs showing an example of a screen displayed by a visualization display unit according to an embodiment of the present technology. Figure 19 is a graph showing the monitoring status of inverter-controlled equipment, showing time-series data of vibration acceleration RMS. In this example, time-series data of vibration acceleration RMS at rotational speeds of 1200 rpm, 1800 rpm, and 2400 rpm is shown.

[0125] Figure 20 shows vibration frequency spectra corresponding to the same time axis as the graph shown in Figure 19. During the learning period S, the failure sign detection system learned operating data at rotational speeds of 1200 rpm, 1800 rpm, and 2400 rpm. The system adapted to the patterns under these new operating conditions, and thereafter, the system no longer detected significant changes in similar situations at these operating speeds. This indicates that the system learned changes in operating conditions, particularly changes in rotational speed, and would no longer detect these changes in similar operating situations.

[0126] With these functions, the fault prediction detection system of this technology can recognize changes in the operating conditions of equipment controlled by inverters as part of normal operation and avoid issuing false warnings. This makes it possible to effectively detect only true signs of faults without mistaking changes in operating conditions themselves for abnormalities, contributing to stable equipment operation and long-term maintenance planning.

[0127] These analysis results are visualized through the visualization display unit 15 and presented in a format that is intuitively easy for equipment operators to understand. By detecting and responding to signs of abnormality at the appropriate time, it is possible to prevent equipment failures and ensure long-term operational stability. Through this process, the failure sign detection system provides an effective means of monitoring the status of equipment and predicting failure signs.

[0128] The above description of the failure sign detection system according to the fourth embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0129] 5. Fifth Embodiment of the Present Technology (Fifth Example of Failure Sign Detection System) A failure sign detection system according to the present technology may further include a report generation unit that automatically generates a diagnostic report based on the detection result. This will be described with reference to Fig. 21 . Fig. 21 is a block diagram showing an example configuration of a failure sign detection system 100 according to an embodiment of the present technology.

[0130] 21 , the failure sign detection system 100 further includes a report generation unit 16 that automatically generates a diagnostic report based on the detection results. The report generation unit 16 analyzes the current state of the equipment and potential causes of failure based on the data collected from the sensors, and presents countermeasures.

[0131] If a symptom is detected, the report generation unit 16 starts processing. The report generation unit 16 generates a diagnostic report that includes the type of abnormality detected, possible causes, the results of a comparison analysis with similar past cases, and recommended countermeasures.

[0132] The processing flow of the report generation unit 16 will be described with reference to Fig. 22. Fig. 22 is a flowchart showing an example of the processing flow of the report generation unit 16 according to an embodiment of the present technology.

[0133] 22, in step S31, the calculation unit identifies the type of abnormality. In this step, the calculation unit analyzes the characteristics and behavior of the abnormality to identify the type of abnormality. Examples of the type of abnormality include imbalance, misalignment, damage, looseness, and cavitation.

[0134] In step S32, the report generator 16 analyzes possible causes by referring to a database that stores vibration acceleration data, operation data, and failure data of the equipment.

[0135] In step S33, the report generator 16 compares the detected anomaly with similar cases recorded in the past to analyze the results. This comparison can reveal patterns and causes behind the anomaly, which can be useful in formulating a solution.

[0136] In step S34, the report generator 16 proposes specific countermeasures based on the results of the cause analysis and case comparison. The proposals include recommended actions for resolving the abnormality, such as repairs, part replacement, or adjustments to operating conditions.

[0137] In step S35, the report generator 16 generates a diagnostic report including the type of abnormality, possible causes, the results of a comparative analysis with similar past cases, and recommended countermeasures.

[0138] The generated diagnostic report is automatically sent via the visual display unit 15, email, SMS, a dedicated web portal, etc. If required, the report is also provided in printed form.

[0139] This system enables early detection of signs of failure and rapid response, contributing to reduced equipment downtime and improved productivity.

[0140] The above description of the failure sign detection system according to the fifth embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0141] 6. Sixth embodiment of the present technology (sixth example of failure sign detection system)

[0142] A failure sign detection system according to an embodiment of the present technology may further include a notification unit that notifies a user of a detection result. This will be described with reference to FIG. 23 . FIG. 23 is a block diagram illustrating an example configuration of a failure sign detection system 100 according to an embodiment of the present technology. As illustrated in FIG. 23 , the failure sign detection system 100 further includes a notification unit 16 that notifies a user of a detection result. The notification unit 16 may be an LED indicator or a liquid crystal display that provides a visual alert, a buzzer or alarm sound that provides an auditory alert, a remote notification system via email or SMS, a notification to a smartphone application, a built-in alert in a control system, or even a notification via a user-customizable dashboard.

[0143] Alternatively, a smart speaker or a voice assistant device may be used as the notification unit 16. In this case, the notification unit 16 directly notifies the user of the equipment failure prediction results by voice. For example, if abnormal vibration is detected from the equipment, the notification unit 16 issues a specific message such as "Caution: Abnormal vibration has been detected in the pump. Please check immediately" to warn the user.

[0144] Users can also ask the voice assistant for more information, such as "What is the cause of the abnormality?" or "Tell me about preventive measures," and the voice assistant can provide additional information.

[0145] The above description of the failure sign detection system according to the sixth embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0146] 7. Seventh Embodiment of the Present Technology (Example of Failure Sign Detection Method) The present technology provides a failure sign detection method that includes calculating an abnormality level of a device based on a vibration acceleration frequency spectrum included in vibration acceleration data of the device, detecting a change in a state of the device based on a vibration acceleration RMS included in the vibration acceleration data, and detecting a sign of failure of the device based on the abnormality level and the state change.

[0147] A failure sign detection method according to this embodiment will be described with reference to Fig. 24. Fig. 24 is a flowchart showing an example of the flow of the failure sign detection method according to an embodiment of the present technology.

[0148] As shown in FIG. 24, in step S1, the calculation unit calculates the degree of abnormality of the equipment based on the vibration acceleration frequency spectrum included in the vibration acceleration data of the equipment.

[0149] In step S2, the calculation unit detects a change in the state of the device based on the vibration acceleration RMS included in the vibration acceleration data, and detects a sign of a failure of the device based on the degree of abnormality and the state change.

[0150] To realize the failure sign detection method according to this embodiment, for example, the above-described failure sign detection system may be applied.

[0151] The above description of the failure sign detection method according to the seventh embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0152] 8. Eighth Embodiment of the Present Technology (Example of Failure Sign Detection Program) The present technology provides a failure sign detection program that causes a computer to calculate an abnormality level of a device based on a vibration acceleration frequency spectrum included in vibration acceleration data of the device, detect a change in a state of the device based on a vibration acceleration RMS included in the vibration acceleration data, and detect a sign of failure of the device based on the abnormality level and the state change.

[0153] This failure sign detection program is designed to execute various processes on a computer. The processes in the failure sign detection program are controlled by a computing unit (e.g., a CPU or GPU) included in the computer. This computing unit quickly executes a series of calculations required for failure sign detection, such as analyzing collected data, calculating the degree of anomaly, and detecting state changes.

[0154] This failure sign detection program can be recorded on a non-volatile computer-readable storage medium (e.g., a flash drive, SSD (Solid State Drive), ROM (Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), etc.), a magnetic storage medium (e.g., a hard disk drive or a floppy disk), or an optical storage medium (e.g., a CD-ROM, a DVD-ROM, or a Blu-ray disk). When this storage medium is installed in a computer system, it enables the computer to execute the failure sign detection program.

[0155] The above description of the failure sign detection program according to the eighth embodiment of the present technology can be applied to other embodiments of the present technology unless there is a particular technical contradiction.

[0156] It should be noted that the embodiments of the present technology are not limited to the above-described embodiments, and various modifications are possible within the scope of the present technology. The specific numerical values, shapes, materials (including compositions), etc. described in each embodiment are merely examples, and the present technology is not limited to these.

[0157] The present technology can also be configured as follows. [1] A failure sign detection system comprising: an abnormality degree calculation unit that calculates an abnormality degree of a device based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the device; and a sign detection unit that detects a change in a state of the device based on a vibration acceleration RMS included in the vibration acceleration data and detects a sign of a failure of the device based on the abnormality degree and the state change. [2] The failure sign detection system described in [1], in which the abnormality degree calculation unit identifies a vibration pattern of the device during normal operation based on time-series data of the vibration acceleration frequency spectrum and calculates an abnormality degree according to the degree of deviation from normal operation. [3] The failure sign detection system described in [1] or [2], in which the abnormality degree calculation unit identifies a vibration pattern of the device during normal operation using statistical analysis. [4] The failure sign detection system described in any one of [1] to [3], in which the abnormality degree calculation unit identifies a vibration pattern of the device during normal operation using a machine learning algorithm. [5] The failure sign detection system according to any one of [1] to [4], wherein the sign detection unit learns the vibration pattern of the equipment during normal operation based on the time-series data of the vibration acceleration RMS, and detects a state change corresponding to a deviation from normal operation. [6] The failure sign detection system according to any one of [1] to [5], wherein the sign detection unit learns the vibration pattern of the equipment during normal operation using statistical analysis. [7] The failure sign detection system according to any one of [1] to [6], wherein the sign detection unit learns the vibration pattern of the equipment during normal operation using a machine learning algorithm. [8] The failure sign detection system according to any one of [1] to [7], further comprising a preprocessing unit that preprocesses the vibration acceleration data, wherein if there is a period during which the value of the vibration acceleration data does not satisfy a reference value, the preprocessing unit excludes the vibration acceleration data for that period. [9] The failure sign detection system according to any one of [1] to [8], further comprising a sensor information acquisition unit that senses and acquires the vibration acceleration data.

[10] The failure sign detection system according to any one of [1] to [9], further comprising an operating information acquisition unit that acquires operating condition information of the equipment, wherein the sign detection unit further detects a sign of failure of the equipment based on the operating condition information.

[11] The failure sign detection system according to

[10] , wherein, when the equipment is a rotating equipment, the operating condition information includes information on the rotation of the rotating equipment.

[12] The failure sign detection system according to any one of [1] to

[11] , further comprising a visualization display unit that visualizes and displays the detection result.

[13] The failure sign detection system according to any one of [1] to

[12] , further comprising a notification unit that notifies a user of the detection result.

[14] The failure sign detection system according to any one of [1] to

[13] , wherein the equipment is a rotating equipment.

[15] A failure sign detection method comprising: calculating a degree of abnormality of a device based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the device; detecting a change in state of the device based on the vibration acceleration RMS included in the vibration acceleration data, and detecting a sign of failure of the device based on the degree of abnormality and the change in state.

[16] A failure sign detection program that causes a computer to execute the following: calculating a degree of abnormality of the device based on the vibration acceleration frequency spectrum included in vibration acceleration data acquired from the device; detecting a change in state of the device based on the vibration acceleration RMS included in the vibration acceleration data, and detecting a sign of failure of the device based on the degree of abnormality and the change in state.

[0158] REFERENCE SIGNS LIST 100 Failure sign detection system 1 Failure sign detection device 11 Calculation unit 111 Abnormality degree calculation unit 112 Sign detection unit 113 Preprocessing unit 114 Operation information acquisition unit 12 Memory 13 Power supply unit 14 User interface 15 Visualization display unit 16 Notification unit 2 Communication interface 3 Sensor information acquisition unit 4 Sensor 5 Equipment S1 Calculating abnormality degree S2 Detecting signs

Claims

1. A failure sign detection system comprising: an abnormality degree calculation unit that calculates the degree of abnormality of equipment based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the equipment; and a sign detection unit that detects changes in the state of the equipment based on the vibration acceleration RMS (Root Mean Square) included in the vibration acceleration data, and detects signs of failure of the equipment based on the abnormality degree and the state change.

2. The failure sign detection system according to claim 1, wherein the anomaly degree calculation unit learns the vibration pattern of the equipment during normal operation based on the time series data of the vibration acceleration frequency spectrum, and calculates the anomaly degree according to the degree of deviation from normal operation.

3. The failure sign detection system according to claim 1, wherein the anomaly degree calculation unit uses statistical analysis to learn the vibration pattern of the equipment during normal operation.

4. The failure sign detection system according to claim 1, wherein the anomaly degree calculation unit uses a machine learning algorithm to learn the vibration pattern of the equipment during normal operation.

5. A failure sign detection system as described in claim 1, wherein the sign detection unit learns the vibration pattern of the equipment during normal operation based on the time series data of the vibration acceleration RMS, and detects state changes that correspond to deviations from normal operation.

6. The failure sign detection system according to claim 1, wherein the sign detection unit uses statistical analysis to learn the vibration pattern of the equipment during normal operation.

7. The failure sign detection system according to claim 1, wherein the sign detection unit uses a machine learning algorithm to learn the vibration pattern of the equipment during normal operation.

8. The failure sign detection system according to claim 1, further comprising a pre-processing unit that pre-processes the vibration acceleration data, wherein if there is a period during which the value of the vibration acceleration data does not satisfy a reference value, the pre-processing unit excludes the vibration acceleration data for that period.

9. The failure sign detection system according to claim 1, further comprising a sensor information acquisition unit that senses and acquires the vibration acceleration data.

10. A failure sign detection system as described in claim 1, further comprising an operating information acquisition unit that acquires operating condition information of the equipment, and wherein the sign detection unit further detects signs of failure of the equipment based on the operating condition information.

11. The failure sign detection system according to claim 10, wherein when the device is a rotating device, the operating condition information includes information relating to the rotation of the rotating device.

12. The failure sign detection system according to claim 1, further comprising a visualization display unit that visualizes and displays the detection results.

13. The failure sign detection system according to claim 1, further comprising a notification unit that notifies a user of the detection result.

14. The failure sign detection system according to claim 1, wherein the device is a rotating device.

15. A failure sign detection method comprising: calculating an abnormality level of a device based on a vibration acceleration frequency spectrum included in vibration acceleration data acquired from the device; detecting a change in the state of the device based on the vibration acceleration RMS included in the vibration acceleration data; and detecting a sign of a failure of the device based on the abnormality level and the change in state.

16. A failure sign detection program that causes a computer to perform the following operations: calculate the degree of abnormality of a device based on the vibration acceleration frequency spectrum included in vibration acceleration data acquired from the device; detect a change in the state of the device based on the vibration acceleration RMS included in the vibration acceleration data; and detect signs of failure of the device based on the degree of abnormality and the change in state.

Images