Lubricating method based on multi-source information fusion and single-point lubricating device

Abstract

The invention discloses a lubrication method and a single-point lubrication device based on multi-source information fusion, and relates to the technical field of lubrication. The method comprises the following steps: acquiring various state data of the rotating equipment, and determining various characteristic values according to the various state data; performing lubrication state judgment on the various characteristic values through a double-channel judgment mechanism to obtain a first judgment result and a second judgment result; the first judgment result and the second judgment result are fused, and the lubricating state of the rotating equipment is obtained; and if the lubrication state is insufficient lubrication, grease injection is conducted on the rotating equipment according to the grease injection strategy. According to the method, the rotating equipment can be accurately lubricated, so that the problem of excessive lubrication or insufficient lubrication of the rotating equipment is avoided.

Description

Technical Field

[0001] The present invention relates to the field of lubrication technology, and in particular to a lubrication method and a single-point lubrication device based on multi-source information fusion. Background Art

[0002] At present, in industrial rotating equipment, such as motors, fans, pumps, compressors, etc., their bearings are in high-speed and high-load conditions for a long time, and the lubrication status directly affects their operating life and failure risk.

[0003] Traditional lubrication methods mostly rely on manual periodic greasing or timed and quantitative automatic greasers for greasing. For example, traditional single-point automatic greasers are usually based on a time control mode, injecting grease at a fixed frequency and injection volume. However, this lubrication method is prone to over-lubrication (causing grease waste, temperature rise, and seal failure) or under-lubrication (leading to problems such as increased friction temperature and increased bearing wear).

[0004] Therefore, there is an urgent need in the prior art for a precise lubrication method to avoid the problem of over-lubrication or under-lubrication of rotating equipment. Summary of the Invention

[0005] Based on this, it is necessary to provide a lubrication method and a single-point lubrication device based on multi-source information fusion to address the above technical problems. This method can accurately lubricate rotating equipment to avoid over-lubrication or under-lubrication of rotating equipment.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides a lubrication method based on multi-source information fusion, which is applied to a single-point lubrication device. The method comprises:

[0008] Collect various state data of rotating equipment and determine various characteristic values ​​based on the various state data;

[0009] The lubrication status is judged by using a dual-channel judgment mechanism on multiple characteristic values ​​to obtain a first judgment result and a second judgment result;

[0010] The first judgment result and the second judgment result are combined to obtain the lubrication state of the rotating equipment;

[0011] If the lubrication status is insufficient, grease the rotating equipment according to the grease injection strategy.

[0012] Optionally, the multiple state data include temperature, vibration signal, and sound signal; the multiple state data of the rotating device are obtained, and multiple characteristic values ​​are determined based on the multiple state data, including:

[0013] The temperature, vibration and sound signals of the rotating equipment are collected by the temperature sensor, vibration sensor and sound sensor installed on the rotating equipment during the preset operation time.

[0014] Obtain the mean and rate of change of temperature, the root mean square, peak-to-peak value and kurtosis of vibration signals, as well as the spectrum energy, main frequency position and transient change indicators of sound signals;

[0015] The mean and change rate of temperature, the root mean square, peak-to-peak value and kurtosis of the vibration signal, as well as the spectral energy, main frequency position and transient change index of the sound signal are determined as multiple characteristic values.

[0016] Optionally, the dual-channel judgment mechanism includes a deviation judgment unit based on an empirical threshold and a classification judgment of a machine learning model; the lubrication state is judged on multiple characteristic values ​​through the dual-channel judgment mechanism to obtain a first judgment result and a second judgment result, including:

[0017] Inputting the multiple feature values ​​into a deviation judgment unit based on an empirical threshold, obtaining multiple key features from the multiple feature values, comparing each key feature with a corresponding key feature reference value, and determining whether the key feature is abnormal; determining a first judgment result based on the number of abnormal key features and the number threshold;

[0018] Multiple characteristic values ​​are input into the machine learning model to obtain a second judgment result; the second judgment result is the probability that the rotating equipment is underlubricated.

[0019] Optionally, the multiple key features include a temperature change rate, a root mean square value of a vibration signal, and a main frequency energy of a sound signal; and comparing each key feature with a corresponding key feature reference value to determine whether the key feature is abnormal includes:

[0020] Comparing the temperature change rate with a temperature change rate threshold, and determining that the temperature change rate is abnormal if a deviation of the temperature change rate exceeds a first deviation threshold;

[0021] Comparing the root mean square of the vibration signal with a root mean square value threshold, and determining that the root mean square of the vibration signal is abnormal if a relative deviation of the root mean square of the vibration signal exceeds a second deviation threshold;

[0022] The main frequency energy of the sound signal is compared with the main frequency energy threshold. If the increase in the main frequency energy of the sound signal exceeds the third deviation threshold, it is determined that the main frequency energy of the sound signal is abnormal.

[0023] Optionally, determining the first judgment result based on the number of abnormal key features and the number threshold includes:

[0024] When the number of abnormal key features is greater than or equal to the number threshold, determining the first judgment result to be 1, where 1 indicates insufficient lubrication;

[0025] When the number of abnormal key features is less than the number threshold, the first judgment result is determined to be 0, where 0 indicates normal lubrication.

[0026] Optionally, the first judgment result and the second judgment result are combined to obtain the lubrication state of the rotating equipment, including:

[0027] Weighting the first judgment result and the second judgment result to obtain a fusion score;

[0028] If the fusion score is greater than or equal to the global judgment threshold, the lubrication state of the rotating equipment is determined to be insufficient lubrication;

[0029] If the fusion score is less than the global judgment threshold, the lubrication state of the rotating equipment is determined to be normal.

[0030] Optionally, grease the rotating equipment according to the grease injection strategy, including:

[0031] Get the historical grease injection amount for rotating equipment within a historical time period;

[0032] If the historical fat injection amount is greater than the fat injection amount threshold, an alarm will be issued;

[0033] If the historical grease injection amount is less than or equal to the grease injection amount threshold, the preset grease injection amount is injected into the rotating equipment and the number of grease injections is increased by 1.

[0034] The present invention provides a single-point lubrication device, which includes a grease injector, a state sensing module, a lubrication state identification module, and an intelligent control module;

[0035] The state sensing module is installed on the rotating equipment and is used to collect various state data of the rotating equipment in real time and determine various characteristic values ​​based on the various state data;

[0036] A lubrication status identification module is used to determine the lubrication status of the rotating equipment by using a dual-channel judgment mechanism based on multiple characteristic values ​​to obtain a first judgment result and a second judgment result; the first judgment result and the second judgment result are combined to determine the lubrication status of the rotating equipment;

[0037] The intelligent control module is used to inject grease into the rotating equipment through a grease injector according to a grease injection strategy if the lubrication state is insufficient lubrication.

[0038] Optionally, the device further includes a power module, a communication module, and a display screen;

[0039] A power module is used to supply power to various components in the device;

[0040] Communication module, used to transmit status data, characteristic values ​​and lubrication status in the network;

[0041] The display screen is used to show the current lubrication status of the equipment, the last grease injection time, the cumulative grease injection amount, the battery power, the signal strength, and the characteristic value.

[0042] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the lubrication method based on multi-source information fusion is implemented.

[0043] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the lubrication method based on multi-source information fusion is implemented.

[0044] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects:

[0045] Collecting multiple status data of rotating equipment and determining multiple characteristic values ​​makes the description of equipment status more comprehensive and accurate. By comprehensively considering multiple aspects of characteristics, the actual operating status of rotating equipment can be more accurately reflected, avoiding the one-sidedness of judgment based on single data. In addition, the dual-channel mechanism is used to judge the lubrication status of multiple characteristics, which is equivalent to introducing double verification. The lubrication status is evaluated in different ways and then the judgment results of the two channels are integrated to improve the accuracy of the judgment and obtain a lubrication status conclusion that is more in line with the actual situation. Then, based on the lubrication status, accurate grease supply is achieved to ensure that the rotating equipment operates under appropriate lubrication conditions and reduce equipment wear and failure caused by improper lubrication. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0047] Figure 1 A schematic structural diagram of a single-point lubrication device provided by the present invention;

[0048] Figure 2 A schematic flow chart of a lubrication method based on multi-source information fusion provided by the present invention;

[0049] Figure 3 A schematic diagram of a computer device for implementing a lubrication method based on multi-source information fusion provided by the present invention.

[0050] Description of reference numerals:

[0051] 1. Grease storage tank; 2. Motor and plunger assembly; 3. Power module; 4. Central control unit; 5. Display screen; 6. Status sensing module. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0053] The traditional lubrication method still has the following problems: (1) Lack of lubrication status perception and feedback loop. Although some high-end grease injection systems on the market have remote control and quantitative grease dispensing functions, they rely on manual judgment and operation. The system itself does not have the ability to perceive the lubrication status and cannot autonomously adjust the lubrication strategy according to the actual working conditions. (2) Failure to effectively integrate multi-source status signals for lubrication judgment. Most of the intelligent devices currently used for lubrication-related status monitoring only collect a single physical quantity (such as temperature) and fail to simultaneously integrate multiple sensitive signals such as temperature, vibration, and sound for fusion analysis, which limits the accurate judgment of the lubrication status. Especially in high-noise and high-interference industrial environments, the reliability of a single signal is low, and multi-source information fusion is urgently needed to improve robustness and diagnostic accuracy. (3) The structural integration is not high, the volume is large, and it is not convenient to install in space-constrained scenarios. Existing intelligent lubrication devices are mostly centralized lubrication systems or multi-point control systems with complex structures, cumbersome installation, and high maintenance costs. They are not suitable for use in equipment with narrow space and inconvenient maintenance, such as wind turbine towers, nuclear power shafts, and mining transportation equipment.

[0054] For example, one approach uses a multi-point lubrication system based on data-driven monitoring that combines vibration and temperature. This approach achieves real-time, on-demand grease injection, improving the accuracy of grease injection. However, this approach is designed for multi-point lubrication systems and is complex, relying on an integrated control terminal and data analysis server. It's not suitable for independent deployment of single-point lubrication. Furthermore, it suffers from the simplicity of signal processing and the fixed ratio (20%, 5%) of grease injection strategies.

[0055] Another approach is to propose an automatic single-point greaser that can be remotely controlled and precisely control the amount of grease delivered, solving the problem of manual grease delivery. However, this automatic single-point greaser relies on counting the number of gear turns to control the amount of grease delivered, without taking into account the actual operating status of the equipment (such as the degree of bearing wear). This lacks intelligence and makes it impossible to deliver grease based on the actual situation, which can easily lead to over- or under-greasing.

[0056] Therefore, the present invention provides a lubrication method and a single-point lubrication device based on multi-source information fusion. The method can accurately lubricate rotating equipment to avoid the problem of over-lubrication or insufficient lubrication of rotating equipment. The device is a single-point greasing device with a compact structure, independent operation, and multi-source status signal acquisition and intelligent analysis capabilities, so as to realize the true sense of the situation-based greasing function.

[0057] The technical solutions provided by various embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0058] In order to clearly illustrate the lubrication method based on multi-source information fusion provided by the present invention, a single-point lubrication device is first described. The device includes a grease injector, a state sensing module, a lubrication state recognition module, and an intelligent control module.

[0059] The state sensing module is installed on the rotating equipment and is used to collect various state data of the rotating equipment in real time and determine various characteristic values ​​based on the various state data.

[0060] The lubrication state identification module is used to judge the lubrication state of multiple characteristic values ​​through a dual-channel judgment mechanism to obtain a first judgment result and a second judgment result; the first judgment result and the second judgment result are integrated to obtain the lubrication state of the rotating equipment.

[0061] The intelligent control module is used to inject grease into the rotating equipment through a grease injector according to a grease injection strategy if the lubrication state is insufficient lubrication.

[0062] Specifically, the grease injector consists of a grease reservoir, a motor, a speed reduction mechanism, and a plunger assembly. The plunger assembly is located below the reservoir. The motor drives the plunger to reciprocate, thereby dispensing grease. Simultaneously, the intelligent control module precisely controls the number of pulses delivered by the drive motor to precisely adjust the plunger's displacement, thereby accurately controlling the amount of grease dispensed.

[0063] The state sensing module primarily includes a temperature sensor, a vibration sensor, an acoustic sensor, a signal conditioning and A / D conversion module, a microcontroller unit, a wireless communication module, a power supply module, and a mechanical external mount. The state sensing module is mounted on the bearing of the rotating equipment via the mechanical external mount. It senses the operating status of the rotating equipment in real time and transmits this data to the lubrication state identification module. The mechanical external mount can be a magnetic design.

[0064] The temperature sensor is a digital temperature sensor used to monitor the operating temperature of rotating equipment. The vibration sensor is used to monitor the vibration amplitude and frequency characteristics of the equipment during operation, with a frequency response range of 1Hz-50kHz. The sound sensor is used to collect structural noise or impact sound during the operation of the rotating equipment to determine the lubrication status or abnormal conditions. The signal conditioning and A / D conversion module includes an amplifier, an anti-aliasing filter, and an analog-to-digital converter (ADC) to convert analog signals into digital signals for processing by the microcontroller unit. The microcontroller unit (MCU) has a built-in data processing algorithm to complete feature value extraction (such as the mean and rate of change of temperature, the root mean square (RMS) of the vibration signal). The system can collect and package the time-domain statistical characteristics of the sound signal (such as square, RMS), peak-to-peak value, kurtosis, etc., as well as the spectrum energy, main frequency position, transient change indicators, etc. of the sound signal; the wireless communication module is a communication module that can be integrated with BLE, LoRa, ZigBee, Wi-Fi or NB-IoT, and sends characteristic data to the lubrication status identification module according to the set period to realize data synchronization and remote analysis; the power supply module supports lithium battery power supply and has a built-in low-power management system to support long-term operation.

[0065] The lubrication status recognition module, housed in the central control unit of a single-point lubrication system, receives characteristic data from temperature, vibration, and acoustic signals collected by the status sensing module. Using a multidimensional fusion judgment method, it assesses the equipment's lubrication status in real time, determining whether there are conditions such as insufficient lubrication, excessive lubrication, or abnormal lubrication, providing a basis for greasing control. Specifically, it compares and analyzes multidimensional features with preset reference values ​​or historical operating condition statistics to identify abnormal conditions such as insufficient lubrication, excessive lubrication, or lubrication failure. The lubrication status recognition module integrates lightweight machine learning models, such as support vector machines (SVMs) and multilayer perceptrons (MLPs), to facilitate fusion and judgment between multidimensional features. Finally, a weighted fusion of the two analysis results is performed to generate a lubrication status judgment result, which is output as a status signal to trigger the intelligent control module, enabling situation-based greasing decision-making and control. The lubrication status recognition module possesses self-learning and parameter updating capabilities, dynamically adjusting judgment thresholds based on long-term equipment operating trends. This improves the accuracy and robustness of lubrication status recognition and provides data-driven decision support for the entire system.

[0066] As the core unit of the single-point lubrication system, the intelligent control module receives the lubrication status of the rotating equipment from the lubrication status recognition module. Combined with the configured greasing strategy, it precisely controls the greaser's startup, operating time, and grease delivery, achieving condition-based greasing control. The intelligent control module integrates a highly reliable microprocessor or control chip with built-in state logic judgment and greasing control algorithms. The intelligent control module sets different greasing response strategies based on the lubrication status (e.g., normal lubrication versus insufficient lubrication). When insufficient lubrication occurs, the module performs a small-dose greasing injection and then reenters the data monitoring and status judgment process, achieving closed-loop regulation. Furthermore, the intelligent control module precisely controls the operation of the motor drive system, adjusting the greasing amount by setting the pulse count to ensure sufficient and complete greasing without overdosing. The intelligent control module also records and compiles data such as greasing amount, greasing interval, and equipment operating time for lubrication maintenance management and data traceability. The single-point lubrication system supports remote parameter configuration and online upgrades, receiving commands via wireless communication to adapt to changing lubrication requirements of different equipment. The overall control logic has a closed-loop capability of state perception-judgment-control-feedback, realizing true on-demand greasing and equipment self-lubrication adjustment functions, improving equipment operation reliability and maintenance efficiency.

[0067] The central control unit supports deep sleep and RTC timed wake-up, and can automatically wake up, collect, process and upload data after setting the sampling period; in the current embodiment, the central control unit supports BLE and Wi-Fi, and can upload lubrication status, sensor data, grease injection records, etc. to the host computer or cloud server wirelessly, and also supports remote parameter updates and firmware OTA upgrades.

[0068] Optionally, the device also includes a power module, a communication module, and a display screen; the power module is used to supply power to each component in the device; the communication module is used to transmit status data, characteristic values ​​and lubrication status in the network; the display screen is used to display the current lubrication status of the equipment, the last grease injection time, the cumulative grease injection amount, the battery power, the signal strength, and the characteristic value.

[0069] The power supply and communication module provides stable energy support and data exchange channels for the intelligent single-point lubrication device, and is the basic guarantee for achieving autonomous operation, remote monitoring and intelligent control. The power supply module uses a high-energy-density lithium battery as the main power supply unit. Combined with the system's low-power design and the microcontroller's sleep / wake-up mechanism, it significantly improves battery life and can stably support the single-point lubrication device to operate continuously for 3 months without an external power supply. The communication module is based on the needs of multi-source information fusion and has high-reliability wireless data transmission capabilities. Depending on the usage scenario, the communication module supports flexible selection of communication methods such as Bluetooth (BLE), ZigBee, LoRa, NB-IoT or Wi-Fi, and can transmit multi-source perception characteristic values ​​such as temperature, vibration, sound and lubrication status assessment results in local networking.

[0070] The single-point lubrication device also includes a housing with a low-power OLED display on top, which displays the device's operating status and key parameters. This information includes the device's current lubrication status (normal / abnormal), the time of the last grease injection, the cumulative grease injection volume, battery charge, signal strength, and sensor data characteristics (such as temperature, RMS vibration, and sound frequency). The display is automatically updated by the control module and can be switched and viewed using buttons.

[0071] To enhance on-site operation, the device's housing features an integrated wake-up button, operating mode switch, and simple interactive menu interface for easy commissioning, maintenance, and emergency control. The display module, combined with the communication module, provides information on remote control status, firmware upgrade progress, and fault alarms, enhancing the device's human-machine interaction and intelligent response capabilities.

[0072] like Figure 1 As shown, Figure 1 This is a schematic diagram of the structure of a single-point lubrication device, which includes a grease reservoir 1, a motor and plunger assembly 2, a power module 3, a central control unit 4, a display 5, and a status sensing module 6. The central control unit 4 includes a lubrication status recognition module, an intelligent control module, and a communication module. The status sensing module 6 includes a temperature sensor, a vibration acceleration sensor, and an acoustic sensor. The central control unit 4 can be an ESP32S3 main control module.

[0073] Based on the above-mentioned single-point lubrication device, a lubrication method based on multi-source information fusion provided by the present invention is described in detail below. Figure 2 As shown, Figure 2 The figure is a flow chart of a lubrication method based on multi-source information fusion in the present invention, which specifically includes the following steps:

[0074] S101, collecting various state data of the rotating equipment, and determining various characteristic values ​​according to the various state data.

[0075] Optionally, the multiple state data include temperature, vibration signals and sound signals; the multiple state data of the rotating equipment are obtained, and multiple characteristic values ​​are determined based on the multiple state data, including: collecting the temperature, vibration signals and sound signals of the rotating equipment for a preset running time through temperature sensors, vibration sensors and sound sensors arranged on the rotating equipment; obtaining the mean and change rate of the temperature, the root mean square, peak-to-peak value and kurtosis of the vibration signal, and the spectral energy, main frequency position and transient change index of the sound signal; determining the mean and change rate of the temperature, the root mean square, peak-to-peak value and kurtosis of the vibration signal, and the spectral energy, main frequency position and transient change index of the sound signal as multiple characteristic values.

[0076] For example, the status signal of the rotating equipment during a preset time period can be collected by a temperature sensor, a vibration sensor, and a sound sensor installed on the rotating equipment, and the corresponding characteristic values ​​can be extracted, including but not limited to the mean and change rate of the temperature, the root mean square, peak-to-peak value, kurtosis and other time domain statistical characteristics of the vibration signal, and the spectral energy, main frequency position, transient change index, etc. of the sound signal, wherein the preset time period can be 30s.

[0077] S102 , performing lubrication status judgment on the multiple characteristic values ​​through a dual-channel judgment mechanism to obtain a first judgment result and a second judgment result.

[0078] Optionally, the dual-channel judgment mechanism includes a deviation judgment unit based on an experience threshold and a classification judgment of a machine learning model; the lubrication status of multiple characteristic values ​​is judged through the dual-channel judgment mechanism to obtain a first judgment result and a second judgment result, including: inputting multiple characteristic values ​​into the deviation judgment unit based on the experience threshold, obtaining multiple key features from the multiple characteristic values, and comparing each key feature with the corresponding key feature reference value to determine whether the key feature is abnormal; determining the first judgment result based on the number of abnormal key features and the quantity threshold; inputting multiple characteristic values ​​into the machine learning model to obtain a second judgment result; the second judgment result is the probability that the rotating equipment is underlubricated.

[0079] Key characteristic reference values ​​are derived from steady-state means and statistical intervals (e.g., mean ±3σ) in historical data, or from average values ​​collected from long-term operation of the equipment in a "healthy lubrication state." These reference values ​​are preset during system installation and commissioning and can be dynamically adjusted based on actual operating conditions.

[0080] In order to improve the effectiveness and sensitivity of the input features in lubrication status judgment, it is preferred to use the principal component analysis (PCA) method to reduce the dimension of the original features and screen the sensitivity, and extract key features from multiple eigenvalues. Specifically, the following steps are included:

[0081] First, the various features extracted from the multi-source signals (temperature, vibration, and sound) collected by the state perception module (including but not limited to the mean and change rate of temperature, the root mean square, peak-to-peak value, kurtosis and other time domain statistical features of the vibration signal, and the spectral energy, main frequency position, transient change index, etc. of the sound signal) are combined into the original feature matrix.

[0082] Secondly, to eliminate the dimensional influence between different features, each column of features is standardized, and the covariance matrix of the standardized feature matrix is ​​calculated. The covariance matrix is ​​then subjected to eigenvalue decomposition to obtain an eigenvalue sequence and the corresponding eigenvector. By calculating the proportion of each eigenvalue in the total eigenvalue, the contribution rate of each principal component is obtained, and the contribution rates are accumulated in sequence, and the first few principal components whose cumulative contribution rates exceed a preset threshold (such as 95%) are selected.

[0083] Finally, based on the eigenvector of each principal component, its contribution in the original feature space is analyzed, and the first few original features with the largest proportion in the eigenvector are selected as key features. The number of key features is not limited and can be set according to actual needs.

[0084] Optionally, multiple key features include the temperature change rate, the root mean square of the vibration signal and the main frequency energy of the sound signal; each key feature is compared with the corresponding key feature reference value to determine whether the key feature is abnormal, including: comparing the temperature change rate with the temperature change rate threshold, if the deviation of the temperature change rate exceeds the first deviation threshold, then determining that the temperature change rate is abnormal; comparing the root mean square of the vibration signal with the root mean square value threshold, if the relative deviation of the root mean square of the vibration signal exceeds the second deviation threshold, then determining that the root mean square of the vibration signal is abnormal; comparing the main frequency energy of the sound signal with the main frequency energy threshold, if the main frequency energy increase of the sound signal exceeds the third deviation threshold, then determining that the main frequency energy of the sound signal is abnormal.

[0085] The absolute value of the difference between the temperature change rate and the temperature change rate threshold represents the deviation of the temperature change rate; the absolute value of the difference between the root mean square (RMS) value of the vibration signal and the RMS value threshold represents the relative deviation of the RMS value of the vibration signal; and the difference between the main frequency energy of the sound signal and the main frequency energy threshold represents the increase in the main frequency energy of the sound signal. The specific values ​​of the first, second, and third deviation thresholds can be set based on extensive experiments or actual needs.

[0086] Optionally, a first judgment result is determined based on the number of abnormal key features and a quantity threshold, including: when the number of abnormal key features is greater than or equal to the quantity threshold, the first judgment result is determined to be 1, where 1 indicates insufficient lubrication; when the number of abnormal key features is less than the quantity threshold, the first judgment result is determined to be 0, where 0 indicates normal lubrication.

[0087] Specifically, when the deviation of the temperature change rate exceeds 1.5℃ / min, or the relative deviation of the vibration root mean square value exceeds 50%, or the main frequency energy of the sound signal is increased by more than 10dB, it is judged that the corresponding characteristic item is abnormal. When the number of abnormal items exceeds 2 (quantity threshold), the first judgment result is 1, indicating insufficient lubrication.

[0088] The machine learning model is a multi-layer perceptron model, a support vector machine (SVM), or a lightweight neural network. Multiple feature values ​​are input into the machine learning model to obtain a second judgment result output by the machine learning model; the second judgment result is the probability that the rotating equipment is underlubricated.

[0089] Optionally, before inputting multiple eigenvalues ​​into the machine learning model, the multiple eigenvalues ​​are first normalized, and then the normalized eigenvalues ​​are input into the machine learning model to obtain a second judgment result.

[0090] Optionally, taking a multilayer perceptron model as an example, the machine learning model includes an input layer, at least one hidden layer, and an output layer. In a preferred embodiment, the multilayer perceptron model includes two hidden layers, each with 8 neurons, using a ReLU activation function, and an output layer with a single neuron that uses a Sigmoid activation function to output a probability value between 0 and 1, indicating the confidence level that the current lubrication state is "insufficient lubrication."

[0091] The machine learning model is trained using supervised learning. Training samples are derived from historical operating data of rotating equipment under normal and insufficient lubrication conditions, with binary labels (normal / inadequate). The training process uses a cross-entropy loss function, optimizing model weights through a backpropagation algorithm. During model deployment, multi-source feature values ​​collected in real time are fed into the machine learning model to generate an output probability value. When this value exceeds 0.5, the equipment is considered to be underlubricated; otherwise, lubrication is considered normal.

[0092] Specifically, after collecting and extracting multi-source feature values, the lubrication status identification module inputs them into two independent judgment channels respectively: one is the empirical threshold deviation channel, which outputs the first judgment result S1 as 0 or 1 (0 represents normal lubrication, 1 represents insufficient lubrication); the other is the machine learning model channel, which outputs the second judgment result as a probability value S2∈[0,1], which represents the confidence of the machine learning model that the current state is "insufficient lubrication".

[0093] S103: The first judgment result and the second judgment result are combined to obtain the lubrication status of the rotating equipment.

[0094] In order to improve the accuracy and robustness of lubrication state identification, a weighted fusion algorithm is introduced in the lubrication state identification module to fuse the first judgment result based on the empirical threshold and the second judgment result based on the machine learning model to form the final lubrication state identification result.

[0095] Optionally, the first judgment result and the second judgment result are fused to obtain the lubrication status of the rotating equipment, including: weighting the first judgment result and the second judgment result to obtain a fusion score; if the fusion score is greater than or equal to the global judgment threshold, the lubrication status of the rotating equipment is determined to be insufficient lubrication; if the fusion score is less than the global judgment threshold, the lubrication status of the rotating equipment is determined to be normal lubrication.

[0096] Specifically, the fusion score S f The calculation formula is: S f =α·S1+β·S2; where S1 is the first judgment result, S2 is the second judgment result, α represents the weight coefficient of empirical judgment, and β represents the weight coefficient of machine learning model judgment, satisfying α+β=1. The specific values ​​of these weight parameters can be set based on actual working conditions, historical accuracy, and field verification results. For example, when the model has sufficient training data and excellent performance, the weight of β can be appropriately increased to enhance the ability to recognize complex patterns. If the equipment environment fluctuates significantly or the model is still in the optimization stage, the proportion of α can be increased to ensure judgment stability.

[0097] According to the fusion score S f and the set global decision threshold T f Compare and determine the final judgment result:

[0098]

[0099] Among them, T f It is usually set between 0.5 and 0.7 and can be dynamically adjusted according to the system sensitivity requirements.

[0100] S104: If the lubrication state is insufficient lubrication, grease the rotating equipment according to the grease injection strategy.

[0101] Optionally, according to the greasing strategy, greasing is performed for the rotating equipment, including: obtaining the historical greasing amount of the rotating equipment within a historical time period; if the historical greasing amount is greater than the situational greasing amount threshold, an alarm is issued; if the historical greasing amount is less than or equal to the situational greasing amount threshold, the rotating equipment is greasing with a preset greasing amount, and the number of greasing times is increased by 1.

[0102] If the lubrication status is determined to be insufficient, the intelligent control module triggers the grease injection control unit to initiate grease injection. This grease injection control utilizes a quantitative grease injection strategy, controlling the number of plunger pump strokes to deliver a set volume of grease. After grease injection is complete, characteristic values ​​are collected and analyzed again to confirm that the lubrication status has returned to normal, thus implementing closed-loop control. 30 minutes after the lubrication status returns to normal, steps S101-S104 are repeated.

[0103] It should be noted that in order to prevent misjudgment caused by accidental events (such as instantaneous fluctuations, environmental interference, measurement errors, etc.), the lubrication state of the rotating equipment can be determined to be insufficient lubrication only when it is judged to be insufficient lubrication for multiple consecutive times. Only when the lubrication state of the rotating equipment is judged to be insufficient lubrication for multiple consecutive times, the rotating equipment is greased according to the grease injection strategy.

[0104] The present invention also provides a lubrication method based on multi-source information fusion, which specifically includes:

[0105] S201, setting the fat injection amount for a single action to x grams, and setting the fat injection amount threshold to Mx grams depending on the situation.

[0106] S202: Use the status perception module to collect information at a set interval of t minutes.

[0107] S203, the lubrication status recognition algorithm is started to determine whether the bearing lubrication status is "insufficient lubrication" for multiple times. If yes, enter S204, if not, return to S202.

[0108] S204 , recording the total number of operations of the single-point lubrication device n=n+1, the number of grease injections of the device K=0, and the total operation time of the single-point lubrication device as n·t.

[0109] S205: After a period of time, the lubrication status identification algorithm is started to determine the bearing lubrication status again.

[0110] S206: If the result of the lubrication status identification algorithm is normal, then return to S202; if the result is insufficient lubrication, then enter S207.

[0111] S207, determine whether the total amount of grease injected K·x (historical grease injection amount) is greater than the threshold value Mx of the grease injection amount according to the situation. If it is, an alarm is issued (the lubrication state does not return after grease injection, which may be a non-lubrication fault). If it is not, proceed to S208.

[0112] S208, the intelligent control module starts the grease injection program, the grease injection amount is x grams, the grease injection times K=K+1 are recorded, and the process returns to S205.

[0113] The present invention effectively reduces the volume of the single-point lubrication device by adopting integrated structural packaging, modular installation interface and miniaturized electric-driven grease injection mechanism technology. The single-point lubrication device can be quickly arranged and fixed in limited space areas such as bearing seats and motor end covers, thereby enhancing the engineering application value of the lubrication device.

[0114] By introducing multi-source information fusion and judgment technology, the collaborative perception and comprehensive analysis of multiple operating status signals such as temperature, vibration, and sound are achieved, effectively improving the accuracy and anti-interference ability of lubrication status identification.

[0115] By constructing a dual-channel judgment mechanism based on empirical thresholds and machine learning models, the combination of stable judgment of lubrication status and nonlinear feature recognition capabilities is achieved, which improves the system's adaptability and judgment accuracy under complex working conditions.

[0116] By designing a weighted fusion algorithm to comprehensively evaluate the two types of judgment results, a fusion decision-making process of data-driven and rule-based judgment is achieved, effectively reducing the probability of misjudgment and missed judgment.

[0117] By setting up an intelligent control module and a quantitative grease injection mechanism, grease injection control based on the lubrication status is achieved, avoiding over-lubrication and under-lubrication, and extending the service life of the equipment.

[0118] When applying the lubrication method based on multi-source information fusion provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0119] The specific definition of a single-point lubrication device can be found in the definition of the lubrication method based on multi-source information fusion above and will not be repeated here. Each module in the aforementioned single-point lubrication device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0120] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 2 A lubrication method based on multi-source information fusion is provided.

[0121] The present invention also provides Figure 3 The structural diagram of the computer equipment shown in FIG. Figure 3As shown in the figure, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 2 A lubrication method based on multi-source information fusion is provided.

[0122] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0123] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A lubrication method based on multi-source information fusion, characterized in that: Applied to a single-point lubrication device, the method comprises: Collect various state data of rotating equipment and determine various characteristic values ​​based on the various state data; The lubrication status is judged by using a dual-channel judgment mechanism on multiple characteristic values ​​to obtain a first judgment result and a second judgment result; The first judgment result and the second judgment result are combined to obtain the lubrication state of the rotating equipment; If the lubrication state is insufficient lubrication, grease is injected into the rotating equipment according to the grease injection strategy.

2. The method according to claim 1, characterized in that The multiple state data include temperature, vibration signal and sound signal; the multiple state data of the rotating device are obtained, and multiple characteristic values ​​are determined based on the multiple state data, including: The temperature, vibration and sound signals of the rotating equipment are collected by the temperature sensor, vibration sensor and sound sensor installed on the rotating equipment during the preset operation time. Obtain the mean and rate of change of temperature, the root mean square, peak-to-peak value and kurtosis of vibration signals, as well as the spectrum energy, main frequency position and transient change indicators of sound signals; The mean and change rate of temperature, the root mean square, peak-to-peak value and kurtosis of the vibration signal, as well as the spectral energy, main frequency position and transient change index of the sound signal are determined as multiple characteristic values.

3. The method according to claim 2, characterized in that The dual-channel judgment mechanism includes a deviation judgment unit based on an empirical threshold and a classification judgment of a machine learning model; the lubrication state is judged on multiple characteristic values ​​through the dual-channel judgment mechanism to obtain a first judgment result and a second judgment result, including: Inputting the multiple feature values ​​into a deviation judgment unit based on an empirical threshold, obtaining multiple key features from the multiple feature values, comparing each key feature with a corresponding key feature reference value, and determining whether the key feature is abnormal; determining a first judgment result based on the number of abnormal key features and the number threshold; Multiple characteristic values ​​are input into the machine learning model to obtain a second judgment result; the second judgment result is the probability that the rotating equipment is underlubricated.

4. The method according to claim 3, characterized in that The multiple key features include the temperature change rate, the root mean square of the vibration signal, and the main frequency energy of the sound signal; each key feature is compared with the corresponding key feature reference value to determine whether the key feature is abnormal, including: Comparing the temperature change rate with a temperature change rate threshold, and determining that the temperature change rate is abnormal if a deviation of the temperature change rate exceeds a first deviation threshold; Comparing the root mean square of the vibration signal with a root mean square value threshold, and determining that the root mean square of the vibration signal is abnormal if a relative deviation of the root mean square of the vibration signal exceeds a second deviation threshold; The main frequency energy of the sound signal is compared with the main frequency energy threshold. If the increase in the main frequency energy of the sound signal exceeds the third deviation threshold, it is determined that the main frequency energy of the sound signal is abnormal.

5. The method according to claim 3, characterized in that Determining the first judgment result based on the number of abnormal key features and the number threshold includes: When the number of abnormal key features is greater than or equal to the number threshold, determining the first judgment result to be 1, where 1 indicates insufficient lubrication; When the number of abnormal key features is less than the number threshold, the first judgment result is determined to be 0, where 0 indicates normal lubrication.

6. The method according to claim 5, characterized in that The step of fusing the first judgment result and the second judgment result to obtain the lubrication state of the rotating equipment includes: Weighting the first judgment result and the second judgment result to obtain a fusion score; If the fusion score is greater than or equal to the global judgment threshold, the lubrication state of the rotating equipment is determined to be insufficient lubrication; If the fusion score is less than the global judgment threshold, the lubrication state of the rotating equipment is determined to be normal.

7. The method according to claim 1, characterized in that The step of injecting grease into the rotating equipment according to the grease injection strategy includes: Get the historical grease injection amount of the rotating equipment within the historical time period; If the historical fat injection amount is greater than the fat injection amount threshold according to the situation, an alarm is issued; If the historical grease injection amount is less than or equal to the situational grease injection amount threshold, the preset grease injection amount is greased for the rotating equipment, and the number of grease injections is increased by 1.

8. A single-point lubrication device, characterized in that: The device includes a grease injector, a state sensing module, a lubrication state identification module, and an intelligent control module; The state sensing module is installed on the rotating equipment and is used to collect various state data of the rotating equipment in real time and determine various characteristic values ​​based on the various state data; A lubrication status identification module is used to determine the lubrication status of the rotating equipment by using a dual-channel judgment mechanism based on multiple characteristic values ​​to obtain a first judgment result and a second judgment result; the first judgment result and the second judgment result are combined to determine the lubrication status of the rotating equipment; The intelligent control module is used to inject grease into the rotating equipment through a grease injector according to a grease injection strategy if the lubrication state is insufficient lubrication.

9. The device according to claim 8, characterized in that The device also includes a power module, a communication module, and a display screen; The power supply module is used to supply power to various components in the device; The communication module is used to transmit status data, characteristic values ​​and lubrication status in the network; The display screen is used to show the current lubrication status of the equipment, the last grease injection time, the cumulative grease injection amount, the battery power, the signal strength, and the characteristic value.

Images