CN106181432A - A kind of lathe turnover plate type workbench with defect recognition function - Google Patents

Abstract

A kind of lathe turnover plate type workbench with defect recognition function of the present invention, including lathe turnover plate type workbench and the cannot-harm-detection device of being connected with lathe turnover plate type workbench, it is characterized in that, described lathe turnover plate type workbench includes: reverse turning bed workbench bracing frame, reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench be may be rotatably mounted on reverse turning bed workbench bracing frame by reverse turning bed workbench turning-over support rotating shaft；Screw pair connects with reverse turning bed workbench bracing frame, reverse turning bed workbench respectively, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；The vertically oriented device of reverse turning bed workbench it is provided with on reverse turning bed workbench, on reverse turning bed workbench bracing frame, device corresponding position vertically oriented with reverse turning bed workbench is provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench bracing frame.The present invention takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves work efficiency.

Description

A kind of lathe turnover plate type workbench with defect recognition function

Technical field

The present invention relates to machine tool field, be specifically related to a kind of lathe turnover plate type workbench with defect recognition function.

Background technology

Lathe refers to manufacture the machine of machine, also known as machine-tool or toolroom machine, is called for short lathe traditionally.It is generally divided into gold Belong to stock-removing machine, forging machine tool and wood craft machines etc..In modern mechanical manufacturing, the method for processing equipment part is a lot: except cutting adds Outside work, also cast, forge, weld, punching press, extruding etc., as long as belonging to, required precision is higher and surface roughness requirements is thinner Part, the most all finally need to process by the method for cutting on lathe.Lathe rises in the construction of national economic modernization Significant role.

Non-Destructive Testing is premised on not destroying measurand internal structure and Practical Performance, to measurand inside or table The physical property in face, state characteristic detect.Electromagnetic nondestructive is changed to basis for estimation with material electromagnetic performance, comes material Material and component implement defects detection and performance test.In correlation technique, although Pulsed eddy current testing technology has obtained deep grinding Study carefully and quickly develop, but still suffer from that testing result is not accurate enough, information excavating is the most deep enough, testing result is had The problems such as effect classification.

Summary of the invention

For the problems referred to above, the present invention provides a kind of lathe turnover plate type workbench with defect recognition function.

The purpose of the present invention realizes by the following technical solutions:

A kind of lathe turnover plate type workbench with defect recognition function, including lathe turnover plate type workbench and turns over lathe The cannot-harm-detection device that board-like workbench is connected, is characterized in that, described lathe turnover plate type workbench includes: reverse turning bed workbench props up Support, reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench is rotatable by reverse turning bed workbench turning-over support rotating shaft It is arranged on reverse turning bed workbench bracing frame；Screw pair connects with reverse turning bed workbench bracing frame, reverse turning bed workbench respectively, turnover panel Workbench is driven by screw pair, and nut rotates, guide screw movement；Reverse turning bed workbench it is provided with vertically fixed on reverse turning bed workbench Position device, on reverse turning bed workbench bracing frame, device corresponding position vertically oriented with reverse turning bed workbench is provided with rotation binding clasp and safety Bolt, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench bracing frame.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

Preferably, described leading screw end is provided with buffering disc spring.

The invention have the benefit that the present invention and can be accurately positioned to vertical state by level upset, workpiece can Positioning and clamping process is completed under level.It is particularly suitable for high-speed and high-efficiency milling class lathe, takes into account vertical processing part dress The convenient advantage with horizontal exfoliation chip removal easy to process of folder, improves work efficiency.

Accompanying drawing explanation

The invention will be further described to utilize accompanying drawing, but the embodiment in accompanying drawing does not constitute any limit to the present invention System, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to obtain according to the following drawings Other accompanying drawing.

Fig. 1 is the schematic diagram of lathe turnover plate type workbench of the present invention.

Fig. 2 is the schematic diagram of the cannot-harm-detection device of the present invention.

Reference:

Based on temporal signatures detection module 1, carry based on frequency domain character detection module 2, comprehensive detection module 3, temporal signatures Take submodule 11, defects detection submodule 12 based on time domain, pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24, defects detection submodule 25 based on frequency domain.

Detailed description of the invention

The invention will be further described with the following Examples.

Application scenarios 1

Seeing Fig. 1, Fig. 2, a kind of lathe turnover plate type workbench with defect recognition function of the present embodiment, including lathe Turnover plate type workbench and the cannot-harm-detection device being connected with lathe turnover plate type workbench, is characterized in that, described lathe turnover plate type Workbench includes: reverse turning bed workbench bracing frame, and reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench passes through turnover panel work Station turning-over support rotating shaft may be rotatably mounted on reverse turning bed workbench bracing frame；Screw pair supports with reverse turning bed workbench respectively Frame, reverse turning bed workbench connect, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；On reverse turning bed workbench It is provided with the vertically oriented device of reverse turning bed workbench, device corresponding position vertically oriented with reverse turning bed workbench on reverse turning bed workbench bracing frame Being provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench and props up On support.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

This preferred embodiment takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves Work efficiency.

Preferably, described leading screw end is provided with buffering disc spring.

It is more stable that this preferred embodiment makes lathe run.

Preferably, the cannot-harm-detection device includes based on temporal signatures detection module 1, based on frequency domain character detection module 2 and Comprehensive detection module 3, particularly as follows:

(1) based on temporal signatures detection module 1, it includes that temporal signatures extracts submodule 11, defect based on time domain inspection Survey submodule 12；Described temporal signatures extracts submodule 11 for using the temporal signatures extracting method of improvement to extract temporal signatures Value；Described defects detection submodule based on time domain 12 is for using the automatic classifying identification method of improvement to lathe turnover plate type work Station defect carries out detection and identifies, to obtain testing result S based on time domain₁；

(2) based on frequency domain character detection module 2, it includes pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24 and defects detection submodule 25 based on frequency domain；Described pretreatment submodule Block 21, for defect area time domain response and reference zone time domain response are carried out fast Fourier transform, obtains defect area frequency Domain response and reference zone frequency domain response, and defect area frequency domain response and reference zone frequency domain response are normalized respectively Carry out difference processing after process again, calculate difference frequency domain response；Described normalized submodule 22 is for difference frequency domain Response is normalized, and then obtains difference normalization frequency domain response；Described frequency domain optimizes submodule 23 for according to collection Skin effect selects the frequency being suitable to detect lathe turnover plate type workbench defect, and returns difference based on the frequency chosen One changes frequency domain response is optimized process；Described frequency domain character extracts the submodule 24 difference normalization frequency after extracting optimization The differential peak spectrum of domain response, characteristic frequency differential amplitude spectrum and difference cross zero frequency as can be used for characterizing lathe turnover plate type work The frequency domain character value of station Material Physics attribute；Described defects detection submodule based on frequency domain 25 is for using the automatic of improvement Classifying identification method carries out detection and identifies, to obtain testing result S based on frequency domain lathe turnover plate type workbench defect₂；

(3) comprehensive detection module 3, for according to testing result S based on time domain₁With testing result S based on frequency domain₂, adopt The defect type of tested lathe turnover plate type workbench it is determined with predetermined defect classifying identification rule.

This preferred embodiment, by the way of temporal signatures detection and frequency domain character detection combine, effectively inhibits lift-off Interference, it is achieved that the accurate detection of lathe turnover plate type workbench defect.

Preferably, described temporal signatures extracting method based on improvement extracts temporal signatures value, including:

(1) use pulse eddy current sensor that lathe turnover plate type workbench defect is detected, adjust impulse eddy current sensing Lift-off distance between device and tested lathe turnover plate type table surface, it is thus achieved that defect area time domain response q (t), chooses tested The time domain response at lathe turnover plate type workbench zero defect position is as reference zone time domain response c (t)；

(2) defect area time domain response q (t) and reference zone time domain response c (t) are carried out difference and normalized, Obtaining difference normalization time domain response S (t), definition process formula is:

$S\left(t\right)=\frac{q\left(t\right)}{{\mathrm{\ξ}}_{1}max\left(q\right(t\left)\right)}-\frac{c\left(t\right)}{{\mathrm{\ξ}}_{2}max\left(c\right(t\left)\right)}$

In formula, ξ₁、ξ₂For the coefficient adjustment factor set, ξ₁、ξ₂Span be [0.9,1.1]；

(3) extract difference normalization time domain response S (t) the differential peak time and difference zero-crossing timing as can be used for table Levy the temporal signatures value of lathe turnover plate type workbench Material Physics attribute.

The automatic classifying identification method of described improvement carries out detection and identifies lathe turnover plate type workbench defect, including:

(1) select gaussian radial basis function kernel function (RBF) as Kernel function, the expression of described gaussian radial basis function kernel function Formula is K (x, y)=exp{-γ ‖ x-y ‖², use particle swarm optimization algorithm that parameter γ of RBF function is optimized；

(2) perform training algorithm, use training data to obtain support vector cassification model；

(3) training data is tested, unknown lathe turnover plate type workbench defect is predicted.

Described predetermined defect classifying identification rule is: use weighted mean method to testing result S based on time domain₁With based on Testing result S of frequency domain₂Process, obtain final detection result, by final detection result with in data base correspond to different The calibration result of degree of impairment compares, and selects the calibration result corresponding with final detection result, according to the damage pre-build Mapping relations between traumatic condition condition and calibration result, obtain the degree of impairment corresponding with described calibration result, and then determine tested The defect type of lathe turnover plate type workbench.

Described degree of impairment includes equivalent size, depth of defect.

Described based on the frequency chosen, difference normalization frequency domain response is optimized process, including:

(1) according to the impulse eddy current response signal data structure data matrix D of multiple lathe turnover plate type workbench defects:

In formula, d_ijExpression i-th defect impulse eddy current response signal value at the frequency that jth is chosen, i=1, 2 ..., p, j=1,2 ..., q；

(2) it is standardized each impulse eddy current response signal value in data matrix D processing, the arteries and veins after definition standardization Rush eddy current response signal value d_ij' computing formula be:

${d_{ij}}^{\′}=\frac{2d_{ij}-\stackrel{\‾}{d_j}-\stackrel{\‾}{d_i}}{\sqrt{\frac{1}{p-1}{\Σ}_{i=1}^p{(d_{ij}-\stackrel{\‾}{d_j})}^2}+\sqrt{\frac{1}{q-1}{\Σ}_{j=1}^q{(d_{ij}-\stackrel{\‾}{d_i})}^2}},(i=1,2,\mathrm{...},p,j=1,2,\mathrm{...},q)$

In formula,

Then the impulse eddy current response at the frequency that jth is chosen of p defect constitutes vector and is:

d_j=(d_1j′,_2j′,…,_pj′)^T

(3) calculate the response of each impulse eddy current and constitute vector d₁,d₂,…,d_qCorrelation coefficient square at q the frequency chosen Battle array C:

In formula, c_mn(m=1,2 ..., q, n=1,2 ..., q) it is d_mAnd d_nCorrelation coefficient；

(4) k optimization frequency N is determined_rReflect the resultant effect of q the frequency chosen, r=1,2 ..., k, k ＜ q, excellent Change frequency matrix to be represented by:

$\left\{\begin{array}{c}{N}_1=h_{11}{d}_1+h_{12}{d}_2+...h_{11}{d}_q\\ N_2=h_{21}{d}_1+h_{22}{d}_2+...h_{2q}{d}_q\\ \mathrm{.........}\\ N_k=h_{k1}{d}_1+h_{k2}{d}_2+...h_{kq}{d}_q\end{array}\right.$

In formula, h_rjRepresenting q the frequency chosen weight coefficient on optimization frequency, weighting coefficient matrix H is expressed as:

Weight coefficient h_rjCalculation be:

1) to characteristic equation ,=0 solves | λ E-C |, asks for each eigenvalue λ_j(j=1,2 ..., q), by each eigenvalue λ_j (j=1,2 ..., q) arrange according to descending order, λ₁＞ λ₂＞ ... ＞ λ_q, and ask for eigenvalue λ_j(j=1,2 ..., Q) character pair vector e_j, it is desirable to | | e_j| |=1, i.e.

2) the r optimization frequency N is defined_rContribution rate G to resultant effect_r:

$G_r=\frac{{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j},(r=1,2,...,k)$

3) k optimization frequency N is calculated_rContribution rate of accumulative total L:

$L=\frac{{\Σ}_{r=1}^k{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j}$

K is the minima meeting L-90% ＞ 0；

4) weight coefficient is calculated:(r=1,2 ..., m；S, j=1,2 ..., q).

Detection data are standardized processing by this preferred embodiment, facilitate different characteristic value to carry out linear combination, improve Calculate speed；Optimized by frequency domain, improve detection efficiency；Comprehensive detection module 3 is set, it is possible to farthest reduce detection by mistake Difference, and improve the Classification and Identification rate to lathe turnover plate type workbench defect, it is simple to follow-up study and the problem of solution, improve product Quality.

This application scene adjusts the lift-off distance between pulse eddy current sensor and tested lathe turnover plate type table surface For 0.4mm, set coefficient adjustment factor ξ₁=0.9, ξ₂=0.9, the Classification and Identification rate of lathe turnover plate type workbench defect is improved 5%.

Application scenarios 2

Seeing Fig. 1, Fig. 2, a kind of lathe turnover plate type workbench with defect recognition function of the present embodiment, including lathe Turnover plate type workbench and the cannot-harm-detection device being connected with lathe turnover plate type workbench, is characterized in that, described lathe turnover plate type Workbench includes: reverse turning bed workbench bracing frame, and reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench passes through turnover panel work Station turning-over support rotating shaft may be rotatably mounted on reverse turning bed workbench bracing frame；Screw pair supports with reverse turning bed workbench respectively Frame, reverse turning bed workbench connect, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；On reverse turning bed workbench It is provided with the vertically oriented device of reverse turning bed workbench, device corresponding position vertically oriented with reverse turning bed workbench on reverse turning bed workbench bracing frame Being provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench and props up On support.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

This preferred embodiment takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves Work efficiency.

Preferably, described leading screw end is provided with buffering disc spring.

It is more stable that this preferred embodiment makes lathe run.

Preferably, the cannot-harm-detection device includes based on temporal signatures detection module 1, based on frequency domain character detection module 2 and Comprehensive detection module 3, particularly as follows:

(1) based on temporal signatures detection module 1, it includes that temporal signatures extracts submodule 11, defect based on time domain inspection Survey submodule 12；Described temporal signatures extracts submodule 11 for using the temporal signatures extracting method of improvement to extract temporal signatures Value；Described defects detection submodule based on time domain 12 is for using the automatic classifying identification method of improvement to lathe turnover plate type work Station defect carries out detection and identifies, to obtain testing result S based on time domain₁；

(2) based on frequency domain character detection module 2, it includes pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24 and defects detection submodule 25 based on frequency domain；Described pretreatment submodule Block 21, for defect area time domain response and reference zone time domain response are carried out fast Fourier transform, obtains defect area frequency Domain response and reference zone frequency domain response, and defect area frequency domain response and reference zone frequency domain response are normalized respectively Carry out difference processing after process again, calculate difference frequency domain response；Described normalized submodule 22 is for difference frequency domain Response is normalized, and then obtains difference normalization frequency domain response；Described frequency domain optimizes submodule 23 for according to collection Skin effect selects the frequency being suitable to detect lathe turnover plate type workbench defect, and returns difference based on the frequency chosen One changes frequency domain response is optimized process；Described frequency domain character extracts the submodule 24 difference normalization frequency after extracting optimization The differential peak spectrum of domain response, characteristic frequency differential amplitude spectrum and difference cross zero frequency as can be used for characterizing lathe turnover plate type work The frequency domain character value of station Material Physics attribute；Described defects detection submodule based on frequency domain 25 is for using the automatic of improvement Classifying identification method carries out detection and identifies, to obtain testing result S based on frequency domain lathe turnover plate type workbench defect₂；

(3) comprehensive detection module 3, for according to testing result S based on time domain₁With testing result S based on frequency domain₂, adopt The defect type of tested lathe turnover plate type workbench it is determined with predetermined defect classifying identification rule.

This preferred embodiment, by the way of temporal signatures detection and frequency domain character detection combine, effectively inhibits lift-off Interference, it is achieved that the accurate detection of lathe turnover plate type workbench defect.

Preferably, described temporal signatures extracting method based on improvement extracts temporal signatures value, including:

(1) use pulse eddy current sensor that lathe turnover plate type workbench defect is detected, adjust impulse eddy current sensing Lift-off distance between device and tested lathe turnover plate type table surface, it is thus achieved that defect area time domain response q (t), chooses tested The time domain response at lathe turnover plate type workbench zero defect position is as reference zone time domain response c (t)；

(2) defect area time domain response q (t) and reference zone time domain response c (t) are carried out difference and normalized, Obtaining difference normalization time domain response S (t), definition process formula is:

$S\left(t\right)=\frac{q\left(t\right)}{{\mathrm{\ξ}}_{1}max\left(q\right(t\left)\right)}-\frac{c\left(t\right)}{{\mathrm{\ξ}}_{2}max\left(c\right(t\left)\right)}$

In formula, ξ₁、ξ₂For the coefficient adjustment factor set, ξ₁、ξ₂Span be [0.9,1.1]；

(3) extract difference normalization time domain response S (t) the differential peak time and difference zero-crossing timing as can be used for table Levy the temporal signatures value of lathe turnover plate type workbench Material Physics attribute.

The automatic classifying identification method of described improvement carries out detection and identifies lathe turnover plate type workbench defect, including:

(1) select gaussian radial basis function kernel function (RBF) as Kernel function, the expression of described gaussian radial basis function kernel function Formula is K (x, y)=exp{-γ ‖ x-y ‖², use particle swarm optimization algorithm that parameter γ of RBF function is optimized；

(2) perform training algorithm, use training data to obtain support vector cassification model；

(3) training data is tested, unknown lathe turnover plate type workbench defect is predicted.

Described predetermined defect classifying identification rule is: use weighted mean method to testing result S based on time domain₁With based on Testing result S of frequency domain₂Process, obtain final detection result, by final detection result with in data base correspond to different The calibration result of degree of impairment compares, and selects the calibration result corresponding with final detection result, according to the damage pre-build Mapping relations between traumatic condition condition and calibration result, obtain the degree of impairment corresponding with described calibration result, and then determine tested The defect type of lathe turnover plate type workbench.

Described degree of impairment includes equivalent size, depth of defect.

Described based on the frequency chosen, difference normalization frequency domain response is optimized process, including:

(1) according to the impulse eddy current response signal data structure data matrix D of multiple lathe turnover plate type workbench defects:

In formula, d_ijExpression i-th defect impulse eddy current response signal value at the frequency that jth is chosen, i=1, 2 ..., p, j=1,2 ..., q；

(2) it is standardized each impulse eddy current response signal value in data matrix D processing, the arteries and veins after definition standardization Rush eddy current response signal value d_ij' computing formula be:

${d_{ij}}^{\′}=\frac{2d_{ij}-\stackrel{\‾}{d_j}-\stackrel{\‾}{d_i}}{\sqrt{\frac{1}{p-1}{\Σ}_{i=1}^p{(d_{ij}-\stackrel{\‾}{d_j})}^2}+\sqrt{\frac{1}{q-1}{\Σ}_{j=1}^q{(d_{ij}-\stackrel{\‾}{d_i})}^2}},(i=1,2,\mathrm{...},p,j=1,2,\mathrm{...},q)$

In formula,

Then the impulse eddy current response at the frequency that jth is chosen of p defect constitutes vector and is:

d_j=(d_1j′,_2j′,…,_pj′)^T

(3) calculate the response of each impulse eddy current and constitute vector d₁,d₂,…,d_qCorrelation coefficient square at q the frequency chosen Battle array C:

In formula, c_mn(m=1,2 ..., q, n=1,2 ..., q) it is d_mAnd d_nCorrelation coefficient；

(4) k optimization frequency N is determined_rReflect the resultant effect of q the frequency chosen, r=1,2 ..., k, k ＜ q, excellent Change frequency matrix to be represented by:

$\left\{\begin{array}{c}{N}_1=h_{11}{d}_1+h_{12}{d}_2+...h_{11}{d}_q\\ N_2=h_{21}{d}_1+h_{22}{d}_2+...h_{2q}{d}_q\\ \mathrm{.........}\\ N_k=h_{k1}{d}_1+h_{k2}{d}_2+...h_{kq}{d}_q\end{array}\right.$

In formula, h_rjRepresenting q the frequency chosen weight coefficient on optimization frequency, weighting coefficient matrix H is expressed as:

Weight coefficient h_rjCalculation be:

1) to characteristic equation ,=0 solves | λ E-C |, asks for each eigenvalue λ_j(j=1,2 ..., q), by each eigenvalue λ_j (j=1,2 ..., q) arrange according to descending order, λ₁＞ λ₂＞ ... ＞ λ_q, and ask for eigenvalue λ_j(j=1,2 ..., Q) character pair vector e_j, it is desirable to | | e_j| |=1, i.e.

2) the r optimization frequency N is defined_rContribution rate G to resultant effect_r:

$G_r=\frac{{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j},(r=1,2,...,k)$

3) k optimization frequency N is calculated_rContribution rate of accumulative total L:

$L=\frac{{\Σ}_{r=1}^k{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j}$

K is the minima meeting L-90% ＞ 0；

4) weight coefficient is calculated:(r=1,2 ..., m；S, j=1,2 ..., q).

Detection data are standardized processing by this preferred embodiment, facilitate different characteristic value to carry out linear combination, improve Calculate speed；Optimized by frequency domain, improve detection efficiency；Comprehensive detection module 3 is set, it is possible to farthest reduce detection by mistake Difference, and improve the Classification and Identification rate to lathe turnover plate type workbench defect, it is simple to follow-up study and the problem of solution, improve product Quality.

This application scene adjusts the lift-off distance between pulse eddy current sensor and tested lathe turnover plate type table surface For 0.6mm, set coefficient adjustment factor ξ₁=1.1, ξ₂=1.1, the Classification and Identification rate of lathe turnover plate type workbench defect is improved 4%.

Application scenarios 3

Seeing Fig. 1, Fig. 2, a kind of lathe turnover plate type workbench with defect recognition function of the present embodiment, including lathe Turnover plate type workbench and the cannot-harm-detection device being connected with lathe turnover plate type workbench, is characterized in that, described lathe turnover plate type Workbench includes: reverse turning bed workbench bracing frame, and reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench passes through turnover panel work Station turning-over support rotating shaft may be rotatably mounted on reverse turning bed workbench bracing frame；Screw pair supports with reverse turning bed workbench respectively Frame, reverse turning bed workbench connect, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；On reverse turning bed workbench It is provided with the vertically oriented device of reverse turning bed workbench, device corresponding position vertically oriented with reverse turning bed workbench on reverse turning bed workbench bracing frame Being provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench and props up On support.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

This preferred embodiment takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves Work efficiency.

Preferably, described leading screw end is provided with buffering disc spring.

It is more stable that this preferred embodiment makes lathe run.

Preferably, the cannot-harm-detection device includes based on temporal signatures detection module 1, based on frequency domain character detection module 2 and Comprehensive detection module 3, particularly as follows:

(1) based on temporal signatures detection module 1, it includes that temporal signatures extracts submodule 11, defect based on time domain inspection Survey submodule 12；Described temporal signatures extracts submodule 11 for using the temporal signatures extracting method of improvement to extract temporal signatures Value；Described defects detection submodule based on time domain 12 is for using the automatic classifying identification method of improvement to lathe turnover plate type work Station defect carries out detection and identifies, to obtain testing result S based on time domain₁；

(2) based on frequency domain character detection module 2, it includes pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24 and defects detection submodule 25 based on frequency domain；Described pretreatment submodule Block 21, for defect area time domain response and reference zone time domain response are carried out fast Fourier transform, obtains defect area frequency Domain response and reference zone frequency domain response, and defect area frequency domain response and reference zone frequency domain response are normalized respectively Carry out difference processing after process again, calculate difference frequency domain response；Described normalized submodule 22 is for difference frequency domain Response is normalized, and then obtains difference normalization frequency domain response；Described frequency domain optimizes submodule 23 for according to collection Skin effect selects the frequency being suitable to detect lathe turnover plate type workbench defect, and returns difference based on the frequency chosen One changes frequency domain response is optimized process；Described frequency domain character extracts the submodule 24 difference normalization frequency after extracting optimization The differential peak spectrum of domain response, characteristic frequency differential amplitude spectrum and difference cross zero frequency as can be used for characterizing lathe turnover plate type work The frequency domain character value of station Material Physics attribute；Described defects detection submodule based on frequency domain 25 is for using the automatic of improvement Classifying identification method carries out detection and identifies, to obtain testing result S based on frequency domain lathe turnover plate type workbench defect₂；

(3) comprehensive detection module 3, for according to testing result S based on time domain₁With testing result S based on frequency domain₂, adopt The defect type of tested lathe turnover plate type workbench it is determined with predetermined defect classifying identification rule.

This preferred embodiment, by the way of temporal signatures detection and frequency domain character detection combine, effectively inhibits lift-off Interference, it is achieved that the accurate detection of lathe turnover plate type workbench defect.

Preferably, described temporal signatures extracting method based on improvement extracts temporal signatures value, including:

(1) use pulse eddy current sensor that lathe turnover plate type workbench defect is detected, adjust impulse eddy current sensing Lift-off distance between device and tested lathe turnover plate type table surface, it is thus achieved that defect area time domain response q (t), chooses tested The time domain response at lathe turnover plate type workbench zero defect position is as reference zone time domain response c (t)；

(2) defect area time domain response q (t) and reference zone time domain response c (t) are carried out difference and normalized, Obtaining difference normalization time domain response S (t), definition process formula is:

$S\left(t\right)=\frac{q\left(t\right)}{{\mathrm{\ξ}}_{1}max\left(q\right(t\left)\right)}-\frac{c\left(t\right)}{{\mathrm{\ξ}}_{2}max\left(c\right(t\left)\right)}$

In formula, ξ₁、ξ₂For the coefficient adjustment factor set, ξ₁、ξ₂Span be [0.9,1.1]；

(3) extract difference normalization time domain response S (t) the differential peak time and difference zero-crossing timing as can be used for table Levy the temporal signatures value of lathe turnover plate type workbench Material Physics attribute.

The automatic classifying identification method of described improvement carries out detection and identifies lathe turnover plate type workbench defect, including:

(1) select gaussian radial basis function kernel function (RBF) as Kernel function, the expression of described gaussian radial basis function kernel function Formula is K (x, y)=exp{-γ ‖ x-y ‖², use particle swarm optimization algorithm that the parameter lambda of RBF function is optimized；

(2) perform training algorithm, use training data to obtain support vector cassification model；

(3) training data is tested, unknown lathe turnover plate type workbench defect is predicted.

Described predetermined defect classifying identification rule is: use weighted mean method to testing result S based on time domain₁With based on Testing result S of frequency domain₂Process, obtain final detection result, by final detection result with in data base correspond to different The calibration result of degree of impairment compares, and selects the calibration result corresponding with final detection result, according to the damage pre-build Mapping relations between traumatic condition condition and calibration result, obtain the degree of impairment corresponding with described calibration result, and then determine tested The defect type of lathe turnover plate type workbench.

Described degree of impairment includes equivalent size, depth of defect.

Described based on the frequency chosen, difference normalization frequency domain response is optimized process, including:

(1) according to the impulse eddy current response signal data structure data matrix D of multiple lathe turnover plate type workbench defects:

In formula, d_ijExpression i-th defect impulse eddy current response signal value at the frequency that jth is chosen, i=1, 2 ..., p, j=1,2 ..., q；

(2) it is standardized each impulse eddy current response signal value in data matrix D processing, the arteries and veins after definition standardization Rush eddy current response signal value d_ij' computing formula be:

${d_{ij}}^{\′}=\frac{2d_{ij}-\stackrel{\‾}{d_j}-\stackrel{\‾}{d_i}}{\sqrt{\frac{1}{p-1}{\Σ}_{i=1}^p{(d_{ij}-\stackrel{\‾}{d_j})}^2}+\sqrt{\frac{1}{q-1}{\Σ}_{j=1}^q{(d_{ij}-\stackrel{\‾}{d_i})}^2}},(i=1,2,\mathrm{...},p,j=1,2,\mathrm{...},q)$

In formula,

Then the impulse eddy current response at the frequency that jth is chosen of p defect constitutes vector and is:

d_j=(d_1j′,d_2j′,…,d_pj′)^T

(3) calculate the response of each impulse eddy current and constitute vector d₁,d₂,…,d_qCorrelation coefficient square at q the frequency chosen Battle array C:

In formula, c_mn(m=1,2 ..., q, n=1,2 ..., q) it is d_mAnd d_nCorrelation coefficient；

(4) k optimization frequency N is determined_rReflect the resultant effect of q the frequency chosen, r=1,2 ..., k, k ＜ q, excellent Change frequency matrix to be represented by:

$\left\{\begin{array}{c}{N}_1=h_{11}{d}_1+h_{12}{d}_2+...h_{11}{d}_q\\ N_2=h_{21}{d}_1+h_{22}{d}_2+...h_{2q}{d}_q\\ \mathrm{.........}\\ N_k=h_{k1}{d}_1+h_{k2}{d}_2+...h_{kq}{d}_q\end{array}\right.$

In formula, h_rjRepresenting q the frequency chosen weight coefficient on optimization frequency, weighting coefficient matrix H is expressed as:

Weight coefficient h_rjCalculation be:

1) to characteristic equation ,=0 solves | λ E-C |, asks for each eigenvalue λ_j(j=1,2 ..., q), by each eigenvalue λ_j (j=1,2 ..., q) arrange according to descending order, λ₁＞ λ₂＞ ... ＞ λ_q, and ask for eigenvalue λ_j(j=1,2 ..., Q) character pair vector e_j, it is desirable to | | e_j| |=1, i.e.

2) the r optimization frequency N is defined_rContribution rate G to resultant effect_r:

$G_r=\frac{{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j},(r=1,2,...,k)$

3) k optimization frequency N is calculated_rContribution rate of accumulative total L:

$L=\frac{{\Σ}_{r=1}^k{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j}$

K is the minima meeting L-90% ＞ 0；

4) weight coefficient is calculated:(r=1,2 ..., m；S, j=1,2 ..., q).

Detection data are standardized processing by this preferred embodiment, facilitate different characteristic value to carry out linear combination, improve Calculate speed；Optimized by frequency domain, improve detection efficiency；Comprehensive detection module 3 is set, it is possible to farthest reduce detection by mistake Difference, and improve the Classification and Identification rate to lathe turnover plate type workbench defect, it is simple to follow-up study and the problem of solution, improve product Quality.

This application scene adjusts the lift-off distance between pulse eddy current sensor and tested lathe turnover plate type table surface For 0.8mm, set coefficient adjustment factor ξ₁=0.9, ξ₂=1.1, the Classification and Identification rate of lathe turnover plate type workbench defect is improved 4.5%.

Application scenarios 4

Seeing Fig. 1, Fig. 2, a kind of lathe turnover plate type workbench with defect recognition function of the present embodiment, including lathe Turnover plate type workbench and the cannot-harm-detection device being connected with lathe turnover plate type workbench, is characterized in that, described lathe turnover plate type Workbench includes: reverse turning bed workbench bracing frame, and reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench passes through turnover panel work Station turning-over support rotating shaft may be rotatably mounted on reverse turning bed workbench bracing frame；Screw pair supports with reverse turning bed workbench respectively Frame, reverse turning bed workbench connect, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；On reverse turning bed workbench It is provided with the vertically oriented device of reverse turning bed workbench, device corresponding position vertically oriented with reverse turning bed workbench on reverse turning bed workbench bracing frame Being provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench and props up On support.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

This preferred embodiment takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves Work efficiency.

Preferably, described leading screw end is provided with buffering disc spring.

It is more stable that this preferred embodiment makes lathe run.

Preferably, the cannot-harm-detection device includes based on temporal signatures detection module 1, based on frequency domain character detection module 2 and Comprehensive detection module 3, particularly as follows:

(1) based on temporal signatures detection module 1, it includes that temporal signatures extracts submodule 11, defect based on time domain inspection Survey submodule 12；Described temporal signatures extracts submodule 11 for using the temporal signatures extracting method of improvement to extract temporal signatures Value；Described defects detection submodule based on time domain 12 is for using the automatic classifying identification method of improvement to lathe turnover plate type work Station defect carries out detection and identifies, to obtain testing result S based on time domain₁；

(2) based on frequency domain character detection module 2, it includes pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24 and defects detection submodule 25 based on frequency domain；Described pretreatment submodule Block 21, for defect area time domain response and reference zone time domain response are carried out fast Fourier transform, obtains defect area frequency Domain response and reference zone frequency domain response, and defect area frequency domain response and reference zone frequency domain response are normalized respectively Carry out difference processing after process again, calculate difference frequency domain response；Described normalized submodule 22 is for difference frequency domain Response is normalized, and then obtains difference normalization frequency domain response；Described frequency domain optimizes submodule 23 for according to collection Skin effect selects the frequency being suitable to detect lathe turnover plate type workbench defect, and returns difference based on the frequency chosen One changes frequency domain response is optimized process；Described frequency domain character extracts the submodule 24 difference normalization frequency after extracting optimization The differential peak spectrum of domain response, characteristic frequency differential amplitude spectrum and difference cross zero frequency as can be used for characterizing lathe turnover plate type work The frequency domain character value of station Material Physics attribute；Described defects detection submodule based on frequency domain 25 is for using the automatic of improvement Classifying identification method carries out detection and identifies, to obtain testing result S based on frequency domain lathe turnover plate type workbench defect₂；

(3) comprehensive detection module 3, for according to testing result S based on time domain₁With testing result S based on frequency domain₂, adopt The defect type of tested lathe turnover plate type workbench it is determined with predetermined defect classifying identification rule.

This preferred embodiment, by the way of temporal signatures detection and frequency domain character detection combine, effectively inhibits lift-off Interference, it is achieved that the accurate detection of lathe turnover plate type workbench defect.

Preferably, described temporal signatures extracting method based on improvement extracts temporal signatures value, including:

(1) use pulse eddy current sensor that lathe turnover plate type workbench defect is detected, adjust impulse eddy current sensing Lift-off distance between device and tested lathe turnover plate type table surface, it is thus achieved that defect area time domain response q (t), chooses tested The time domain response at lathe turnover plate type workbench zero defect position is as reference zone time domain response c (t)；

(2) defect area time domain response q (t) and reference zone time domain response c (t) are carried out difference and normalized, Obtaining difference normalization time domain response S (t), definition process formula is:

$S\left(t\right)=\frac{q\left(t\right)}{{\mathrm{\ξ}}_{1}max\left(q\right(t\left)\right)}-\frac{c\left(t\right)}{{\mathrm{\ξ}}_{2}max\left(c\right(t\left)\right)}$

In formula, ξ₁、ξ₂For the coefficient adjustment factor set, ξ₁、ξ₂Span be [0.9,1.1]；

(3) extract difference normalization time domain response S (t) the differential peak time and difference zero-crossing timing as can be used for table Levy the temporal signatures value of lathe turnover plate type workbench Material Physics attribute.

The automatic classifying identification method of described improvement carries out detection and identifies lathe turnover plate type workbench defect, including:

(1) select gaussian radial basis function kernel function (RBF) as Kernel function, the expression of described gaussian radial basis function kernel function Formula is K (x, y)=exp{-γ ‖ x-y ‖², use particle swarm optimization algorithm that parameter γ of RBF function is optimized；

(2) perform training algorithm, use training data to obtain support vector cassification model；

(3) training data is tested, unknown lathe turnover plate type workbench defect is predicted.

Described predetermined defect classifying identification rule is: use weighted mean method to testing result S based on time domain₁With based on Testing result S of frequency domain₂Process, obtain final detection result, by final detection result with in data base correspond to different The calibration result of degree of impairment compares, and selects the calibration result corresponding with final detection result, according to the damage pre-build Mapping relations between traumatic condition condition and calibration result, obtain the degree of impairment corresponding with described calibration result, and then determine tested The defect type of lathe turnover plate type workbench.

Described degree of impairment includes equivalent size, depth of defect.

Described based on the frequency chosen, difference normalization frequency domain response is optimized process, including:

(1) according to the impulse eddy current response signal data structure data matrix D of multiple lathe turnover plate type workbench defects:

In formula, d_ijExpression i-th defect impulse eddy current response signal value at the frequency that jth is chosen, i=1, 2 ..., p, j=1,2 ..., q；

(2) it is standardized each impulse eddy current response signal value in data matrix D processing, the arteries and veins after definition standardization Rush eddy current response signal value d_ij' computing formula be:

${d_{ij}}^{\′}=\frac{2d_{ij}-\stackrel{\‾}{d_j}-\stackrel{\‾}{d_i}}{\sqrt{\frac{1}{p-1}{\Σ}_{i=1}^p{(d_{ij}-\stackrel{\‾}{d_j})}^2}+\sqrt{\frac{1}{q-1}{\Σ}_{j=1}^q{(d_{ij}-\stackrel{\‾}{d_i})}^2}},(i=1,2,\mathrm{...},p,j=1,2,\mathrm{...},q)$

In formula,

Then the impulse eddy current response at the frequency that jth is chosen of p defect constitutes vector and is:

d_j=(d_1j′,d_2j′,…,d_pj′)^T

(3) calculate the response of each impulse eddy current and constitute vector d₁,d₂,…,d_qCorrelation coefficient square at q the frequency chosen Battle array C:

In formula, c_mn(m=1,2 ..., q, n=1,2 ..., q) it is d_mAnd d_nCorrelation coefficient；

(4) k optimization frequency N is determined_rReflect the resultant effect of q the frequency chosen, r=1,2 ..., k, k ＜ q, excellent Change frequency matrix to be represented by:

$\left\{\begin{array}{c}{N}_1=h_{11}{d}_1+h_{12}{d}_2+...h_{11}{d}_q\\ N_2=h_{21}{d}_1+h_{22}{d}_2+...h_{2q}{d}_q\\ \mathrm{.........}\\ N_k=h_{k1}{d}_1+h_{k2}{d}_2+...h_{kq}{d}_q\end{array}\right.$

In formula, h_rjRepresenting q the frequency chosen weight coefficient on optimization frequency, weighting coefficient matrix H is expressed as:

Weight coefficient h_rjCalculation be:

1) to characteristic equation ,=0 solves | λ E-C |, asks for each eigenvalue λ_j(j=1,2 ..., q), by each eigenvalue λ_j (j=1,2 ..., a) arrange according to descending order, λ₁＞ λ₂＞ ... ＞ λ_q, and ask for eigenvalue λ_j(j=1,2 ..., Q) character pair vector e_j, it is desirable to | | e_j| |=1, i.e.

2) the r optimization frequency N is defined_rContribution rate G to resultant effect_r:

$G_r=\frac{{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j},(r=1,2,...,k)$

3) k optimization frequency N is calculated_rContribution rate of accumulative total L:

$L=\frac{{\Σ}_{r=1}^k{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j}$

K is the minima meeting L-90% ＞ 0；

4) weight coefficient is calculated:(r=1,2 ..., m；S, j=1,2 ..., q).

Detection data are standardized processing by this preferred embodiment, facilitate different characteristic value to carry out linear combination, improve Calculate speed；Optimized by frequency domain, improve detection efficiency；Comprehensive detection module 3 is set, it is possible to farthest reduce detection by mistake Difference, and improve the Classification and Identification rate to lathe turnover plate type workbench defect, it is simple to follow-up study and the problem of solution, improve product Quality.

This application scene adjusts the lift-off distance between pulse eddy current sensor and tested lathe turnover plate type table surface For 1.0mm, set coefficient adjustment factor ξ₁=1.1, ξ₂=0.9, the Classification and Identification rate of lathe turnover plate type workbench defect is improved 5.6%.

Application scenarios 5

Seeing Fig. 1, Fig. 2, a kind of lathe turnover plate type workbench with defect recognition function of the present embodiment, including lathe Turnover plate type workbench and the cannot-harm-detection device being connected with lathe turnover plate type workbench, is characterized in that, described lathe turnover plate type Workbench includes: reverse turning bed workbench bracing frame, and reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench passes through turnover panel work Station turning-over support rotating shaft may be rotatably mounted on reverse turning bed workbench bracing frame；Screw pair supports with reverse turning bed workbench respectively Frame, reverse turning bed workbench connect, and reverse turning bed workbench is driven by screw pair, and nut rotates, guide screw movement；On reverse turning bed workbench It is provided with the vertically oriented device of reverse turning bed workbench, device corresponding position vertically oriented with reverse turning bed workbench on reverse turning bed workbench bracing frame Being provided with rotation binding clasp and safety cock, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench and props up On support.

Preferably, the vertically oriented device of reverse turning bed workbench described reverse turning bed workbench arranged is 4 and is separately positioned on and turns over The both sides of plate workbench upper and lower；Described rotation binding clasp is 4, respectively device pair vertically oriented with reverse turning bed workbench Should install；Described safety cock is 2, respectively device pair vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top Should install.

This preferred embodiment takes into account the convenient advantage with horizontal exfoliation chip removal easy to process of vertical processing parts fixation, improves Work efficiency.

Preferably, described leading screw end is provided with buffering disc spring.

It is more stable that this preferred embodiment makes lathe run.

Preferably, the cannot-harm-detection device includes based on temporal signatures detection module 1, based on frequency domain character detection module 2 and Comprehensive detection module 3, particularly as follows:

(1) based on temporal signatures detection module 1, it includes that temporal signatures extracts submodule 11, defect based on time domain inspection Survey submodule 12；Described temporal signatures extracts submodule 11 for using the temporal signatures extracting method of improvement to extract temporal signatures Value；Described defects detection submodule based on time domain 12 is for using the automatic classifying identification method of improvement to lathe turnover plate type work Station defect carries out detection and identifies, to obtain testing result S based on time domain₁；

(2) based on frequency domain character detection module 2, it includes pretreatment submodule 21, normalized submodule 22, frequency domain Optimize submodule 23, frequency domain character extracts submodule 24 and defects detection submodule 25 based on frequency domain；Described pretreatment submodule Block 21, for defect area time domain response and reference zone time domain response are carried out fast Fourier transform, obtains defect area frequency Domain response and reference zone frequency domain response, and defect area frequency domain response and reference zone frequency domain response are normalized respectively Carry out difference processing after process again, calculate difference frequency domain response；Described normalized submodule 22 is for difference frequency domain Response is normalized, and then obtains difference normalization frequency domain response；Described frequency domain optimizes submodule 23 for according to collection Skin effect selects the frequency being suitable to detect lathe turnover plate type workbench defect, and returns difference based on the frequency chosen One changes frequency domain response is optimized process；Described frequency domain character extracts the submodule 24 difference normalization frequency after extracting optimization The differential peak spectrum of domain response, characteristic frequency differential amplitude spectrum and difference cross zero frequency as can be used for characterizing lathe turnover plate type work The frequency domain character value of station Material Physics attribute；Described defects detection submodule based on frequency domain 25 is for using the automatic of improvement Classifying identification method carries out detection and identifies, to obtain testing result S based on frequency domain lathe turnover plate type workbench defect₂；

(3) comprehensive detection module 3, for according to testing result S based on time domain₁With testing result S based on frequency domain₂, adopt The defect type of tested lathe turnover plate type workbench it is determined with predetermined defect classifying identification rule.

This preferred embodiment, by the way of temporal signatures detection and frequency domain character detection combine, effectively inhibits lift-off Interference, it is achieved that the accurate detection of lathe turnover plate type workbench defect.

Preferably, described temporal signatures extracting method based on improvement extracts temporal signatures value, including:

(1) use pulse eddy current sensor that lathe turnover plate type workbench defect is detected, adjust impulse eddy current sensing Lift-off distance between device and tested lathe turnover plate type table surface, it is thus achieved that defect area time domain response q (t), chooses tested The time domain response at lathe turnover plate type workbench zero defect position is as reference zone time domain response c (t)；

(2) defect area time domain response q (t) and reference zone time domain response c (t) are carried out difference and normalized, Obtaining difference normalization time domain response S (t), definition process formula is:

$S\left(t\right)=\frac{q\left(t\right)}{{\mathrm{\ξ}}_{1}max\left(q\right(t\left)\right)}-\frac{c\left(t\right)}{{\mathrm{\ξ}}_{2}max\left(c\right(t\left)\right)}$

In formula, ξ₁、ξ₂For the coefficient adjustment factor set, ξ₁、ξ₂Span be [0.9,1.1]；

(3) extract difference normalization time domain response S (t) the differential peak time and difference zero-crossing timing as can be used for table Levy the temporal signatures value of lathe turnover plate type workbench Material Physics attribute.

The automatic classifying identification method of described improvement carries out detection and identifies lathe turnover plate type workbench defect, including:

(1) select gaussian radial basis function kernel function (RBF) as Kernel function, the expression of described gaussian radial basis function kernel function Formula is K (x, y)=exp{-γ ‖ x-y ‖², use particle swarm optimization algorithm that parameter γ of RBF function is optimized；

(2) perform training algorithm, use training data to obtain support vector cassification model；

(3) training data is tested, unknown lathe turnover plate type workbench defect is predicted.

Described predetermined defect classifying identification rule is: use weighted mean method to testing result S based on time domain₁With based on Testing result S of frequency domain₂Process, obtain final detection result, by final detection result with in data base correspond to different The calibration result of degree of impairment compares, and selects the calibration result corresponding with final detection result, according to the damage pre-build Mapping relations between traumatic condition condition and calibration result, obtain the degree of impairment corresponding with described calibration result, and then determine tested The defect type of lathe turnover plate type workbench.

Described degree of impairment includes equivalent size, depth of defect.

Described based on the frequency chosen, difference normalization frequency domain response is optimized process, including:

(1) according to the impulse eddy current response signal data structure data matrix D of multiple lathe turnover plate type workbench defects:

In formula, d_ijExpression i-th defect impulse eddy current response signal value at the frequency that jth is chosen, i=1, 2 ..., p, j=1,2 ..., q；

(2) it is standardized each impulse eddy current response signal value in data matrix D processing, the arteries and veins after definition standardization Rush eddy current response signal value d_ij' computing formula be:

${d_{ij}}^{\′}=\frac{2d_{ij}-\stackrel{\‾}{d_j}-\stackrel{\‾}{d_i}}{\sqrt{\frac{1}{p-1}{\Σ}_{i=1}^p{(d_{ij}-\stackrel{\‾}{d_j})}^2}+\sqrt{\frac{1}{q-1}{\Σ}_{j=1}^q{(d_{ij}-\stackrel{\‾}{d_i})}^2}},(i=1,2,\mathrm{...},p,j=1,2,\mathrm{...},q)$

In formula,

Then the impulse eddy current response at the frequency that jth is chosen of p defect constitutes vector and is:

d_j=(d_1j′,d_2j′,…,d_pj′)^T

(3) calculate the response of each impulse eddy current and constitute vector d₁,d₂,…,d_qCorrelation coefficient square at q the frequency chosen Battle array C:

In formula, c_mn(m=1,2 ..., q, n=1,2 ..., q) it is d_mAnd d_nCorrelation coefficient；

(4) k optimization frequency N is determined_rReflect the resultant effect of q the frequency chosen, r=1,2 ..., k, k ＜ q, excellent Change frequency matrix to be represented by:

$\left\{\begin{array}{c}{N}_1=h_{11}{d}_1+h_{12}{d}_2+...h_{11}{d}_q\\ N_2=h_{21}{d}_1+h_{22}{d}_2+...h_{2q}{d}_q\\ \mathrm{.........}\\ N_k=h_{k1}{d}_1+h_{k2}{d}_2+...h_{kq}{d}_q\end{array}\right.$

In formula, h_rjRepresenting q the frequency chosen weight coefficient on optimization frequency, weighting coefficient matrix H is expressed as:

Weight coefficient h_rjCalculation be:

1) to characteristic equation ,=0 solves | λ E-C |, asks for each eigenvalue λ_j(j=1,2 ..., q), by each eigenvalue λ_j (j=1,2 ..., q) arrange according to descending order, λ₁＞ λ₂＞ ... ＞ λ_q, and ask for eigenvalue λ_j(j=1,2 ..., Q) character pair vector e_j, it is desirable to | | e_j| |=1, i.e.

2) the r optimization frequency N is defined_rContribution rate G to resultant effect_r:

$G_r=\frac{{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j},(r=1,2,...,k)$

3) k optimization frequency N is calculated_rContribution rate of accumulative total L:

$L=\frac{{\Σ}_{r=1}^k{\mathrm{\λ}}_r}{{\Σ}_{j=1}^q{\mathrm{\λ}}_j}$

K is the minima meeting L-90% ＞ 0；

4) weight coefficient is calculated:(r=1,2 ..., m；S, j=1,2 ..., q).

Detection data are standardized processing by this preferred embodiment, facilitate different characteristic value to carry out linear combination, improve Calculate speed；Optimized by frequency domain, improve detection efficiency；Comprehensive detection module 3 is set, it is possible to farthest reduce detection by mistake Difference, and improve the Classification and Identification rate to lathe turnover plate type workbench defect, it is simple to follow-up study and the problem of solution, improve product Quality.

This application scene adjusts the lift-off distance between pulse eddy current sensor and tested lathe turnover plate type table surface For 1.2mm, set coefficient adjustment factor ξ₁=1, ξ₂=1, the Classification and Identification rate of lathe turnover plate type workbench defect is improve 4%.

Last it should be noted that, above example is only in order to illustrate technical scheme, rather than the present invention is protected Protecting the restriction of scope, although having made to explain to the present invention with reference to preferred embodiment, those of ordinary skill in the art should Work as understanding, technical scheme can be modified or equivalent, without deviating from the reality of technical solution of the present invention Matter and scope.

Claims (3)

1. there is a lathe turnover plate type workbench for defect recognition function, including lathe turnover plate type workbench and with lathe turnover panel The cannot-harm-detection device that formula workbench is connected, is characterized in that, described lathe turnover plate type workbench includes: reverse turning bed workbench supports Frame, reverse turning bed workbench bracing frame is fixed on ground, and reverse turning bed workbench is rotatably pacified by reverse turning bed workbench turning-over support rotating shaft It is contained on reverse turning bed workbench bracing frame；Screw pair connects with reverse turning bed workbench bracing frame, reverse turning bed workbench respectively, turnover panel work Station is driven by screw pair, and nut rotates, guide screw movement；Reverse turning bed workbench it is provided with vertically oriented on reverse turning bed workbench Device, on reverse turning bed workbench bracing frame, device corresponding position vertically oriented with reverse turning bed workbench is provided with rotation binding clasp and safety Bolt, safety cock is connected with safety cock cylinder, and safety cock cylinder is arranged on reverse turning bed workbench bracing frame.

A kind of lathe turnover plate type workbench with defect recognition function the most according to claim 1, is characterized in that, described The vertically oriented device of reverse turning bed workbench arranged on reverse turning bed workbench is 4 and is separately positioned on reverse turning bed workbench upper and lower Both sides；Described rotation binding clasp is 4, and device vertically oriented with reverse turning bed workbench is corresponding respectively installs；Described safety cock Being 2, device vertically oriented with the reverse turning bed workbench being arranged on reverse turning bed workbench top is corresponding respectively installs.

A kind of lathe turnover plate type workbench with defect recognition function the most according to claim 2, is characterized in that, described Leading screw end is provided with buffering disc spring.