JP5360742B2 - Observation data processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To facilitate pretreatment for analyzing observation data. <P>SOLUTION: A variation component (second observation data) of other physical quantities affecting the physical quantity of first observation data is recorded along with a first observation data group obtained by observing a long-term physical quantity. A data analysis is performed by an analysis means after performing pretreatment for interpolation by selectively applying a spline interpolation method or a linear interpolation method to fit the time scale of the respective observation data to the analysis. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a method for processing observation data obtained by observing a physical phenomenon in time series.

  Observation data obtained by observing physical phenomena such as temperature, humidity, tide, atmospheric pressure, water pressure, wind speed, and wind direction are affected by multiple fluctuation factors. For example, it is known that pore water pressure observation data in underground rock mass is affected by atmospheric pressure and tides.

  Analyzing the observation data of the observation target and separating the tidal fluctuation component, atmospheric pressure fluctuation component, and noise component from the observation data affected by tidal and atmospheric pressure fluctuation as an analysis technique for separating the influence of these fluctuation factors Tidal analysis programs are known.

JP 2003-75205 A Institute of Statistical Mathematics Vol. 32, 71-85 Makio Ishiguro, Tadahiro Sato, Yoshiaki Tamura, Masaaki Oe Introduction of Tidal Data Analysis Program BAYTAP-G Computer Science Monograph, 22, 56-117 M.I. Ishiguro, Y. Tamura “BAYTAP-G” in TIMSAC-84, The Institute of Mathematics, Tokyo, Japan Geophysical Journal International, 104, 507-516, Y. Tamura, t. Sato, M.M. Ooe, M.C. Ishiguro “A procedure for tidal analysis with a Bayesian information criterion” Earth tidal data analysis, time series analysis II (Koji Akaike, Genshiro Kitagawa), Yoshiaki Tamura Asakura Shoten, 151-165

  However, in order to analyze the observation data to be observed and to separate the influence of the fluctuation factors from this observation data, it is necessary to convert the accumulated observation data into data at time intervals that can be handled by the analysis program. A pre-processing technique for adjusting (interpolating) the time scale is required.

  Further, in order to grasp the output result, a technique for visualizing the main factors of the fluctuation component by displaying the analysis result data and the input data on the display is also necessary.

  Conventionally, when processing an enormous amount of observation data accumulated over a long period of time, it takes a long time to process these data.

  One object of the present invention is to propose a preferred interpolation method for interpolating observation data so as to be suitable for analysis means.

  Another object of the present invention is to propose an observation data processing method capable of preventing an analysis deviating from an actual phenomenon even if the observation data has a missing measurement for a predetermined long period of time.

  Still another object of the present invention is to propose an observation data processing method that can be quickly executed without requiring learning by executing a predetermined interpolation method for missing observation data or the like.

  Furthermore, another object of the present invention is to propose an observation data processing method in which various data related to observation data processing and various data after tide analysis are linked on the same screen and can be enlarged and confirmed as necessary. There is to do.

  The present invention records time-series observation data obtained by observing physical phenomena over a long period of time, performs pre-processing to interpolate the time scale of this observation data to suit the analysis program, and then performs a predetermined analysis. A method and system capable of performing a series of observation data processing in a short time by applying observation data calculation processing for analyzing fluctuation components affecting the observation data by means (analysis program) It is.

  By performing batch processing to analyze the fluctuation components that affect the observation data from a huge amount of observation data accumulated over a long period of time, it is possible to incorporate preprocessed data into an analysis program. This significantly reduces the time required to transfer data for importing into the processing program for visualizing the analysis results.

Specifically, other physical quantities (second observation data) that affect the physical quantities of the first observation data are recorded together with the first observation data group obtained by observing the physical quantities over a long period of time. After performing pre-processing for interpolation by selectively applying a spline interpolation method or a linear interpolation method so that the time scale of the observation data is suitable for analysis, the physical quantity of the first observation data and the physical quantity of the first observation data are The fluctuation component of the physical quantity of the second observation data affecting the physical quantity of the first observation data is analyzed, and the fluctuation component due to the physical quantity of the second observation data affecting the physical quantity of the first observation data is separated. calculating the physical quantity of data of the first observation data, is configured to perform display or storage of various data, the first in the preprocessing, data interpolation of the second observation data, the are recorded first The data interval in the period covered by the data interpolation of the second observation data, carried out by the spline interpolation method when it is less data interval of the data groups in the vicinity of, by linear interpolation when the exceeds the data interval of the data set in the vicinity It is characterized by performing.

  According to the present invention, it is possible to construct a method and a system for automating a series of processes for separating a fluctuation component affecting the observation data from a huge amount of observation data accumulated over a long period of time. it can.

Also, in the preprocessing that interpolates the time scale of the observation data, the spline interpolation method is applied when the data interval in the data interpolation target period of the recorded observation data is less than or equal to the data interval of the neighboring data group. By applying the linear interpolation method when the data interval of the data group is exceeded, the observation data can be efficiently interpolated in a short time in the form close to the actual phenomenon when creating the input data of the analysis program.

  In addition, by executing a predetermined interpolation method for missing observation data or the like, processing of observation data can be realized easily and quickly without requiring learning.

  In addition, various data related to observation data processing and various data after tide analysis are linked on the same screen, and the display that can be enlarged and confirmed as necessary can be achieved between each observation data or tide analysis results. The relationship between data can be confirmed.

  When collecting observation data of physical phenomena, if there is a missing measurement during the observation period, or if the sampling interval is switched during observation, the time series interval of the observation data becomes irregular (see FIG. 1). When the spline interpolation method is uniformly applied to the interpolation of observation data with such irregular time series intervals, the data interval is larger than that of the neighboring data group (before or after long-term missing measurement or sampling interval switching) In FIG. 2, the interpolation line may be greatly bent, resulting in interpolation data greatly deviating from the actual phenomenon (see FIG. 2).

  Therefore, data interpolation to obtain data suitable for the analysis program is basically performed by the spline interpolation method. However, for locations where the data interval is larger than the neighboring data group (before or after long-term measurement or sampling interval switching) A linear interpolation method is applied (see FIG. 3).

  For example, basically, in the observation data group sampled at 15-minute intervals, if there is a missing day for a certain period, the data interpolation for this missing day period is handled by linear interpolation, Other periods are handled by the spline interpolation method.

  In addition, the selection when the spline interpolation method and the linear interpolation method are selectively applied to observation data with periodicity, as described above, is the relative difference between the data interval of the target period and the neighboring data interval. In addition to the determination criterion, the period of the observation data can also be used as a criterion. For example, tides have the largest half-day (12-hour) cycle variation. Therefore, using the tide period as a guideline, a standard is set such as a quarter period of 3 hours, a half period of 6 hours, and a period of 12 hours. The spline interpolation method is applied, and the interpolation is performed by applying the linear interpolation method in a period in which the data interval is larger than the reference. This is because when the data interval or missing period is longer than the dominant period for the observed data, interpolation results that are close to the actual phenomenon cannot always be obtained even if interpolation is performed using the spline interpolation method. is there. Therefore, for such a period, by applying a linear interpolation method that is easy to process, it is possible to reduce the load required for the process, and it is possible to prevent the interpolation data from deviating from the actual phenomenon.

  FIG. 4 is a block diagram of the observation data processing system.

  1 is a groundwater pressure sensor that detects groundwater pressures at a plurality of depths as first observation data to be observed, and 2 to 4 are second observation data that affects the physical quantity of the first observation data. A sensor for detection, 2 is a tide level sensor, 3 is an atmospheric pressure sensor, and 4 is another sensor.

  Reference numeral 5 denotes an observation data recording device that acquires detection data output from each of the sensors 1 to 4 at a predetermined time interval and records it as observation data.

  6 is a small computer, 6a is a main body, 6b is an input device, 6c is a display device, and 6d is an external storage device. The small computer 6 incorporates a data processing program for reading observation data from the observation data recording device 5 and processing it.

  FIG. 5 is a flowchart showing a data processing procedure in the observation data processing system.

Step S1
The observation data recording device 5 records the detection data output from each of the sensors 1 to 4 in the observation data recording file in a file format determined as observation data at an arbitrary time interval. In order to make it easier to find necessary data later, this observation data recording file is stored with a unique observation data file name.

Step S2
The small computer 6 executes processing in the following procedure. First, observation data is read from the observation data recording device 5. Details will be described later.

Step S3
A display area is set for each observation data, and each observation data and edit data are displayed in a graph. This data display processing includes a processing function for enlarging / reducing the display range, switching the display unit, a processing function for enlarging / reducing the time axis of the display area in response to the enlargement / reduction of the display range, Provide a processing function to display observation data and edited data in contrast.

Step S4
Determine the necessity of data editing and branch processing. Here, the necessity of data editing is input by visually determining whether or not there is inappropriate data (data not suitable for analysis of observation data) included in the displayed data.

Step S5
This is a data editing process performed when inappropriate data is included, and selectively deletes or replaces (edits) the inappropriate data. Details will be described later.

Step S6
The necessity of data interpolation with respect to observation data or editing observation data is determined, and the process is branched. Data interpolation is necessary when the data format is not suitable for data analysis. Data interpolation is required when inappropriate data is deleted from the observation data for editing, or when the observation data includes missing periods or the data recording interval is too long.

Step S7
Interpolation data is created and the observation data for editing is interpolated (edited) into a data format suitable for analysis. Details will be described later.

Step S8
Executes analysis (tide analysis) processing of edited observation data or edited observation data edited by interpolation. This analysis processing itself is executed by arbitrarily applying a known analysis processing program as described above.

Step S9
A display area is set for each data, and the analysis results (data) and the observation data are arranged and displayed in a graph. This data display processing has a processing function for enlarging / reducing the display range and switching the display unit, and a processing function for enlarging / reducing the display area in conjunction with the time axis corresponding to the enlargement / reduction of the display range. Make it.

Step S10
Save the analysis result (data) in the analysis work data file.

Step S11
The analysis results (data) are integrated into an integrated file that integrates the analysis work data of multiple times.

  FIG. 6 is a flowchart showing a specific processing procedure for reading observation data in step S2.

Step S2001
In order to read each target observation data from the observation data recording device 5, an observation data file designation input screen is displayed.

Step S2002
The borehole name and observation data file name input from the input device 6b are acquired.

Step S2003
Read borehole information from observation data recording file.

Step S2004
Read observation data from specified observation data recording file.

Step S2005
Create an observation data file for editing to hold the read observation data and edit it.

Step S2006
The read observation data is written and saved in the observation data file for editing as observation data for editing.

  Specific operations related to steps S2001 to S2006 will be described with reference to FIG.

  First, the "new creation" dialog is designated from the display screen of the display device 6c of the small computer main body 6a, and an observation data file designation input screen [Fig. 7a]] for reading observation data is displayed.

  Next, input the “borehole name” and “observation data file name” to be read into the observation data file designation input screen using the input device 6b, and specify the observation data recording device 5 by clicking the button. The observation data is read from the recorded observation data recording file.

  FIG. 7 b) is an example of a display screen that displays pore water pressure and atmospheric pressure data among the observation data read from the designated observation data recording file.

  FIG. 8 is a flowchart showing a specific procedure of the data editing process in step S5.

  Observation data that is inconvenient to use for analysis is generated due to a defect in the data logger in each observation section of the borehole. This process is a process for deleting or replacing inconvenient data in the observation data for editing.

Step S5001
The edit process selection instruction input screen is displayed. In this embodiment, as an editing process to be selected, an inappropriate data deletion process and a replacement process are selectively displayed.

Step S5002
The process selection input information input from the input device 6b is acquired.

Step S5003
The process branches according to the acquired process selection input information.

Step S5004
When the deletion process is selected, a deletion data designation input screen is displayed. In order to specify deletion data, items for setting a data type (for example, water pressure data, atmospheric pressure data) and a data deletion period are displayed.

Step S5005
Delete data designation input information input from the input device 6b is acquired.

Step S5006
Delete the data corresponding to the deletion specified input information from the observation data for editing.

Step S5007
Write the edited observation data with the inappropriate data deleted to the file.

Step S5008
When the replacement process is selected, a replacement data designation input screen is displayed. In order to designate replacement data, items for setting a data type (for example, water pressure data, atmospheric pressure data), a data replacement period, and an observation data file name used for replacement are displayed.

Step S5009
Setting information input from the input device 6b is acquired.

Step S5010
Read replacement observation data used for replacement according to the setting information.

Step S5011
The specified observation data for editing is replaced with the observation data for replacement, and the process proceeds to step S5007 to write the edited observation data in which inappropriate data is replaced to the file.

  Specific operations related to steps S5001 to S5011 will be described with reference to FIG.

  FIG. 9A shows the pore water pressure in the observation data read from the specified observation data recording file in FIG. 7B.

  Here, the display data is visually checked to confirm whether observation data that is inconvenient for analysis is generated in each observation section of the borehole due to a defect in the data logger.

  If there is an abnormal value that does not normally occur in the observation data, or data that cannot be read is generated, the data is deleted for that part (cut and displayed on the screen of FIG. 9 a). Execute.

  On the other hand, if the observation data of a certain section could not be observed (recorded) at all due to a failure of the data logger, for example, if the atmospheric pressure data can be replaced with the observation data of the near-field pressure data, it could not be observed A replacement process for replacing the atmospheric pressure data of the section with atmospheric pressure data that can be replaced is executed.

  These data deletion and replacement operations can be specified by dragging the time axis (horizontal axis) section of a) in FIG. 9 with the mouse cursor, etc., and selecting the processing contents on the screen that appears by right-clicking. Data deletion (cutting on the screen) or replacement is selectively specified.

  After that, in the case of data deletion, a confirmation screen (partial deletion and display on the screen) as shown in FIG. 9B is displayed for the contents specified by the mouse cursor described above, and data deletion is performed by clicking the OK button. Is executed.

  On the other hand, in the case of data replacement, a confirmation file (replacement of data on the screen) as shown in FIG. 9C is displayed for the content specified by the mouse cursor described above, and the storage file storing the data to be replaced Specify the name and click the OK button to execute data replacement.

  FIG. 10 is a flowchart showing a specific procedure of the interpolation data creation process in step S7.

  Since the input data for tide analysis needs to be in the form of data at regular time intervals, 30-minute interval data (interpolation data) is created from the observation data for editing using the cubic spline interpolation method. Save to data file. As the interpolation data file, an atmospheric pressure data interpolation file and a water pressure data interpolation file for each observation section are created. Cubic spline interpolation for the long-term missing period of observation data for editing that has been lost due to missing or editing (deletion) of observation data due to malfunction of the data logger installed in each observation section of the borehole Since it is difficult to create preferable interpolation data by the method, interpolation data is generated by a linear interpolation method for a portion having a long missing period.

Step S7001
The process branches after confirming the presence of unprocessed editing observation data (water pressure data and atmospheric pressure data for each observation section).

Step S7002
If unprocessed editing observation data exists, the head data of the editing observation data is read.

Step S7003
The processing branches after confirming whether the next data exists.

Step S7004
Read next data.

Step S7005
The process branches after confirming whether the observation date and time of the two read data is separated from the threshold.

Step S7006
When the observation date / time of two data is separated from the threshold value, the two data are interpolated with linear interpolation data at a predetermined interval (for example, every 30 minutes).

Step S7007
In the confirmation in step S7003, if there is no next data, interpolation data at a predetermined time interval (for example, every 30 minutes) is created for the observation data for editing by the cubic spline interpolation method.

Step S7008
Write the interpolation data to the interpolation data file and save it.

  In steps S7001 to S7008, the missing observation data caused by the malfunction of the data logger and the long-term missing interpolation of the observation data for editing lost due to the editing process (deletion) are described. The data interpolation to be performed will be described next.

  Since the data for performing the analysis process is analyzed based on data at a predetermined time interval (for example, every 30 minutes), for example, as shown in FIG. When data is required, observation data for that date does not always exist.

  In that case, the data of the required date and time are interpolated from the observation data of two each before and after the required date and time. In the example of FIG. 11, an interpolation function is derived from a total of four of the previous observation data A and B and the subsequent observation data C and D, and the necessary date and time (in this case “2008/06/15 Interpolate the X value of 14:00 "

It is a graph which shows observation data. It is a graph which shows the edited observation data which interpolated the observation data shown in FIG. 1 using the spline interpolation method. It is a graph which shows the edited observation data which interpolated the observation data shown in FIG. 1 selectively using the spline interpolation method and the linear interpolation method. It is a block diagram of the observation data processing system which is an Example of this invention. It is a flowchart which shows the data processing procedure in the observation data processing system shown in FIG. It is a flowchart which shows the specific procedure of the observation data reading in the data processing procedure shown in FIG. FIG. 7 is an example of an input / output screen of a small computer related to reading of observation data of FIG. 6 and display of the read data. It is a flowchart which shows the specific procedure of the data editing process in the data processing procedure shown in FIG. FIG. 9 is a specific input / output screen example regarding the deletion processing and replacement processing of the data editing processing shown in FIG. 8. FIG. It is a flowchart which shows the specific procedure of the interpolation data creation process in the data processing procedure shown in FIG. It is an example of the data interpolation of the observation data with respect to the date and time required for tide analysis.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Groundwater pressure sensor, 2 ... Tidal sensor, 3 ... Atmospheric pressure sensor, 4 ... Other sensors, 5 ... Observation data recording device, 6 ... Small computer, 6a ... Main body, 6b ... Input device, 6c ... Display device, 6d ... External storage device.

Claims (4)

Along with the first observation data group obtained by observing long-term physical quantities, the fluctuation components (second observation data) of other physical quantities that affect the physical quantities of the first observation data are recorded, and each of the observation data The prescale processing is performed by selectively applying the spline interpolation method and the linear interpolation method so that the time scale is adapted to the analysis means, and then the physical quantity of the first observation data and the first amount are analyzed by a predetermined analysis means. The fluctuation component of the physical quantity of the second observation data affecting the physical quantity of the first observation data is analyzed, and the fluctuation component due to the physical quantity of the second observation data affecting the physical quantity of the first observation data is separated. In the observation data processing method for calculating the physical quantity data of one observation data and displaying or storing various data,
The first in the preprocessing, data interpolation of the second observation data, recording to have the first data interval in the period covered by the data interpolation of the second observation data, the following data interval of the data set in the vicinity An observation data processing method characterized in that it is carried out by a spline interpolation method when it is, and is carried out by a linear interpolation method when the data interval of neighboring data groups is exceeded. In claim 1,
The physical quantity of the first observation data is groundwater pressure at a plurality of underground depths,
The observation data processing method, wherein the physical quantity of the second observation data affecting the physical quantity is one or both of a tidal tide level and an atmospheric pressure physical quantity. In claim 1 or 2,
The observation data processing method, wherein the predetermined analysis means is a tide analysis program. In one of Claims 1-3,
An observation data processing method characterized by displaying the observation data before analysis and analysis data on a display so that the main factors of the fluctuation component can be visualized and grasped.

Images