JP5950508B2 - External force discrimination method in acceleration sensor response - Google Patents

Description

  The present invention relates to an external force discrimination method in response to an acceleration sensor applied to a structure such as a bridge pier.

Health monitoring of structures is to extract meaningful information from long-term observation results by sensors. Conventionally, what has been put into practical use as monitoring is operated for the purpose of knowing the magnitude and danger of external force, such as an earthquake observation system that issues an emergency earthquake warning.
The method of predicting damage to structures such as measured seismic intensity and SI value is an evaluation method that considers the characteristics of the target structure as a given one. Health monitoring, on the other hand, focuses on extracting not only the characteristics of external forces but also the characteristics of structures. The sensor response includes physical phenomena such as microtremors, earthquakes, hydrodynamic forces (water, wind, etc.), train vibration, etc., as well as various electrical noises, etc., for health monitoring. Therefore, it is necessary to evaluate the external force characteristics and the structure characteristics using signal processing or the like.

Normally, the sensor is selected so as to correspond to the target external force. However, the recent technological improvements in sensor devices and AD conversion are remarkable, and the sensor can respond to various external forces in the amplitude level and the target frequency range. Has been developed. Various signal processing methods have also been developed, and evaluation methods in accordance with the characteristics of external forces are being studied.
For this reason, when an external force is known in operation, it is possible to make the monitoring system multi-functional by applying an evaluation method for the external force, and a high added value system is constructed.

  For example, an earthquake observation system for issuing an earthquake early warning is known that has an automatic detection algorithm for seismic waves and daily noise (see, for example, Patent Document 1).

By the way, when a sensor is installed in a civil engineering structure and the response is observed, responses due to various external forces are observed. When discriminating the external force, it is possible to determine what kind of phenomenon the response is by using not only the sensor response but also an image or the like, but in this case, it is always necessary for a person to monitor. Therefore, as in Patent Document 1, in the monitoring of natural external forces such as earthquakes and earth and sand detection, there is a method in which a low-pass filter is applied to observation data, the amplitude of artificial vibration is reduced, and an event with a large amplitude is detected as an earthquake. Has been used. Furthermore, an algorithm that automatically discriminates seismic waves and daily noise from the characteristics of acceleration waveform rise and seismic frequency components in seismic waves is being studied.
In addition, in order to detect landslides, methods for distinguishing from earthquake response have been studied, and it is said that floods, earthquakes, and noise can be discriminated by using the difference in dominant frequency. In such a method, the main focus is on extracting only the desired response.

JP 2009-103672 A

  However, in health monitoring of railway bridges and road bridges, they are exposed to train vibrations accompanying the passage of trains and vehicles, and the response acceleration is greater than responses due to natural external forces such as small and medium earthquakes, floods, and strong winds. . For this reason, in health monitoring, it is important to identify and take out various external forces other than seismic force from the viewpoint of structural performance evaluation, such as how to eliminate the response of train vibration. There was room for improvement.

  The present invention has been made in view of the above-described problems. Since various external forces can be identified and extracted, evaluation according to these external force characteristics can be performed. It is an object of the present invention to provide an external force discrimination method in an acceleration sensor response that can be constructed.

In order to achieve the above object, an external force determination method in response to an acceleration sensor according to the present invention is an external force determination method for determining external force characteristics acting on a structure based on the response of the acceleration sensor, which is attached to the structure. The step of acquiring an acceleration response from the acceleration sensor and the first discrimination method based on the peak factor for the acceleration response are used to determine the presence or absence of spike noise, and when the spike noise is determined, it is removed. And a second discriminating method based on the frequency component for the acceleration response not having the spike noise, and a step of discriminating the vibration frequency component of the ground motion and taking out the seismic external force. in the method, earthquakes outside included in the frequency component from the ratio of the standard deviation processing before and after the digital filter obtained by using a digital filter It is characterized by determining the presence or absence of.

In the present invention, spike noise is removed from the acceleration response acquired by the acceleration sensor by the peak factor of the first discrimination method, and the frequency component of the ground motion is detected by the second discrimination method for the acceleration response that does not have the spike noise. And only the seismic external force can be taken out. Then, it is possible to analyze and evaluate the extracted specific seismic external force. Furthermore, by extracting only the seismic external force in the second discrimination method, it is possible to discriminate acceleration responses other than the seismic external force. It is possible to identify the external force, and the structure can be evaluated according to the external force characteristics.
Furthermore, in the present invention, since data that has been continuously measured can be analyzed in real time by using a digital filter, the duration can be increased or decreased by the target external force as in the conventional method using Fourier transform. For example, there is an advantage that it is difficult to determine the recording length.

  Further, in the external force discrimination method in the acceleration sensor response according to the present invention, the train generated when the train passes by the third discrimination method based on the magnitude of the response to the other acceleration response from which the earthquake external force is removed by the second discrimination method. It is preferable to have a step of discriminating between vibration and fine vibration and taking out external force due to traffic.

  Further, in the external force discrimination method in response to the acceleration sensor according to the present invention, the presence or absence of drift is discriminated by the fourth discrimination method based on the non-stationary average for the fine motion vibration from which the external force due to traffic is removed by the third discrimination method. , If there is a drift, the step of removing the drift, and the acceleration response from which the drift has been removed by the fourth discriminating method are always discriminated by the fifth discriminating method based on the magnitude of the fine motion response. It is preferable to include a step of taking out the fine movement when it has fine movement and determining that there is an external force of increased water or strong wind external force when there is no constant fine movement.

  Further, in the external force discrimination method in the acceleration sensor response according to the present invention, the response characteristic at the time of water increase and the response at strong wind are obtained by the sixth discrimination method based on the fine motion spectrum ratio with respect to the acceleration response determined by the fifth discrimination method. It is preferable to have a step of discriminating the characteristics and separating out the increased water external force and the strong wind external force.

In this case, for the acceleration response other than the seismic external force classified by the second discrimination method, the train vibration is extracted by the third discrimination method, the external force (train vibration) by this specific traffic is analyzed, and the structure is It becomes possible to evaluate. Further, the fine discrimination vibration from which external force due to traffic is removed by the third discrimination method is used to remove the drift by the fourth discrimination method, and then the fine response is always fine by the fifth discrimination method for the acceleration response from which the drift is removed. It is possible to analyze the external force due to this specific microtremor and evaluate the structure. Furthermore, with respect to the acceleration response determined to have an external force by the fifth discrimination method, the water increase external force and the strong wind external force are extracted by the sixth discrimination method, the specific water increase external force and the strong wind external force are analyzed, and the structure is evaluated. It becomes possible.
In this way, by determining the order from the first discrimination method to the sixth discrimination method, the above-described earthquake external force, external force due to train vibration, external force due to constant tremor, increased water external force, and strong wind external force are reliably discriminated and taken out. Therefore, it is possible to evaluate the structure with high accuracy by analyzing these individual external force characteristics using an appropriate analysis means.

  According to the external force discrimination method in response to the acceleration sensor of the present invention, it is possible to evaluate various external forces by identifying and extracting various external forces, so that a multifunctional and excellent monitoring system is constructed. can do.

It is a figure which shows the operation | movement flow of the external force discrimination method by embodiment of this invention. It is a figure for demonstrating the discrimination method based on a peak factor, Comprising: It is a figure which shows an example of the abnormal waveform of a peak factor. It is a figure for demonstrating the method using a Fourier transform, Comprising: It is a figure which shows an example of the characteristic of an earthquake spectrum. It is a figure for demonstrating the discrimination method based on a spectrum ratio, Comprising: It is a figure which shows the example of a determination. It is a figure for demonstrating the method using a digital filter, Comprising: It is a figure which shows an example of the standard deviation ratio before and behind digital filter application. It is a figure for demonstrating the discrimination method based on the magnitude | size of a response, Comprising: It is a figure which shows an example at the time of a train passage. It is the figure which showed the example of discrimination | determination of a train vibration and a fine motion response. It is a figure for demonstrating the discrimination | determination method based on an unsteady average, Comprising: It is a figure which shows an example of a drift waveform. It is a figure for demonstrating the discrimination method based on the magnitude | size of a fine movement response, Comprising: It is a figure which shows an example of the fine movement waveform of the structure at the time of a strong wind. It is a figure which shows an example of the magnitude | size of a fine movement RMS response. It is a figure for demonstrating the discrimination | determination method based on a microtremor spectrum ratio, Comprising: It is a figure which shows an example of the spectral characteristic difference between water increase and a strong wind.

  Embodiments of an external force determination method in response to an acceleration sensor according to the present invention will be described below with reference to the drawings.

  As shown in FIG. 1, the external force determination method in the acceleration sensor response according to the present embodiment is a method of determining the external force characteristics acting on the pier (structure) of the railway bridge on which the train travels based on the response of the acceleration sensor. is there. Here, the pier according to the present embodiment is intended for a bridge built over a river, and the pier is also provided in the river. The acceleration sensor is attached to an appropriate position on the upper part of the pier.

In the external force determination method according to the present embodiment, various external forces are automatically determined using an external force automatic determination method (not shown).
As shown in FIG. 1, as an external force determination method, first, an acceleration response is acquired in step S1.

  In step S2, the presence / absence of spike noise is determined by the first determination method based on the peak factor for the acceleration response. If spike noise is determined, it is removed. That is, based on the peak factor, it is determined whether the phenomenon is a physical phenomenon or an electrical noise of the sensor. At this time, if it is a physical phenomenon, the process proceeds to step S3, and if it is caused by spike noise, the process proceeds to step S4. Note that spike noise is targeted as electrical noise. However, when spike noise with a large value occurs, a large response is generated, and when it is determined based on the maximum value or the like, a false alarm is caused.

  Next, in step S3, the frequency component of the seismic motion is discriminated by the second discriminating method based on the frequency component for the acceleration response having no spike noise, and the seismic external force is taken out. That is, this second discrimination method is a discrimination method for extracting an earthquake based on the spectrum ratio, and discriminates between an earthquake and other external forces. When an earthquake is determined in step S3, the earthquake can be analyzed based on the extracted earthquake external force (step S5), and the acceleration response determined to be other than the earthquake external force proceeds to the next step S6. become.

  In addition, since earthquakes of various magnitudes occur and cannot be determined simply by the magnitude of the response, determination is made using the difference in frequency components in step S3. When comparing the frequency components, a method using a general Fourier transform or a method using a digital filter can be used. In order to apply the Fourier transform, it is necessary to secure a certain waveform length (number of samples). On the other hand, the method using a digital filter is effective because it can be discriminated in real time.

  Furthermore, in step S6, with respect to other acceleration responses from which the seismic external force has been removed by the second discrimination method, the train vibration and the fine vibration generated at the time of passing the train are discriminated by the third discrimination method based on the magnitude of the response, Take out external force (train vibration) caused by traffic. In other words, the response other than the earthquake determined in step S3 described above is divided into train vibrations that increase the vibration level and other fine vibrations. Here, since the train vibration has a greater response than the fine movement regardless of the weight or speed of the load, it can be distinguished relatively easily. In step S7, train vibration analysis can be performed using the train vibration determined and extracted in step S6. Note that fine movements other than train vibrations proceed to step S8.

  Subsequently, in step S8, the presence or absence of drift is determined by the fourth determination method based on the non-stationary average for the fine vibration from which the external force due to traffic has been removed by the third determination method. Remove. If there is no drift, the process proceeds to step S9. If there is a drift, the process proceeds to step S10.

  In the fourth discrimination method, the condition for using the response for the fine tremor waveform other than the train vibration is that the signal is determined by the sensor noise and the self-noise determined by the measurement environment in the S / N ratio that is the ratio of the signal and noise. Is also big. When there is non-stationarity such that the average of the response waveform changes, the waveform is considered to have drifted, and self-noise increases. The drift phenomenon of the acceleration sensor may be affected by temperature characteristics, but if a large drift occurs, it will affect not only the DC value but also the high frequency, so it may not be possible to accurately measure. It is necessary to. That is, when there is a drift, the waveform is confirmed in step S10. And, it is known that the signal will increase at the time of water increase and strong wind, and the characteristics of the structure will be clearly shown. Therefore, if the determination of water increase and strong wind can be made by sensor response, it is more reliable. Analysis can be performed.

  Further, in step S9, the acceleration response from which the drift has been removed by the fourth determination method is determined by the fifth determination method based on the magnitude of the fine movement response. At the same time, it is determined that there is an external force of increased water or strong wind when there is no slight movement. That is, the acceleration response determined to have no drift in step S8 is divided into a constantly fine one and an external force one. In step S9, it is possible to perform a fine movement analysis using the determined fine movement. In addition, the thing with external force other than that progresses to step S12.

  Subsequently, in step S12, with respect to the acceleration response determined to have an external force by the fifth determination method, the response characteristic at the time of water increase and the response characteristic at the time of strong wind are determined by the sixth determination method based on the fine motion spectral ratio, Separate and extract external force and strong wind external force. That is, if it is determined that the vibration is increased, a vibration analysis is performed in step S13, and if it is determined that it is a strong wind vibration, the strong wind vibration analysis is performed in step S14.

(First discrimination method)
Next, the discrimination method based on the peak factor performed in step S2 shown in FIG. 1 will be described in detail.
That is, as shown in FIG. 2, since large spike noise affects a statistical index such as the maximum value and average and needs to be removed, it is determined whether spike noise has occurred in the waveform due to the peak factor. Further, when trigger measurement is performed, the determination in step S2 is performed because the trigger is abnormal and causes an abnormal waveform.

  Specifically, the recorded acceleration waveform is set to an interval of T seconds, and the maximum value and the standard deviation of m (= T / dt) samples in the p interval are calculated by Equations (1) and (2).

Here, || represents an absolute value, and max [] and rms [] are procedures for obtaining a maximum value and a standard deviation, respectively. The peak factor is obtained from the ratio between the standard deviation and the maximum value from the equation (3). As the threshold, it sets the _{e p} (4) as equation.
However, e _p is the previously set value from the temperature characteristics of the sensor. FIG. 2 shows an example of an abnormal waveform in which the horizontal axis indicates time (seconds), the vertical axis indicates acceleration (m / s), and the peak factor increases.

(Second discrimination method)
Next, the determination method based on the spectrum ratio performed in step S3 shown in FIG. 1 will be described in detail.
In this discriminant method, a method applied to a system using an existing trigger measurement (method using Fourier transform) and a method applied to real-time evaluation (method using a digital filter)
And can be adopted.
The trigger measurement method has restrictions on the analysis speed and capacity of the recording system, so as a data collection method for monitoring, measurement starts when a response exceeding a certain acceleration occurs, and a waveform of a certain length is analyzed. It has become a way to do. In this case, it is difficult to determine the recording length because the duration is long or short depending on the target external force. On the other hand, the method applied to the real-time evaluation is more preferable than the method based on the trigger measurement because the data continuously measured can be analyzed in real time.

(Method using Fourier transform)
As shown in FIG. 3, the method using this Fourier transform is different in the frequency band included in the spectrum due to the earthquake response and the spectrum due to the train vibration. This is a method for making a determination.
The acceleration time-series response in the p interval is Fourier transformed, a Fourier spectrum is calculated using equation (5), and a power spectrum is calculated using equation (6).

A cumulative power spectrum is calculated as a cumulative power spectrum based on equation (7).
Further, the cutoff frequency ωel of the high-pass filter is set, and in the cumulative spectrum of the equation (7), a spectrum having a higher frequency than ωel is accumulated based on the equation (8).
Further, how much high frequency is included in the entire spectrum is calculated as in equation (9) using equations (7) and (8).

Then, set the q _et as a threshold, to determine the following criteria.
When q _et, p _{<q et,} earthquake force there q _et waveform in p _interval, when _p ≧ _{q et,} earthquake external force as in the determination example shown in the waveform in the p interval without earthquake external force diagram 4 there If it is determined, the seismic waveform is created by connecting all the sections with the nearby seismic external force, and the seismic vibration analysis is performed. The threshold is not so important because it differs depending on the order, but is set in consideration of the characteristics of the structure.

(Method using digital filter)
Next, a method for determining an earthquake in real time using a digital filter will be described.
In the case of a digital filter, since Fourier transform is not used, a certain length of waveform (number of samples) is required in order to secure the frequency step of the spectrum, and it may be difficult to determine in real time. There is no advantage.
According to Parseval's theorem, the power spectrum and the mean square (dispersion) of the time series waveform have the following relationship.

Using this relationship, a waveform is discriminated from the ratio of the standard deviation before and after the digital filter processing by the same concept as the method using the Fourier transform described above.
The application of this method as a digital filter design method is not limited to a specific digital filter, but the following low-pass digital filter is used for ease of application.
A sink function with a cutoff frequency B is defined as in equation (11).
Here, τ = −S, −S + dt,..., 0,..., −S−dt, S, S = ½B, and a = 2S. The sink function has a characteristic as shown in equation (12).

  The acceleration waveform after applying the filter can be subjected to digital filter processing as shown in equation (13) using equation (11).

Here, df = 1 / dt.
Although the filter characteristics are affected by S, it is the number of samples m when the standard deviation is taken after the filter processing that greatly affects the time required for determination. If the interval P (number of samples M) is determined in consideration of the stability of the standard deviation of the acceleration response, the movement ratio of the ratio of the standard deviation before and after the filter can be calculated as in equation (14).

In this case, the time required for the determination is M × dt. Q _rt is determined as a threshold value, and is determined according to the following criteria.
When q _{r, p} <q _rt , there is an earthquake external force in the waveform in the P section. When q _{r, p} ≧ q _rt , there is no earthquake external force in the waveform in the P section. FIG. 5 shows an applied example. Since train vibration has a high frequency component, the response after filtering is relatively small. However, for the tremor section that is neither an earthquake nor a train, the numerator value of the equation (14) is used to make a judgment in combination with the following method based on the response magnitude.

(Third discrimination method)
Next, the determination method based on the response magnitude performed in step S6 shown in FIG. 1 will be described in detail.
As shown in FIG. 6, in a structure such as a bridge, the response due to train vibration or earthquake is larger than the fine motion response. Therefore, a method for discriminating a fine movement response using a difference in response magnitude will be described.
The standard deviation of the acceleration waveform time series can be obtained by the above-described equation (2). If the threshold value et is determined, it can be discriminated as in the equations (15) and (16).
In order to further improve the reliability, a section that cannot be determined is defined as _eqt , and determination is not possible as in Expression (17).

  If it is determined that there is a traffic external force, all adjacent sections with a traffic external force are connected to create a traffic (train) waveform, and a train vibration analysis is performed. FIG. 7 shows an example of the determination.

(Fourth discrimination method)
Next, the determination method based on the nonstationary average performed in step S8 shown in FIG. 1 will be described in detail.
That is, a waveform in which a large drift occurs as an unusable waveform of the fine movement waveform is removed by the following method.
The average value of the P section of the fine motion acceleration waveform is obtained as follows.
Also, the drift amount is defined as in Eq. (19) in ^∃ l, k section.

Judgment is performed as shown in equation (20) using the threshold value _qzt .
However, q _zt is a value determined in advance from the temperature characteristics of the sensor.
The drifted waveform shown in FIG. 8 is not used, and the waveform is checked to deal with sensor failures and the like.

(Fifth discrimination method)
Next, the determination method based on the magnitude of the fine motion response performed in step S9 shown in FIG. 1 will be described in detail.
The response at the time of water increase or strong wind is about 2 orders of magnitude lower than the response to the train vibration of FIG. 6 as in the case of the strong wind shown in FIG. Result in a large response. If the fine movement is increased, the S / N ratio is increased and the characteristics of the structure are clearly shown, so that a highly reliable analysis can be performed.
When a fine acceleration waveform is obtained, the standard deviation is calculated using the equation (2) on a continuous time axis.
The reference vibration level is determined as e _at and e _bt, and the determination as shown in equation (21) is performed.

If it is the fine movement determination, the analysis using the fine movement is performed. If there is an external force, response analysis is performed when water is increased or wind is strong. The threshold value e _at is set in consideration of system performance such as sensor self-noise and quantization resolution. The threshold value e _bt is determined by evaluating the environment and structure characteristics in the field.

(Sixth discrimination method)
Next, the determination method based on the fine motion spectrum analysis performed in step S12 shown in FIG. 1 will be described in detail.
When there is a possibility of both a water increase response and a strong wind response in the waveform determined as having an external force by the fine motion response discrimination method in step S9, a method of determining which external force is used to amplify the fine motion is used.

  When the excited frequency bands are different, it is important when evaluating the characteristics of the structure. FIG. 11 shows a comparison of the response spectrum during water increase and the response spectrum during strong wind in the same structure. Thus, it can be seen that the response at the time of water increase has more high frequency components than the response at the time of strong wind. Since this is considered to be due to the difference in viscosity, it is possible to apply the discrimination method based on the spectral ratio, which is the second discrimination method described above.

Next, the effect | action of the external force discrimination method in the acceleration sensor response mentioned above is demonstrated in detail.
In the external force determination method of the present embodiment, spike noise is removed from the acceleration response acquired by the speed sensor by the peak factor of the first determination method, and the second determination method is used for the acceleration response that does not have the spike noise. The frequency component of the seismic motion can be discriminated by and only the seismic external force can be extracted. Then, it is possible to analyze and evaluate the extracted specific seismic external force. Furthermore, by extracting only the seismic external force in the second discrimination method, it is possible to discriminate acceleration responses other than the seismic external force. It is possible to identify the external force, and the structure can be evaluated according to the external force characteristics.

  In addition, for the acceleration response other than the seismic external force classified by the second discrimination method, the train vibration is extracted by the third discrimination method, the external force (train vibration) by this specific traffic is analyzed, and the structure is evaluated. It becomes possible to do. Further, the fine discrimination vibration from which external force due to traffic is removed by the third discrimination method is used to remove the drift by the fourth discrimination method, and then the fine response is always fine by the fifth discrimination method for the acceleration response from which the drift is removed. It is possible to analyze the external force due to this specific microtremor and evaluate the structure. Furthermore, with respect to the acceleration response determined to have an external force by the fifth discrimination method, the water increase external force and the strong wind external force are extracted by the sixth discrimination method, the specific water increase external force and the strong wind external force are analyzed, and the structure is evaluated. It becomes possible.

  In this way, by determining the order from the first discrimination method to the sixth discrimination method, the above-described earthquake external force, external force due to train vibration, external force due to constant tremor, increased water external force, and strong wind external force are reliably discriminated and taken out. Therefore, it is possible to evaluate the structure with high accuracy by analyzing these individual external force characteristics using an appropriate analysis means.

  In the external force discrimination method in the acceleration sensor response according to the first embodiment described above, various external forces are identified and extracted, and evaluation according to these external force characteristics becomes possible. A monitoring system can be constructed.

As mentioned above, although embodiment of the external force discrimination method in the acceleration sensor response by this invention was described, this invention is not limited to said embodiment, In the range which does not deviate from the meaning, it can change suitably.
For example, in the present embodiment, the first discrimination method in step S2 to the sixth discrimination method in step S12 are performed, and the seismic external force, the external force due to train vibration, the external force due to microtremors, the increased external force, and the strong wind external force are extracted. Although it is possible to analyze and evaluate the external force, it is not limited to performing up to the sixth discrimination method. For example, it is assumed that the third determination method in step S6 performs determination until the external force of train vibration is extracted, and step S8 (fourth determination method) and the subsequent steps can be omitted.

  In addition, it is possible to appropriately replace the components in the above-described embodiments with known components without departing from the spirit of the present invention.

Claims (4)

An external force determination method for determining external force characteristics acting on a structure based on a response of an acceleration sensor,
Obtaining an acceleration response from the acceleration sensor attached to the structure;
Determining whether or not spike noise is generated by a first determination method based on a peak factor for the acceleration response, and removing the spike noise when it is determined;
A step of discriminating the vibration frequency component of the seismic motion by the second discrimination method based on the frequency component for the acceleration response not having the spike noise, and extracting the seismic external force;
Have
The second discriminating method uses a digital filter and discriminates the presence or absence of an earthquake external force included in the frequency component from the ratio of the standard deviation before and after the processing of the digital filter obtained based on the following mathematical formula. To discriminate external force in response to acceleration sensor.

  For other acceleration responses from which the seismic external force is removed by the second discrimination method, the third discrimination method based on the magnitude of the response is used to discriminate between train vibrations and fine vibrations that occur when the train passes and The external force discrimination method in the acceleration sensor response according to claim 1, further comprising a step of taking out. A step of determining the presence or absence of drift by a fourth determination method based on non-stationary average for fine vibrations from which external force due to traffic has been removed by the third determination method, and removing the drift if it has the drift;
With respect to the acceleration response from which the drift has been removed by the fourth determination method, the presence or absence of the constant fine movement is determined by the fifth determination method based on the magnitude of the fine movement response, and when there is the constant fine movement, this is taken out, A step of determining that there is an external force of increased water force or strong wind external force when there is no constant tremor;
The external force determination method in the acceleration sensor response according to claim 2, wherein:   With respect to the acceleration response determined to have external force by the fifth discrimination method, the response characteristic at the time of water increase and the response characteristic at the time of strong wind are discriminated by the sixth discrimination method based on the fine movement spectral ratio, and the water increase external force and the strong wind external force are separated. The external force determination method in the acceleration sensor response according to claim 3, further comprising a step of taking out separately.

Images