CN-121977155-A - Multi-feature trend fused motor lubrication distinguishing and intelligent oiling method and system

CN121977155ACN 121977155 ACN121977155 ACN 121977155ACN-121977155-A

Abstract

The invention relates to the technical field of motor state monitoring and intelligent maintenance, in particular to a motor lubrication distinguishing and intelligent oiling method and system with multi-feature trend fusion. The method comprises the steps of collecting vibration acceleration time domain signals of a motor, obtaining vibration speed time domain signals through frequency domain integration, calculating lubrication state characteristics such as acceleration effective values, speed effective values, kurtosis, high-frequency energy duty ratio and the like, constructing a characteristic history sequence, obtaining characteristic trend quantities through a sliding window, adaptively updating a dynamic base line through a forgetting factor strategy, executing multi-characteristic trend fusion judgment according to deviation relation between the characteristic trend quantities and the dynamic base line, carrying out fusion calculation on oil injection quantity coefficients when judging poor lubrication, controlling an oil injection executing mechanism to inject oil according to requirements, verifying characteristic regression conditions in a monitoring window period after oil injection, and carrying out secondary oil injection when no regression exists. The system comprises an acquisition and speed acquisition module, a characteristic acquisition module, a trend and baseline construction module, a lubrication judgment module and an oiling control module.

Inventors

ZANG TINGPENG
ZHANG WENGANG
ZHANG MINGXIA
CHENG BENJUN
CAO JIN

Assignees

浙江舜云互联技术有限公司

Dates

Publication Date: 20260505
Application Date: 20260123

Claims (10)

1. The motor lubrication distinguishing and intelligent oiling method with multi-feature trend fusion is characterized by comprising the following steps of: Step S1, collecting vibration acceleration time domain signals in the running process of a motor, and obtaining vibration speed time domain signals based on the vibration acceleration time domain signals; S2, acquiring a lubrication state feature set based on the vibration acceleration time domain signal and the vibration speed time domain signal, wherein the lubrication state feature set at least comprises an acceleration effective value, a speed effective value, kurtosis and a high-frequency energy duty ratio, and the high-frequency energy duty ratio represents the duty ratio of the energy of the vibration acceleration time domain signal in a preset high-frequency band relative to the energy of a full frequency band; step S3, establishing a historical sequence for each feature in the lubrication state feature set, calculating to obtain a corresponding feature trend quantity based on the historical sequence, simultaneously establishing a dynamic baseline for each feature, and adaptively updating the dynamic baseline based on a preset updating strategy; S4, outputting a lubrication judgment result according to a preset multi-feature trend fusion judgment rule according to a deviation relation of feature trend quantity of at least part of features in the lubrication state feature set relative to a corresponding dynamic base line; and S5, when the lubrication judging result indicates poor lubrication, carrying out fusion calculation according to the deviation degree of at least part of the characteristics in the lubrication state characteristic set to obtain an oil injection quantity coefficient, and outputting an oil injection control instruction to an oil injection executing mechanism so that the oil injection executing mechanism carries out oil injection on the motor according to the oil injection quantity determined by the oil injection quantity coefficient.
2. The multi-feature trend fused motor lubrication judging and intelligent oiling method according to claim 1 is characterized in that a vibration speed time domain signal is obtained based on the vibration acceleration time domain signal by performing frequency domain integration processing on the vibration acceleration time domain signal.
3. The multi-feature trend fused motor lubrication judging and intelligent oiling method according to claim 1, wherein obtaining the high-frequency energy duty ratio comprises the steps of carrying out high-pass filtering of a preset cut-off frequency on the vibration acceleration time domain signal to obtain a high-frequency component signal, respectively calculating an effective value of the vibration acceleration time domain signal and an effective value of the high-frequency component signal, and taking the ratio of the effective value of the high-frequency component signal and the effective value of the vibration acceleration time domain signal as the high-frequency energy duty ratio.
4. The multi-feature trend fused motor lubrication judging and intelligent oiling method according to claim 1 is characterized in that calculating the corresponding feature trend quantity based on the history sequence comprises calculating the feature trend quantity by moving average of the history sequence of each feature in the lubrication state feature set by adopting a sliding window with a preset length.
5. The multi-feature trend fused motor lubrication distinguishing and intelligent oil filling method according to claim 1, wherein the preset updating strategy comprises a forgetting factor updating strategy, and for each feature in the lubrication state feature set, when the deviation degree of the feature trend quantity of the feature relative to the corresponding dynamic baseline is smaller than a preset health threshold, the dynamic baseline of the feature is adaptively updated according to the following formula: Wherein, the method comprises the steps of, Is forgetting factor and meets , As a dynamic baseline for the last moment in time, And when the deviation degree is greater than or equal to the preset health threshold value, keeping the corresponding dynamic baseline not updated.
6. The multi-feature trend fused motor lubrication judging and intelligent oiling method according to claim 1, wherein the multi-feature trend fused judging rule comprises respectively calculating the ratio of the characteristic trend quantity corresponding to the acceleration effective value, the speed effective value, the kurtosis and the high-frequency energy ratio to the dynamic base line, and at least meeting the following conditions when judging the poor lubrication: the ratio of the characteristic trend quantity corresponding to the acceleration effective value to the dynamic base line is larger than a first threshold coefficient; the ratio of the characteristic trend quantity corresponding to the effective speed value to the dynamic base line is positioned in a preset interval; The ratio of the characteristic trend quantity corresponding to the kurtosis to the dynamic base line is larger than a second threshold coefficient; the ratio of the characteristic trend quantity corresponding to the high-frequency energy duty ratio to the dynamic base line is larger than a third threshold coefficient.
7. The multi-feature trend fused motor lubrication judging and intelligent oiling method according to claim 1 is characterized in that the fusion calculation of the oiling amount coefficient comprises the steps of respectively calculating relative deviation amounts of at least two features in the lubrication state feature set, wherein the relative deviation amounts are ratios of differences of corresponding feature trend amounts and corresponding dynamic base lines to the dynamic base lines, weighting and fusion are conducted on the basis of the relative deviation amounts to obtain the oiling amount coefficient, the oiling amount is set to be the product of the oiling amount coefficient and a single oiling reference amount of an oiling actuator, or oiling duration is set to be the product of the oiling amount coefficient and the reference oiling duration, and accordingly self-adaptive adjustment of the oiling amount along with poor lubrication degree is achieved.
8. The multi-feature trend fused motor lubrication discrimination and intelligent oiling method according to claim 1, further comprising step S6: Continuously acquiring vibration acceleration time domain signals in a preset monitoring window period, acquiring a lubrication state feature set, calculating feature trend amounts of all features, and comparing the feature trend amounts with corresponding dynamic baselines; and determining that oiling is effective when at least a preset number of features in the lubrication state feature set meet a regression criterion.
9. The multi-feature trend fused motor lubrication judging and intelligent oil filling method according to claim 8, wherein the regression criterion comprises that the ratio of the feature trend quantity corresponding to the acceleration effective value, the kurtosis and the high-frequency energy duty ratio to the corresponding dynamic base line is reduced to be within respective preset regression thresholds, the ratio of the feature trend quantity corresponding to the speed effective value to the corresponding dynamic base line is kept within a preset interval, and secondary oil filling processing or alarm processing is triggered when the regression criterion is not met at the end of the preset monitoring window period.
10. Motor lubrication discrimination and intelligent oiling system that many characteristic trend fused, its characterized in that, the system includes: The acquisition and speed acquisition module is used for acquiring vibration acceleration time domain signals in the running process of the motor and acquiring vibration speed time domain signals based on the vibration acceleration time domain signals; The device comprises a vibration acceleration time domain signal, a characteristic acquisition module, a lubrication state characteristic set and a control module, wherein the vibration acceleration time domain signal and the vibration speed time domain signal are used for acquiring the lubrication state characteristic set; The trend and baseline construction module is used for establishing a historical sequence for each feature in the lubrication state feature set, calculating the corresponding feature trend quantity based on the historical sequence, establishing a dynamic baseline for each feature, and adaptively updating the dynamic baseline based on a preset updating strategy; The lubrication judgment module outputs a lubrication judgment result according to a preset multi-feature trend fusion judgment rule according to the deviation relation of the feature trend quantity of at least part of the features in the lubrication state feature set relative to the corresponding dynamic base line; And when the lubrication judging result indicates poor lubrication, the oiling control module performs fusion calculation according to the deviation degree of at least part of the characteristics in the lubrication state characteristic set to obtain an oiling rate coefficient, and outputs an oiling control instruction to an oiling executing mechanism so that the oiling executing mechanism carries out oiling on the motor according to the oiling rate determined by the oiling rate coefficient.

Description

Multi-feature trend fused motor lubrication distinguishing and intelligent oiling method and system Technical Field The invention relates to the technical field of motor state monitoring and intelligent maintenance, in particular to a motor lubrication distinguishing and intelligent oiling method and system with multi-feature trend fusion. Background The development of equipment lubrication and fault diagnostics has generally undergone an evolving path from manual empirical maintenance to automated lubrication to predictive maintenance driven by condition monitoring. The early industrial site mainly relies on manual periodical greasing or oiling, and as the requirements of large-scale and continuous production of equipment are improved, a centralized lubrication and automatic lubrication system gradually appears, and a plurality of lubrication points are connected into a system through a metering device, a pipeline and a distribution element, so that manual operation is reduced, the consistency of grease supply is improved, and a mature system is formed in the engineering field by related product forms. Meanwhile, the concept of "maintenance according to state" is gradually popularized, state parameters such as vibration are widely used for equipment health assessment, and a standardized state monitoring and diagnosis framework is formed, for example, ISO 17359 provides a general guideline for monitoring and diagnosis by using parameters such as vibration, temperature, tribology and the like. Entering the industrial internet stage, IIoT and edge calculation push state monitoring to develop towards the on-line, real-time and closed-loop decision making directions, and the data is processed on the site side and drives maintenance actions to become one of important forms of intelligent factories. In the prior art, a timing lubrication strategy is still commonly adopted for lubrication and maintenance of a motor and rotating equipment, excessive lubrication and insufficient lubrication are easy to occur, and grease waste, rising maintenance cost and lubrication failure risks are caused. In order to improve the diagnostic capability, some more advanced schemes rely on performing spectrum analysis, such as FFT, on the vibration signal to extract the fault characteristic frequency, but such methods are complex in calculation, have high computational requirements on the processing unit, and often need to continuously upload high-sampling-rate vibration data to the cloud for processing in engineering. Under the links of data uploading, cloud processing and instruction issuing, a loop is longer, control delay is caused, real-time control requirements are difficult to meet, meanwhile, high-sampling-rate data is continuously uploaded, bandwidth is occupied, communication and platform cost is increased, and independent work is difficult under a scene that network conditions are limited or unstable. Therefore, a motor lubrication distinguishing and intelligent oiling method and system with multiple feature trend fusion are needed to solve the problems. Disclosure of Invention (One) solving the technical problems Aiming at the defects of the prior art, the invention provides a multi-feature trend fused motor lubrication distinguishing and intelligent oiling method and system, which solve the problems. (II) technical scheme In order to achieve the purpose, the invention provides the following technical scheme that the motor lubrication distinguishing and intelligent oiling method with multi-feature trend fusion comprises the following steps: Step S1, collecting vibration acceleration time domain signals in the running process of a motor, and obtaining vibration speed time domain signals based on the vibration acceleration time domain signals; S2, acquiring a lubrication state feature set based on the vibration acceleration time domain signal and the vibration speed time domain signal, wherein the lubrication state feature set at least comprises an acceleration effective value, a speed effective value, kurtosis and a high-frequency energy duty ratio, and the high-frequency energy duty ratio represents the duty ratio of the energy of the vibration acceleration time domain signal in a preset high-frequency band relative to the energy of a full frequency band; step S3, establishing a historical sequence for each feature in the lubrication state feature set, calculating to obtain a corresponding feature trend quantity based on the historical sequence, simultaneously establishing a dynamic baseline for each feature, and adaptively updating the dynamic baseline based on a preset updating strategy; S4, outputting a lubrication judgment result according to a preset multi-feature trend fusion judgment rule according to a deviation relation of feature trend quantity of at least part of features in the lubrication state feature set relative to a corresponding dynamic base line; and S5, when the lubrication judging result indicates poo