CN105841645A - Thickness measuring method based on electro-magnetic ultrasonic wave - Google Patents

Abstract

The invention relates to a thickness measuring method based on an electro-magnetic ultrasonic wave. Compared with routine piezo-electric ultrasonic detection, electro-magnetic ultrasonic detection can realize noncontact detection and can be adapted with certain lift-off. Detection can be realized on coatings and corrosive layers. Furthermore electro-magnetic ultrasonic detection can be well adapted to a high-temperature detection environment, thereby preventing wastage of a large amount of coupling agent required for coupling. According to the thickness measuring method, through a set of embedded digital acquisition system, a field-programmable gate array performs AD sampling, compression detection, correlation averaging and buffering on a digital signal, a central processor successively performs data correlation algorithm, averaging algorithm, peak value retaining algorithm and automatic gain on a digital waveform signal, thereby determining a current echo number-of-times, a small-segment RF data extraction algorithm and an over-zero-point thickness measuring algorithm and obtaining a final thickness result, and finally a measuring result is obtained. According to the thickness measuring method, digital processing is performed on waveform data with relatively low signal-to-noise ratio, thereby preventing tedious operation in adjusting a gate, and realizing workpiece testing without manual parameter operation. Not only can the waveform be displayed in real time, but also the thickness detecting result can be visually observed in a direct manner.

Description

A kind of thickness measuring method based on electromagnetic acoustic

Technical field

This algorithm is based on electromagnetic acoustic to steel pipe, and steel plate carries out Thickness sensitivity, is different from Conventional piezoelectric ultrasonic, for high temperature State, it is not necessary to the detection environment of couplant has good Detection results.This algorithm core content is primarily directed to electromagnetism and surpasses The characteristic of sound detection echo processes, and obtains undistorted, the detection echo waveform that signal to noise ratio is good, thus calculates Thickness measuring result.

Background technology

At present, domestic market and the most conventional thickness measuring are all to come steel pipe, steel plate with piezoelectric supersonic excitation ultrasound ripple Etc. carrying out Thickness sensitivity, this thickness measuring mode determines to must assure that during thickness measuring and is always maintained between probe and workpiece Good coupling, and the incident angle popped one's head in can not have big skew.And during on-line real-time measuremen, I Find to meet this 2 conditions, the precision and stability of plant machinery device is all had higher requirements, and to week The environment (such as temperature) enclosed also there are certain requirements.And the mode that electromagnetic ultrasonic thickness measuring is used is a kind of vertical incidence Shear wave, this ultrasound wave is to produce at inside workpiece, so when the angle of probe changes a lot, the transmission of ripple Remain the wall thickness direction vertical incidence along workpiece, do not interfere with the accuracy of testing result；This detection mode is visited Need not couplant between head and workpiece couple, therefore when probe occurs shake and lift-off during detection, will not produce Biopsy surveys error；Due to the ultrasonic shear wave that cannot produce vertical incidence of Conventional piezoelectric, therefore take is compressional wave thickness measuring mode, Because transverse wave velocity is less one times than longitudinal wave velocity, double so accuracy of detection is more ultrasonic than Conventional piezoelectric under equal frequencies.

But the piezoelectric supersonic detection that electromagnetic acoustic detection contrast is the most ripe, has sensitivity relatively low, makes an uproar at random The inferior positions such as sound is the most, and detection echo is relatively unstable, the present invention i.e. extracts directly perceived effective for this some inferior position Detection echo, is calculated thickness measuring result.

Summary of the invention

In order to effectively utilize the detection advantage of electromagnetic ultrasonic thickness measuring, peak value in electromagnetic acoustic detection can be solved again unstable, Random noise is higher, the technical difficulty that sensitivity is relatively low, object of the present invention is to provide a kind of effective based on electricity The thickness measuring method that magnetic is ultrasonic, the method Center Scheme based on storage zero crossing correspondence horizontal level, to-noise ratio is relatively low Original waveform data after respective handling, obtain accurate thickness measuring result.

The technical solution used in the present invention is:

A kind of thickness measuring method based on electromagnetic acoustic, sets up on the basis of a set of digital acquisition circuit system, and numeral is adopted Collector system includes: electromagnetic ultrasonic probe, analog circuit, digital sampling circuitry, field programmable gate array, in Central processor, USB communicating circuit, LED display, electromagnetic ultrasonic probe is connected with analog circuit by probe wire, mould Intend circuit be connected with digital sampling circuitry, field programmable gate array respectively with analog circuit, digital sampling circuitry, in Central processor is connected, and central processing unit is connected with USB communicating circuit, LCD display respectively；Described analog circuit bag Containing amplifying circuit and filter circuit；It is characterized in that sequentially including the following steps:

Analog circuit receive the faint alternation induced signal that detects of electromagnetic ultrasonic probe amplify through frequency-selecting, after filtering, Digital waveform signal is obtained via the analog digital conversion of digital sampling circuitry；Being input to field programmable gate array, scene can Programming gate array carries out data exchange by data/address bus, by central processing unit to digital waveform signal with central processing unit Carrying out data correlation algorithm, average algorithm, peak value reservation algorithm successively, peak value retains algorithm and obtains the peak value of waveform, In order to obtain measurement result accurately, peak value must reach certain requirement, so increasing wave amplitude feedback function the most automatically Gain, automatic gain is that the peak computational value of feedback obtained according to peak value reservation algorithm acts on analog circuit, makes peak value Reach corresponding requirement, be determined current echo times, segment rf data extraction algorithm, zero crossing thickness measuring the most again Algorithm, obtain final thickness result, download the thickness results of central processing unit by USB communicating circuit and print report Table, by LCD display display waveform.

Thickness measuring flow process and calculating concretely comprise the following steps:

(1) sound velocity calibration and zero must be carried out before thickness measuring partially to calibrate, the rf wave figurate number obtained according to digital sampling circuitry According to, after field programmable gate array carries out detection compression to radio frequency Wave data, according to peak value reservation method and determine current The peak level time location of the effective echo of the record of peaking without gate of echo times, peak value reservation is designated as T1, T1 and is The number count value of the rf data current AD sample corresponding to present peak value of field programmable gate array storage；

(2) velocity of sound V is remembered via after reference block sound velocity calibration；1.5mm before T1 is efficiently sampling original position, In the case of the velocity of sound is known, tries to achieve the leveled time position of the rf data first sampling point of required record zero point, be designated as T2；In like manner using 1.5mm after T1 as end position, it is designated as T3；

(3) sampled using T2 as Sampling starting point obtain for original radio frequency Wave data, through field programmable gate array Sampling caching after do streamlined averagely, the waveform averagely obtained transfer to after being cached by field programmable gate array central authorities process Device processes；Central processing unit records the corresponding horizontal level correspondence in this section of Wave data of positive and negative threshold value respectively For T4, T5；The most required zero point position T6 corresponding in this section of rf data is (T4+T5)/2；

(4) T6 corresponds to a bit of rf data rather than the Sampling starting point of whole frame data, and required dead-center position pair Should be T2+T6 in the position of whole frame data, be designated as T7；

(5) try to achieve the time location of zero point needed for correspondence, but not can determine that this is the zero point which echo is corresponding； By peak value reservation method and determine that the peaking without gate of current echo times tries to achieve the peak that the peak value of corresponding T1 is adjacent T7, trying to achieve N corresponding to T1, N according to T1/ | T1-T7 | is the echo times that T1 is corresponding；

(6) the number of samples count value that T7 refers to, it is known that AD sample frequency is f, the velocity of sound is V, then thickness

D=(T6*V)/2fN, certainty of measurement is 1/f.

The feature of this algorithm is without using gate to entangle waveform peak, it is not necessary to manually adjust waveform position, velocity of sound school Thickness measuring result can be obtained after probe contacts workpiece in 2 seconds intuitively after standard.

Accompanying drawing explanation

Fig. 1 is the circuit diagram of the present invention.

Fig. 2 is the FB(flow block) of the present invention.

Fig. 3 is the logic diagram of the data correlation algorithm of the present invention.

Fig. 4 is the average algorithm logic chart of the present invention.

Fig. 5 is for striving peak phenomenon schematic diagram.

Fig. 6 is for determining current echo times schematic diagram.

Fig. 7 is zero crossing thickness measuring algorithm schematic diagram.

Detailed description of the invention

As shown in Figure 1, 2, the present invention sets up on the basis of a set of digital acquisition circuit system, digital collection electricity Road system includes: electromagnetic ultrasonic probe, analog circuit, digital sampling circuitry, field programmable gate array, centre Reason device, USB communicating circuit, LED display, electromagnetic ultrasonic probe is connected with analog circuit by probe wire, simulation Circuit is connected with digital sampling circuitry, field programmable gate array respectively with analog circuit, digital sampling circuitry, central authorities Processor is connected, and central processing unit is connected with USB communicating circuit, LCD display respectively；Described analog circuit bag Containing amplifying circuit and filter circuit；It is characterized in that sequentially including the following steps:

Analog circuit receive the faint alternation induced signal that detects of electromagnetic ultrasonic probe amplify through frequency-selecting, after filtering, Digital waveform signal is obtained via the analog digital conversion of digital sampling circuitry；Being input to field programmable gate array, scene can Programming gate array carries out data exchange by data/address bus, by central processing unit to digital waveform signal with central processing unit Carrying out data correlation algorithm, average algorithm, peak value reservation algorithm successively, peak value retains algorithm and obtains the peak value of waveform, In order to obtain measurement result accurately, peak value must reach certain requirement, so increasing wave amplitude feedback function the most automatically Gain, automatic gain is that the peak computational value of feedback obtained according to peak value reservation algorithm acts on analog circuit, makes peak value Reach corresponding requirement, be determined current echo times, segment rf data extraction algorithm, zero crossing thickness measuring the most again Algorithm, obtain final thickness result, download the thickness results of central processing unit by USB communicating circuit and print report Table, by LCD display display waveform.

1. data correlation algorithm.

If the digital waveform obtained after analog digital conversion is without any process, transfer to central processing unit real-time on liquid crystal display screen Display, the random noise of digital waveform is very big, and even some moment has exceeded the height of effective detection echo, therefore can Erroneous effects is caused to extracting useful signal.

Described data correlation algorithm concretely comprises the following steps: several frame data correspondence in same position to Coutinuous store Numerical value carries out correspondence and compares, and extracts the minima after comparing on each position, and the new data of recomposition one frame are as waveform The master data of display；Owing to random noise is irregular random appearance and the sampling baseline value higher than standard, so After data correlation algorithm, most random noise all can be filtered out.

Fig. 3 is the logic diagram of data correlation algorithm, wherein T1-two digit counter, M1, M2, M3, M4- Being 512 × 8Bit memorizer, C1, C2, C3-8 position is less than comparator.

From the figure 3, it may be seen that front and back 4 frame sampling data, compare two-by-two at each clock sampling point, finally To the minimum Wave data of each sampling clock point, form the minimum Wave data that a frame is new, the base shown as waveform Notebook data.

2. average algorithm

Data after data correlation algorithm filter major part random noise, but the degree of stability of waveform bad, Peak Jitter obtains terribly, in order to lower the mistake in computation that waveform instability is brought as far as possible, in addition it is also necessary to the ripple after being correlated with Graphic data carry out streamlined averagely.

Average algorithm is streamlined average algorithm, concretely comprises the following steps: first set up the buffer of average time, each frame Data write each buffer the most from front to back, then to average after the data investigation in these buffers, and lock Deposit, such as Fig. 4.Wherein T1-tri-digit counter, M1, M2, M3, M4, M5, M6, M7, M8-are 512 × 8Bit memorizer, A1-8 inputs 8 adders.

As shown in Figure 4, adder that field programmable gate array is embedded, front and back 8 frame sampling data are added and carry out Latching, be uploaded to CPU, CPU do division, obtain 8 meansigma methodss, then CPU carries out next step calculating again.

Streamlined average benefit is averagely not interfere with repetition rate, so display to waveform, the speed of algorithm, The fluency of system is the most beneficial.After Wave data after average algorithm is stored by the hardware bottom layer of central processing unit Carry out subsequent treatment.

3. peak value reservation method

Peak value reservation method particularly step is: finds peak value in any one frame Wave data, records this peak level Position, is worth centered by this peak level position, respectively extends 1.5mm, and 3mm is as sampling interval altogether；Because After scope is stretched it can be seen that, the width of each echo is all without more than 3mm, so with the 3mm work comprising peak value For sampling interval, then in zero signal certainly exists this sampling interval.Previously mentioned, the waveform after detection there will be to be striven Peak phenomenon, can cause peak value sampling to fluctuate, see that Fig. 5, Ha and Hb there will be fluctuated, cause peak value corresponding Horizontal level switches among Aa and Ab, and thus sampling interval is unstable, causes result inaccurate.Therefore connect at probe Touch after workpiece, using first peak value occurring as measuring point, in the interval of the 3mm of each 1.5mm in measuring point left and right Inside occurring that peak value is all ignored, occur peak value just record outside 3mm, the peak value at this moment recorded is next echo Peak value.

4. determine current echo times

Conventional calibrator is all to entangle the peak value of adjacent two echoes to read the position letter that peak value is corresponding with double-gate Breath, and the peak value reservation method that this algorithm relates to is also required to obtain the peak value of detection echo, walks to reduce the operation of instrument Suddenly, faster, obtain thickness measuring result more intuitively, spy employs without gate algorithm.

Because the existence of initial ripple, detection there may be certain blind area, and first echo beyond blind area is effectively Echo；Electromagnetic acoustic detects without couplant, in different steel test blocks after experiment, finds that blind area is basically unchanged, with Outside blind area near blind area a little as starting point, Wave data behind asks for peak value；Terminating point following sampling Terminal changes and changes, and is analogous to gate, and simply gate starting point is constant, and terminating point is servo-actuated；Through different test blocks Test experience after find, might not be the weakest between detection echo and echo times, it is possible that front Phenomenon low after height, therefore peak value not necessarily obtains on first echo.

After starting point determines, finding according in peak value reservation method described previously Wave data behind of central processing unit Peak value stored waveform height are designated as H1, and peak is designated as A1, with 1.5mm before this peak or rear 1.5mm again As starting point, find forward or backward first height maximum point higher than H1/2 (6dB), be designated as A2, A2 Corresponding height is H2；Carried above, owing to striving peak phenomenon, so with front or rear each 1.5mm after finding first peak value Avoid storing another peak value of same echo as starting point, cause result mistake；A2 is and the echo peak of adjacent A1 Value, in order to ensure that A2 with A1 is adjacent, and is effective echo, therefore determines at A1, seek as datum line using H1/2 When looking for A2, it is necessary to assure higher than on the datum line of H1/2, Wave data is first on the rise, rise to fixed point After A2, follow downward trend closely, and H2/2 can be dropped to, it is ensured that this A2 is the peak value of efficient neighbor echo, sees Fig. 6. And during seeking A2, the priority searched forward with A1 peak value is more than the priority searched with A1 peak value backward, I.e. as found A2 forward, find the most backward, as do not found A2, then find backward.

After trying to achieve A1, A2, can not determine that A1, A2 are which corresponding time echoes, owing to electromagnetic acoustic detection is not required to Couplant, therefore the zero of waveform calibration partially is only affected by digital sample time delay, and the impact of the most examined environment.Cause By A1 and | A1-A2 |, this can determine which time echo A1 is.The one-tenth-value thickness 1/10 that | A1-A2 | is general for detecting test block, simply Due to strive peak and can not be accurate, and the peak value that A1 is certain detection echo, A1-0 is about this detection corresponding to echo Distance, with (A1-0)/| A1-A2 | i.e. determine A1 corresponding be which time echo.

5. zero crossing thickness measuring algorithm

The thickness measuring method of conventional is all the thickness being calculated workpiece by the peak value of two effective echoes, referred to as peak Value thickness measuring method.And this application by display waveform it will be seen that after detection the waveform of same position have two phases Nearly peak value, and constantly striving peak, see Fig. 5.If owing to the unstable thickness measuring computational methods by routine of peak value can be to causing peak The error of value position correspondence horizontal level record, affects final result of calculation.Therefore intend a kind of new record zero point of sampling Algorithm, referred to as zero crossing thickness measuring method.It can be seen that waveform after from the point of view of arbitrary detection echo is amplified under radio frequency state Form, it can be seen that within certain a period of time, waveform is monotonic decreasing and must be the baseline that have passed through sampling, sees Fig. 7.When recording first such zero point in arbitrary neighborhood two detection echo, peak value thickness measuring method can be analogous to Calculate thickness measuring result.The advantage that zero crossing thickness measuring in contrast to peak value thickness measuring method: peak value thickness measuring method wants peaking at horizontal position Put and must stablize, otherwise can cause error result, and this application after detection owing to striving after positive and negative two peak detections There are 2 peak values in summit, in order to meet during the waveform under arbitrary detection state shows the meter that can be carried out thickness measuring result Calculate, it is therefore apparent that zero crossing thickness measuring algorithm is better than peak value thickness measuring method.

6. segment rf data extraction algorithm

Although the waveform after data correlation algorithm has filtered most random noise, it can be seen that some needs The detection echo sensitivity carrying out recording is the lowest, substantially similar with baseline noise, improves signal while gain The most all improving with noise.Owing to effective waveform cannot be differentiated where, zero crossing thickness measuring algorithm will be made Become impact.Due to during data acquisition, for the requirement of detection range and system data quantitative limitation, greatly Wave data in certain detection range all gets through overcompression, is i.e. to penetrate original under current sampling frequency Frequently signal has carried out regular certain discrete sampling, so the Wave data that bottom software is used for processing can not be determined accurately Radio frequency waveform dead-center position place, position.Therefore, zero crossing thickness measuring is to be original radiofrequency signal based on Wave data.

Being previously mentioned the data quantitative limitation due to detection range and digital display circuit, digital acquisition system cannot collect whole All original radio frequency signal in individual detection range.But zero crossing thickness measuring method is again to build on the basis of original radio frequency signal On, therefore, can only be by the Wave data after detection be extracted wherein one section of feature comprising radiofrequency signal zero point Data, based on the reason that primary signal random noise is higher, segment rf data extraction algorithm is exactly by penetrating of collecting Frequently signal demand again through streamlined average after latch (see above-mentioned average algorithm), the most again to this section of waveform number latched The number of it is believed that carries out sampling and recording obtaining zero signal.

7. store and calculate zero point correspondence horizontal level

One section of original radio frequency data after Ping Jun have obtained, symmetry based on waveform, with AD sampling centrage as base Standard, positive and negative all as a reference point with certain waveform height ratio, improve gain and make the peak value of waveform more than or equal to this The height of individual reference point.In underlying algorithm, using positive and negative the two reference point as two threshold values, with in threshold value Wave data monotonic decreasing as basis for estimation, extract the time location of positive and negative two threshold values, the two threshold value Centre position be dead-center position.Dead-center position found rear and store, being used for carrying out THICKNESS CALCULATION and (specifically calculate public affairs The THICKNESS CALCULATION detailed step that formula sees below), see Fig. 7.

8. automatic gain

Owing to this algorithm exists positive and negative reference value, Wave data must reach this benchmark could effectively read zero point position Put.Difference based on workpiece, the sensitivity of waveform is discrepant, differs sometimes and reaches positive and negative threshold surely Value, in order to reduce operation, rapid and convenient, so at the scene before programmable gate array stored waveform data, to all of Data carry out automatic gain.Automatic gain refers to peak height as basis for estimation, and no matter current form data peaks exists Where, the most all being controlled by signaling, DA conversion carries out gain control by controlling amplifier, risen to or It is down to the reference point (ensureing that the height of this reference point is more than positive and negative threshold value) set, when the waveform peak of display on LCD When highly reaching this setting reference value, automatic gain stops, and otherwise continues.Thus can guarantee that and can reach to ask for The requirement of zero point, is shown in Fig. 7.

Arrange above is the algorithm steps that some is concrete, hereafter supplements detailed step with regard to whole thickness measuring flow process and calculating process:

Must carry out sound velocity calibration and zero before thickness measuring partially to calibrate, zero is to be partially due to what whole digital acquisition system algorithm caused , there is not coupling factor due to electromagnetic ultrasonic thickness measuring in system time delay, so zero is the most relevant with the constant time lag of system, and Owing to zero crossing thickness measuring algorithm is based on zero point, so zero is also inclined to come school zero based on zero point time inclined.

(1) the radio frequency waveform data obtained according to hardware AD sampling, radio frequency Wave data is entered by field programmable gate array After row detection compression, according to the peak level time location of the previously described effective echo of method record of peaking without gate, Peak value retains and is designated as rf data current AD sample corresponding to present peak value that T1, T1 are field programmable gate array storage Number count value；

(2) velocity of sound V is remembered via after reference block sound velocity calibration；By described previously, before T1,1.5mm is as efficiently sampling Original position, in the case of known to the velocity of sound, when trying to achieve the level of the required rf data first sampling point recording zero point Between position, be designated as T2；In like manner using 1.5mm after T1 as end position, it is designated as T3；

(3) sampled using T2 as Sampling starting point obtain for original radio frequency Wave data, through field programmable gate array Sampling caching after do streamlined averagely, the waveform averagely obtained transfer to after being cached by field programmable gate array central authorities process Device processes；Central processing unit records the corresponding horizontal level correspondence in this section of Wave data of positive and negative threshold value respectively For T4, T5；The most required zero point position T6 corresponding in this section of rf data is (T4+T5)/2；

(4) T6 corresponds to a bit of rf data rather than the Sampling starting point of whole frame data, and required dead-center position pair Should be T2+T6 in the position of whole frame data, be designated as T7；

(5) above-mentioned steps has tried to achieve the time location of corresponding required zero point, but not can determine which echo pair this is The zero point answered；By the peak T7 adjacent without trying to achieve the peak value of corresponding T1 described in gate peaking, foundation T1/ | T1-T7 | tries to achieve the N that T1 is corresponding, and N is the echo times that T1 is corresponding；

(6) the number of samples count value that T7 refers to, it is known that AD sample frequency is f, the velocity of sound is V, then thickness D=(T6*V)/2fN, certainty of measurement is 1/f.

Claims (2)

1. a thickness measuring method based on electromagnetic acoustic, sets up on the basis of a set of digital acquisition circuit system, numeral Acquisition Circuit system includes: electromagnetic ultrasonic probe, analog circuit, digital sampling circuitry, field programmable gate array, Central processing unit, USB communicating circuit, LED display, electromagnetic ultrasonic probe is connected with analog circuit by probe wire, Analog circuit is connected with digital sampling circuitry, field programmable gate array respectively with analog circuit, digital sampling circuitry, Central processing unit is connected, and central processing unit is connected with USB communicating circuit, LCD display respectively；Described analog circuit Comprise amplifying circuit and filter circuit；It is characterized in that sequentially including the following steps:

Analog circuit receive the faint alternation induced signal that detects of electromagnetic ultrasonic probe amplify through frequency-selecting, after filtering, Digital waveform signal is obtained via the analog digital conversion of digital sampling circuitry；Being input to field programmable gate array, scene can Programming gate array carries out data exchange by data/address bus, by central processing unit to digital waveform signal with central processing unit Carrying out data correlation algorithm, average algorithm, peak value reservation algorithm successively, automatic gain is to retain algorithm according to peak value to obtain To peak computational value of feedback act on analog circuit, make peak value reach corresponding requirement, be determined current returning the most again Ripple number of times, segment rf data extraction algorithm, zero crossing thickness measuring algorithm, obtain final thickness result, by USB communication Circuit downloads thickness results the printing reports of central processing unit, by LCD display display waveform.

Thickness measuring method based on electromagnetic acoustic the most according to claim 1, it is characterised in that thickness measuring flow process and meter Concretely comprise the following steps:

(1) sound velocity calibration and zero must be carried out before thickness measuring partially to calibrate, the rf wave figurate number obtained according to digital sampling circuitry According to, after field programmable gate array carries out detection compression to radio frequency Wave data, according to peak value reservation method and determine current The peak level time location of the effective echo of the record of peaking without gate of echo times, peak value reservation is designated as T1, T1 and is The number count value of the rf data current AD sample corresponding to present peak value of field programmable gate array storage；

(2) velocity of sound V is remembered via after reference block sound velocity calibration；1.5mm before T1 is efficiently sampling original position, In the case of the velocity of sound is known, tries to achieve the leveled time position of the rf data first sampling point of required record zero point, be designated as T2；In like manner using 1.5mm after T1 as end position, it is designated as T3；

(3) sampled using T2 as Sampling starting point obtain for original radio frequency Wave data, through field programmable gate array Sampling caching after do streamlined averagely, the waveform averagely obtained transfer to after being cached by field programmable gate array central authorities process Device processes；Central processing unit records the corresponding horizontal level correspondence in this section of Wave data of positive and negative threshold value respectively For T4, T5；The most required zero point position T6 corresponding in this section of rf data is (T4+T5)/2；

(4) T6 corresponds to a bit of rf data rather than the Sampling starting point of whole frame data, and required dead-center position pair Should be T2+T6 in the position of whole frame data, be designated as T7；

(5) try to achieve the time location of zero point needed for correspondence, but not can determine that this is the zero point which echo is corresponding； By peak value reservation method and determine that the peaking without gate of current echo times tries to achieve the peak that the peak value of corresponding T1 is adjacent T7, trying to achieve N corresponding to T1, N according to T1/ | T1-T7 | is the echo times that T1 is corresponding；

(6) the number of samples count value that T7 refers to, it is known that AD sample frequency is f, the velocity of sound is V, then thickness

D=(T6*V)/2fN, certainty of measurement is 1/f.