CN103217366A - Method for processing online oil path grain signal - Google Patents

Method for processing online oil path grain signal Download PDF

Info

Publication number
CN103217366A
CN103217366A CN2013101162117A CN201310116211A CN103217366A CN 103217366 A CN103217366 A CN 103217366A CN 2013101162117 A CN2013101162117 A CN 2013101162117A CN 201310116211 A CN201310116211 A CN 201310116211A CN 103217366 A CN103217366 A CN 103217366A
Authority
CN
China
Prior art keywords
abrasive particle
data
signal
point
oil circuit
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.)
Granted
Application number
CN2013101162117A
Other languages
Chinese (zh)
Other versions
CN103217366B (en
Inventor
赵辉
古军
杨坚
杨红宇
张放
詹惠琴
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.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
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 University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201310116211.7A priority Critical patent/CN103217366B/en
Publication of CN103217366A publication Critical patent/CN103217366A/en
Application granted granted Critical
Publication of CN103217366B publication Critical patent/CN103217366B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Investigating Or Analyzing Materials By The Use Of Magnetic Means (AREA)
  • Finish Polishing, Edge Sharpening, And Grinding By Specific Grinding Devices (AREA)

Abstract

The invention provides a method for processing an online oil path grain signal. The method comprises the steps of: firstly, carrying out alternative cache and reading on the acquired online oil path grain signal through two cache regions with storage depth of N so that continuous acquisition is realized and a condition is provided for continuous processing; secondly, circularly splicing the online oil path grain signal data acquired by the two front and rear cache regions to obtain the online oil path grain signal data with length of 2N, and then carrying out overlapping redundancy FFT (Fast Fourier Transform)-denoising-IFFT (Inverse Fast Fourier Transform) frequency domain filtering to ensure that the signal data with frequency higher than a set frequency fset is filtered; and thirdly, extracting a dynamic average, and extracting an effective grain signal according to the constraint condition. According to the method, the signal data such as a carrier signal with the frequency higher than the set frequency fset can be effectively filtered so as to avoid the interference, low-frequency components generated by power voltage low-frequency fluctuation, sensor temperature drift and the like can be filtered, and the effective grain signal is accurately obtained according to the constraint condition, and thus the size of grains is accurately known, and the accuracy of monitoring the grains of an online oil path is improved.

Description

A kind of disposal route of online oil circuit abrasive particle signal
Technical field
The invention belongs to the wear particle monitoring technical field, more specifically say, relate to a kind of disposal route of online oil circuit abrasive particle signal.
Background technology
Wearing and tearing are the one of the main reasons that cause component of machine to lose efficacy, abrasive particle in oil circuit fluid is the important information carrier of relevant engine scuffing state, abrasive particle is carried out to quantity statistics and size fractionation can judge the mechanical degree of wear and running status well, predict potential danger.
The method of carrying out wear particle monitoring in oil circuit fluid at present both at home and abroad mainly contains spectrum, iron spectrum, grain count and magnetic barrier method, and these methods belong to off-line measurement, and real-time performance is limited.
The applicant has proposed a kind of online wear particle monitoring method, utilize electromagnetic induction principle, adopt inductance type oil circuit abrasive particle sensor to be detected abrasive particle, during pipeline in the fluid with abrasive particle is flowed through inductance type oil circuit abrasive particle sensor, change the Distribution of Magnetic Field of inductance type oil circuit abrasive particle cell winding by magnetization, and then the equivalent inductance of change coil, produce corresponding induced signal, this induced signal is amplified, can obtain online oil circuit abrasive particle signal after the important conditioning process such as detection.
Fig. 1 is a kind of hardware system of online oil circuit abrasive particle signal acquisition and processing.
As shown in Figure 1, this hardware system can detect and the processes sensor induced signal in real time, is to take input and the treatment circuit that DSP is core.
Each module and function are mainly as follows:
DSP and memory module: this part mainly realizes adjusting control, signals collecting and the signal processing of exciting signal source, realizes that RS232 communicates by letter with host computer simultaneously, receives the parameter configuration of host computer.Wherein, signal after AD1 pick-up transducers magnetic test coil induced signal amplifies, it is the prime output signal, and as the foundation of balance adjustment, AD2 amplifies the final stage output signal of exporting with conditioning module for collection signal, be online oil circuit abrasive particle signal, the raw data of processing as signal on software.
Exciting signal source: be mainly used in providing inductance type oil circuit abrasive particle sensor required excitation signals U1 and U2, wherein U1 amplitude and phase place are fixed, and the amplitude of an other road U2 is adjustable and adjustable with respect to the phase place of U1.In a basic balance for the magnetic field that the pumping signal that makes inductance type oil circuit abrasive particle sensor both sides drive coil produces, must be regulated phase place and the amplitude of pumping signal U2.Particularly, DSP utilizes spi bus to send the amplitude/phase code value and realizes amplitude modulation and phase modulation to exciting signal source.Particularly, amplitude modulation is to realize by DA, and phase modulation is realized by FPGA.
Signal amplifies and conditioning module: mainly the induced signal on inductance type oil circuit abrasive particle sensor intermediate detection coil is carried out to prime amplification, detection filter and amplify and process, extract online oil circuit abrasive particle signal and deliver to DSP for the AD2 sampling and do signal and process.Whether inductance type oil circuit abrasive particle sensor two ends pumping signal balance, can embody from the amplitude of middle magnetic test coil induction gained induced signal, specifically, detect prime and amplify the output induced signal, if amplitude is greater than designing requirement, DSP utilizes spi bus to send the amplitude/phase code value and realizes amplitude modulation and phase modulation to exciting signal source, finally makes the induced signal of output meet design requirement, and the magnetic field that pumping signal produces is in a basic balance.And online oil circuit abrasive particle signal directly gathers by AD2 the PPI mouth that enters DSP.
DSP carries out analyzing and processing to online oil circuit abrasive particle signal, obtains the parameters such as the size of abrasive particle in fluid and material, and these parameters reflect the abrasion condition of monitored plant equipment, thereby obtain in real time the health status of equipment under test, realize the monitoring of online oil circuit abrasive particle.
But comprise some undesired signals in online oil circuit abrasive particle signal, how to obtain effective abrasive particle signal, making online oil circuit abrasive particle monitoring is more accurately to need the major issue solved.
Summary of the invention
The object of the present invention is to provide a kind of disposal route of online oil circuit abrasive particle signal, with the undesired signal in the online oil circuit abrasive particle of effective filtering signal, to obtain effective abrasive particle signal, thereby obtain more exactly the size cases of abrasive particle, improve the accuracy of online oil circuit abrasive particle monitoring.
For realizing above purpose, the disposal route of the online oil circuit abrasive particle of the present invention signal, is characterized in that, comprises the following steps:
(1), the online oil circuit abrasive particle of double buffering signals collecting
The data that collected for online oil circuit abrasive particle signal, the buffer zone Buffer1 that is N with two storage depths and Buffer2 replace buffer memory and read it: after buffer zone Buffer1 is filled with, generation is interrupted, allow DSP read the data of this buffer zone Buffer1, simultaneously, DMA(Direct Memory Address, direct memory access (DMA)) controller deposits the data of collection to buffer zone Buffer2, after buffer zone Buffer2 is filled with, generation is interrupted, allow DSP read the data of this buffer zone Buffer2, simultaneously, DMA deposits the data of collection to buffer zone Buffer1, so alternately, on the software processing speed, need to meet: within the time that gathers a full buffer zone, DSP must complete reading and processing data in another buffer zone,
(2), FFT frequency domain filtering
At first the N data of the N data of front another buffer zone once read and the current buffer zone read are spliced, obtain the online oil circuit abrasive particle signal data of counting as 2N, then call the FFT(Fourier transform) function, it is transformed into to frequency domain, obtains counting as the online oil circuit abrasive particle signal data of 2N of frequency domain;
By in the online oil circuit abrasive particle signal data of counting as 2N of frequency domain, represent setpoint frequency f setabove spectral line value directly is set to 0, and then call the IFFT(Fourier inversion) function, by counting as the online oil circuit abrasive particle of 2N signal data is transformed into time domain of frequency domain, and remove front N/2 point data, remove back N/2 point data, N point data in the middle of only retaining;
The middle N point data that each IFFT is retained are stitched together, and obtain filtered online oil circuit abrasive particle signal data;
(3), the dynamically extraction of average line
3.1), find out the extreme point of the online oil circuit abrasive particle signal data obtained after filtering;
3.2), get successively three extreme points of next-door neighbour as summit by the horizontal ordinate size, form a triangle, computing center's point: horizontal ordinate is directly got the horizontal ordinate of that middle point, ordinate is asked the ordinate mean value of three points, the sampling order value that wherein horizontal ordinate is online oil circuit abrasive particle signal data, the sampled value that ordinate of orthogonal axes is online oil circuit abrasive particle signal data;
3.3), three central points getting successively the next-door neighbour who calculates by the horizontal ordinate size are as triangular apex, form a new triangle, by 3.2 method, calculate new central point;
3.4), repeating step 3.3), until reach set point number;
3.5), every two adjacent central points are got up with segment link, do the mean value smoothing filtering of 3～5 points, obtain dynamic average line;
(4), the extraction of active grain signal
4.1), on the amplitude size, set a threshold value Δ A, as | y (x) | > during Δ A, find data point A corresponding in the online oil circuit abrasive particle signal data s (x) obtained after filtering, y (x)=s (x)-a (x) wherein, a (x) is dynamic average line, and x is horizontal ordinate, i.e. the sampling order value of data;
4.2), then according to the monotonicity of online oil circuit abrasive particle signal data s (x), find successively extreme point B, | y (x) | corresponding data point C during<Δ A, with the intersection point M of dynamic average line, | y (x) | > during Δ A corresponding data point D, extreme point E and | y (x) | corresponding data point F during<Δ A;
4.3) if meet following constraint condition:
1., BD slope over 10 k1, CE slope over 10 k2:
|(|k1|-|k2|)|<20%×min(|k1|,|k2|)；
2., the slope absolute value of CD section is all larger than the absolute value of BC section, DE slope over 10;
3., the horizontal ordinate x of B, M, tri-data points of E b, x mand x ebetween meet:
|(|x M-x B|-|x E-x M|)|<20%×min(|x M-x B|,|x E-x M|)；
4., the amplitude that B, two data points of E depart from dynamic average line meets:
|(|s(x B)-a(x B)|-|sa(x E)|)|<20%×min(|s(x B)-a(x B)|,|s(x E)-a(x E)|)；
Think that this section abrasive particle signal data is the active grain signal, and it is extracted;
(5), abrasive particle kind and size determines
5.1), the determining of abrasive particle kind: the phase place of the active grain signal of extraction and the phase place of demarcating the magnetic mill are compared, if homophase, abrasive particle is the magnetic substance abrasive particle, if anti-phase, abrasive particle is the diamagnetism abrasive particle;
5.2), the determining of abrasive particle size
1., first by the A data point, find forward the intersection point X1 with dynamic average line, by F point, find backward the intersection point X2 with dynamic average line, note D=X2-X1, i.e. the shared sampling number of active grain signal, then according to the area area of the active grain signal that calculates extraction:
area = &Sigma; X 1 X 2 | s ( x ) - a ( x ) | , x∈(X1,X2)；
2., area area is done to normalization, normalize to and account for a M 0corresponding area during sampled point:
Figure BDA00003014129300042
3., according to the area area after normalization normalizationdetermine the grade of abrasive particle.
The object of the present invention is achieved like this:
The disposal route of the online oil circuit abrasive particle of the present invention signal, at first by two storage depths, be that the N buffer zone replaces buffer memory and reads the online oil circuit abrasive particle signal data gathered, realized continuous acquisition, and for the condition that provides is provided continuously, then the online oil circuit abrasive particle signal data of adjacent twice collection is spliced, the online oil circuit abrasive particle signal data that obtains length and be 2N carries out the FFT frequency domain filtering, will be higher than setpoint frequency f setfiltering, then extract dynamic average line, and, according to constraint condition, extract the active grain signal.The present invention not only effectively in the online oil circuit abrasive particle of filtering signal higher than setpoint frequency f setdisturb, can also filtering because of the low-frequency component that supply voltage low-frequency fluctuation, sensor temperature drift etc. are former thereby produce, and, according to constraint condition, accurately obtain effective abrasive particle signal, thereby obtain more exactly the size cases of abrasive particle, improve the accuracy of online oil circuit abrasive particle monitoring.
The accompanying drawing explanation
Fig. 1 is a kind of hardware system of online oil circuit abrasive particle signal acquisition;
Fig. 2 is a kind of embodiment process flow diagram of disposal route of the online oil circuit abrasive particle of the present invention signal;
Fig. 3 is that online oil circuit abrasive particle signal shown in Fig. 1 is through the filtered waveform schematic diagram of FFT;
Fig. 4 is the result schematic diagram of continuously online oil circuit abrasive particle signal being carried out to FFT-denoising-IFFT filtering;
Fig. 5 is data splicing schematic diagram after IFFT;
Fig. 6 is that online oil circuit abrasive particle signal shown in Fig. 1 is through the filtered schematic diagram of the waveform with low-frequency disturbance of FFT;
Fig. 7 is dynamic average line extracting method schematic diagram;
Fig. 8 is active grain method for extracting signal schematic diagram;
Fig. 9 is active grain signal area normalization derivation schematic diagram;
Figure 10 is abrasive particle size fractionation schematic diagram.
Embodiment
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described, so that those skilled in the art understands the present invention better.Requiring particular attention is that, in the following description, when perhaps the detailed description of known function and design can desalinate main contents of the present invention, these are described in here and will be left in the basket.
Fig. 2 is a kind of embodiment process flow diagram of disposal route of the online oil circuit abrasive particle of the present invention signal.
In the present embodiment, as shown in Figure 2, the disposal route of the online oil circuit abrasive particle of the present invention signal comprises the online oil circuit abrasive particle of double buffering signals collecting 1, FFT frequency domain filtering 2, the extraction 3 of dynamic average line and several steps such as extraction 4 of active grain signal, below in conjunction with embodiment, each step is elaborated.
1, the online oil circuit abrasive particle of double buffering signals collecting
If pumping signal U1 and U2 frequency are f, the main carrier wave composition of signal of being sensed by the magnetic test coil in the middle of inductance type oil circuit abrasive particle sensor so is also f.In actual applications, abrasive particle is through out-of-date, and the abrasive particle signal is modulated onto on this carrier wave, abrasive particle signal frequency apparent inductance formula oil circuit abrasive particle sensor construction, oil circuit flow velocity and determine.Induced signal is by preposition amplification, synchronous detection again after the filtering direct current, enter frequency content the highest in the signal of parallel A/D converter AD2 and substantially can not surpass f, therefore, according to nyquist sampling theorem, in conjunction with the requirement of frequency resolution, parallel A/D converter AD2 chooses 3 times to the sampling rate of f.
In order to realize the continuous acquisition of abrasive particle signal and to process continuously, in the present embodiment, the buffer zone Buffer1 and the Buffer2 that have opened up two bit wide 16bit, storage depth N=512 at DSP in L1memory replace buffer memory and read, realize the PingPong operation, specifically realize by the dma descriptor table in DSP.When in buffer zone Buffer1, data acquisition reaches at 512, produce and interrupt, at the mid-bit flag variable of break in service function buffer1_ready=true, master routine detects buffer1_ready, true if, read the data of Buffer1, simultaneously, dma controller, according to descriptor table, is deposited the data of collection to Buffer2 automatically; In like manner, when in Buffer2, data acquisition reaches at 512, produce and interrupt, at the mid-bit flag variable of break in service function buffer2_ready=true, master routine detects buffer2_ready, true if, read the data of Buffer2, simultaneously, dma controller, according to descriptor table, is deposited the data of collection to Buffer1 automatically.So alternately;
On processing speed, need to meet: within the time that gathers full Buffer1, DSP must complete reading and processing data in Buffer2; Within the time that gathers full Buffer2, DSP must complete reading and processing data in Buffer1.
In the present embodiment, in master routine, utilize a while circulation constantly to detect data full scale will buffer1_ready and these two signs of buffer2_ready, if detect as true, read the online oil circuit abrasive particle signal data of corresponding buffer zone, after reading, data full scale will is updated to false.
2, FFT frequency domain filtering
2.1, the active grain signal
Abrasive Flow, after the pipeline in inductance type oil circuit abrasive particle sensor, has a monocycle sinusoidal modulation signal on the intermediate detection coil, can detect this monocyclic sinusoidal signal after synchronous detection, is referred to as the active grain signal.Its frequency is relevant with inductance type oil circuit abrasive particle sensor parameters with oil flow: F ≈ v/ (2 * d), and in formula, v is oil flow, d is the distances of drive coil two ends to center.In conjunction with actual inductance type oil circuit abrasive particle sensor parameters d and the oil flow v in rig-site utilization, determine the frequency of the active grain signal of the maximum that may exist, be designated as F max.
2.2, the FFT frequency domain filtering
In the present embodiment, for 2 * F in online oil circuit abrasive particle signal maxabove frequency component, can convert by FFT, and at frequency domain, by the even zero setting that decays of its spectral line, then IFFT transforms to time domain, can obtain online oil circuit abrasive particle signal curve preferably.After having captured separately one section 1024 the online oil circuit abrasive particle signal datas block-IFFT that is FFT-and processing, obtain the exemplary waveforms as Fig. 3, now to 2 * F maxfiltration result with upper frequency is comparatively desirable.In the present embodiment, used base-2FFT conversion.
In order to realize the continuous processing to signal, the present invention adopts the mode of overlapping redundancy to carry out block-IFFT of FFT-and processes, as shown in Figure 4.The N point that is last Buffer for the online oil circuit abrasive particle of the 2N=1024 point signal data of FFT each time and the N point data of current Buffer are spliced and are obtained by sampling order, put into array DataForFFT[1024], call Fourier transform function, it is transformed into to frequency domain, the frequency domain after FFT conversion count for the online oil circuit abrasive particle of 2N=1024 signal data be that the spectral line value is also returned in this array.By DataForFFT[1024] in represent setpoint frequency f set=2 * F maxabove frequency content directly is set to 0.Directly block that the continuity that may cause back IFFT inverse transformation result is bad a saltus step, but in order to guarantee to remove f setabove ripple, this mode of blocking is also one to be selected relatively preferably, and continuity can solve by other ways.Then call Fourier inversion by array DataForFFT[1024] in data carry out the IFFT conversion, remove front N/2=256 point data in result, remove back N/2=256 point data, 512 data in the middle of only retaining.Do like this, be equivalent to each reality and effectively processed N=512 point data, and be also to be just in time the N=512 point in the buffering of each sampling, just realized real-time sampling, processed in real time.
The middle N=512 point data that each IFFT retains are stitched together, and saltus step as shown in Figure 5 may occur.At this time only need to find the median point of splicing place trip point a, b, then cross median point and do straight line, hand over one section of front in a c, hand over one section of back in a d, connecting line segment cd, and with the discrete point on this line segment as filtering after signal in a c, the data of point between d, thereby guaranteed continuity, the flatness of online oil circuit abrasive particle signal.It is consistent with the ca section that the slope of cd section need keep, as long as realize allowing filtered signal continuously smooth.
3, the dynamically extraction of average line
As shown in Figure 6, the online oil circuit abrasive particle signal curve that filtering obtains has flatness preferably, but because supply voltage low-frequency fluctuation, sensor temperature drift or other reasons, the component that some frequency is very low appears in final filtered online oil circuit abrasive particle signal, show as the drift of baseline, namely mean value is constantly swinging up and down, the amplitude sometimes swung is larger, the extraction of this signal of active grain to back is very unfavorable, if can find the composition of this low frequency, or dynamically find an average line to replace just improving accuracy of detection.This curve is " dynamically average line " just as shown in Figure 6.In order to find this dynamic average line, in the present invention, taked following methods:
(1), find out the extreme point of the online oil circuit abrasive particle signal data obtained after filtering, these extreme points comprise maximum value and minimal value, by the time order and function occurred, sequentially arrange, as P1, P2, P3, P4, P5, P6, P7, P8, P9 ...
(2), three extreme points getting successively the next-door neighbour by horizontal ordinate sequence are as summit, form a triangle, computing center's point: horizontal ordinate directly get in the middle of the horizontal ordinate of that point, ordinate is got the ordinate average of three points.As Fig. 7, obtain successively central point M1, M2, M3, M4, M5, M6, M7 ... wherein M1 is the leg-of-mutton central point that extreme point P1, P2, P3 form, M2 is the leg-of-mutton central point that extreme point P2, P3, P4 form, M3 is the leg-of-mutton central point that extreme point P3, P4, P5 form, and the rest may be inferred.
(3), by the horizontal ordinate sequence, get successively three central points of the next-door neighbour who calculates as triangular apex, form a new triangle, by the method in (2), computing center's point A1, A2, A3, A4 ..., as shown in Figure 7, wherein, put M1, M2 centered by A1,, the leg-of-mutton central point that forms of M3, put the leg-of-mutton central point that M2, M3, M4 form centered by A2, put the leg-of-mutton central point that M3, M4, M5 form centered by A3.
(4), repeating step (3), until reach set point number, in this example, for convenience of explanation, set point number is 0, no longer repeats.
(5), by central point A1, A2, A3, A4, A5 ... these sequence of points, every two adjacent points get up with segment link, and this segmentation straight line has just relatively approached dynamic average line.If obtain better average line, can also continue the mode of top step (3), then ask several times central value line section again.After connecting line segment, also can do the mean value smoothing filtering of one time 3～5.
5, the extraction of active grain signal
In oil circuit, there is the fluid with abrasive particle to pass through, the active grain signal has appearred, the amplitude of active grain signal is necessarily many greatly than common interference waveform, on the amplitude size, set a threshold value Δ A, as shown in Figure 8, as | y (x) | > during Δ A, just mean that the active grain signal may occur.Y (x)=s (x)-a (x) wherein, a (x) is dynamic average line, x is that horizontal ordinate is data point sampling order value.
Mark is several important points as shown in Figure 8: A, B, C, D, E, F, M, wherein, A is | y (x) | find data point corresponding to online oil circuit abrasive particle signal data s (x) obtained after filtering during Δ A, according to monotonicity, find successively extreme point B, | y (x) | corresponding data point C during<Δ A, with the intersection point M of dynamic average line, | y (x) | > during Δ A corresponding data point D, extreme point E and | y (x) | corresponding data point F during<Δ A.
These points have following relation:
(1) if the A point exists, be bound to exist B, C 2 points;
(2) if the D point exists, be bound to exist E, F 2 points;
(3) if 3 existence of ABC may not have DEF 3 points (being caused by interference).
For these situations, the present invention provides four constraint conditions:
1., BD slope over 10 k1, CE slope over 10 k2:
|(|k1|-|k2|)|<20%×min(|k1|,|k2|)；
2., the slope absolute value of CD section is all larger than the absolute value of BC section, DE slope over 10;
3., the horizontal ordinate x of B, M, tri-data points of E b, x mand x ebetween meet:
|(|x M-x B|-|x E-x M|)|<20%×min(|x M-x B|,|x E-x M|)；
4., the amplitude that departs from dynamic average line of B, two data points of E meets:
|(|s(x B)-a(x B)|-|s(x E)-a(x E)|)|<20%×min(|s(x B)-a(x B)|,|s(x E)-a(x E)|)；
Wherein, min(*, *) mean to select less numerical value in two numerical value.
When meeting above-mentioned this four conditions, can conclude and an active grain signal detected, and it is extracted.Certainly because each active grain signal likely only has tens points, mean, some is sparse, the A found out, B, C, D, E, F, M these points and some difference of ideal position, but enough for judgement shape and size.
5, abrasive particle kind and size determines
5.1, the determining of abrasive particle kind: the phase place of the active grain signal of extraction and the phase place of demarcating the magnetic mill are compared, if homophase, abrasive particle is the magnetic substance abrasive particle, if anti-phase, abrasive particle is the diamagnetism abrasive particle.
Differentiation for magnetic substance abrasive particle and diamagnetism abrasive particle.After concrete circuit and inductance type oil circuit abrasive particle sensor are determined, the direction of the active grain signal that magnetic substance abrasive particle and diamagnetism abrasive particle induce is determined.If the first half first appears in the active grain signal, magnetic substance abrasive particle; If negative half one first occurs, it is the diamagnetism abrasive particle.This is not absolute, be mapped with concrete circuit and inductance type oil circuit abrasive particle sensor, does a demarcation and gets final product.Program can judge and record the active grain signal shape (positive or anti-phase) of each extraction, with the phase place of demarcating the magnetic mill, compares, if homophase, abrasive particle is the magnetic substance abrasive particle, if anti-phase, abrasive particle is the diamagnetism abrasive particle.
5.2, the determining of abrasive particle size
For speed, in fact, the sampling number that each waveform accounts for, shown speed.Determine flow velocity with inductance type oil circuit abrasive particle sensor, can be substantially determine by the point of 2 of B in Fig. 8, E is poor:
v = d t = d ( x E - x B ) / f s = df s x E - x B
Wherein, d is distance between the drive coil central point of inductance type oil circuit abrasive particle sensor two ends, and t means that abrasive Flow is through this segment distance of d time used.By counting between E, B, in conjunction with sample rate f sjust can determine the time, and then definite speed.
Determining the size of abrasive particle, is mainly to calculate with the area that dynamically average line surrounds by calculating the active grain signal.Feasibility analysis and correlation technique are as follows:
Two sinusoidal signal y0=sin (t) as shown in Figure 9, y1=sin (t/2).Calculate respectively the area of its one-period.
Area 0 = 2 * &Integral; 0 &pi; sin ( t ) dt
Area 1 = 2 * &Integral; 0 2 &pi; sin ( t / 2 ) dt = 2 * 2 * &Integral; 0 2 &pi; sin ( t / 2 ) d ( t / 2 )
Make x=t/2, Area 1 = 4 * &Integral; 0 &pi; sin ( x ) d ( x ) = 2 * Area 0
According to simply calculating: the dash area area of second figure is the twice of first figure dash area area.Further, if two signal y0=sin (t) are arranged, y1=sin (t/a) can draw: monocycle area A real1=a*Areal0, and frequency is dwindled a doubly, and area just enlarges a doubly, draws the area of active grain signal and the relation of frequency:
Area &Proportional; T = 1 F
Wherein, the cycle that T is the active grain signal.
Because the pipeline that same abrasive particle is flowed through in inductance type oil circuit abrasive particle sensor with friction speed, the frequency of the active grain signal of generation is different again:
F∝v
And according to experimental phenomena, the speed of abrasive particle is little on the amplitude impact of characteristic signal, substantially can ignore, only relevant with the relative position that is in pipeline with the abrasive particle diameter.So:
Area∝T∝1/F∝1/v
Hence one can see that, the area that the pipeline that same abrasive particle is flowed through in inductance type oil circuit abrasive particle sensor with friction speed produces also just may change, since amplitude and speed are substantially irrelevant, as long as by the normalization that concerns of area and speed, just can judge according to the area of useful signal the grade of abrasive particle diameter so.
The method of normalizing is as follows:
V0:M 0individual sampled point, area is A 0
V1:M 1individual sampled point, area is A 1
V2:M 2individual sampled point, area is A 2
Unification is arrived M to area normalization so 0area value during individual sampled point:
A 10=A 1×M 0/M 1
A 20=A 2×M 0/M 2
A now 10, A 20the area of representative, rejected the impact of speed, all normalized to the area of this speed of v0.
Specifically, for the definite of abrasive particle size, comprise the following steps:
1., first by the A data point, find forward the intersection point X1 with dynamic average line, by F point, find backward the intersection point X2 with dynamic average line, note D=X2-X1, i.e. the shared sampling number of active grain signal, then according to the area area of the active grain signal that calculates extraction:
area = &Sigma; X 1 X 2 | s ( x ) - a ( x ) | , x∈(X1,X2)；
2., area area is done and normalizes to M 0corresponding area during individual sampled point:
Figure BDA00003014129300112
3., according to the area area after normalization normalizationdetermine the grade of abrasive particle size.
Area A after normalization has embodied the size of abrasive particle substantially.But, under actual conditions, there are much relations the amplitude of active grain signal and the abrasive particle position in inductance type oil circuit abrasive particle sensor interior conduit.The closer to tube wall, signal amplitude is larger, and more in the pipeline middle, amplitude is less.So under non-uniform magnetic-field, the size that can't determine abrasive particle according to amplitude or the area of active grain signal.Can only carry out a general grade classification to abrasive particle.Rough foundation of dividing is: less abrasive particle, even pass through against tube wall, the active grain characteristic signal amplitude produced no more than a certain large abrasive particle from pipeline center by the time characteristic signal amplitude that produces large, so according to this reason, still can be by the abrasive particle classification.
Concrete Hierarchical Detailed, as shown in figure 10.If diameter is D1, D2, D3 ... abrasive particle when the sensor, the normalized area scope of its active grain characteristic signal be expressed as respectively (Area0, Area1), (Area2, Area3), (Area4, Area5) ...
At first, corresponding (Area0, the Area1) of abrasive particle that to record by experiment diameter be D1, as known conditions, then, find the abrasive particle that suitable diameter is D2 to make Area2=Area1 by experiment, determined diameter D2 and area A rea3.Continue to do experiment and find the abrasive particle that suitable diameter is D3 and make Area4=Area3, determined diameter D3 and area A rea5, by that analogy, the diameter of determining each abrasive particle with and the areal extent of abrasive particle signal.When abrasive particle signal area is obtained come and normalization after, successively with area A rea0, Area3, Area5 ... compare: iff meeting area normalizationarea1, more than abrasive particle being classified as to diameter D1; Iff meeting area normalizationarea3, abrasive particle is classified as and is greater than diameter D2 mono-class; Iff meeting area normalizationarea5, abrasive particle is classified as and is greater than diameter D3 mono-class; The rest may be inferred by analogy.The roughly grade of abrasive particle can be distinguished according to this mode, but only distinguished lower limit, do not determined the upper limit.
Although the above is described the illustrative embodiment of the present invention; so that those skilled in the art understand the present invention; but should be clear; the invention is not restricted to the scope of embodiment; to those skilled in the art; as long as various variations appended claim limit and definite the spirit and scope of the present invention in, these variations are apparent, all innovation and creation that utilize the present invention to conceive are all at the row of protection.

Claims (4)

1. the disposal route of an online oil circuit abrasive particle signal, is characterized in that, comprises the following steps:
(1), the online oil circuit abrasive particle of double buffering signals collecting
The data that collected for online oil circuit abrasive particle signal, the buffer zone Buffer1 that is N with two storage depths and Buffer2 replace buffer memory and read it: after Buffer1 is filled with, generation is interrupted, allow DSP read the data of this buffer zone Buffer1, simultaneously, DMA(Direct Memory Address, direct memory access (DMA)) controller deposits the data of collection to buffer zone Buffer2, after Buffer2 is filled with, generation is interrupted, allow DSP read the data of this buffer zone Buffer2, simultaneously, DMA deposits the data of collection to buffer zone Buffer1, so alternately, on the software processing speed, need to meet: within the time that gathers a full buffer zone, DSP must complete reading and processing data in another buffer zone,
(2), FFT frequency domain filtering
At first the N data of the N data of front another buffer zone once read and the current buffer zone read are spliced, obtain the online oil circuit abrasive particle signal data of counting as 2N, then call the FFT(Fourier transform) function, it is transformed into to frequency domain, obtains counting as the online oil circuit abrasive particle signal data of 2N of frequency domain;
By in the online oil circuit abrasive particle signal data of counting as 2N of frequency domain, represent setpoint frequency f setabove spectral line value directly is set to 0, and then call the IFFT(Fourier inversion) function, to counting as the online oil circuit abrasive particle of 2N signal data is transformed into time domain of frequency domain, and remove front N/2 point data, remove back N/2 point data, N point data in the middle of only retaining;
The middle N point data that each IFFT is retained are stitched together, and obtain filtered online oil circuit abrasive particle signal data;
(3), the dynamically extraction of average line
3.1), find out the extreme point of the online oil circuit abrasive particle signal data obtained after filtering;
3.2), three extreme points getting successively the next-door neighbour by the horizontal ordinate size are as summit, form a triangle, computing center's point: the horizontal ordinate of that point of horizontal ordinate in the middle of directly getting, ordinate is asked the ordinate mean value of three points;
3.3), by the horizontal ordinate size get successively calculate the next-door neighbour three central points as triangular apex, form a new triangle, calculate new central point by 3.3 method;
3.4), repeating step 3.3), until reach set point number;
3.5), every two adjacent central points are got up with segment link, do the mean value smoothing filtering of 3～5 points, obtain dynamic average line;
(4), the extraction of active grain signal
4.1), on the amplitude size, set a threshold value Δ A, as | y (x) | > during Δ A, find data point A corresponding in the online oil circuit abrasive particle signal data s (x) obtained after filtering, y (x)=s (x)-a (x) wherein, a (x) is dynamic average line, and x is horizontal ordinate;
4.2), then according to the monotonicity of online oil circuit abrasive particle signal data s (x), find successively extreme point B, | y (x) | corresponding data point C during<Δ A, with the intersection point M of dynamic average line, | y (x) | > during Δ A corresponding data point D, extreme point E and | y (x) | corresponding data point F during<Δ A;
4.3) if meet following constraint condition:
1., BD slope over 10 k1, CE slope over 10 k2:
|(|k1|-|k2|)|<20%×min(|k1|,|k2|)；
2., the slope absolute value of CD section is all larger than the absolute value of BC section, DE slope over 10;
3., the horizontal ordinate x of B, M, tri-data points of E b, x mand x ebetween meet:
|(|x M-x B|-|x E-x M|)|<20%×min(|x M-x B|,|x E-x M|)；
4., the amplitude that B, two data points of E depart from dynamic average line meets:
|(|s(x B)-a(x B)|-|s(x E)-a(x E)|)|<20%×min(|s(x B)-a(x B)|,|s(x E)-a(x E)|)；
Think that this section abrasive particle signal data is the active grain signal, and it is extracted;
(5), abrasive particle kind and size determines
5.1), the determining of abrasive particle kind: the phase place of the active grain signal of extraction and the phase place of demarcating the magnetic mill are compared, if homophase, abrasive particle is the magnetic substance abrasive particle, if anti-phase, abrasive particle is the diamagnetism abrasive particle;
5.2), the determining of abrasive particle size
1., first by the A data point, find forward the intersection point X1 with dynamic average line, by F point, find backward the intersection point X2 with dynamic average line, note D=X2-X1, i.e. the shared sampling number of active grain signal, then according to the area area of the active grain signal that calculates extraction:
area = &Sigma; X 1 X 2 | s ( x ) - a ( x ) | , x∈(X1,X2)；
2., area area is done to normalization, normalize to M 0corresponding area during individual sampled point:
3., according to the area area after normalization normalizationdetermine the grade of abrasive particle.
2. the disposal route of online oil circuit abrasive particle signal according to claim 1, is characterized in that, in step (1), further comprising the steps of:
When buffer zone Buffer1 is filled with rear generation interruption, in break in service, data full scale will is updated to buffer1_ready=true, for master routine, is used for reading buffer zone Buffer1 data;
When buffer zone Buffer1 is filled with rear generation interruption, in break in service, data full scale will is updated to buffer1_ready=true, for master routine, is used for reading buffer zone Buffer1 data;
The online oil circuit abrasive particle signal data flow process that described DSP reads this buffer zone Buffer1, Buffer1 is:
In master routine, utilize a while circulation constantly to detect data full scale will buffer1_ready and these two signs of buffer2_ready, if detect as true, read the online oil circuit abrasive particle signal data of corresponding buffer zone, after reading, data full scale will is updated to false.
3. the disposal route of online oil circuit abrasive particle signal according to claim 1, is characterized in that, further comprising the steps of in step (2):
Find the median point of splicing place trip point a, b in filtered online oil circuit abrasive particle signal data, then cross median point and do straight line, hand over one section of front in a c, hand over one section of back in a d, connecting line segment cd, and with the discrete point on this line segment as filtering after signal in a c, the data of point between d, thereby guaranteed continuity, the flatness of online oil circuit abrasive particle signal, it is consistent with the ca section that the slope of cd section need keep, as long as realize allowing filtered signal continuously smooth.
4. the disposal route of online oil circuit abrasive particle signal according to claim 1, is characterized in that, in step (5), described according to the area area after normalization normalizationthe grade of determining abrasive particle is:
Diameter is D1, D2, D3 ... abrasive particle when the sensor, the normalized area scope of its active grain characteristic signal be expressed as respectively (Area0, Area1), (Area2, Area3), (Area4, Area5),
At first, (Area0 corresponding to abrasive particle that to record by experiment diameter be D1, Area1), as known conditions, then, find by experiment the abrasive particle that suitable diameter is D2 to make Area2=Area1, determine diameter D2 and area A rea3, continued to do the abrasive particle that the suitable diameter of experiment searching is D3 and make Area4=Area3, determined diameter D3 and area A rea5, by that analogy, the diameter of determining each abrasive particle with and the areal extent of abrasive particle signal;
When abrasive particle signal area is obtained come and normalization after, successively with area A rea0, Area3, Area5 ... compare: iff meeting area normalizationarea1, more than abrasive particle being classified as to diameter D1; Iff meeting area normalizationarea3, abrasive particle is classified as and is greater than diameter D2 mono-class; Iff meeting area normalizationarea5, abrasive particle is classified as and is greater than diameter D3 mono-class, the rest may be inferred by analogy.
CN201310116211.7A 2013-04-07 2013-04-07 Method for processing online oil path grain signal Expired - Fee Related CN103217366B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310116211.7A CN103217366B (en) 2013-04-07 2013-04-07 Method for processing online oil path grain signal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310116211.7A CN103217366B (en) 2013-04-07 2013-04-07 Method for processing online oil path grain signal

Publications (2)

Publication Number Publication Date
CN103217366A true CN103217366A (en) 2013-07-24
CN103217366B CN103217366B (en) 2015-06-24

Family

ID=48815355

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310116211.7A Expired - Fee Related CN103217366B (en) 2013-04-07 2013-04-07 Method for processing online oil path grain signal

Country Status (1)

Country Link
CN (1) CN103217366B (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104390893A (en) * 2014-11-24 2015-03-04 电子科技大学 Oil mill grain sensor based on printed coil
CN105547932A (en) * 2015-12-08 2016-05-04 中国航空工业集团公司北京长城航空测控技术研究所 Engine lubricating oil metal particle on-line monitoring signal demodulation method
CN105823711A (en) * 2016-04-18 2016-08-03 上海电机学院 Online monitoring method for oil abrasive particles
CN109154586A (en) * 2016-04-27 2019-01-04 株式会社岛津制作所 Data processing method and device
CN109648478A (en) * 2019-01-30 2019-04-19 华侨大学 A kind of method and device thereof for monitoring wheel passivatio state on-line
CN109839334A (en) * 2019-03-12 2019-06-04 西安交通大学 Signal recognition method towards unicoil magnetic inductive abrasive grain detection sensor
CN111504859A (en) * 2020-04-28 2020-08-07 沈阳顺义科技有限公司 System and method for online monitoring and evaluation of lubricating oil abrasive particles
US10823700B2 (en) 2019-02-15 2020-11-03 Raytheon Technologies Corporation Active oil debris monitor particle detection and monitoring system
CN113029879A (en) * 2021-03-19 2021-06-25 爱德森（厦门）电子有限公司 Method for multi-frequency online monitoring of metal particles in oil and detection device thereof
US11137383B2 (en) 2019-02-18 2021-10-05 Raytheon Technologies Corporation Active oil debris monitor particle detection and monitoring system
US11536706B2 (en) 2019-02-18 2022-12-27 Raytheon Technologies Corporation Active oil debris monitor phase angle calculation and monitoring system
CN117629041A (en) * 2023-11-27 2024-03-01 山西科达自控股份有限公司 Oil abrasive particle size detection method and system

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101650706A (en) * 2009-06-30 2010-02-17 重庆重邮信科通信技术有限公司 Method and device for calculating FFT branch

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101650706A (en) * 2009-06-30 2010-02-17 重庆重邮信科通信技术有限公司 Method and device for calculating FFT branch

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
胡祥胜: "基于DSP的油路磨粒检测***信号处理模块设计", 《中国硕士学位论文全文数据库》, 15 January 2013 (2013-01-15), pages 32 - 47 *
胡祥胜等: "基于DSP的油路磨粒检测***的设计", 《仪表技术与传感器》, no. 3, 15 March 2012 (2012-03-15) *

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104390893A (en) * 2014-11-24 2015-03-04 电子科技大学 Oil mill grain sensor based on printed coil
CN105547932A (en) * 2015-12-08 2016-05-04 中国航空工业集团公司北京长城航空测控技术研究所 Engine lubricating oil metal particle on-line monitoring signal demodulation method
CN105547932B (en) * 2015-12-08 2018-06-01 中国航空工业集团公司北京长城航空测控技术研究所 A kind of signal demodulator method of motor oil metallic particles on-line monitoring
CN105823711A (en) * 2016-04-18 2016-08-03 上海电机学院 Online monitoring method for oil abrasive particles
CN109154586A (en) * 2016-04-27 2019-01-04 株式会社岛津制作所 Data processing method and device
CN109154586B (en) * 2016-04-27 2020-10-16 株式会社岛津制作所 Data processing method and device
CN109648478A (en) * 2019-01-30 2019-04-19 华侨大学 A kind of method and device thereof for monitoring wheel passivatio state on-line
CN109648478B (en) * 2019-01-30 2024-02-23 华侨大学 Method and device for on-line monitoring grinding wheel grinding state
US10823700B2 (en) 2019-02-15 2020-11-03 Raytheon Technologies Corporation Active oil debris monitor particle detection and monitoring system
US11137383B2 (en) 2019-02-18 2021-10-05 Raytheon Technologies Corporation Active oil debris monitor particle detection and monitoring system
US11536706B2 (en) 2019-02-18 2022-12-27 Raytheon Technologies Corporation Active oil debris monitor phase angle calculation and monitoring system
CN109839334B (en) * 2019-03-12 2020-08-14 西安交通大学 Signal identification method for single-coil magnetic induction type abrasive particle detection sensor
CN109839334A (en) * 2019-03-12 2019-06-04 西安交通大学 Signal recognition method towards unicoil magnetic inductive abrasive grain detection sensor
CN111504859B (en) * 2020-04-28 2021-04-06 沈阳顺义科技有限公司 System and method for online monitoring and evaluation of lubricating oil abrasive particles
CN111504859A (en) * 2020-04-28 2020-08-07 沈阳顺义科技有限公司 System and method for online monitoring and evaluation of lubricating oil abrasive particles
CN113029879A (en) * 2021-03-19 2021-06-25 爱德森（厦门）电子有限公司 Method for multi-frequency online monitoring of metal particles in oil and detection device thereof
CN113029879B (en) * 2021-03-19 2024-05-24 爱德森（厦门）电子有限公司 Method for multi-frequency on-line monitoring of metal particles in oil and detection device thereof
CN117629041A (en) * 2023-11-27 2024-03-01 山西科达自控股份有限公司 Oil abrasive particle size detection method and system

Also Published As

Publication number Publication date
CN103217366B (en) 2015-06-24

Similar Documents

Publication Publication Date Title
CN103217366B (en) Method for processing online oil path grain signal
US10698106B2 (en) Obstacle detecting method and apparatus, device and storage medium
CN103926276B (en) A kind of online oil liquid abrasive grain monitoring device and detection method
CN103217365A (en) Online oil way abrasive particle monitoring device
CN105738061B (en) A kind of image analysis method of vibration signal
CN103196989A (en) ACFM different-angle crack detection system based on rotating magnetic field
CN104408927B (en) Model classification method based on frequency modulation continuous wave radar
CN102109451A (en) Non-contact conductive gas/liquid two-phase flow pattern identifying device and method
CN103345842A (en) Road vehicle classification system and method
CN104299417A (en) Vehicle identification method based on waveform detection
CN102230872A (en) Device and method for on-line testing of grade of magnetic iron in flowing ore pulp
CN104281738A (en) Evaluation system and method for trunk line coordination control scheme of trunk line road
Wang et al. Analysis and experiments of vehicle detection with magnetic sensors in urban environments
CN103997315A (en) Self-adaptive filtering device and method for pipeline magnetic leakage signals
CA2418938A1 (en) Method and device for classifying vehicles
CN102944773A (en) Method for detecting and classifying power disturbances based on space conversion
CN105092691A (en) Quantitative detection method and detector for detecting accumulated oxide scales in pipelines
CN108896454A (en) A kind of multichannel wear particle detection method and device based on time-division multiplex technology
CN107749174A (en) Parking detection method, device, equipment and the storage medium of anti-subway interference
WO2016095557A1 (en) Sampling method and sampling apparatus for anti-counterfeiting information about cash note
CN106841830B (en) Early-warning for high pressure method, apparatus and system based on electric field intensity signal detection
CN202793421U (en) Signal processing device for passive wheel sensor
CN102945360A (en) Method for detecting headwind zone of Doppler radar image based on morphological and logic operations
CN206975199U (en) A kind of railway system&#39;s low-frequency AC magnetic fields test system
CN109164342A (en) A kind of diagnostic method, device and the electronic equipment of three-phase inverter open-circuit fault

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20150624

Termination date: 20190407