JP7216238B1 - Grout filling situation evaluation system and grout filling situation evaluation program - Google Patents

Abstract

An object of the present invention is to analyze the non-destructive inspection data by artificial intelligence to evaluate the grout filling state with high accuracy. An acquisition means for acquiring waveform data obtained by performing a non-destructive inspection on a cable sheath for evaluating grout filling conditions, and a restoration that compresses input waveform data and restores the compressed waveform data. output means for outputting restored data based on the waveform data acquired by said acquisition means with reference to a restoration model for outputting data; and restoration data outputted by said output means and waveform data acquired by said acquisition means; and evaluation means for evaluating the grout filling situation based on. [Selection drawing] Fig. 1

Description

The present invention relates to a grout filling condition evaluation system and a grout filling condition evaluation program for evaluating the grout filling condition into the cable sheath of a PC structure by non-destructive inspection.

Conventionally, when injecting grout into a cable sheath of PC structures such as post-tension bridges, viaducts, and buildings, a pressure feed pump is connected to the grout injection side of the cable sheath. Then, the compressing pump is operated to fill the cable sheath with grout.

Incidentally, in recent years, there have been occasional incidents of tendon corrosion and breakage caused by insufficient filling of grout into cable sheaths in PC structures. In such cases, it is common to drill a hole and directly observe the inside as a method to grasp the grout filling condition inside the cable sheath. It is desired to evaluate by

For example, in Patent Document 1, three or more degrees of association between a past waveform image obtained by nondestructive inspection of a cable sheath and the determination result of the grout filling state for the past waveform image are stored in advance, and newly stored. A waveform image obtained by performing a non-destructive inspection on the cable sheath to determine the grout filling status is input, the degree of association is referenced, and the grout filling status in the cable sheath is determined based on the input waveform image. A discriminating grout filling status determination system and program is disclosed. As a result, Patent Document 1 discloses a grout filling state determination system and a grout filling state determination program that can accurately determine the grout filling state by analyzing non-destructive inspection data with artificial intelligence. disclosed.

Japanese Patent No. 6367506

On the other hand, FIG. 9 is a diagram showing the relationship between the frequency and intensity of waveforms detected when elastic waves are excited for each filling degree of grout. According to the findings so far, as shown in FIG. 9, focusing on the dominant frequency between the dashed lines T corresponding to the cable sheath position, when the filling degree of grout in the cable sheath is 0%, the cable sheath position The frequency corresponding to the depth was dominant due to the resonance action. However, as shown in FIG. 9, when focusing on the dominant frequency between the dashed lines T corresponding to the cable sheath position, even when the filling degree of grout in the cable sheath is 100%, the depth of the cable sheath position is Corresponding frequencies dominate due to resonance effects.

For these reasons, in the technique disclosed in Patent Document 1, in order to determine whether or not the frequency corresponding to the assumed depth is dominant, even if the filling degree of grout in the cable sheath is 100%, If the frequency is dominant, it is determined that there is a gap in the grout filling condition. For this reason, the technology disclosed in Patent Document 1 has a problem that it is difficult to accurately determine the state of filling in the cable sheath.

10 is a diagram showing a cross section 107a of the PC structure 107. As shown in FIG. FIG. 10 is a diagram of applying elastic waves to a point P on the surface 107b of the PC structure 107. As shown in FIG. Further, in FIG. 10, points A, B, C, D, E, and F indicate interface points between the cable sheath 171 and the PC structure 107, respectively. 10, X indicates the depth from point P to cable sheath 171. In FIG. FIG. 11 is a graph showing intensity for each frequency when high frequency is input. Frequencies a, b, c, d, e, and f in FIG. 11 indicate the intensity for each frequency at points A, B, C, D, E, and F in FIG. 10, respectively. FIG. 12 is a graph showing intensity for each frequency when low frequencies are input. The frequencies a', b', c', d', e', and f' in FIG. indicates

A cross section 171a of the PC structure 107 is generally provided with a plurality of cable sheaths 171, as shown in FIG. 10, for example. In such a case, it is difficult to evaluate the grouting state of the cable sheath 171 having a depth from the point P because the resonance frequency is dispersed simply by exciting the elastic wave. For example, as shown in FIG. 11, when a high frequency is input and excited, frequency a corresponding to point A with short depth X is dominant, and frequency f corresponding to point F with long depth X is dominant. It's getting harder.

Also, as shown in FIG. 12, when a low-frequency excitation is input, it becomes possible to ignore the frequency a' corresponding to the point A where the depth X is short, and the point F where the depth X is long can be evaluated. The corresponding frequency f' tends to be dominant.

Therefore, when evaluating the state of grout filling in the PC structure 7 in which a plurality of cable sheaths 171 are arranged, it is necessary to select and input an input frequency corresponding to the depth of the cable sheath 171 to be focused on.

On the other hand, the technique disclosed in Patent Document 1 does not assume that an input frequency is selected and input according to the depth of the cable sheath 171 to be focused on. Therefore, the technique disclosed in Patent Document 1 has a problem that it is difficult to determine the grout filling state with high accuracy when a plurality of cable sheaths 171 are arranged in the PC structure 107 .

Therefore, the present invention has been devised in view of the above-described problems, and an object of the present invention is to provide a grout filling state evaluation system and a grouting state evaluation system that can evaluate the grout filling state with high accuracy. To provide a filling status evaluation program.

A grout filling condition evaluation system according to a first invention is a grout filling condition evaluation system for evaluating the grout filling condition in a cable sheath of a PC structure, in which nondestructive inspection is performed on the cable sheath for evaluating the grout filling condition. and an acquisition means for acquiring waveform data obtained by performing the above , and generated by machine learning using past waveform data or pseudo-generated waveform data as learning data, and limiting the frequencies having characteristics in the input waveform data , output means for outputting restored data based on the waveform data obtained by said obtaining means with reference to a restoration model for outputting restored data obtained by restoring waveform data based on said limited characteristics ; evaluation means for evaluating the grout filling condition based on the restored data and the waveform data acquired by the acquisition means.

A grout filling status evaluation system according to a second invention is the first invention, wherein the output means uses past waveform data obtained by non-destructive inspection of a cable sheath having a grout filling level equal to or higher than a threshold as learning data. It is characterized by referring to the restoration model.

A grout filling status evaluation system according to a third invention is the first invention, wherein the output means is a non-destructive inspection that inputs elastic waves with different frequencies according to the depth from the surface of the PC structure to the cable sheath. It is characterized by referring to the restoration model using acquired waveform data as learning data.

A grout filling status evaluation system according to a fourth invention is the first invention, wherein the output means uses waveform data having an intensity of 70% to 130% of the intensity for each frequency of the past waveform data as learning data. It is characterized by referring to the restoration model.

A grout filling status evaluation system according to a fifth aspect of the present invention is the first aspect, wherein the output means generates the restoration model using waveform data having a frequency with a difference of 200 Hz or less from the frequency of the past waveform data as learning data. Characterized by referencing.

A grout filling situation evaluation system according to a sixth invention is the first invention, wherein the acquisition means acquires configuration information related to constituent members of a PC structure having a cable sheath for which the grout filling situation is to be newly evaluated, and the output means refers to a restoration model associated with the configuration information acquired by the acquisition means.

A grout filling situation evaluation system according to a seventh invention is, in the first invention, wherein the acquisition means newly acquires shape information related to a cross-sectional shape of a PC structure having a cable sheath for evaluating the grout filling situation, and the output means is characterized by referring to a restoration model associated with the shape information acquired by the acquisition means.

A grout filling situation evaluation system according to an eighth invention is the first invention, wherein the acquisition means acquires depth information about a depth from the surface of the PC structure to the cable sheath for newly evaluating the grout filling situation, The output means is characterized by referring to the restoration model linked to the depth information acquired by the acquisition means.

A grout filling situation evaluation system according to a ninth invention is the first invention, wherein the acquisition means acquires incidental information about a cable sheath for newly evaluating the grout filling situation from the surface of the PC structure, and the output means It is characterized by referring to a restoration model linked to the incidental information acquired by the acquisition means.

A grout filling situation evaluation program according to a tenth invention is a grout filling situation evaluation program for evaluating the grout filling situation in a cable sheath of a PC structure. An acquisition step of acquiring waveform data obtained by conducting an inspection, and generating by machine learning using past waveform data or pseudo-generated waveform data as learning data, and limiting the frequencies having characteristics in the input waveform data. an output step of outputting the restored data based on the waveform data obtained in the obtaining step with reference to a restoration model for outputting restored data obtained by restoring the waveform data based on the limited characteristics ; and an evaluation step of evaluating the grout filling state based on the restored data and the waveform data obtained in the obtaining step.

The grout filling condition evaluation program according to the eleventh invention is a grout filling condition evaluation program for evaluating the grout filling condition in the cable sheath of the PC structure, and the impact elasticity of the cable sheath for evaluating the grout filling condition is An acquisition step of acquiring waveform data obtained by performing the wave method, and generating by machine learning using past waveform data or pseudo-generated waveform data as learning data, and obtaining a characteristic frequency in the input waveform data. an output step of outputting the restored data based on the waveform data obtained by the obtaining step, with reference to a restoration model for outputting the restored data obtained by restoring the waveform data based on the limited features ; A computer is caused to execute an evaluation step of evaluating the grout filling state based on the output restoration data and the waveform data obtained in the obtaining step.

According to the first invention to the eleventh invention, the waveform data obtained by performing a non-destructive inspection on the cable sheath for which the grout filling situation is newly evaluated is input to the restoration model, and the output restoration data and waveform data to evaluate the grout filling situation. As a result, the waveform data obtained by performing a non-destructive inspection on the cable sheath for which the grout filling status is newly evaluated is input to the restoration model, and based on the output restoration data and waveform data, a new Abnormal values can be evaluated when the obtained waveform data and the waveform data obtained when the grout is at a predetermined filling degree are compared. This makes it possible to evaluate the grout filling state with high accuracy even when the grout filling state is evaluated in a structure in which a plurality of cable sheaths are arranged, or when the frequency is dominant. In addition, the output means refers to a restoration model generated by machine learning using past waveform data obtained by non-destructive inspection of the cable sheath as learning data. Using the data as learning data, extracting feature data, performing a non-destructive inspection on the cable sheath to newly evaluate the grout filling situation, inputting the waveform data obtained by inputting the waveform data into the restoration model, and using the output restoration data Based on the waveform data, it is possible to evaluate abnormal values when comparing the newly acquired waveform data with the waveform data when the grout is at a predetermined filling degree. This makes it possible to evaluate the grout filling state with higher accuracy.

In particular, according to the second invention, the database stores a restoration model using past waveform data obtained by non-destructive inspection of a cable sheath having a grout filling level equal to or higher than a threshold as learning data. Thereby, it becomes possible to use the waveform data when the grout filling degree is high as the learning data. For example, by using the waveform data obtained when the grout filling degree is 100% as the learning data, it is possible to evaluate abnormal values when comparing the newly obtained waveform data and the learning data. . As a result, the grout filling condition can be evaluated with higher accuracy.

In particular, according to the third invention, the database uses waveform data acquired by non-destructive inspection in which elastic waves with different frequencies are input according to the depth from the surface of the PC structure to the cable sheath as learning data. Store the restoration model. As a result, it is possible to use the feature data extracted from the optimum waveform data according to the depth as the learning data, and to evaluate the grout filling state with higher accuracy.

In particular, according to the fourth invention, the database generates waveform data having an intensity that is 70% to 130% of the intensity for each frequency of the past waveform data, and creates a restoration model using the generated waveform data as learning data. Remember. As a result, when evaluating the grout filling state, it is possible to perform accurate evaluation even when there is variation in the intensity of the acquired waveform data for each frequency.

In particular, according to the fifth aspect, the database generates waveform data having a frequency difference of 200 Hz or less from the frequency of past waveform data, and stores a restoration model using the generated waveform data as learning data. As a result, when evaluating the grout filling state, it is possible to perform accurate evaluation even when there is variation in the frequency of waveform data to be acquired.

In particular, according to the sixth invention, the output means refers to the restoration model linked to the configuration information. As a result, it is possible to refer to appropriate restoration models according to the configuration information, and it is possible to accurately evaluate the grouting state even when the constituent members of the PC structure are different.

In particular, according to the seventh invention, the output means refers to the restoration model linked to the shape information. As a result, it becomes possible to refer to an appropriate restoration model according to the shape information, and it is possible to accurately evaluate the grout filling state even when the cross-sectional shape of the PC structure is different.

In particular, according to the eighth invention, the output means refers to the restoration model linked to the depth information. This makes it possible to refer to the appropriate restoration model according to the depth information, and to accurately evaluate the grout filling situation even when the depth from the surface of the PC structure to the cable sheath is different. becomes possible.

In particular, according to the ninth invention, the output means refers to the restoration model linked to the incidental information regarding the cable sheath. As a result, it is possible to refer to the appropriate restoration model according to the incidental information. steel material), information on grout filling in the cable sheath, information on concrete placed in the PC structure where the cable sheath is installed, information on test conditions for non-destructive inspection of the cable sheath, cable sheath Even when the surface condition of the PC structure to be provided is different, it is possible to accurately evaluate the grouting condition.

FIG. 1 is a block diagram showing the overall configuration of a grout filling condition evaluation system to which the present invention is applied. FIG. 2 is a diagram showing an example of waveform data. FIG. 3 is a diagram showing a specific configuration example of the evaluation device. FIG. 4 is a flow chart showing the processing operation of the grout filling condition evaluation system to which the present invention is applied. FIG. 5 is a schematic diagram of a restoration model. FIG. 6A is a diagram showing waveform data and restored data obtained when the grout filling degree is 100%. FIG. 6B is a diagram showing waveform data and restored data acquired when the grout filling degree is 0%. FIG. 7 is a schematic diagram of a concrete test piece. FIG. 8 is a diagram showing evaluation results of the grout filling condition evaluation system to which the present invention is applied. FIG. 9 is a diagram showing waveform data for each grout filling degree obtained by nondestructive inspection. FIG. 10 is a diagram showing a cross section of a PC structure. FIG. 11 is a graph showing intensity for each frequency when high frequency is input. FIG. 12 is a graph showing intensity for each frequency when low frequencies are input.

Hereinafter, a grout filling condition evaluation system to which the present invention is applied will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the overall configuration of a grout filling condition evaluation system 1 to which the present invention is applied. The grout filling condition evaluation system 1 evaluates the grout filling condition in the cable sheath 71 of the PC structure 7 by non-destructive inspection. The grout filling status evaluation system 1 includes a nondestructive inspection unit 8, a conversion device 9 connected to the nondestructive inspection unit 8, an evaluation device 2 connected to the conversion device 9, and a database 3 connected to the evaluation device 2. and

The PC structure 7 is a bridge, a viaduct, a building, etc. in which a PC steel material and a cable sheath 71 are arranged.

The cable sheath 71 is provided with a PC steel rod, a large number of PC steel wires, and other PC steel materials (not shown) in a tensioned state and separated from the inner wall surface of the cable sheath 71 . By the way, in this embodiment, since the post-tension type PC structure 7 is targeted, in such a case, after the cable sheath 71 is placed in the PC structure 7, the concrete is filled and hardened, and then A PC steel material (not shown) is inserted into the cable sheath 71 to apply a tensile stress. After that, the cable sheath 71 is filled with grout and hardened.

The non-destructive inspection unit 8 applies elastic waves to the PC structure 7 by a non-destructive inspection method such as an impact elastic wave method, an impact echo method, or an ultrasonic method, and detects the reflected waveform as inspection data. . The nondestructive inspection unit 8 transmits the detected inspection data to the conversion device 9 .

The conversion device 9 is configured by an electronic device such as a PC (personal computer), a smart phone, a tablet terminal, or a wearable terminal. This conversion device 9 converts the acquired inspection data into waveform data including data indicating at least the strength and frequency for each frequency for evaluating the filling state of grout in the cable sheath 71, and evaluates the converted waveform data. Send to device 2.

The conversion device 9 may transmit the obtained inspection data to the evaluation device 2 as waveform data without any particular analysis. For example, when waveform data indicating frequencies and intensities for each frequency is acquired as inspection data, the conversion device 9 may transmit the inspection data as it is to the evaluation device 2 as waveform data without performing any particular analysis.

Further, the conversion device 9 may analyze the obtained inspection data, convert it into waveform data, and transmit the waveform data to the evaluation device 2 . For example, when the waveform numerical data indicating the relationship between time and amplitude is acquired as inspection data, the conversion device 9 analyzes this, converts it into waveform data indicating the frequency and the intensity of each frequency, and evaluates it. 2 can be sent. The conversion device 9 analyzes the inspection data consisting of the waveform numerical data, and appropriately processes it into the optimum one for evaluating the filling degree of the grout in the cable sheath 71 .

For example, the conversion device 9 performs FFT (Fast Fourier Transform) on the acquired waveform numerical data indicating the relationship between time and amplitude, thereby converting it into waveform data indicating the relationship between frequency and intensity for each frequency. may be converted. In addition, the conversion device 9 performs Fourier transform on the acquired waveform numerical data indicating the relationship between time and amplitude, converts it into waveform numerical data indicating the relationship between time and frequency, and converts it into waveform numerical data indicating the relationship between time and frequency. may be removed, and an inverse Fourier transform may be applied to convert to waveform numerical data representing the relationship between time and amplitude while leaving the specific frequency region. Further, the conversion device 9 further performs a Fourier transform on the waveform numerical data indicating the relationship between the time and amplitude of the specific frequency region left, thereby obtaining the frequency and the intensity of each frequency leaving the specific frequency region. may be converted into waveform data indicating the relationship between . The conversion device 9 may convert the acquired waveform numerical data indicating the relationship between time and amplitude into waveform data indicating the relationship between intensity and frequency, such as a power spectrum.

The conversion device 9 may apply wavelet transform to the acquired waveform numerical data as inspection data to convert it into waveform numerical data indicating the relationship between time and frequency. By applying the wavelet transform, it is possible to leave the time characteristics that are lost during the Fourier transform. Therefore, by applying wavelet transform, it is transformed into waveform numerical data indicating the relationship between time and frequency.

The conversion device 9 takes the logarithm of the value of the power spectrum obtained by Fourier transforming the acquired waveform numerical data as inspection data, and further converts it into waveform numerical data representing a cepstrum obtained by inverse Fourier transformation. may Further, the conversion device 9 may further extract waveform numerical data indicating a spectral envelope, which is a low-order cepstrum, or a spectral fine structure, which is a high-order cepstrum, from waveform numerical data indicating a cepstrum. For example, the waveform numerical data representing the spectral envelope may be extracted by determining the cepstrum order. This cepstrum order can take any value, for example 20, 100, and so on. Furthermore, the coefficients of each cepstrum order may be extracted from the waveform numerical data representing the spectral envelope. Further, the conversion device 9 may convert the amplitude extracted from a certain time domain in the acquired waveform image into waveform numerical data representing a formant, which is a peak when the amplitude is converted into the frequency domain. When the first formant, the second formant, . may indicate the relationship of Further, the conversion device 9 may apply AFTE (Auditory filterbank temporal envelope) conversion to the acquired waveform numerical data as inspection data.

In this manner, the conversion device 9 may convert the acquired inspection data into two-dimensional waveform data as shown in FIG. The two-dimensional waveform data may indicate two relationships among, for example, time, amplitude, frequency, intensity, spectrum, cepstrum, formant, and the like. Alternatively, the reciprocals of these values may be taken.

Further, the conversion device 9 may apply a spectrogram or the like to convert the acquired inspection data into three-dimensional waveform numerical data. The three-dimensional waveform numerical data may indicate three relationships among, for example, time, amplitude, frequency, intensity, spectrum, cepstrum, formant, and the like.

Incidentally, the conversion device 9 can display each inspection data via a display unit such as a display (not shown). Further, the conversion device 9 can store these data in a storage, display these data on a display unit, or write these data in a portable memory based on a user's command. The user can remove this portable memory from the conversion device 9 and carry it freely. Furthermore, the conversion device 9 can transfer these data to other electronic equipment via a public communication network.

Note that the configuration of the conversion device 9 is not essential in the present invention, and may be omitted. In such a case, the inspection data output from the non-destructive inspection unit 8 will be directly transmitted to the evaluation device 2 .

Various types of information are stored in the database 3 . The database 3 stores information sent via a public communication network or information input by a user of this system. The database 3 also transmits this accumulated information to the evaluation device 2 based on a request from the evaluation device 2 .

The evaluation device 2 is composed of an electronic device such as a personal computer (PC), for example, but in addition to the PC, it can be embodied in any other electronic device such as a mobile phone, a smart phone, a tablet terminal, a wearable terminal, etc. It may be one that is made into. The user can evaluate whether or not the cable sheath 71 is sufficiently filled with grout by obtaining the evaluation result of the grout filling state as the search solution by the evaluation device 2 . When the cable sheath 71 is insufficiently filled with grout, an unillustrated compressing pump is operated to fill the cable sheath 71 with grout.

FIG. 3 shows a specific configuration example of the evaluation device 2. As shown in FIG. The evaluation apparatus 2 performs wired or wireless communication with a control unit 24 for controlling the entire evaluation apparatus 2 and an operation unit 25 for inputting various control commands via operation buttons, a keyboard, or the like. , a search unit 27 for searching for optimum design conditions, and a storage unit 28, represented by a hard disk or the like, for storing programs for performing searches to be executed are connected to the internal bus 21, respectively. It is Further, the internal bus 21 is connected with a display unit 23 as a monitor for actually displaying information.

The control unit 24 is a so-called central control unit for controlling each component implemented in the evaluation device 2 by transmitting control signals via the internal bus 21 . In addition, the control unit 24 transmits commands for various controls through the internal bus 21 according to operations through the operation unit 25 .

The operation unit 25 is embodied by a keyboard or a touch panel, and a user inputs an execution command for executing a program. The operation unit 25 notifies the control unit 24 when an execution command is input by the user. Upon receiving this notification, the control unit 24 cooperates with each component including the search unit 27 to execute desired processing operations. Also, the operation unit 25 may receive various types of information from the user. The operation unit 25 receives, for example, configuration information about constituent members of the PC structure 7, shape information about the cross-sectional shape of the PC structure 7, or the depth from the surface of the PC structure 7 to the cable sheath for newly evaluating the grouting state. Additional information may be input by the user, including depth information about the .

The search unit 27 searches for the evaluation result of the grout filling state. The search unit 27 reads various information stored in the storage unit 28 and various information stored in the database 3 as necessary information in executing the search operation. The search unit 27 may be controlled by artificial intelligence. This artificial intelligence may be based on any known artificial intelligence technology.

The display unit 23 is composed of a graphic controller that creates a display image based on control by the control unit 24 . The display unit 23 is implemented by, for example, a liquid crystal display (LCD).

When the storage unit 28 is configured with a hard disk, predetermined information is written to each address based on the control by the control unit 24, and this information is read as necessary. The storage unit 28 also stores a program for executing the present invention. This program is read and executed by the control unit 24 .

The operation of the grout filling condition evaluation system 1 configured as described above will be described.

In the grout filling condition evaluation system 1, inspection data obtained by performing a non-destructive inspection on a cable sheath whose grout filling condition is to be evaluated is converted into waveform data, and the converted waveform data is input to a restoration model and restored. Data is output, and grout filling status is evaluated based on the output restoration data and waveform data. FIG. 4 shows a processing operation flow of this grout filling condition evaluation system 1. As shown in FIG. Detailed processing in each step in FIG. 4 will be described below.

First, in step S11, the non-destructive inspection unit 8 detects inspection data of the cable sheath 71 by a non-destructive inspection method such as an impact elastic wave method, an impact echo method, an ultrasonic method, or the like. In such a case, the frequency of the elastic wave to be excited may be changed according to the depth Y from the surface of the PC building 7 to the cable sheath 71 as shown in FIG. For example, when evaluating a cable sheath 71 with a large depth Y, a low frequency is input. The nondestructive inspection unit 8 outputs the detected inspection data to the conversion device 9 . Alternatively, an elastic wave having a frequency f or more shown in Equation (1) of Equation 1 below may be excited. In such case, V denotes the propagation velocity of the elastic wave.

Next, in step S12, the conversion device 9 performs various analyzes as necessary on the inspection data detected by the non-destructive inspection unit 8. Process the waveform data. Incidentally, this step S12 can be omitted when various analyzes and processing are not performed. The conversion device 9 outputs the converted waveform data to the evaluation device 2 .

Moreover, in step S12, the conversion device 9 may normalize the waveform data. The conversion device 9 may normalize the waveform data, for example, by setting the maximum value of intensity for each frequency to 1 and the minimum value to 0. In addition, the conversion device 9 may normalize the strength of the frequency portion common to the frequencies other than the frequency corresponding to the depth Y as 1 and the minimum value as 0.

Next, in step S13, the input waveform data is compressed, a restoration model for outputting restored data obtained by restoring the compressed waveform data is referred to, and in step S12 restored data based on the waveform data output from the conversion device 9 is extracted. to output

FIG. 5 is a diagram showing a restoration model. The restored model is a model that compresses input waveform data and outputs restored data obtained by restoring the compressed waveform data. The restoration model has an encoder that compresses the input waveform and a decoder that restores the compressed waveform data. Also, the restored model may be a learned model generated by machine learning using past waveform data obtained by non-destructive inspection of the cable sheath as learning data. In such a case, the restored model is generated by compressing the waveform data of the learning data, extracting features of the compressed waveform data, and accumulating the features of the waveform data. As a result, the restoration model can compress the input waveform data and restore the compressed waveform data based on the features of the learning data. The restoration model may be generated by machine learning such as AutoEncoder. A feature of the waveform data is information that can reproduce the input waveform data. Also, the compression of waveform data is to convert the waveform data into lower dimensional data. For example, waveform data may be represented by frequency and intensity for each frequency, and in such a case, compressing the waveform data means limiting the characteristic frequencies. Restoration of waveform data is conversion of waveform data into data of a higher dimension. For example, waveform data may be represented by frequency and intensity of frequency, in which case restoration of the waveform data is the transformation of the data from the characterized frequency back to the original frequency.

For example, past waveform data obtained by non-destructive inspection of the cable sheath 71 may be used as the waveform data to be the learning data. For this waveform data, a plurality of past waveform data obtained by non-destructive inspection by exciting elastic waves with frequencies of 10 Hz from 2500 to 50000 Hz, for example, may be used. Also, the waveform data may be artificial waveform data generated in a pseudo fashion by assuming the sheath and elastic wave velocities, for example, from 2500 to 50000 Hz at intervals of 10 Hz.

Also, the learning data may be classified into normal data and abnormal data according to the grout filling state of the cable sheath 71 when non-destructive inspection is performed. For example, as learning data, waveform data obtained when the grout filling condition of the cable sheath 71 when non-destructively inspected is a filling level equal to or higher than a preset threshold is set as normal data, and at a filling level below the threshold, Waveform data obtained when there is an error may be used as abnormal data. Specifically, the waveform data obtained when the filling degree is 100% may be treated as normal data, and the waveform data obtained when the filling degree is less than 100% may be treated as abnormal data. In addition, the learning data is a plurality of waveform data obtained when the grout filling state of the cable sheath 71 when the nondestructive inspection is performed has a plurality of types of filling degrees equal to or higher than a preset threshold value. The waveform data obtained when there are multiple types of filling degrees less than or equal to 100 degrees may be used as the abnormal data. Specifically, when the threshold is 80% and the waveform data obtained when the filling degrees are 100% and 90% are normal data, and when the filling degrees are 70% and 50% Each waveform data obtained in 1 may be used as abnormal data. Also, the learning data may be normalized waveform data.

In addition, the restoration model may generate new waveform data by varying the intensity and frequency band of past waveform data for each frequency, and use the generated waveform data as learning data. For example, the restoration model may generate waveform data having an intensity that is 70% to 130% of the intensity for each frequency of waveform data actually measured in the past, and the generated waveform data may be used as learning data. This is a value that takes into consideration the variation in the input of the device that inputs elastic waves. In addition to the past waveform data, the restoration model may use simulated waveform data as learning data. The pseudo-generated waveform data is waveform data having a virtual peak calculated using an assumed sheath depth and an assumed elastic wave velocity. Also, the pseudo-generated waveform data includes artificially created waveform data. Further, the restoration model may newly generate waveform data having a frequency with a frequency difference of 200 Hz or less from the frequency of past waveform data, and acquire the generated waveform data as learning data. This is a value considering that the elastic wave velocity to be evaluated is not uniform in the concrete structure and that it has variations, and that there is an error in the design and construction of the cable sheath position. As a result, when evaluating the grout filling state, it is possible to perform accurate evaluation even when there is variation in intensity and frequency for each frequency of waveform data to be acquired.

The restored model has the same number of nodes 61 in the input layer and the output layer, and the number of nodes 61 in the intermediate layer is smaller than the number of nodes 61 in the input layer. For example, the number of nodes 61 in the input layer and the output layer may be 4750, and the number of nodes 61 in the intermediate layer may be 8. In the restored model, each node 61 of the input layer is connected to each node 61 of the intermediate layer by three or more degrees of association, and each node 61 of the intermediate layer is connected to each node 61 of the output layer by three or more degrees of association. Connected by relevance. Also, a plurality of hidden layers may be provided between the input layer and the intermediate layer, or between the intermediate layer and the output layer. Each node 61 in the hidden layer is connected to each node 61 in the adjacent layer with three or more degrees of association.

The degree of association indicates the accuracy between each node 61 . For example, with respect to the input layer node 61a to which the waveform data A is input, the nodes 61b and 61c in the hidden layer are connected with a degree of relevance of 30% and a degree of relevance of 60%, respectively. In such a case, the node 61c with a higher degree of association is closer to accurate determination than the node 61b. This degree of association may be indicated by, for example, 5 levels or percentage.

The grout filling condition evaluation system 1 may generate the restoration model described above in advance, store it in the database 3 , and output the restoration model to the evaluation device 2 in response to a request from the evaluation device 2 . In step S13, the evaluation device 2 refers to the restoration model described above and outputs restoration data based on the waveform data output in step S12.

Further, in step S13, the evaluation device 2 may refer to a restoration model generated using only normal data as learning data, and output restoration data. As a result, the restored data is restored based on the characteristics of the normal data, and thus becomes waveform data based on the characteristics of the normal data. As a result, waveform data and normal data can be compared, and the grout filling state can be evaluated with high accuracy.

Further, in step S13, the evaluation device 2 may refer to the restoration model generated using only the abnormal data as the learning data, and output the restoration data. As a result, since the restored data is restored based on the characteristics of the abnormal data, it becomes waveform data based on the characteristics of the abnormal data. As a result, the waveform data and the abnormal data can be compared, and the grout filling condition can be evaluated with high accuracy.

Also, the database 3 may store a plurality of restoration models in association with the features of the learning data. For example, the restoration model may be stored in association with the configuration information regarding the configuration members of the PC structure 7 input by the user when the learning data was acquired. In such a case, in step S13, the evaluation device 2 acquires the configuration information input by the user via the operation unit 25, for example, and refers to the restoration model linked to the configuration information that is the same as or corresponds to the acquired configuration information. , output the restored data. As a result, it is possible to refer to appropriate restoration models according to the configuration information, and it is possible to accurately evaluate the grouting state even when the components of the PC structure 7 are different. The structural information about the constituent members of the PC structure 7 means information indicating structural types such as T-girders, I-girders, and box girders, and parts such as overhanging floor slabs, webs, and flanges.

Further, the restoration model may be stored in association with the shape information about the cross-sectional shape of the PC structure 7 input by the user when the learning data is acquired. In this case, in step S13, the evaluation device 2 acquires the shape information input by the user, for example, via the operation unit 25, refers to the restoration model linked to the shape information that is the same as or corresponds to the shape information, Output restored data. Accordingly, it is possible to refer to an appropriate restoration model according to the shape information, and it is possible to accurately evaluate the grouting state even when the cross-sectional shape of the PC structure 7 is different. The shape information about the cross-sectional shape of the PC structure 7 is, for example, the thickness, width, length, etc. of the member, including the distance from the end of the member.

Further, the restoration model may be stored in association with the depth information from the surface of the PC structure 7 input by the user to the cable sheath 71 for newly evaluating the grout filling state when the learning data is acquired. good. In this case, in step S13, the evaluation device 2 acquires the depth information input by the user via the operation unit 25, for example, and creates a restoration model linked to the depth information that is the same as or corresponds to the depth information. Browse and output restored data. As a result, it is possible to refer to the appropriate restoration model according to the depth information, and even if the depth from the surface of the PC structure 7 to the cable sheath is different, the grouting condition can be evaluated with high accuracy. becomes possible.

Further, the restoration model may be stored in association with the supplementary information input by the user when the learning data is acquired and the restoration model. The incidental information includes configuration information of the cable sheath 71, arrangement information of the cable sheath 71, tendon information on tendons (PC steel materials) provided in the PC structure 7 in which the cable sheath 71 is provided, and information on filling in the cable sheath 71. grout information about the grout to be installed, concrete information placed in the PC structure in which the cable sheath 71 is provided, test condition information for non-destructive inspection of the cable sheath 71, surface condition of the PC structure 7 in which the cable sheath 71 is provided information, and configuration information within the PC structure 7 in which the cable sheath 71 is provided.

The form information of the cable sheath 71 includes the length of the cable sheath 71, the diameter of the cable sheath 71, the material of the cable sheath, the degree of corrosion of the cable sheath, and the disappearance of the cable sheath due to corrosion. The arrangement information of the cable sheath 71 includes all information regarding the form of arrangement of the cable sheath 71 within the PC structure 7 . Examples of the arrangement information of the cable sheaths 71 include, for example, whether or not the cable sheaths are arranged in parallel when viewed from the front, and information regarding the spacing between the cable sheaths 71 and the like. The tension material information includes the length of the PC steel material, the type of the PC steel material, the tension force, and the like. The grout information is Young's modulus of grout and the like. The concrete information includes the Young's modulus of the concrete placed in the PC structure 7, the presence or absence of cover concrete, and the like. In the case of the impact elastic wave method, the test condition information for the non-destructive inspection includes information about the impact force of the steel ball striking the inspection object. In the case of the impact echo method, the test condition information for the non-destructive inspection includes information regarding the diameter of the impact steel ball for impacting the inspection target. In the case of the ultrasonic method, the test condition information for the non-destructive inspection includes information regarding the frequency of ultrasonic waves applied to the inspection object. The surface condition information of the PC structure 7 includes information such as the presence or absence of unevenness, the condition of cracks, and the condition of dryness and humidity. The configuration information in the PC structure 7 includes, for example, the arrangement intervals of reinforcing bars.

FIG. 6A is a diagram showing restored data obtained by restoring the compressed waveform data by referring to the waveform data acquired when the grout filling degree is 100% and the restoration model generated using the normal data as learning data. is. FIG. 6B is a diagram showing restored data obtained by restoring compressed waveform data by referring to a restoration model generated using waveform data acquired when the grout filling degree is 0% and normal data as learning data. is.

In step S14, as shown in FIGS. 6A and 6B, the evaluation device 2 determines the grout filling status based on the waveform data converted in step S12 and the restored data output in step S13. Evaluate. The evaluation device 2 may calculate an abnormal value, for example, based on the difference between the waveform data converted in step S12 and the restored data output in step S13. In such a case, for example, the abnormal value may be the sum of the values obtained by squaring the intensity difference values for each frequency between the waveform data and the restored data. Also, the evaluation device 2 may calculate an abnormal value based on, for example, the ratio between the waveform data converted in step S12 and the restored data output in step S13. In such a case, for example, the abnormal value may be the sum of the values obtained by adding -1 to the ratio of the intensity of each frequency between the waveform data and the restored data and squaring them. Since this abnormal value increases according to the difference between the waveform data and the restored data, it serves as an index indicating the difference between the waveform data and the restored data. Since the restored data is waveform data reflecting the characteristics of the learning data of the restoration model referred to in step S13, a large abnormal value indicates that the waveform data does not include the characteristics of the learning data. . For example, as shown in FIGS. 6A and 6B, when referring to a restoration model generated using normal data as learning data, waveform data obtained when the grout filling degree is 100% is input. Compared with FIG. 6(a), the abnormal value in FIG. 6(b) obtained by inputting the waveform data acquired when the grout filling degree is 100% is larger. This makes it possible to evaluate the degree of similarity between waveform data and normal data, for example, by using normal data as learning data, and evaluate the degree of similarity between waveform data and abnormal data by using abnormal data as learning data. can be evaluated. As a result, it is possible to evaluate the grouting state with high accuracy even when the frequency is dominant or when evaluating the grouting state in a structure in which a plurality of cable sheaths 71 are arranged. Become.

Further, in step S14, the evaluation device 2 calculates an abnormal value based on a comparison between a plurality of waveform data and each restored data obtained from the waveform data, and averages the calculated plurality of abnormal values. may be regarded as an outlier.

Next, the process proceeds to step S15, and the status of filling the cable sheath 71 with grout evaluated in step S14 is displayed via the display unit 23. FIG. Accordingly, by visually recognizing the display unit 23, the user can immediately grasp the state of filling the cable sheath 71 with grout.

Furthermore, in the present invention, the degree of association described above may be updated. This update may reflect information provided via a public communication network such as the Internet, for example. If new knowledge about the relationship between input parameters and output solutions (evaluation results of grout filling status) is discovered through site information and writings that can be obtained from public communication networks, according to the knowledge Increase or decrease the degree of association.

In addition to the information that can be obtained from the public communication network, this degree of association is updated manually by the system or the user based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. or automatically. Artificial intelligence may be used in these update processes.

Next, evaluation results of the grout filling condition evaluation system 1 to which the present invention is applied will be described with reference to the drawings.

FIG. 7 is a schematic diagram of a concrete specimen 70. As shown in FIG. The concrete test piece 70 is a test piece made of a concrete member simulating a building. The concrete specimen 70 is, for example, a rectangular parallelepiped having a rectangular cross section with short sides of 400 mm and long sides of 1000 mm. The concrete specimen 70 has a cable sheath 71 inside. Y indicates the depth from the lower surface 70a of the concrete test piece 70 to the cable sheath 71;

Waveform data obtained by non-destructive inspection in which elastic waves are applied to the lower surface 70a of the concrete specimen 70 from the Z direction are input to the grout filling condition evaluation system 1 to which the present invention is applied. At this time, the depth Y is set to 120 mm, and the evaluation is performed using the waveform data acquired for each filling degree of the cable sheath 71 with the grout filling degree of 0% and 100%. Also, at this time, a restoration model is used in which waveform data acquired when the grout filling degree is 100% is generated as learning data.

FIG. 8 is a graph showing outliers against the number of data for each degree of grout filling. In FIG. 8 , triangular legends indicate abnormal values when the grout filling degree is 100%, and circular legends indicate abnormal values when the grout filling degree is 0%. Also, the legend indicates the average of the outliers for each degree of grouting.

As shown in FIG. 8, when the grout filling degree is 0% and 100%, the abnormal value is lower when the grout filling degree is 100%, and the grout filling degree is 0%. Then, the outlier becomes large. Therefore, the restored data is restored based on the characteristics of the normal data obtained when the grout filling degree is 100%. Thus, by using the grout filling condition evaluation system 1 to which the present invention is applied, it is possible to evaluate the grout filling condition with high accuracy.

1 Grout filling status evaluation system 2 Evaluation device 3 Database 7 PC structure 8 Non-destructive inspection unit 9 Conversion device 21 Internal bus 23 Display unit 24 Control unit 25 Operation unit 26 Communication unit 27 Search unit 28 Storage unit 61 Node 70 Concrete specimen 71 cable sheath 107 PC structure 171 cable sheath

Claims (11)

In a grout filling situation evaluation system for evaluating the grout filling situation in the cable sheath of a PC structure,
Acquisition means for acquiring waveform data obtained by performing a non-destructive inspection on the cable sheath for evaluating the grout filling situation;
Generated by machine learning using past waveform data or pseudo-generated waveform data as learning data , limiting the frequencies having characteristics in the input waveform data, and restoring the waveform data based on the limited characteristics. output means for outputting restored data based on the waveform data acquired by said acquisition means, with reference to the restored model to be output;
A grout filling condition evaluation system comprising: evaluation means for evaluating the grout filling condition based on the restored data output by the output means and the waveform data acquired by the acquisition means. The grout filling according to claim 1 , wherein the output means refers to the restoration model using past waveform data obtained by non-destructive inspection of a cable sheath having a grout filling level equal to or higher than a threshold as learning data. situation assessment system. The output means refers to the restoration model whose learning data is waveform data obtained by non-destructive inspection in which elastic waves having different frequencies are input according to the depth from the surface of the PC structure to the cable sheath. The grout filling condition evaluation system according to claim 1 , characterized in that. The grout according to claim 1 , wherein the output means refers to the restoration model using waveform data having an intensity of 70% to 130% of the intensity of each frequency of the past waveform data as learning data. Filling status evaluation system. The grout filling condition evaluation according to claim 1 , wherein the output means refers to the restoration model using waveform data having a frequency of 200 Hz or less as learning data from the frequency of the past waveform data. system. The acquisition means newly acquires configuration information about a component of a PC structure having a cable sheath for evaluating grout filling status,
The grout filling status evaluation system according to claim 1, wherein the output means refers to a restoration model linked to the configuration information acquired by the acquisition means. The acquisition means newly acquires shape information about a cross-sectional shape of a PC structure having a cable sheath for evaluating grout filling status,
The grout filling status evaluation system according to claim 1, wherein the output means refers to a restoration model linked to the shape information acquired by the acquisition means. The acquisition means acquires depth information about the depth from the surface of the PC structure to the cable sheath for newly evaluating the grout filling situation,
The grout filling status evaluation system according to claim 1, wherein the output means refers to a restoration model linked to the depth information acquired by the acquisition means. The acquisition means acquires incidental information about the cable sheath for newly evaluating the grout filling situation from the surface of the PC structure,
The grout filling status evaluation system according to claim 1, wherein the output means refers to a restoration model linked to the incidental information acquired by the acquisition means. In the grout filling situation evaluation program for evaluating the grout filling situation in the cable sheath of the PC structure,
an acquiring step of acquiring waveform data obtained by performing a non-destructive inspection on the cable sheath for evaluating the grout filling status;
Generated by machine learning using past waveform data or pseudo-generated waveform data as learning data , limiting the frequencies having characteristics in the input waveform data, and restoring the waveform data based on the limited characteristics. an output step of referencing the restored model to be output and outputting restored data based on the waveform data obtained by the obtaining step;
A grout filling condition evaluation program characterized by causing a computer to execute an evaluation step of evaluating the grout filling condition based on the restored data output by the output step and the waveform data acquired by the acquisition step. In the grout filling situation evaluation program for evaluating the grout filling situation in the cable sheath of the PC structure,
an acquisition step of acquiring waveform data obtained by performing an elastic shock wave method on the cable sheath for evaluating the grout filling state;
Generated by machine learning using past waveform data or pseudo-generated waveform data as learning data , limiting the frequencies having characteristics in the input waveform data, and restoring the waveform data based on the limited characteristics. an output step of referencing the restored model to be output and outputting restored data based on the waveform data obtained by the obtaining step;
A grout filling condition evaluation program that causes a computer to execute an evaluation step of evaluating the grout filling condition based on the restored data output by the output step and the waveform data acquired by the acquisition step.

Images