CN117494485B - Method, apparatus and storage medium for predicting failure probability of rotating apparatus - Google Patents

Method, apparatus and storage medium for predicting failure probability of rotating apparatus Download PDF

Info

Publication number
CN117494485B
CN117494485B CN202410003478.3A CN202410003478A CN117494485B CN 117494485 B CN117494485 B CN 117494485B CN 202410003478 A CN202410003478 A CN 202410003478A CN 117494485 B CN117494485 B CN 117494485B
Authority
CN
China
Prior art keywords
data
frequency
time domain
fault
domain data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410003478.3A
Other languages
Chinese (zh)
Other versions
CN117494485A (en
Inventor
窦浩
戴林杉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
KSB Shanghai Pump Co Ltd
Original Assignee
KSB Shanghai Pump Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by KSB Shanghai Pump Co Ltd filed Critical KSB Shanghai Pump Co Ltd
Priority to CN202410003478.3A priority Critical patent/CN117494485B/en
Publication of CN117494485A publication Critical patent/CN117494485A/en
Application granted granted Critical
Publication of CN117494485B publication Critical patent/CN117494485B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD

Abstract

Embodiments of the present invention relate to a method, apparatus, and storage medium for predicting failure probability of a rotating apparatus. Acquiring working condition data about rotating equipment, equipment parameter information and acceleration detection data about each detection position of the rotating equipment; processing the acquired acceleration detection data to acquire time domain data, frequency domain data and axis track data about the rotating equipment; calculating morphological similarity according to the time domain data, the frequency domain data and the axis track data based on the trained characteristic calculation model so as to determine the time domain morphological feature, the frequency spectrum morphological feature and the axis track morphological feature corresponding to the current acceleration detection data; and calculating a probability of failure with respect to the rotating device based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature, and axis trace morphology feature. Thus, the present invention can realize prediction of the failure probability for the rotating equipment.

Description

Method, apparatus and storage medium for predicting failure probability of rotating apparatus
Technical Field
Embodiments of the present invention relate generally to the field of industrial control, and more particularly, to a method, apparatus, and storage medium for predicting failure probability of rotating equipment.
Background
Conventional methods for detecting a failure of a rotating device (e.g., without limitation, a rotating type mechanical device such as a pump, a motor, etc.) are, for example: the failure judgment is performed by detecting whether the spectrum has a peak value at a specific frequency based on a mechanism model. The mechanism model mainly judges whether faults occur according to mechanism rules.
The conventional method for detecting the faults of the rotating equipment does not consider how to perform fault detection under the condition that the specific frequency domain peak value cannot be identified due to factors such as resolution, energy leakage and the like, and cannot give out the probability of fault occurrence, and the manual maintenance cannot be performed in a targeted manner due to the diversity of fault types and the unquantifiable characteristic of the probability of fault occurrence, so that a great deal of manpower resources and other costs are wasted.
In summary, the conventional manner for predicting the failure probability of the rotating apparatus has the following disadvantages: the probability of occurrence of various faults of the rotating equipment cannot be calculated.
Disclosure of Invention
In view of the above problems, the present invention provides a method, apparatus, and storage medium for predicting failure probability of a rotating apparatus, capable of calculating probability of occurrence of various failures of the rotating apparatus.
According to a first aspect of the present invention, there is provided a method for predicting a failure probability of a rotating apparatus, comprising: acquiring working condition data, equipment parameter information and acceleration detection data of each detection position of the rotating equipment; processing the acquired acceleration detection data to acquire time domain data, frequency domain data and axis track data about the rotating equipment; calculating morphological similarity according to the time domain data, the frequency domain data and the axis track data based on the trained characteristic calculation model so as to determine the time domain morphological feature, the frequency spectrum morphological feature and the axis track morphological feature corresponding to the current acceleration detection data; and calculating a probability of failure with respect to the rotating device based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature, and axis trace morphology feature.
In some embodiments, the method for predicting failure probability of a rotating device further comprises: determining a predetermined failure frequency associated with a failure of the rotating equipment based on the acquired operating condition data, equipment parameter information, frequency domain data, and predetermined mechanism model data; and determining a fault frequency energy characteristic for the predetermined fault frequency within the predetermined range based on the predetermined range; and calculating a probability of failure with respect to the rotating equipment based on the determined failure frequency energy feature, time domain morphology feature, frequency spectrum morphology feature, and axis locus morphology feature.
In some embodiments, calculating the probability of failure with respect to the rotating equipment includes: based on at least one of the determined fault frequency energy characteristic, time domain morphology characteristic, frequency spectrum morphology characteristic, and axis trace morphology characteristic, a probability of occurrence of at least one fault type with respect to the rotating equipment is determined via a predetermined fuzzy control system calculation.
In some embodiments, the predetermined fuzzy control system is generated via the steps of: determining a fuzzy control variable and a variable range corresponding to any one of a time domain morphological feature, a frequency spectrum morphological feature, an axis track morphological feature and/or a fault frequency energy feature related to the fault; determining a membership function for the fuzzy control variable based on the determined fuzzy control variable and variable range for the fault; and determining a fuzzy rule based at least on the fuzzy control variable and the membership function for the fuzzy control variable.
In some embodiments, the time domain morphology includes at least shock waves, beat waves, and non-periodic waveforms; the spectral morphology includes at least harmonics, sidebands and bottom ridges; the axle center track form at least comprises banana shape, ellipse shape, inner splayed shape, multiple splayed shape and outer splayed shape.
In some embodiments, acquiring time domain data, frequency domain data, and axis trajectory data about the rotating device includes: performing Fourier transform on the obtained acceleration detection data, performing time domain integration based on the acceleration detection data after Fourier transform, and performing inverse Fourier transform to obtain speed time domain data about the rotating equipment; performing Fourier transform on the obtained speed time domain data, performing time domain integration based on the speed time domain data after Fourier transform, and performing inverse Fourier transform to obtain displacement time domain data about the rotating equipment, and performing frequency spectrum transform based on the obtained acceleration time domain data, speed time domain data and displacement time domain data respectively to obtain acceleration frequency domain data, speed frequency domain data and displacement frequency domain data about the rotating equipment.
In some embodiments, calculating morphological similarity for the acquired time domain data, frequency domain data, and axis trajectory data based on the trained feature calculation model further comprises: normalizing the acquired acceleration time domain data, velocity time domain data and displacement time domain data based on a preset sampling period to eliminate dimension gaps of the time domain data about the rotating equipment acquired under different working conditions; and normalizing the acquired acceleration frequency domain data, velocity frequency domain data and displacement frequency domain data based on a preset fundamental frequency to eliminate dimension gaps of the acquired frequency domain data about the rotating equipment under different working conditions.
In some embodiments, the trained feature computation model includes a fault morphology similarity computation model and a fault frequency energy computation model, the fault morphology similarity computation model obtained via: acquiring multiple sets of training time domain data and corresponding time domain forms, multiple sets of training frequency domain data and corresponding frequency domain forms, multiple sets of training axle center track data and corresponding axle center track forms of rotating equipment; and determining the time domain form corresponding to the training time domain data, the frequency domain form corresponding to the training frequency domain data and the axle center track form corresponding to the training axle center track data as training data, and training at least one preset neural network model to obtain a trained fault form similarity calculation model.
In some embodiments, the rotational device is a pump, and the acceleration detection data includes: three-way acceleration detection data of at least one end of the pump driving end bearing, the pump non-driving end side, the motor driving end and the motor non-driving end; the three-way acceleration detection data includes acceleration detection data in the vertical direction, acceleration detection data in the horizontal direction, and acceleration detection data in the axial direction. In some embodiments, the operating condition data is at least one of rotational speed, temperature, and flow rate; the equipment parameter information at least comprises the number of blades, resonance frequency and bearing fault coefficients; and the predetermined fault frequency at least comprises a predetermined frequency multiplication, an outer ring fault frequency and a harmonic frequency thereof, an inner ring fault frequency and a harmonic frequency thereof, and a fundamental frequency sideband frequency.
According to a second aspect of the present invention there is provided a computing device comprising: at least one processing unit; at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions when executed by the at least one processing unit cause the apparatus to perform the steps of the method according to the first aspect.
According to a third aspect of the present invention, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a machine, implements a method according to the first aspect.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
The above and other features, advantages and aspects of embodiments of the present invention will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements.
Fig. 1 shows a schematic diagram of a system for implementing a method for predicting failure probability of a rotating device according to an embodiment of the present invention.
Fig. 2 shows a flow chart of a second method for predicting the failure probability of a rotating device according to an embodiment of the invention.
Fig. 3 shows a flow chart of a method three for acquiring time domain data, frequency domain data, about a rotating device, according to an embodiment of the invention.
FIG. 4 illustrates a flow chart of a method four for fault morphology similarity calculation model training in accordance with an embodiment of the present invention.
Fig. 5 shows a flow chart of a method five for fuzzy control system construction according to an embodiment of the present invention.
Fig. 6 shows a schematic representation of a time domain morphology according to an embodiment of the present invention.
Fig. 7 shows a schematic diagram of a frequency domain morphology according to an embodiment of the invention.
FIG. 8 illustrates a schematic diagram of an axial trace configuration according to an embodiment of the invention.
Fig. 9 shows a schematic diagram of a fault frequency domain morphology according to an embodiment of the present invention.
Fig. 10 schematically illustrates a block diagram of a second system implementing a method for predicting failure probability of a rotating equipment according to further embodiments of the invention.
Fig. 11 shows a block diagram of an electronic device according to an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, in which various details of the embodiments of the present invention are included to facilitate understanding, and are to be considered merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
The term "comprising" and variations thereof as used herein means open ended, i.e., "including but not limited to. The term "or" means "and/or" unless specifically stated otherwise. The term "based on" means "based at least in part on". The terms "one example embodiment" and "one embodiment" mean "at least one example embodiment. The term "another embodiment" means "at least one additional embodiment". The terms "first," "second," and the like, may refer to different or the same object. Other explicit and implicit definitions are also possible below.
As described above, the conventional means for predicting the probability of failure of the rotating equipment cannot calculate the probability of occurrence of various failures of the rotating equipment.
To at least partially solve one or more of the above-mentioned problems and other potential problems, an exemplary embodiment of the present invention proposes a scheme for predicting a failure probability of a rotating apparatus, in which operating condition data about the rotating apparatus, apparatus parameter information, and acceleration detection data about respective detection positions of the rotating apparatus are acquired; processing the acquired acceleration detection data to acquire time domain data, frequency domain data and axis track data about the rotating equipment; calculating morphological similarity according to the time domain data, the frequency domain data and the axis track data based on the trained characteristic calculation model so as to determine the time domain morphological feature, the frequency spectrum morphological feature and the axis track morphological feature corresponding to the current acceleration detection data; and calculating a probability of failure with respect to the rotating apparatus based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature, and axis trace morphology feature; therefore, the present invention can acquire time-frequency domain data and axis locus data about the rotating apparatus through the acceleration detection data of the rotating apparatus. In addition, the form similarity is calculated through the time-frequency domain data and the axis track data to determine the frequency domain form data, the time domain form data and the axis track form data of the rotating equipment, so that the probability of faults of the rotating equipment is calculated based on the characteristics of various form data. Thus, the present invention can realize prediction of the failure probability for the rotating equipment.
The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, but not all embodiments, and they should not be construed as limiting the protection scope of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 shows a schematic diagram of a system 100 for implementing a method for predicting failure probability of a rotating device according to an embodiment of the invention. As shown in fig. 1, the system 100 includes a computing device 110 and a rotating device 130 and an acceleration detection apparatus 140. In some embodiments, the computing device 110, the rotation device 130, and the acceleration detection apparatus 140 are communicatively connected for data interaction directly or via a network.
As for the rotation device 130, it is, for example and not limited to, a rotation device driven by variable frequency or fixed frequency, such as a pump, a water pump, or the like.
As for the computing device 110, it is used to calculate the probability of occurrence of a failure of the rotating device 130. Specifically, the computing device 110 is configured to acquire operating condition data about the rotating device, device parameter information, and acceleration detection data about each detection position of the rotating device; the acquired acceleration detection data is processed to acquire time domain data, frequency domain data, and axis locus data about the rotating device. The computing device 110 is further configured to calculate morphology similarity for the time domain data, the frequency domain data, and the axis trajectory data based on the trained feature calculation model, to determine a time domain morphology feature, a frequency spectrum morphology feature, and an axis trajectory morphology feature corresponding to the current acceleration detection data; and calculating a probability of failure with respect to the rotating device based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature, and axis trace morphology feature.
In some embodiments, computing device 110 may have one or more processing units, including special purpose processing units such as GPUs (Graphics Processing Unit, graphics processors), FPGAs (Field Programmable Gate Array, field programmable gate arrays), and ASICs (Application Specific Integrated Circuit, application specific integrated circuits), as well as general purpose processing units such as CPUs. In addition, one or more virtual machines may also be running on each computing device 110. In some embodiments, computing device 110 includes, for example: acceleration detection data and working condition data acquisition unit 112, time-frequency domain and axis trajectory data acquisition unit 114, morphological feature calculation unit 116, and failure probability calculation unit 118.
An acceleration detection data and condition data acquisition unit 112 for acquiring condition data about the rotating equipment, equipment parameter information, and acceleration detection data about the respective detection positions of the rotating equipment.
A time-frequency domain and axis trace data acquisition unit 114 for processing the acquired acceleration detection data to acquire time-domain data, frequency-domain data, and axis trace data about the rotating device.
Regarding the morphology feature calculation unit 116, it is used for calculating morphology similarity for time domain data, frequency domain data and axis trace data based on the trained feature calculation model to determine the time domain morphology feature, the frequency spectrum morphology feature and the axis trace morphology feature corresponding to the current acceleration detection data.
Regarding the failure probability calculation unit 118, it is used for calculating the probability of failure occurring regarding the rotating equipment based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature and axis locus morphology feature.
Fig. 2 shows a flow chart of a second method for predicting the failure probability of a rotating device according to an embodiment of the invention. Method two may be performed by computing device 110 as shown in fig. 1 or at electronic device 1100 as shown in fig. 11. It should be understood that method two may also include additional steps not shown and/or may omit steps shown, the scope of the present invention being not limited in this respect.
In step 202, the computing device 110 obtains operating condition data about the rotating device, device parameter information, and acceleration detection data about various detected positions of the rotating device.
As for the rotating device, it is for example, but not limited to, a pump, a motor, etc. driven at a fixed frequency or at a variable frequency.
The operating condition data of the rotating device is, for example, at least one of the rotational speed, the temperature, and the flow rate.
Device parameter information about the rotating device including, for example, basic parameter information of the rotating device such as the number of blades, resonance frequency, and bearing failure coefficient; and further comprising the predetermined fault frequency including at least a predetermined doubling of frequency, an outer ring fault frequency and its harmonic frequency, an inner ring fault frequency and its harmonic frequency, and a fundamental sideband frequency.
Acceleration detection data concerning the respective detection positions of the rotating apparatus, which are, for example, acceleration waveform data acquired via an acceleration detection device (such as an acceleration sensor) for the respective detection positions of the rotating apparatus; with respect to an acceleration sensor, such as a three-axis acceleration sensor, a three-way acceleration signal of one detection position can be acquired at the same time.
For example, when the rotating device is a pump, the acceleration detection data includes: three-way acceleration detection data of at least one end of the pump driving end bearing, the pump non-driving end side, the motor driving end and the motor non-driving end; the three-way acceleration detection data includes acceleration detection data in a vertical direction, acceleration detection data in a horizontal direction, and acceleration detection data in an axial direction.
At step 204, the computing device 110 processes the acquired acceleration detection data to acquire time domain data, frequency domain data, and axis trace data about the rotating device.
Acquiring time domain data, frequency domain data and axis locus, for example, when the acceleration sensor acquires a three-way acceleration detection signal about any detection position of the rotating device, acquiring time domain and frequency domain data, speed time domain and frequency domain data and displacement time domain and frequency domain data about acceleration of the rotating device in each direction of the detection position according to the acquired three-way (such as horizontal, vertical and axial) acceleration signals; and acquiring axis locus data at the detection position. The time domain data, the frequency domain data and the axis trace are obtained in detail with reference to fig. 3, and will not be described here again.
Regarding a method of generating a time domain feature, a frequency domain feature, it includes, for example: the computing device 110 determines whether the acquired acceleration detection data is abnormal, the abnormality including one or more of the following: the failure rate is higher than a predetermined failure rate threshold, the abnormality rate is higher than a predetermined abnormality rate threshold, etc.; if the obtained acceleration data is determined to be abnormal, eliminating the abnormal acceleration data. Regarding the abnormality rate, for example, the difference value of the acceleration detection data is abnormal, the variance of the acceleration detection data is abnormal, and the like.
In some embodiments, the computing device 110 determines a predetermined failure frequency associated with a failure of the rotating equipment based on the obtained operating condition data, the equipment parameter information, and the frequency domain data, and the predetermined mechanism model data; determining a fault frequency energy characteristic of the predetermined fault frequency within a predetermined range based on the predetermined range; and calculating a probability of failure with respect to the rotating equipment based on the determined failure frequency energy feature, time domain morphology feature, frequency spectrum morphology feature, and axis locus morphology feature. With respect to the predetermined range, for example, a failure frequency + -10 Hz range.
Therefore, quality monitoring of acceleration detection data can be achieved, abnormal data are removed, time domain data, frequency domain data, axis track data and fault frequency energy characteristics of the rotating equipment are obtained more accurately, and therefore more accurate probability of faults of the rotating equipment is obtained conveniently.
In step 206, the computing device 110 calculates morphology similarity for the time domain data, the frequency domain data, and the axis trace data based on the trained feature calculation model to determine time domain morphology features, frequency spectrum morphology features, and axis trace morphology features corresponding to the current acceleration detection data.
With respect to the trained feature calculation model, a fault morphology similarity calculation model and a fault frequency energy calculation model are included. The fault form similarity calculation model is used for calculating the obtained form similarity of the time domain form, the frequency domain form, the axis locus and the fault form of the rotating equipment aiming at the time domain data, the frequency domain data and the axis locus data. With respect to the failure frequency energy calculation model, frequency energy of a predetermined failure frequency within a predetermined range is calculated based on the frequency domain data to acquire energy characteristics with respect to each failure frequency of the rotating apparatus to calculate the similarity of the frequency energy characteristics. The training method of the fault morphology similarity calculation model will be described in detail with reference to fig. 4, and will not be described herein.
Regarding the calculation of the form similarity, feature similarity calculation is performed based on the fault form, for example, the similarity of various time domain forms of acceleration, speed and displacement and the fault form is calculated respectively aiming at the calculation of the time domain form similarity; for example, calculating the similarity of frequency domain forms, and respectively calculating the similarity of various frequency domain forms of acceleration, speed and displacement and fault forms; for example, the similarity of the axle center track forms is calculated, and the similarity of various axle center track forms is calculated respectively. Therefore, through the form similarity calculation, the similarity between the current time domain form, the current frequency domain form and the current axis track of the rotating equipment and the preset fault form can be determined based on the acceleration detection data and the working condition, so that the fault type possibly occurring in the rotating equipment can be determined.
Referring to fig. 6, fig. 6 shows a schematic diagram of a time domain morphology according to an embodiment of the present invention. As shown in fig. 6, the waveform representation of the time domain waveform includes at least a shock wave (6 a), a beat wave (6 b), and an aperiodic waveform (6 c), for example. Regarding shock waves, problems such as gear meshing, loosening, or bumping can result in shock signals that suddenly increase and decrease in the waveform; regarding the beat wave, for example, the occurrence of various faults of the rolling bearing may cause superposition of two or more fault frequencies, and the beat wave may occur in the time domain; with respect to non-periodic waveforms, for example, a turbomachine failure may cause the waveform to become irregular, non-periodic, or random.
Referring to fig. 7, fig. 7 shows a schematic diagram of a frequency domain morphology according to an embodiment of the present invention. As shown in fig. 7, the waveform representation concerning the frequency domain morphology includes, for example, at least harmonics (7 a), sidebands (7 b), and bottom ridges (7 c). Fig. 9 shows a schematic diagram of a fault frequency domain morphology according to an embodiment of the present invention. For example, please refer to fig. 9 for two fault frequency domain morphologies of the embodiment of the present invention, 9a in fig. 9 illustrates an outer ring fault (rule: outer ring fault frequency and its harmonics), and 9b in fig. 9 illustrates an inner ring fault (inner ring fault frequency and axial frequency sidebands). In the morphological illustrations of FIGS. 7 and 9, the abscissa is in Hertz (Hz) and the ordinate is in various choices, such as m/s 2 (m/s square), m 2 S (meter square/second), um (microns), etc., as the present solution fully takes into account the various frequency domain morphologies of acceleration, velocity, and displacement.
Referring to fig. 8, fig. 8 is a schematic diagram illustrating an axial trace configuration according to an embodiment of the invention. As shown in FIG. 8, the axial trace comprises at least banana shape, oval shape, inside splayed shape, multi-splayed shape and outside splayed shape.
At step 208, the computing device 110 calculates a probability of failure with respect to the rotating device based on the determined at least one of temporal morphology feature, spectral morphology feature, and axial trace morphology feature.
With respect to calculating the probability of occurrence of a fault with respect to the rotating equipment, for example, via a predetermined fuzzy control system calculation, to determine the probability of occurrence of at least one fault type with respect to the rotating equipment. For example, the determined time domain morphology feature, frequency spectrum morphology feature and axis locus morphology feature are used as input data of the fuzzy control system to calculate the probability of failure of the rotating equipment.
In some embodiments, the computing device 110 also calculates a failure frequency energy signature for the predetermined failure frequency within a predetermined range, calculates a probability of failure with respect to the rotating device, such as based on at least one of the determined failure frequency energy signature, time domain morphology signature, frequency spectrum morphology signature, and axis trace morphology signature, via a predetermined fuzzy control system, to determine a probability of occurrence of at least one failure type with respect to the rotating device.
By adopting the means, the invention can determine the probability of various fault types of the rotating equipment aiming at different rotating equipment and different working conditions thereof, and realize the accurate prediction of the fault types and the fault probability of the rotating equipment.
The construction of the fuzzy control system comprises the steps of defining fuzzy control variables, defining fuzzy sets and membership functions, defining fuzzy rules, selecting a fuzzy solution mode and calculating probability based on fault forms and fault rules. The construction of the fuzzy control system will be described in detail with reference to fig. 5, and will not be described here.
In the above scheme, for the rotating equipment, the acceleration raw data detected and output by the acceleration sensor can be converted into time domain data, frequency domain data, axis track and other data, the time domain form, the frequency domain form, the axis track form and other features are constructed through time domain fault form, frequency domain form and axis track form and other features, the time domain form feature, the frequency domain form and the axis track form are combined with the energy feature of the fault frequency, and the time domain form feature, the axis track form feature and the fault frequency energy feature related to the rotating equipment are converted into the fault probability of the rotating equipment through the fuzzy control system.
Fig. 3 shows a flow chart of a method three for acquiring time domain data, frequency domain data, about a rotating device, according to an embodiment of the invention. It should be appreciated that method three may be performed, for example, at the electronic device 1100 depicted in fig. 11. May also be performed at the computing device 110 depicted in fig. 1. It should be understood that method three may also include additional actions not shown and/or may omit actions shown, the scope of the invention being not limited in this respect.
In step 302, the computing device 110 performs a fourier transform on the acquired acceleration detection data, performs a time domain integration based on the fourier transformed acceleration detection data, and performs an inverse fourier transform to acquire speed time domain data about the rotating device.
In step 304, the computing device 110 performs a fourier transform on the acquired velocity time domain data, performs a time domain integration based on the fourier transformed velocity time domain data, and performs an inverse fourier transform to acquire displacement time domain data about the rotating device.
Regarding the acquisition of acceleration time domain data, for example, a calculus technique is employed to convert the acquired acceleration signal (acceleration detection data) into a velocity signal (velocity time domain data), a displacement signal (displacement time domain data). Firstly, carrying out Fourier transform on an acceleration signal to be processed, processing a transform result, and then carrying out inverse Fourier transform to obtain an acceleration time domain signal after integration or differentiation, wherein the transform principle is as follows:
the acceleration signal at any frequency is set as:
（1）
in the above formula (1), a (t) is a fourier component of the acceleration signal at the frequency ω, t represents a time variable of the acceleration signal, a is a coefficient corresponding to a (t), j is an imaginary number, and e is a natural constant.
When the initial velocity component is 0, the velocity signal component can be obtained by performing time domain integration on the acceleration signal component, namely:
（2）
in the above formula (2), v (t) is the fourier component of the velocity signal at the frequency ω; v is a coefficient corresponding to V (t), a (τ) is the time domain integration of a (t), τ represents the time variable in the integration process.
When the initial velocity and the initial displacement component are both 0, the displacement signal component can be obtained by performing time domain integration on the acceleration signal component twice, namely:
（3）
in the above formula (3), s (t) is the fourier component of the displacement signal at the frequency ω; s is a coefficient corresponding to S (t), d τ Minute units, d, being time intervals of integration λ Is an integral variable parameter.
In step 306, the computing device 110 performs a spectral transformation based on the acquired acceleration time domain data, velocity time domain data, and displacement time domain data, respectively, to acquire acceleration frequency domain data, velocity frequency domain data, and displacement frequency domain data for the rotating device.
With respect to spectral transformations, such as filtering, fast fourier transform techniques to calculate an acceleration spectrum, a velocity spectrum, a displacement spectrum of the signal data, it should be understood that the acceleration spectrum is acceleration frequency domain data, the velocity spectrum is velocity frequency domain data, and the displacement spectrum is displacement frequency domain data; it should be understood that when the obtained frequency domain data is presented in the form of a coordinate image, the horizontal axis of the obtained acceleration spectrum, velocity spectrum and displacement spectrum is frequency (in Hz) and the vertical axis is amplitude (in dB).
Regarding filtering, for example, an IIR digital filter is employed, the expression is as follows:
（4）
in the above-mentioned formula (4),、/>respectively as inputAnd an output time domain signal sequence, n being an nth point in the time sequence; m is zero point number of the transfer function of the filter system, N is pole point number of the transfer function of the filter system, k1 represents any natural number from 1 to M, and k2 represents any natural number from 1 to N; c k1 、b k2 Are filter coefficients.
Thus, by the above-described processing, the obtained acceleration detection signal can be converted into acceleration time domain data, velocity time domain data, and displacement time domain data, and further acceleration frequency domain data, velocity time domain data, and displacement time domain data can be obtained by spectral transformation. It should be appreciated that at each detection position, the acceleration detection signal for each direction may obtain its acceleration time and frequency domain data, velocity time and frequency domain data, and displacement time and frequency domain data.
In some embodiments, calculating morphological similarity for the acquired time domain data, frequency domain data, and axis trajectory data based on the trained feature calculation model further comprises: normalizing the collected time domain morphology data and frequency domain morphology data to eliminate dimension gaps of the data under different working conditions, and converting the normalized data into a black-and-white image to generate the time domain morphology and the frequency domain morphology.
For example, for the acquired acceleration time domain data, velocity time domain data, and displacement time domain data, normalization is performed based on a predetermined sampling period to eliminate dimensional gaps of the acquired time domain data about the rotating device under different conditions.
For example, normalization is performed on the basis of a predetermined fundamental frequency for the acquired acceleration frequency domain data, velocity frequency domain data, and displacement frequency domain data to eliminate dimensional differences of the acquired frequency domain data about the rotating device under different conditions. For example, the fundamental frequency of the obtained frequency domain data one is 10Hz, the fundamental frequency of the obtained frequency domain data two is 60Hz, and even if the two have the same morphological characteristics, the difference of the waveform patterns displayed by the frequency domain morphology is large due to the dimension difference caused by the fundamental frequency, so that the waveform patterns are difficult to be identified as the same morphological characteristics.
Therefore, the invention can remove the dimension difference by frequency normalization, such as stretching or shrinking the abscissa (frequency Hz) of the frequency domain data with different fundamental frequencies, so that the obtained multi-position frequency domain data of a plurality of rotating devices can be subjected to similarity calculation after the dimension difference is removed, more accurate similarity calculation results can be obtained, and the training of the fault form similarity model can be more efficient.
Regarding normalization, for example, not limited to the manner in which Z-score normalization is employed, the manner in which normalization is exemplified below in conjunction with equation (5):
（5）
in the above formula (5), z represents normalized data. q represents the weighted features before normalization. μ represents the mean of the weighted features before normalization. σ represents the standard deviation of the weighted features prior to normalization.
Therefore, by adopting the means, the method can normalize the time domain data of the acceleration, the speed and the displacement obtained by processing the acceleration detection data acquired under different rotating equipment and different working conditions based on the sampling period, normalize the frequency domain data of the acceleration, the speed and the displacement based on the preset fundamental frequency, so that the acquired time domain form and the frequency domain form can eliminate dimension gaps, the fault form can be identified efficiently and accurately, the form features of various time domain data (the acceleration, the speed and the displacement), the form features of various frequency domain data (the acceleration, the speed and the displacement), the form features of the axle center track and the energy features of the fault frequency are fully considered, and various form features are synthesized, so that a more accurate fault probability prediction result is achieved.
FIG. 4 illustrates a flow chart of a method four for fault morphology similarity calculation model training in accordance with an embodiment of the present invention. It should be appreciated that method four may be performed, for example, at the electronic device 1100 depicted in fig. 11. May also be performed at the computing device 110 depicted in fig. 1. It should be appreciated that method four may also include additional actions not shown and/or may omit actions shown, the scope of the invention being not limited in this respect.
In step 402, the computing device 110 obtains sets of training time domain data and corresponding time domain morphology, sets of training frequency domain data and corresponding frequency domain morphology, sets of training axis trajectory data and corresponding axis trajectory morphology with respect to the rotating device.
For the collection of training data, for example, common time domain forms and frequency spectrum forms are collected, and various forms are marked. Regarding the time domain morphology, for example, a data morphology in which one sampling period is collected is taken as one training sample, and regarding the frequency domain morphology, for example, data of ±5 Hz is collected as one sample in combination with a specific frequency.
In step 404, the time domain form corresponding to the training time domain data, the frequency domain form corresponding to the training frequency domain data, and the axis trace form corresponding to the training axis trace data are determined as training data, and training is performed on at least one predetermined neural network model to obtain a trained fault form similarity calculation model.
With respect to the fault morphology similarity calculation model, it includes at least one predetermined neural network model, such as a first neural network model for calculating time domain morphology features, frequency domain morphology features, and a second neural network model for calculating the axis trajectory. For example, the time domain form corresponding to the training time domain data and the frequency domain form corresponding to the training frequency domain data are determined as verification data of the first neural network, the training time domain data and the training frequency domain data are determined as training data of the first neural network, and the first neural network model is trained. For example, the axis locus shape corresponding to the training axis locus data is determined as verification data, the training axis locus data is determined as training data, and the second neural network model is trained.
With respect to predetermined neural Network models, they are, for example, the oxford university visual geometry group (Visual Geometry Group, VGG) model, dense connectivity convolutional Network (Densely Connected Convolutional Networks, densely), residual Network (ResNet), and the like.
Regarding the calculation of the fault frequency energy characteristics, the fault characteristic frequency can be determined according to the acquired acceleration frequency spectrum (frequency domain data), velocity frequency spectrum (frequency domain data) and displacement frequency spectrum (frequency domain data), for example, based on a mechanism model, and common fault characteristics can be collected by combining expert experience or fault rules. In consideration of the fact that the amplitude of a specific frequency cannot be found accurately due to different resolutions of the frequency spectrum, and in addition, the energy leakage factor needs to be considered, the scheme provided by the embodiment of the invention is different from the traditional method for calculating the amplitude, and the energy of the fault characteristic frequency in the preset range of the fault frequency is calculated.
The characteristic frequency energy calculation method refers to formula (6):
（6）
in the above-mentioned formula (6),representing the energy of the frequency of the fault,frepresenting the frequency of the failure,f i representing the magnitude of the ith frequency point in the fault frequency range, and R represents all frequency points in the predetermined range.
Regarding the determination of the failure frequency, for example, the frequency multiplication feature (1 frequency multiplication, 2 frequency multiplication, 3 frequency multiplication, … …), for example, the outer ring failure feature (outer ring failure frequency (outer ring failure coefficient×fundamental frequency) and 1 to 10 harmonic frequencies thereof), for example, the inner ring failure feature (inner ring failure frequency (inner ring failure coefficient×fundamental frequency) and 1 to 10 harmonic frequencies thereof, fundamental frequency sideband frequencies thereof). It will be appreciated that the frequency of failure may also be obtained from manufacturer testing of the rotating equipment, and typically the manufacturer will provide the frequency of failure in a product manual or parameter description of the rotating equipment.
Fig. 5 shows a flow chart of a method five for fuzzy control system construction according to an embodiment of the present invention. It should be appreciated that method five may be performed, for example, at the electronic device 1100 depicted in fig. 11. May also be performed at the computing device 110 depicted in fig. 1. It should be appreciated that method five may also include additional actions not shown and/or may omit actions shown, the scope of the invention being not limited in this respect.
At step 502, the computing device 110 determines fuzzy control variables and variable ranges corresponding to any of the temporal morphology feature, the spectral morphology feature, the axial trace morphology feature, and/or the fault frequency energy feature for the fault.
In some embodiments, a time domain morphology feature, a frequency spectrum morphology feature, an axis trace morphology feature, and/or a fault frequency energy feature is determined with respect to the fault based on the fault rules.
Regarding the failure rule, for example, when the rotating apparatus is a water pump, the failure rule regarding the water pump is obtained based on expert experience and a mechanism model. For example, the types of faults included in the fault rules are rolling bearing inner ring faults, rolling bearing outer ring faults, sliding bearing oil film instability/whirl/oscillation, coupling misalignment, rotor shaft bending, etc.
The fault rule includes a time domain shape (time domain waveform) feature, a frequency domain shape (frequency spectrum) feature, a shape of an axis trace feature and a fault frequency energy feature, a feature of a value, a similarity range, and the like, which are related to various faults. For example, regarding a rolling bearing inner ring failure: the acceleration frequency spectrum has inner ring fault frequency, harmonic wave and axial frequency as sidebands; with respect to slide bearing oil film instability/whirl/oscillation failure: the axis track is 8-shaped, and the frequency multiplication of the speed spectrum is 0.3X-0.7X, and the axis track has peak protrusions; large rules regarding plain bearing clearances: the axis track is a hollow perfect circle, the phase is approaching 90 degrees, and the speed spectrum appears 1X frequency multiplication or the peak protrusion of the lobe number frequency multiplication of the bearing bush.
Regarding determining the fuzzy control variable and variable range, for example, a fuzzy control variable is created for each feature (time domain morphology feature, frequency spectrum morphology feature, axis trace morphology feature, and/or fault frequency energy feature), for example, a similarity range of 0-1 for the shockwave waveform in the time domain morphology.
At step 504, the computing device 110 determines a membership function for the fuzzy control variable based on the determined fuzzy control variable and variable range for the fault.
For example, taking the shock waveform fuzzy control variable as an example, the fuzzy set is defined as: the shockwave similarity (small) =n1, shockwave similarity (medium) =m1, shockwave similarity (large) =p1, shockwave similarity can be converted into membership by determining membership functions, for example, membership functions select triangular membership functions, for example, membership functions can be linear or nonlinear.
Therefore, through the means, the method can convert the relatively fuzzy expression of the variable form in the fault rule into a clear functional relation; such as large shock waves, small shock waves, and what ranges should be respectively included, such as the degree of protrusion (what degree of protrusion is large); the probability that the value of each feature belongs to the membership function is determined by determining the membership function to map the value of each morphological feature or energy feature into the membership function, so that the time domain morphological feature, the frequency spectrum morphological feature, the axis track morphological feature and the fault frequency energy feature are further quantized, and fault matching and fault probability calculation are conveniently carried out.
At step 506, the computing device 110 determines a fuzzy rule based at least on the fuzzy control variable and the membership function for the fuzzy control variable.
Regarding the fuzzy rule, for example, based on the determined fault rule, the fuzzy rule is determined based on the fuzzy control variable and the membership function of the fuzzy control variable so as to realize the fault matching and the fault probability calculation of the feature data of the input fuzzy system through the fuzzy rule. It should be appreciated that, for fault rules, a plurality of fuzzy rules are generated in the fuzzy system for matching the calculation of the plurality of fault types and the occurrence probabilities thereof.
Regarding the feature data of the input fuzzy system, including the time domain morphology feature, the frequency spectrum morphology feature, the axis track morphology feature and the fault frequency energy feature, it should be understood that the rotating device may not necessarily use all the feature data when acquiring the fault and the fault probability, and the morphology characterization of some faults may only involve the change of a certain morphology feature. In the scheme provided by the embodiment of the invention, the fault type of the rotating equipment cannot be known in advance before fault matching and fault probability calculation are carried out, so that the acquired time domain morphological characteristics, frequency spectrum morphological characteristics, axis track morphological characteristics and fault frequency energy characteristics are input into a fuzzy system, and a more comprehensive fault prediction result can be matched.
Regarding the method of resolving blur, for example, a blur value of a control amount is calculated using the gravity center method, the calculation method is as follows formula (7):
（7）
in the above formula (7), U represents the position of the center of gravity of the fuzzy set, h i Representing the value of variable h, u N (h i ) Is a membership function representing element h i Belonging to the degree of fuzzy aggregation N.
For example, there are hundreds of fuzzy rules in the fuzzy system, and based on the input time domain morphological feature, frequency spectrum morphological feature, axis track morphological feature and fault frequency energy feature of the rotating equipment A, calculation is performed based on a fuzzy solution method to determine that the probability of first faults is 50%, the probability of second faults is 90% and the probability of other faults is 0.
Therefore, by the means, the method takes the mechanism model as a starting point, performs characteristic construction on common fault forms (time domain waveforms and frequency spectrum forms) in the mechanism model, and acquires a fault form similarity calculation model; converting the data acquired by the acceleration sensor into time domain data and frequency domain data of the rotating equipment so as to calculate the fault form of the rotating equipment through a fault form similarity calculation model; converting the acquired data of the acceleration sensor into an axle center track, and constructing an axle center track identification model to judge the form of the axle center track; and introducing the obtained time domain morphological characteristics, frequency domain morphological characteristics, axis track characteristics and fault frequency energy characteristics into a fuzzy control system, and finally quantifying the fault occurrence probability and intuitively reflecting possible faults.
Regarding the construction of the fault morphology similarity calculation model, a specific frequency spectrum morphology or an axis track morphology can be converted into a pre-training model in advance through a pre-training method and the like, the morphology similarity can be directly calculated through a calling model, and quantized morphology features can be directly input into the model as parameters to carry out auxiliary judgment. Meanwhile, the acceleration output by the acceleration sensor can be converted into speed and displacement time domain data, and the axis track is obtained by adopting technologies such as filtering, fourier transformation and the like.
In addition, the fuzzy control system is suitable for the problems of insensitive response time, complex system model or difficult accurate description, and can process nonlinear, multivariable and dynamic systems, so that the problems of high complexity and high accuracy requirement are solved while the system is simplified. Even if an unknown fault cause occurs, the fuzzy control system can still control, and feedback the unknown faults to technicians with the help of related programs, so that a fault library is expanded, the identification capability is continuously developed and improved, and a virtuous circle is formed.
Fig. 10 schematically illustrates a block diagram of a second system implementing a method for predicting failure probability of a rotating equipment according to further embodiments of the invention. The second system includes, for example, a basic information acquisition module 1002, a data acquisition module 1004, a data quality monitoring module 1006, a signal conversion module 1008, a fault morphology similarity training module 1010, an axis trajectory morphology similarity training module 1012, a feature calculation module 1014, a fault rule extraction module 1016, and a fuzzy control system 1018.
Regarding the basic information collection module 1002, it is mainly used for collecting the working condition data and the equipment parameter information regarding the rotating equipment.
As for the data acquisition module 1004, it is mainly used to acquire acceleration detection data regarding each detection position of the rotating apparatus.
With respect to the data quality monitoring module 1006, it is mainly used to determine whether the acquired acceleration detection data has an abnormality, where the abnormality includes one or more of the following abnormality conditions of the acceleration detection data: the failure rate is higher than a predetermined failure rate threshold, the abnormality rate is higher than a predetermined abnormality rate threshold, etc.; if the obtained acceleration data is determined to be abnormal, eliminating the abnormal acceleration data.
Regarding the signal conversion module 1008, it is used to perform signal conversion and spectrum conversion for the acceleration detection data, for example, it is used to process for the acquired acceleration detection data to acquire time domain data, frequency domain data, and axis locus data regarding the rotating device. For example, fourier transform is performed on the obtained acceleration detection data, time-domain integration is performed based on the acceleration detection data after fourier transform, and inverse fourier transform is performed to obtain speed time-domain data about the rotating device; performing Fourier transform on the obtained speed time domain data, performing time domain integration based on the speed time domain data after Fourier transform, and performing inverse Fourier transform to obtain displacement time domain data about the rotating equipment; based on the acquired acceleration time domain data, velocity time domain data, and displacement time domain data, frequency spectrum transformation is performed, respectively, to acquire acceleration frequency domain data, velocity frequency domain data, and displacement frequency domain data with respect to the rotating device.
A fault morphology similarity training module 1010 for training the fault morphology data to calculate the similarity of the acquired frequency domain data, time domain data and the fault morphology; the method is used for morphological data collection, morphological data preprocessing and morphological recognition model training.
The axle center track form similarity training module 1012 is used for training the axle center track form data so as to calculate the axle center track similarity of the obtained acceleration detection data.
With respect to the feature computation module 1014, it is primarily used to compute fault frequency energy and to compute fault morphology similarity, including fault frequency energy computation and fault morphology similarity computation. For example, a predetermined failure frequency associated with a failure of the rotating equipment is determined based on the acquired operating condition data, equipment parameter information, and frequency domain data, and predetermined mechanism model data; determining a fault frequency energy characteristic of the predetermined fault frequency within a predetermined range based on the predetermined range; and calculating a probability of failure with respect to the rotating equipment based on the determined failure frequency energy feature, time domain morphology feature, frequency spectrum morphology feature, and axis locus morphology feature. For example, based on the trained feature computation model, morphological similarity is computed for time domain data, frequency domain data, and axis trajectory data.
With respect to the fault rule extraction module 1016, it is configured to generate a plurality of fault rules related to the fault morphology based on the mechanism model in combination with expert experience.
Regarding the fuzzy control system 1018, it is used for defining fuzzy control variables, defining fuzzy sets and membership functions, defining fuzzy rules, selecting a fuzzy solution mode and calculating fault probability; for determining a probability of occurrence of at least one fault type with respect to the rotating equipment based on at least one of the determined fault frequency energy characteristic, time domain morphology characteristic, frequency spectrum morphology characteristic and axis trajectory morphology characteristic.
Fig. 11 shows a schematic step diagram of an example electronic device 1100 that may be used to implement embodiments of the present description. For example, computing device 110 as shown in FIG. 1 may be implemented by electronic device 1100. As shown, the electronic device 1100 includes a Central Processing Unit (CPU) 1101 that can perform various suitable actions and processes in accordance with computer program instructions stored in a Read Only Memory (ROM) 1102 or loaded from a storage unit 1108 into a Random Access Memory (RAM) 1103. In the random access memory 1103, various programs and data necessary for the operation of the electronic device 1100 can also be stored. The central processing unit 1101, the read only memory 1102 and the random access memory 1103 are connected to each other through a bus 1104. An input/output interface (I/O) 1105 is also connected to bus 1104.
A number of components in the electronic device 1100 are connected to the input/output interface 1105, including: an input unit 1106 such as a keyboard, mouse, microphone, etc.; an output unit 1107 such as various types of displays, speakers, and the like; a storage unit 1108, such as a magnetic disk, optical disk, etc.; and a communication unit 1109 such as a network card, modem, wireless communication transceiver, or the like. The communication unit 1109 allows the device 1100 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.
The various processes and treatments described above, e.g., methods two through five, may be performed by the central processing unit 1101. For example, in some embodiments, methods two through five may be implemented as computer software programs tangibly embodied on a machine-readable medium, such as storage unit 1108. In some embodiments, some or all of the computer programs may be loaded and/or installed onto the device 1100 via the read only memory 1102 and/or the communication unit 1109. One or more of the acts of methods two through five described above may be performed when a computer program is loaded into random access memory 1103 and executed by central processing unit 1101.
The present invention relates to methods, apparatus, systems, electronic devices, computer readable storage media and/or computer program products. The computer program product may include computer readable program instructions for carrying out aspects of the present invention.
The computer readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: portable computer disks, hard disks, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static Random Access Memory (SRAM), portable compact disk read-only memory (CD-ROM), digital Versatile Disks (DVD), memory sticks, floppy disks, mechanical coding devices, punch cards or in-groove structures such as punch cards or grooves having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media, as used herein, are not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., optical pulses through fiber optic cables), or electrical signals transmitted through wires.
The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a respective computing/processing device or to an external computer or external storage device over a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge computing devices. The network interface card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing/processing device.
Computer program instructions for carrying out operations of the present invention may be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, c++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing electronic circuitry, such as programmable logic circuitry, field Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), with state information for computer readable program instructions, which can execute the computer readable program instructions.
Various aspects of the present invention are described herein with reference to flowchart illustrations and/or step diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or step diagrams, and combinations of blocks in the flowchart illustrations and/or step diagrams, can be implemented by computer-readable program instructions.
These computer readable program instructions may be provided to a processing unit of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or step diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium having the instructions stored therein includes an article of manufacture including instructions which implement the function/act specified in the flowchart and/or step diagram step or steps.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or step diagram block or blocks.
The flowcharts and step diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block of the flowchart or step diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the steps may occur out of the order noted in the figures. For example, two consecutive steps may actually be performed substantially in parallel, and they may sometimes be performed in reverse order, depending on the function involved. It will also be noted that each step of the step diagrams and/or flowchart illustration, and combinations of steps in the step diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The foregoing description of embodiments of the invention has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the technical improvements in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims (11)

1. A method for predicting failure probability of a rotating device, comprising:
acquiring working condition data, equipment parameter information and acceleration detection data of each detection position of the rotating equipment;
processing the acquired acceleration detection data to acquire time domain data, frequency domain data and axis track data about the rotating equipment;
calculating morphological similarity according to the time domain data, the frequency domain data and the axis track data based on the trained characteristic calculation model so as to determine the time domain morphological feature, the frequency spectrum morphological feature and the axis track morphological feature corresponding to the current acceleration detection data; and
Calculating a probability of failure with respect to the rotating device based on at least one of the determined time domain morphology feature, frequency spectrum morphology feature, and axis trace morphology feature;
wherein the obtaining of time domain data, frequency domain data, and axis trajectory data about the rotating device includes:
performing Fourier transform on the obtained acceleration detection data, performing time domain integration based on the acceleration detection data after Fourier transform, and performing inverse Fourier transform to obtain speed time domain data about the rotating equipment;
performing Fourier transform on the obtained speed time domain data, performing time domain integration based on the speed time domain data after Fourier transform, and performing inverse Fourier transform to obtain displacement time domain data about the rotating equipment; and
based on the obtained acceleration time domain data, velocity time domain data, and displacement time domain data, frequency spectrum transformation is performed, respectively, to obtain acceleration frequency domain data, velocity frequency domain data, and displacement frequency domain data with respect to the rotating device.
2. The method as recited in claim 1, further comprising:
determining a predetermined failure frequency associated with a failure of the rotating equipment based on the acquired operating condition data, equipment parameter information, frequency domain data, and predetermined mechanism model data;
Determining a fault frequency energy characteristic of the predetermined fault frequency within a predetermined range based on the predetermined range; and
and calculating the probability of the fault of the rotating equipment based on the determined fault frequency energy characteristic, time domain morphological characteristic, frequency spectrum morphological characteristic and axis track morphological characteristic.
3. The method of claim 2, wherein calculating a probability of failure with respect to the rotating equipment comprises:
based on at least one of the determined fault frequency energy characteristic, time domain morphology characteristic, frequency spectrum morphology characteristic, and axis trace morphology characteristic, a probability of occurrence of at least one fault type with respect to the rotating equipment is determined via a predetermined fuzzy control system calculation.
4. A method according to claim 3, characterized in that the predetermined fuzzy control system is generated via the steps of:
determining a fuzzy control variable and a variable range corresponding to any one of a time domain morphological feature, a frequency spectrum morphological feature, an axis track morphological feature and/or a fault frequency energy feature related to the fault;
determining a membership function for the fuzzy control variable based on the determined fuzzy control variable and variable range for the fault; and
A fuzzy rule is determined based at least on the fuzzy control variable and a membership function for the fuzzy control variable.
5. The method of claim 1, wherein the step of determining the position of the substrate comprises,
the time domain morphology includes at least shock waves, beat waves and non-periodic waveforms;
the spectral morphology includes at least harmonics, sidebands and bottom ridges; and
the axle center track form at least comprises banana shape, ellipse shape, inner splayed shape, multiple splayed shape and outer splayed shape.
6. The method of claim 5, wherein prior to calculating morphological similarity for the acquired time domain data, frequency domain data, and axis trajectory data based on the trained feature calculation model, further comprising:
normalizing the acquired acceleration time domain data, velocity time domain data and displacement time domain data based on a preset sampling period to eliminate dimension gaps of the time domain data about the rotating equipment acquired under different working conditions; and
and normalizing the acquired acceleration frequency domain data, velocity frequency domain data and displacement frequency domain data based on a preset fundamental frequency to eliminate dimension differences of the acquired frequency domain data about the rotating equipment under different working conditions.
7. The method according to any one of claims 1 to 6, wherein the trained feature calculation model comprises a fault morphology similarity calculation model and a fault frequency energy calculation model, the fault morphology similarity calculation model being obtained via:
acquiring multiple sets of training time domain data and corresponding time domain forms, multiple sets of training frequency domain data and corresponding frequency domain forms, multiple sets of training axle center track data and corresponding axle center track forms of rotating equipment; and
and determining the time domain form corresponding to the training time domain data, the frequency domain form corresponding to the training frequency domain data and the axle center track form corresponding to the training axle center track data as training data, and training at least one preset neural network model to obtain a trained fault form similarity calculation model.
8. The method according to any one of claims 1 to 6, wherein the rotating device is a pump, and the acceleration detection data includes:
three-way acceleration detection data of at least one end of the pump driving end bearing, the pump non-driving end side, the motor driving end and the motor non-driving end;
the three-way acceleration detection data includes acceleration detection data in a vertical direction, acceleration detection data in a horizontal direction, and acceleration detection data in an axial direction.
9. The method of any one of claims 1 to 6, wherein the operating condition data is at least one of a rotational speed, a temperature, and a flow rate;
the equipment parameter information at least comprises the number of blades, resonance frequency and bearing fault coefficients; and
the predetermined fault frequency at least comprises a predetermined multiplication frequency, an outer ring fault frequency and harmonic frequency thereof, an inner ring fault frequency and harmonic frequency thereof, and a fundamental frequency sideband frequency.
10. A computing device, comprising:
at least one processing unit;
at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, which when executed by the at least one processing unit, cause the apparatus to perform the steps of the method according to any one of claims 1 to 9.
11. A computer readable storage medium having stored thereon a computer program which, when executed by a machine, implements the method according to any of claims 1 to 9.
CN202410003478.3A 2024-01-02 2024-01-02 Method, apparatus and storage medium for predicting failure probability of rotating apparatus Active CN117494485B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410003478.3A CN117494485B (en) 2024-01-02 2024-01-02 Method, apparatus and storage medium for predicting failure probability of rotating apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410003478.3A CN117494485B (en) 2024-01-02 2024-01-02 Method, apparatus and storage medium for predicting failure probability of rotating apparatus

Publications (2)

Publication Number Publication Date
CN117494485A CN117494485A (en) 2024-02-02
CN117494485B true CN117494485B (en) 2024-03-29

Family

ID=89671299

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410003478.3A Active CN117494485B (en) 2024-01-02 2024-01-02 Method, apparatus and storage medium for predicting failure probability of rotating apparatus

Country Status (1)

Country Link
CN (1) CN117494485B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114165474A (en) * 2022-02-11 2022-03-11 蘑菇物联技术（深圳）有限公司 Method, apparatus and computer storage medium for detecting a fault condition of an air compressor
CN115593880A (en) * 2022-10-10 2023-01-13 国能神东煤炭集团有限责任公司（Cn） Fault early warning method and fault early warning device for belt conveyor

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114165474A (en) * 2022-02-11 2022-03-11 蘑菇物联技术（深圳）有限公司 Method, apparatus and computer storage medium for detecting a fault condition of an air compressor
CN115593880A (en) * 2022-10-10 2023-01-13 国能神东煤炭集团有限责任公司（Cn） Fault early warning method and fault early warning device for belt conveyor

Also Published As

Publication number Publication date
CN117494485A (en) 2024-02-02

Similar Documents

Publication Publication Date Title
Pan et al. LiftingNet: A novel deep learning network with layerwise feature learning from noisy mechanical data for fault classification
Kumar et al. Optimization of VMD using kernel-based mutual information for the extraction of weak features to detect bearing defects
Cheng et al. A deep learning-based remaining useful life prediction approach for bearings
Lu et al. In situ motor fault diagnosis using enhanced convolutional neural network in an embedded system
Wang et al. A novel statistical time-frequency analysis for rotating machine condition monitoring
Yongbo et al. Review of local mean decomposition and its application in fault diagnosis of rotating machinery
Soualhi et al. Pattern recognition method of fault diagnostics based on a new health indicator for smart manufacturing
Wu et al. Rub-impact fault diagnosis of rotating machinery based on 1-D convolutional neural networks
Heyns et al. Combining synchronous averaging with a Gaussian mixture model novelty detection scheme for vibration-based condition monitoring of a gearbox
Popescu et al. Fault detection of rolling element bearings using optimal segmentation of vibrating signals
Cong et al. Research of singular value decomposition based on slip matrix for rolling bearing fault diagnosis
Chaabi et al. Fault diagnosis of rolling bearings in non-stationary running conditions using improved CEEMDAN and multivariate denoising based on wavelet and principal component analyses
Ding et al. Sparsity-based algorithm for condition assessment of rotating machinery using internal encoder data
Kim et al. Deep learning-based explainable fault diagnosis model with an individually grouped 1-D convolution for three-axis vibration signals
Xu et al. Deep coupled visual perceptual networks for motor fault diagnosis under nonstationary conditions
Liu et al. An online bearing fault diagnosis technique via improved demodulation spectrum analysis under variable speed conditions
Ayad et al. Search of a robust defect signature in gear systems across adaptive Morlet wavelet of vibration signals
Carbajal-Hernández et al. Classification of unbalance and misalignment in induction motors using orbital analysis and associative memories
Zhao et al. Underdetermined blind source extraction of early vehicle bearing faults based on EMD and kernelized correlation maximization
Zhao et al. Intelligent fault identification for rolling element bearings in impulsive noise environments based on cyclic correntropy spectra and LSSVM
Tang et al. Bearing fault diagnosis based on the maximum squared-enveloped multipoint kurtosis morphological deconvolution
Yang et al. An explainable intelligence fault diagnosis framework for rotating machinery
Cannarile et al. Elastic net multinomial logistic regression for fault diagnostics of on-board aeronautical systems
Mubaraali et al. Intelligent fault diagnosis in microprocessor systems for vibration analysis in roller bearings in whirlpool turbine generators real time processor applications
Akbar et al. State-of-the-Art Techniques for Fault Diagnosis in Electrical Machines: Advancements and Future Directions

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant