Data calculation program, data calculation device, and data calculation method

Abstract

To provide a data calculation program, a data calculation device, and a data calculation method which can remove noise components from data showing a change of a state of a sample under contradicting conditions and can precisely calculate a desired physical quantity.SOLUTION: The data calculation device has a control unit for calculating characteristic data showing characteristics of a sample by using a dataset in which plural first data showing properties of the sample under a first condition and plural second data showing properties of the sample under a second condition contradictory to the first condition are alternately lined in chronological order. The control unit generates a differential operator vector representing a time derivative at an order one less than the total number of data in the dataset, based on a measurement time of each of all data in the dataset, and obtains characteristic data by dividing an inner product of the differential operator vector and a numerical vector in which all data in the dataset are arranged in chronological order, by the sum of absolute values of each element of the differential operator vector.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a data calculation program, a data calculation device, and a data calculation method for calculating data related to characteristics of a sample, which is an object to be analyzed. [Background technology]

[0002] In the past, when analyzing the characteristics of a sample, data showing changes in the state of the sample under contradictory conditions have been measured and used. However, the data showing changes in the state of the sample contains noise components caused by the measurement environment. Therefore, in all fields, a method for accurately removing noise components is desired. For example, when measuring the magnetic permeability of a magnetic material, it is necessary to extract the desired data from minute changes in the signal under different magnetic fields, so how to remove the noise components is the key.

[0003] Various methods, such as the inductance method and the yoke method, are used to measure the magnetic permeability of a magnetic material. For example, Patent Document 1 discloses a method of measuring S parameters (hereinafter referred to as set S parameters) of a jig containing a sample in a magnetic field below the saturation threshold and S parameters (hereinafter referred to as saturation S parameters) in a magnetic field equal to or higher than the saturation threshold, and substituting these into a predetermined calculation formula to obtain the magnetic permeability. This calculation formula is configured to obtain the magnetic permeability using the difference between one set S parameter and one saturation S parameter. However, in a method using only one set of S parameters, the accuracy of the data is directly linked to the measurement accuracy of the magnetic permeability. Therefore, a method using multiple sets of S parameters is desired to mitigate the effect of data variability and increase the reliability of the data. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-159886 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, in actual measurements, various low-frequency noises are added to the S-parameters, and the amount of change in this low-frequency noise is much larger than the difference between the S-parameters in the symmetrical magnetic field. Therefore, even if the simple average or standard deviation of multiple S-parameters in a given magnetic field is used, the noise components cannot be effectively removed. In this way, when calculating some physical quantity from the change in multiple data measured under two different conditions in an environment where the data are relatively noisy, the conventional method cannot effectively remove the noise, and the calculation accuracy cannot be improved.

[0006] The present disclosure has been made to solve the above-mentioned problems, and aims to provide a data calculation program, a data calculation device, and a data calculation method for removing noise components from data showing changes in the state of a sample under conflicting conditions and accurately determining a desired physical quantity. [Means for solving the problem]

[0007] A data calculation program according to one embodiment of the present disclosure causes a computer to function as: a vector generation means for generating a differential operator vector representing a time derivative at an order one less than the total number of data in a dataset, based on the measurement times of each of the first data and second data in a dataset in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition opposite to the first condition are alternately arranged in a chronological order; and a calculation processing means for dividing the inner product of the differential operator vector and a numerical vector in which the plurality of first data and the plurality of second data are arranged in a chronological order, by the sum of the absolute values ​​of each element of the differential operator vector, to obtain characteristic data indicating the characteristics of the sample.

[0008] A data calculation program according to one embodiment of the present disclosure causes a computer to function as a vector generating means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a chronological order, from first data and second data in a data set in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition opposite to the first condition are alternately arranged in a chronological order, by shifting the shift by an even number less than the total number of data in the data group, and for each of the extracted plurality of data groups, generating a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of each of the first data and the second data in the data group, and adding up the generated plurality of group differential operator vectors to obtain a differential operator vector, and a calculation processing means for dividing the inner product of the differential operator vector and a numerical vector in which a plurality of first data and a plurality of second data are arranged in a chronological order, by the sum of the absolute values ​​of each element of the differential operator vector to obtain characteristic data indicating the characteristics of the sample.

[0009] A data calculation program according to one embodiment of the present disclosure causes a computer to function as: a vector generation means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a chronological order, from first data and second data in a dataset in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition, by shifting the shift by an even number less than the total number of data in the data group; and for each of the extracted plurality of data groups, generating a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of each of the first data and second data in the data group; and a calculation processing means for obtaining group characteristic data for each of the plurality of data groups by dividing the inner product of the generated group differential operator vector and a numerical vector in which the first data and second data in the data group are arranged in chronological order, by the sum of the absolute values ​​of each element of the group differential operator vector, and taking the average of the obtained plurality of group characteristic data to obtain characteristic data indicating the characteristics of the sample.

[0010] A data calculation device according to one embodiment of the present disclosure includes a control unit that calculates characteristic data indicating characteristics of the sample using a data set in which a plurality of first data indicating properties of the sample under a first condition and a plurality of second data indicating properties of the sample under a second condition opposite to the first condition are arranged alternately in a chronological order, and the control unit includes a vector generating means that generates a differential operator vector representing a time derivative at an order one less than the total number of data in the data set, based on the measurement times of the first data and the second data in the data set, and a calculation processing means that obtains the characteristic data by dividing the inner product of the differential operator vector and a numerical vector in which the plurality of first data and the plurality of second data are arranged in a chronological order, by the sum of the absolute values ​​of each element of the differential operator vector.

[0011] A data calculation device according to an embodiment of the present disclosure includes a control unit that calculates characteristic data indicating characteristics of the sample using a dataset in which a plurality of first data indicating characteristics of the sample under a first condition and a plurality of second data indicating characteristics of the sample under a second condition opposite to the first condition are alternately arranged in a chronological order, and the control unit includes a vector generating means that extracts a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a chronological order from the first data and the second data in the dataset, by shifting the number of shifts by an even number less than the total number of pieces of data in the data group, and generates, for each of the extracted plurality of data groups, a group differential operator vector representing a time derivative at an order one less than the total number of pieces of data in the data group, based on the measurement times of each of the first data and the second data in the data group, and adds up the generated plurality of group differential operator vectors to obtain a differential operator vector; and a calculation processing means that divides an inner product of the differential operator vector and a numerical vector in which a plurality of first data and a plurality of second data are arranged in a chronological order by a sum of absolute values ​​of each element of the differential operator vector to obtain characteristic data indicating characteristics of the sample.

[0012] A data calculation device according to one embodiment of the present disclosure includes a control unit that calculates characteristic data indicating characteristics of the sample using a dataset in which a plurality of first data indicating properties of the sample under a first condition and a plurality of second data indicating properties of the sample under a second condition opposite to the first condition are arranged alternately along a time series, and the control unit includes a vector generation means that extracts a plurality of data groups, each of which is a combination of two or more pieces of data arranged along a time series from the first data and the second data in the dataset, by shifting the number of shifts being an even number less than the total number of data in the data group, and generates, for each of the extracted plurality of data groups, a group differential operator vector that represents a time derivative at an order one less than the total number of data in the data group, based on the measurement times of each of the first data and the second data in the data group; and a calculation processing means that, for each of the plurality of data groups, obtains group characteristic data by dividing the inner product of the group differential operator vector generated by the vector generation means and a numerical vector in which the first data and the second data in the data group are arranged along a time series, by the sum of the absolute values ​​of each element of the group differential operator vector, and obtains the characteristic data by taking an average of the obtained plurality of group characteristic data.

[0013] A data calculation method according to one embodiment of the present disclosure includes one or more processors that generate a differential operator vector representing a time derivative at an order one less than the total number of data in a dataset, based on measurement times of each of a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition, in a dataset in which the first data and the second data are arranged alternately in a chronological order, and divide the inner product of the differential operator vector and a numerical vector in which the plurality of first data and the plurality of second data are arranged in a chronological order by the sum of the absolute values ​​of each element of the differential operator vector to obtain characteristic data indicating the characteristics of the sample.

[0014] A data calculation method according to one aspect of the present disclosure includes one or more processors extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a chronological order, from first data and second data in a data set in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition opposite to the first condition are alternately arranged in a chronological order, by shifting the data by an even number less than the total number of data in the data group, generating, for each of the plurality of data groups, a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of each of the first data and the second data in the data group, obtaining a differential operator vector by adding up the plurality of group differential operator vectors thus generated, and dividing the inner product of the obtained differential operator vector and a numerical vector in which the plurality of first data and the plurality of second data are arranged in a chronological order by the sum of the absolute values ​​of each element of the differential operator vector, to obtain characteristic data indicating characteristics of the sample.

[0015] A data calculation method according to one embodiment of the present disclosure includes one or more processors extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a chronological order, from first data and second data in a dataset in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition, by shifting the data by an even number less than the total number of data in the data groups, generating, for each of the plurality of data groups, a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of each of the first data and second data in the data group, generating group characteristic data by dividing the inner product of the generated group differential operator vector and a numerical vector in which the first data and second data in the data group are arranged in chronological order, by the sum of the absolute values ​​of each element of the group differential operator vector, and taking an average of the group characteristic data for each of the plurality of data groups to obtain characteristic data indicating the properties of the sample. Effect of the Invention

[0016] In the present invention, for a data set in which first data and second data indicating the properties of a sample under opposing conditions are arranged alternately in a time series, an operation is performed to divide the inner product of a differential operator vector based on the measurement time of each data and a numerical vector in which each data is arranged by the sum of the absolute values ​​of each element of the differential operator vector, thereby obtaining characteristic data indicating the properties of the sample. Since the differential operator vector indicates the differentiation of each data in the data set at an order that is one less than the total number of each data, by taking the inner product with the numerical vector, it is possible to accurately remove noise components contained in each data in the data set. Therefore, by this method, it is possible to eliminate many noise components contained in multiple data indicating changes in the state of the sample under opposing conditions and accurately obtain a desired physical quantity. [Brief description of the drawings]

[0017] [Figure 1] 1 is a configuration diagram illustrating a data calculation system according to a first embodiment of the present invention. [Diagram 2] FIG. 2 is a plan view showing a schematic configuration of the jig shown in FIG. [Diagram 3] 3 is a schematic cross-sectional view illustrating a cross section taken along line AA in FIG. 2. [Figure 4] FIG. 3 is an exploded perspective view showing a main part of the jig in FIG. 2. [Diagram 5] 2 is a functional block diagram illustrating a functional configuration of the data processing device of FIG. 1. [Figure 6] 6 is a flowchart illustrating an operation of the data calculation device of FIG. 5 regarding acquisition of first data and second data in a standard mode. [Figure 7] 6 is a flowchart illustrating an operation of the data calculation device of FIG. 5 regarding acquisition of first data and second data in high-speed mode. [Figure 8] 6 is an example of a graph in which all the data in the data set of FIG. 5 is plotted on a coordinate system. [Figure 9]FIG. 6 is an explanatory diagram illustrating a schematic example of a graph (a) in which all data in the data set in FIG. 5 is plotted on a coordinate system, the graph being divided into true value components (b) and noise components (c). [Figure 10] 4 is a flowchart illustrating a process flow for calculating characteristic data from a data set in the data calculation method according to the first embodiment. [Figure 11] FIG. 6 is an explanatory diagram illustrating each data item, a graph relating to the polynomial, and amplitudes, etc., on the assumption that the noise components contained in the data (total number of data items=n+1) in the data set of FIG. 5 can be expressed by an n-1 degree polynomial. [Figure 12] 11 is an explanatory diagram illustrating an example of a difference in differential operator when the number of data items to be differentiated is changed. FIG. [Figure 13] FIG. 11 is an explanatory diagram illustrating an example of each process of a data calculation method according to a second embodiment of the present invention. [Figure 14] FIG. 11 is an explanatory diagram illustrating roughly another example of each process of the data calculation method according to the second embodiment. [Figure 15] 13 is a flowchart illustrating a process flow for calculating characteristic data from a data set in the data calculation method according to the second embodiment. [Figure 16] 13 is a flowchart illustrating a process flow for calculating characteristic data from a data set in a data calculation method according to a modified example of the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Embodiment 1 An example of the overall configuration of a data calculation device 10 and a data calculation system 100 according to the first embodiment of the present invention will be described with reference to Fig. 1 to Fig. 5. In each figure, some of the reference numerals may be omitted to avoid the complexity of the drawings. First, an example of the overall configuration of the data calculation system 100 will be described with reference to Fig. 1.

[0019] As shown in Fig. 1, the data calculation system 100 includes a data calculation device 10, a signal processing device 20, an amplifier 25, a magnetic field control device 30, a pair of electromagnets 35, and a jig 50. The jig 50 is for measuring the magnetic permeability of a magnetic body. That is, Fig. 1 illustrates a magnetic permeability measurement system as the data calculation system 100.

[0020] The data calculation device 10 is configured, for example, by a PC (Personal Computer) or the like, and controls the signal processing device 20 and the magnetic field control device 30. The PC includes tablet PCs, notebook PCs, desktop PCs, and the like. The data calculation device 10 receives S parameters (scattering parameters) as data indicating the properties of the sample 85 from the signal processing device 20, and calculates the magnetic permeability (complex magnetic permeability). The data calculation device 10 may be configured by a cloud server based on cloud computing, an on-premise physical server, or a system combining these. In this case, the data calculation system 100 may have an information terminal that cooperates with the data calculation device 10 and relays communication between the data calculation device 10 and the signal processing device 20 and the magnetic field control device 30.

[0021] Here, the S-parameter is one of the circuit network parameters used to express the characteristics of high-frequency electronic circuits and high-frequency electronic components. The S-parameter is divided into four characteristics, S11, S22, S21, and S12, and the permeability of the sample 85 is calculated using the S-parameter indicating the signal transmission characteristic (S21), that is, the transmission characteristic from the input side to the output side of the jig 50. The S-parameter in the present embodiment 1 corresponds to the ratio of the magnetization signal input to the signal processing device 20 to the excitation signal output by the signal processing device 20. The magnetization signal includes a signal output by the interaction of the excitation signal with the magnetization of the sample 85, a signal of the excitation signal that transmits through the jig 50, and a noise component.

[0022] The signal processing device 20 generates a predetermined excitation signal and outputs the generated excitation signal to the input waveguide 60 of the jig 50. The signal processing device 20 has a function of controlling the power and frequency of the generated excitation signal, and can generate a high-frequency excitation signal. In addition, a magnetization signal amplified by an amplifier 25 is input to the signal processing device 20. The signal processing device 20 measures the S-parameters using the output excitation signal and the magnetization signal input corresponding to the output excitation signal.

[0023] The signal processing device 20 is composed of, for example, a vector network analyzer (VNA). The amplifier 25 is composed of, for example, a low noise amplifier (LNA) and has a function of amplifying a weak signal to a required power. However, if the magnetization signal sent from the detection unit 72b is at a level that does not require amplification, the data calculation system 100 may be configured without providing the amplifier 25. The data calculation system 100 may be configured without providing the amplifier 25 by providing the signal processing device 20 with a function equivalent to that of the amplifier 25.

[0024] The magnetic field control device 30 controls the pair of electromagnets 35 to generate a magnetic field to be applied to the sample 85 in the jig 50 arranged between the pair of electromagnets 35. The magnitude of the magnetic field applied by the pair of electromagnets 35 is appropriately set and changed depending on the type, size, shape, etc. of the sample 85. The data calculation system 100 may be configured to have a pair of permanent magnets instead of the pair of electromagnets 35, and in this case, for example, a plurality of permanent magnets having different residual magnetic flux densities may be appropriately combined and used.

[0025] Next, a specific configuration example of the jig 50 will be described with reference to Figs. 2 to 4. As shown in Figs. 2 to 4, the jig 50 has an input waveguide 60 and an output waveguide 70. The input waveguide 60 transmits an excitation signal generated in the signal processing device 20. In the jig 50, an excitation magnetic field generated by the excitation signal output from the signal processing device 20 is applied to the sample 85, and the applied excitation magnetic field causes a movement in the magnetization of the sample 85, causing a change in magnetic flux. The output waveguide 70 transmits a magnetization signal indicating the movement of the magnetization of the sample 85 and a back electromotive force induced by a change in magnetic flux caused by the excitation signal flowing through the input waveguide 60. In the jig 50, the input waveguide 60 and the output waveguide 70 are electrically insulated from each other, but are magnetically coupled to each other to form a transformer.

[0026] As shown in Fig. 2 and Fig. 4, the input waveguide 60 and the output waveguide 70 each have a coplanar line structure. That is, the input waveguide 60 has a signal line 62 and a ground line (ground line) 63 on a substrate 61 made of a dielectric material. The signal line 62 is disposed in the center of the substrate 61, and the ground line 63 is formed so as to sandwich the signal line 62 from both sides. The output waveguide 70 has a signal line 72 and a ground line (ground line) 73 on a substrate 71 made of a dielectric material. The signal line 72 is disposed in the center of the substrate 71, and the ground line 73 is formed so as to sandwich the signal line 72 from both sides.

[0027] More specifically, the signal line 62 extends from the input section 62a, where the excitation signal is input, to the short-circuit end 62s. The short-circuit end 62s is a portion that is coupled to the ground line 63. The portion of the signal line 62 at the short-circuit end 62s becomes the excitation section 62b that generates a magnetic field to be applied to the sample 85. The sample 85 is placed on the excitation section 62b. The excitation section 62b extends with a predetermined width along the transmission direction in which the excitation signal is transmitted. The transmission direction corresponds to the positive X-axis direction in each figure.

[0028] The width of the signal line 62 gradually narrows from the input section 62a side toward the excitation section 62b. Since the signal line 91a of the coaxial cable 91 is connected to the signal line 62, the width of the signal line 62 on the coaxial cable 91 side is wider than the width of the excitation section 62b. Although each figure shows an example in which the width of the signal line 62 narrows in two stages toward the excitation section 62b, the present invention is not limited to this, and the signal line 62 can have various shapes within a range in which the signal line 91a can be connected. For example, the width of the signal line 62 may be narrowed in one stage toward the excitation section 62b, or may be narrowed in three or more stages. The signal line 62 may be formed so as to gradually narrow toward the excitation section 62b. The ground lines 63 are provided on both sides of the signal line 62 on the input section 63a side, and the ground lines 91b of the coaxial cable 91 are connected to the ground lines 63, respectively. A gap of a predetermined width is provided between the side of the ground line 63 and the side of the signal line 62. The ground line 63 is coupled from both sides to the short-circuited end 62s of the magnetic excitation portion 62b.

[0029] Similarly, the signal line 72 extends from the output section 72a side where the magnetization signal is output to the short-circuit end 72s. The short-circuit end 72s is a portion that is coupled to the ground line 73. The short-circuit end 72s portion of the signal line 72 serves as a detection section 72b that detects the back electromotive force generated by the application of a magnetic field to the sample 85 and the excitation signal flowing through the input waveguide 60. The detection section 72b is disposed parallel to and facing the excitation section 62b, and the sample 85 is disposed between the excitation section 62b and the detection section 72b. The detection section 72b extends along the transmission direction with a predetermined width. The transmission direction is also the direction in which the magnetization signal based on the back electromotive force detected by the detection section 72b is transmitted.

[0030] The signal line 72 is formed so that its width expands from the detection section 72b toward the output section 72a. Since the signal line 92a of the coaxial cable 92 is connected to the signal line 72, the width of the signal line 72 is wider than the width of the detection section 72b at that portion. Although each figure shows an example in which the width of the signal line 72 narrows in two stages toward the detection section 72b, the present invention is not limited to this, and the signal line 72 can have various shapes within a range in which the signal line 92a can be connected. For example, the signal line 72 may have a width that narrows in one stage toward the detection section 72b, or may have a width that narrows in three or more stages. The signal line 72 may also be formed so that it gradually narrows toward the detection section 72b. The ground lines 73 are provided on both sides of the signal line 72 on the input section 73a side, and the ground lines 92b of the coaxial cable 92 are connected to the ground lines 73, respectively. A gap of a predetermined width is provided between the side of the ground line 73 and the side of the signal line 72. The ground line 73 is coupled to both sides of the short-circuited end 72s of the magnetic excitation portion 62b.

[0031] From the viewpoint of improving the measurement sensitivity, the length of the excitation unit 62b along the transmission direction (length in the X-axis direction) is preferably 10 mm or less. Similarly, from the viewpoint of improving the measurement sensitivity, the length of the detection unit 72b along the transmission direction is preferably 10 mm or less. From the viewpoint of arranging the sample 85 and facilitating the overlapping of the excitation unit 62b and the detection unit 72b, the length of each of the excitation unit 62b and the detection unit 72b along the transmission direction is preferably 1 mm or more. However, the length of the excitation unit 62b and the detection unit 72b along the transmission direction may be 10 nm or more and less than 100 nm. If a microfabrication technique (e.g., a semiconductor process technique) is used in the fabrication of the excitation unit 62b and the detection unit 72b and the arrangement of the sample, the magnetic vibration (magnetization fluctuation) of the extremely small sample 85 having a size of several nm can be detected with high sensitivity, and the magnetic permeability can be measured with high accuracy.

[0032] The signal line 62, the signal line 72, the ground line 63, and the ground line 73 are made of a conductive material such as copper, and are formed to have a thickness of, for example, 35 μm. The substrate 61 and the substrate 71 are made of, for example, a glass cloth-based epoxy resin laminate (glass epoxy), and are formed to have a thickness of, for example, 0.8 μm. However, the thickness of the substrate 61 and the substrate 71 can be changed as appropriate. The input waveguide 60 and the output waveguide 70 are formed to have a transfer impedance (characteristic impedance) of, for example, 50 Ω. As illustrated in FIGS. 2 to 4, the jig 50 may have a nonmagnetic insulating layer 80 between the excitation unit 62b and the detection unit 72b, separating them by a predetermined distance, and the sample 85 may be fixed on the excitation unit 62b by the nonmagnetic insulating layer 80.

[0033] Next, a functional configuration example of the data calculation device 10 will be described with reference to Fig. 5. As shown in Fig. 5, the data calculation device 10 has a communication unit 11, a control unit 12, a storage unit 13, an input unit 14, and a display unit 15. The communication unit 11 is an interface for the control unit 12 to perform wired or wireless communication with an external device.

[0034] The storage unit 13 stores various information in addition to operation programs of the control unit 12, such as a data calculation program 13p. The data calculation program 13p is software that cooperates with a calculation device, such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), to realize various functions. The storage unit 13 is composed of a RAM (Random Access Memory), a ROM (Read Only Memory), a PROM (Programmable ROM) such as a flash memory, an SSD (Solid State Drive), or an HDD (Hard Disk Drive), etc.

[0035] The input unit 14 includes, for example, a keyboard and a pointing device such as a mouse or a trackball. The input unit 14 accepts an input operation by a user and transmits an operation signal according to the accepted content to the control unit 12. The display unit 15 includes, for example, a liquid crystal display (LCD) and displays various information according to an instruction from the control unit 12. The data calculation device 10 may have a touch panel including a display panel that displays characters or images, and a detection means that is stacked on the display panel and detects a touch operation, instead of the input unit 14 and the display unit 15. The data calculation device 10 may have both a touch panel and the input unit 14.

[0036] The control unit 12 obtains characteristic data indicating the characteristics of the sample 85. The control unit 12 of the present embodiment 1 calculates characteristic data indicating the characteristics of the sample 85 using a data set D in which a plurality of first data indicating the properties of the sample 85 under a first condition and a plurality of second data indicating the properties of the sample 85 under a second condition are arranged alternately in a chronological order. The number of first data and second data in the data set D may be the same or may differ by one. When the first data and the second data are arranged in a chronological order, the data set D may start with the first data or may start with the second data.

[0037] The first condition and the second condition are contradictory conditions. In the first embodiment, the first condition is that a magnetic field less than the saturation threshold is applied to the sample 85, and the second condition is that a magnetic field equal to or greater than the saturation threshold is applied to the sample 85. The first data are S parameters (set S parameters) that indicate the transmission characteristics of the jig 50 containing the sample 85 when a magnetic field less than the saturation threshold is applied, and the second data are S parameters (saturation S parameters) that indicate the transmission characteristics of the jig 50 containing the sample 85 when a magnetic field equal to or greater than the saturation threshold is applied.

[0038] The saturation threshold is the value of the magnetic field at which the magnetization of a magnetic material becomes saturated. In other words, if a magnetic material is placed in a magnetic field and the magnetization is observed, the magnetization increases as the strength of the magnetic field increases, and above a certain magnetic field, it becomes a constant value. This state is called magnetic saturation, and the magnitude of magnetization at this point is called saturation magnetization. Magnetization here is a physical quantity that expresses the degree of magnetization of a magnetic material. Hereinafter, a magnetic field below the saturation threshold is also called the set magnetic field, and when these are referred to collectively, the symbol "H M " is used. A magnetic field above the saturation threshold is also called a saturation magnetic field, and is represented by the symbol "H S " is used. The set magnetic field and the saturation magnetic field are set in advance to match the sample 85. In the data set D, the set S-parameters and the saturation S-parameters are arranged alternately.

[0039] In the first embodiment, a plurality of set magnetic fields H M Since it is assumed that j (j = 1, 2, ..., m / m is an arbitrary natural number) is set, in Figure 5, each set magnetic field H M In this example, the data set D corresponding to each of the magnetic fields H j is stored in the storage unit 13. However, since it is possible to calculate the characteristic data with only one data set D, the magnetic field H M Only one may be set.

[0040] The control unit 12 has an input processing means 12a, an output processing means 12b, a magnetic field processing means 12c, an acquisition processing means 12d, a vector generating means 12e, and an arithmetic processing means 12f. The input processing means 12a accepts a user's operation via the input unit 14, and outputs an operation signal indicating the content of the operation to the output processing means 12b. In addition, the input processing means 12a stores new information in the storage unit 13 or updates the information in the storage unit 13 according to the content of the user's input. The output processing means 12b reads out necessary information from the storage unit 13 according to the input operation signal, and displays various screens, images, etc. based on the read information on the display unit 15.

[0041] The magnetic field processing means 12c generates a set magnetic field H M and the saturation magnetic field H SThe data calculation device 10 has two modes (standard mode and high-speed mode) for acquiring S parameters. The magnetic field processing means 12c applies a plurality of set magnetic fields H M j The signal to indicate the application of each and the saturation magnetic field H S and a signal instructing application of the magnetic field to the magnetic field control device 30.

[0042] The acquisition processing means 12d acquires various data from an external device via the communication unit 11, and stores the acquired data in the storage unit 13. The acquisition processing means 12d acquires a plurality of first data and a plurality of second data together with information on the respective measurement times from the signal processing device 20. That is, the acquisition processing means 12d acquires a plurality of set magnetic fields H M The set S parameters corresponding to each j are obtained as the first data, and the saturation magnetic field H S The acquisition processing means 12d acquires a plurality of saturation S parameters corresponding to the plurality of set magnetic fields H M A data set D is constructed in which a plurality of first data and a plurality of second data are arranged in time series in association with each other for each j, and is stored in the storage unit 13.

[0043] Each setting magnetic field H M For each of the first and second data, in both the standard and high-speed modes, the measurement of the first data and the measurement of the second data are performed alternately, so that when the first data and the second data in the data set D are arranged in order of measurement time, the first data and the second data are arranged alternately one by one. The acquisition processing means 12d may have a function of instructing the signal processing device 20 to output an excitation signal.

[0044] The vector generating means 12e generates a differential operator vector [a] representing a time differential at an order one less than the total number of data in the data set D based on the measurement time of each of all data in the data set D. All data in the data set D refers to a plurality of first data and a plurality of second data included in one data set D. The total number of data in the data set D refers to the sum of the total number of first data and the total number of second data included in one data set D. The calculation processing means 12f obtains characteristic data by dividing the inner product of the differential operator vector [a] generated by the vector generating means 12e and the numerical vector [y] by the sum of the absolute values ​​of each element of the differential operator vector [a]. The numerical vector [y] is a vector in which a plurality of first data and a plurality of second data in the data set D are arranged in a chronological order.

[0045] Specific functional configurations of the vector generating means 12e and the arithmetic processing means 12f will be described later with reference to Fig. 8, Fig. 9, and formulas (1) to (9). The control unit 12 is configured with one or more processors. That is, the control unit 12 is configured with a calculation processing device such as a CPU or a GPU, and various functions of the control unit 12 are realized by the calculation processing device and a data calculation program 13p.

[0046] Next, a method of acquiring S parameters as the first data and the second data from among the data calculation methods in the first embodiment will be described with reference to the flowcharts of Fig. 6 and Fig. 7. Fig. 6 and Fig. 7 show an example of operation in which the number of first data and the number of second data constituting a data set D are equal. Therefore, for one data set D, the total number of first data and second data is set to 2N, and N corresponding to the number of first data and the number of second data is called a prescribed number. The prescribed number N is preset in the storage unit 13 and can be changed as appropriate.

[0047] First, the flow of operations of the data calculation device 10 and its associated devices related to acquisition of S parameters in standard mode will be described with reference to FIG. MA plurality of first data (set S parameters) regarding the jig 50 containing the sample 85 when a saturation magnetic field H S A data set D in which a plurality of second data (saturation S parameters) for the jig 50 containing the sample 85 when the sample 85 is placed in the jig 50 and a plurality of second data (saturation S parameters) for the jig 50 are arranged alternately in a time series is generated by applying a plurality of different set magnetic fields H M j (j=1, 2, ..., m / number of groups m is an arbitrary natural number) is constructed. In other words, the storage unit 13 stores the number of groups (m) of different set magnetic fields H M It is assumed that j is registered. The data set D is constructed by the acquisition processing means 12d, and is used for calculation processing by the vector generation means 12e and the calculation processing means 12f.

[0048] The control unit 12 starts the process of acquiring the S parameters in response to a user's operation on the input unit 14 or at a preset timing. That is, the magnetic field processing means 12c outputs a set magnetic field H M The magnetic field control device 30 instructs the application of a set magnetic field H M 1 (step S101). After that, in step S101, the magnetic field processing means 12c applies a set magnetic field H M The magnetic field control device 30 is instructed to apply j (j=1, 2, . . . , m).

[0049] The magnetic field processing means 12c is a set magnetic field H M The magnetic field processing means 12c waits until j satisfies the convergence condition (step S102 / No). The convergence condition is a condition that indicates that the magnetic field has stabilized, that is, that the magnetic field has converged to a certain value. The magnetic field processing means 12c may set the lapse of a preset standby time as the convergence condition, or may determine the convergence of the actual magnetic field in cooperation with a sensing device that measures the magnetic field.

[0050] The magnetic field processing means 12c generates a set magnetic field H MWhen j satisfies the convergence condition (step S102 / Yes), the acquisition command is output to the acquisition processing means 12d. In response to the acquisition command, the acquisition processing means 12d instructs the signal processing device 20 to output an excitation signal. The signal processing device 20 outputs an excitation signal to the jig 50, measures the complex amplitude of the magnetization signal input in response to the output excitation signal, and measures the set S-parameter as the first data based on the excitation signal and the magnetization signal. At that time, the signal processing device 20 records the measurement time together with the set S-parameter. The signal processing device 20 transmits a measurement completion signal to the data calculation device 10 indicating that the measurement of the set S-parameter has been completed. The signal processing device 20 may transmit the measured set S-parameter and its measurement time together with the measurement completion signal to the data calculation device 10 (step S103).

[0051] When the magnetic field processing means 12c receives the measurement completion signal, the magnetic field control device 30 generates a saturation magnetic field H S The magnetic field control device 30 instructs the application of a saturation magnetic field H S The magnetic field processing means 12c applies the saturation magnetic field H S satisfies the convergence condition (step S105 / No), and the saturation magnetic field H S satisfies the convergence condition (step S105 / Yes), the acquisition command is output to the acquisition processing means 12d. Then, the acquisition processing means 12d instructs the signal processing device 20 to output an excitation signal. The signal processing device 20 outputs the excitation signal to the jig 50, measures the complex amplitude of the magnetization signal input according to the output excitation signal, and measures the saturated S-parameter as the second data based on the excitation signal and the magnetization signal. At that time, the signal processing device 20 records the measurement time together with the saturated S-parameter. The signal processing device 20 transmits a measurement completion signal to the data calculation device 10 indicating that the measurement of the saturated S-parameter is completed. The signal processing device 20 may transmit the measured saturated S-parameter and its measurement time together with the measurement completion signal to the data calculation device 10 (step S106).

[0052] Next, the acquisition processing means 12d judges whether the number of measurements k has reached a specified number N. In the standard mode, the number of measurements k is equal to one set magnetic field H M The number of times k is measured is the number of times that the set S-parameter corresponding to j and the saturated S-parameter associated with it are measured. The saturated S-parameter associated with the set S-parameter is the saturated S-parameter that is measured alternately with the set S-parameter. The number of measurements k is initially set to "1" (step S107).

[0053] The acquisition processing means 12d acquires one set magnetic field H M The signal processing device 20 alternately measures a combination of the set S-parameter corresponding to j and the saturation S-parameter associated therewith a specified number N times. If the number of measurements k is less than the specified number N (step S107 / No), the acquisition processing means 12d increments the number of measurements k. For example, when the first measurement of the set S-parameter and the saturation S-parameter is completed, the acquisition processing means 12d updates the number of measurements k to "2" (step S108). Then, the acquisition processing means 12d measures the set magnetic field H M The magnetic field processing means 12c outputs a sustain request signal requesting application of j (j=1, 2, . . . , m) to the magnetic field processing means 12c (step S101).

[0054] Here, the number of S parameters acquired by the control unit 12 in the series of processes from steps S101 to S108, that is, the number of S parameters included in one data set D, is defined as n+1 (=2N), and the symbol "y" is used for the S parameter. Then, all the S parameters in the data set D are expressed as "y1, y2, y3, ..., y n ,y n+1 " and each setting S-parameter can be expressed as "y1, y3, ..., y 2N-1 " and each saturated S-parameter is expressed as "y2, y4, ..., y 2N The subscripts on these "y"s correspond to the measurement order, or time series.

[0055] That is, the control unit 12 executes a series of processes in steps S101 to S108 until the number of measurements k reaches the specified number N, and sets the set S parameters (y1, y3, ..., y 2N-1 ) and saturation S-parameters (y2, y4, …, y 2N ) is acquired. The acquisition processing means 12d may acquire a certain amount of setting S-parameters and saturation S-parameters and their respective measurement times collectively from the signal processing device 20. The acquisition processing means 12d generates a data set D in which the S-parameters acquired in the series of processes in steps S101 to S108 correspond to the respective measurement times, and stores the data set D in the storage unit 13 in a readable state.

[0056] When the number of measurements k reaches the specified number N (step S108 / Yes), the acquisition processing means 12d judges whether the setting number j reaches the number of sets m (step S109). If the setting number j is less than the number of sets m (step S109 / No), the acquisition processing means 12d increments the setting number j (step S110). The control unit 12 calculates the set magnetic field H corresponding to the updated setting number j. M For j, the series of processes from steps S101 to S108 are repeatedly executed until the number of measurements k reaches the specified number N. While the setting number j is less than the number of sets m, the control unit 12 repeatedly executes the series of processes from steps S101 to S108, step S109 / No, and step S110. When the setting number j reaches the number of sets m (step S109 / Yes), the control unit 12 ends the process of acquiring S parameters in the standard mode.

[0057] Next, the flow of operations of the data calculation device 10 and its associated devices related to the acquisition of S parameters in the high-speed mode will be described with reference to FIG. M j is registered. The set magnetic field H M The larger the value of the setting number j, the larger the value of the magnetic field. Contents similar to those in the standard mode will be omitted or simplified as appropriate.

[0058] The magnetic field processing means 12c generates a set magnetic field H MThe set magnetic field H, which is the smallest magnetic field among j (j=1, 2, …, m) M The magnetic field control device 30 is instructed to apply the set magnetic field H 1 to the sample 85 in the jig 20. M 1 (step S201). After that, in step S201, the magnetic field processing means 12c applies the set magnetic field H M The magnetic field control device 30 is instructed to apply j (j=1, 2, . . . , m) in sequence.

[0059] Next, the magnetic field processing means 12c outputs an acquisition command to the acquisition processing means 12d, and the acquisition processing means 12d instructs the signal processing device 20 to output an excitation signal in response to the acquisition command. The signal processing device 20 outputs an excitation signal to the jig 50, measures the complex amplitude of the magnetization signal input in response to the output excitation signal, and measures the set S-parameter as the first data based on the excitation signal and the magnetization signal. At that time, the signal processing device 20 records the measurement time together with the set S-parameter. The signal processing device 20 transmits a measurement completion signal to the data calculation device 10 indicating that the measurement of the set S-parameter has been completed. The signal processing device 20 may transmit the measured set S-parameter and its measurement time together with the measurement completion signal to the data calculation device 10 (step S202).

[0060] The acquisition processing means 12d judges whether the setting number j has reached the number of sets m (step S203). If the setting number j is less than the number of sets m (step S203 / No), the acquisition processing means 12d increments the setting number j in the storage unit 13 and outputs a magnetic field change command to the magnetic field processing means 12c (step S204). In response to the magnetic field change command, the magnetic field processing means 12c generates the next largest set magnetic field H M j is applied to the sample 85 via the magnetic field control device 30 (step S201).

[0061] The control unit 12 repeatedly executes a series of processes in steps S201 to S204 until the setting number j reaches the number of sets m. When the setting number j reaches the number of sets m (step S203 / Yes), the magnetic field processing means 12c controls the magnetic field control device 30 to generate a saturation magnetic field H SThe magnetic field control device 30 instructs the application of a saturation magnetic field H S is applied (step S205).

[0062] Next, the magnetic field processing means 12c outputs an acquisition command to the acquisition processing means 12d, and the acquisition processing means 12d instructs the signal processing device 20 to output an excitation signal in response to the acquisition command. The signal processing device 20 outputs an excitation signal to the jig 50, measures the complex amplitude of a magnetization signal input in response to the output excitation signal, and measures saturated S-parameters as second data based on the excitation signal and the magnetization signal. At that time, the signal processing device 20 records the measurement time together with the saturated S-parameters. The signal processing device 20 transmits a measurement completion signal to the data calculation device 10 indicating that the measurement of the saturated S-parameters has been completed (step S206).

[0063] Next, the acquisition processing means 12d judges whether the number of measurements k has reached a specified number N (step S207). In the high-speed mode, the number of measurements k is equal to the number of measurements m of the set magnetic field H M j, and the number of times that the set S-parameters corresponding to them and the saturation S-parameters associated with them have been measured, that is, the number of times that the series of processes from step S201 to S207 have been performed. M The signal processing device 20 measures the combination of the set S-parameter corresponding to each of j and the saturation S-parameter associated therewith a specified number N times.

[0064] If the number of measurements k is less than the specified number N (step S207 / No), the acquisition processing means 12d increments the number of measurements k (step S208) and outputs a reset command to the magnetic field processing means 12c (step S209). In response to the reset command, the magnetic field processing means 12c returns the setting number j to 1 and instructs the magnetic field control device 30 to set the magnetic field H MIn step S201, the control unit 12 instructs application of voltage 1. That is, the control unit 12 executes a series of processes of steps S201 to S206, S207 / No, step S208, and step S209 until the number of measurements k reaches the specified number N, and when the number of measurements k reaches the specified number N, ends the process of acquiring S parameters in the high-speed mode.

[0065] The acquisition processing means 12d acquires a plurality of set magnetic fields H M A data set D is constructed for each of the samples j. However, in the high-speed mode, the control unit 12 applies a magnetic field to the sample 85 at a minimum magnetic field H M Set magnetic field H from 1 to maximum M m and then the saturation magnetic field H S The process repeats a series of steps of changing the magnetic field to the set magnetic field H M Construct a dataset D for each j.

[0066] The above operation description is an example of the processing of the control unit 12 in the standard mode or the high-speed mode, and the processing contents of the control unit 12 may be appropriately changed within a range in which the data set D can be constructed. The data calculation device 10 may be configured to generate only one data set D. In the standard mode, the control unit 12 can measure the S-parameters at regular intervals by adjusting the convergence conditions, but in the high-speed mode, the intervals of the measurement times of the S-parameters become non-equidistant. When constructing a data set D in which the first data is more than the second data, the control unit 12 only needs to perform one less measurement process of the second data, and when constructing a data set D in which the second data is more than the first data, the control unit 12 only needs to perform one less measurement process of the first data.

[0067] Next, with reference to Fig. 8, Fig. 9 and formulas (1) to (9), the principle of removing noise components from data set D that is constructed by measuring in standard mode (Fig. 6) or high-speed mode (Fig. 7) will be described. Note that in each formula, an arrow symbol (→) is attached to the vector, but this is omitted in the specification, and a symbol is enclosed in [ ] to indicate that it is a vector. Here, the n+1 S-parameters (n is any natural number) of data set D are defined as "y1, y2, ..., y n+1 " and these measurement times are expressed as "t1, t2, ..., t n+1 The combination of S-parameters and their measurement time is called measurement data. When measurement data is expressed in a coordinate system, it is expressed as (t i ,y i ) for i = 1, 2, …, n + 1.

[0068] If each measurement data is plotted on a coordinate system with the measurement time on the horizontal axis and the S-parameters on the vertical axis, it will look like Figure 8. In data set D, the set S-parameters and the saturated S-parameters are arranged alternately, with the odd-numbered S-parameters (y1, y3, ...) being the set S-parameters and the even-numbered S-parameters (y2, y4, ...) being the saturated S-parameters.

[0069] If there is no noise component in the data set D, the difference between the set S-parameters and the saturated S-parameters (|y i -y i+1 |) should always be equal. Therefore, as shown in Figure 9, if each S-parameter (actual measurement value) containing a noise component is separated into the noise component and the true value component excluding the noise component, the true value component will alternate between positive and negative values ​​with a constant amplitude.

[0070] Figure 9(a) shows the S-parameters (y1, y2, ..., y n+1 ) is a plot of the true value components (s1, s2, ..., s n+1) is plotted, and oscillates with an amplitude P. In other words, the true value components contained in each S-parameter can be normalized to alternate between -P and +P. Figure 9(c) shows the noise components (d1, d2, ..., d n+1 9(b) is obtained by removing the noise components in FIG. 9(c) from each S-parameter measured in FIG. 9(a), and this amplitude P corresponds to the magnetic permeability of sample 85.

[0071] The noise contained in the S-parameters during actual measurement is mainly linear drift, 1 / f noise, and 1 / f 2 Since the noise components are composed of various low-frequency noises such as noise, it is presumed that they show larger fluctuations than the components of the true values. Therefore, in consideration of the fluctuation tendency of the noise components, the inventor of the present invention has developed a method for removing noise components from each S-parameter, assuming that "the noise components are polynomials of degree n-1" for the number n of S-parameter measurements. As a result, it has been confirmed that this method can remove noise components with extremely high accuracy. The specific method is as follows.

[0072] First, since a polynomial that passes through n+1 points is of degree n, the S parameter y in the graph of FIG. 8 can be expressed as a function of time t, as in the following equation (1).

[0073]

number

[0074] Here, finding the n-th order coefficient Cn in equation (1) is synonymous with differentiating equation (1) n times with respect to time. This is because when equation (1) is differentiated n times with respect to time to find the n-th order derivative, the terms from 0th order to n-1th order are eliminated. Assuming that the noise component is expressed by an n-1-th order polynomial, the noise component can be eliminated by performing a calculation to eliminate the terms from 0th order to n-1th order on the right-hand side of equation (1).

[0075] Then, if a differential operator vector [a] that satisfies the following formula (2) is found, the coefficient Cn can be found from the numerical vector [y] in which all S-parameters are arranged in time series. The differential operator vector [a] represents the time derivative with an order that is one less than the total number of data in the data set D.

[0076]

number

[0077] The numerical vector [y] is an (n+1)-dimensional column vector in which all S-parameters are arranged vertically, and is expressed as in the following formula (3). The differential operator vector [a] is expressed as in the following formula (4). The coefficient Cn is given by the inner product of the differential operator vector [a] and the numerical vector [y], as in formula (2), and can be expressed specifically as in the following formula (5). The differential operator vector [a] is the coefficient a on the right-hand side of formula (5). i It is an n+1-dimensional column vector with (i=1,2,…,n+1) arranged vertically.

[0078]

number

[0079]

number

[0080]

number

[0081] The differential operator vector [a] is the sum of all S-parameters at measurement time t i Based on (i=1,2,...,n+1), it can be calculated using the following formula (6). This differential operator vector [a] can derive the coefficient Cn by taking the inner product with the numerical vector [y] as in formula (5), and is essentially an operator that differentiates formula (1) n-th with respect to time.

[0082]

number

[0083] T in Equation (6) -1 is the inverse matrix of the time power matrix T. The time power matrix T is expressed as the inverse matrix of the measurement time t i The time vector (t1,t2,…,t n+1 ) is a matrix in which the multiplier for each element is increased by 1 from 0 to the total number of data (n+1) and arranged vertically. In other words, the time power matrix T is i For each, it is a matrix in which vectors (n+1-dimensional column vectors) are arranged vertically from the 0th power to the nth power, and then arranged horizontally in chronological order.

[0084]

number

[0085] In other words, the differential operator vector [a] is the vector of all S-parameters at measurement time t i and performs a time differentiation (n-th order differentiation) on the vector with which the inner product is taken, with an order that is one less than the total number of S-parameters. More specifically, the differential operator vector [a] is expressed as the product of the inverse matrix of the time power matrix T and an elimination vector (n+1-dimensional column vector), which is a column vector that eliminates data with orders less than the total number of data (n+1), as shown in formula (6). The elimination vector performs a process equivalent to an n-th order differentiation and division by the factorial of n (n!) on the vector with which the inner product is taken.

[0086] The differential operator vector [a] is equivalent to dividing the n-th order differential operator by the factorial of n. In other words, the coefficient Cn is obtained by dividing the n-th order differential of the numerical vector [y] by the factorial of n. Therefore, as shown in the following formula (8), if the noise component [d] is an n-1 degree polynomial, the value obtained by dividing the inner product of the differential operator vector [a] and the numerical vector [y] by the sum of the absolute values ​​of each element of the differential operator vector [a] becomes the amplitude P in FIG. 9(b).

[0087]

number

[0088] Next, a flow of a process for calculating characteristic data from a data set in the data calculation method according to the first embodiment will be described with reference to the flowchart of FIG. M The flow of obtaining individual characteristic data from the data set D corresponding to each j will be described.

[0089] First, the vector generating means 12e acquires one data set D from the storage unit 13 (step S301). Next, the vector generating means 12e generates a differential operator vector [a] representing a time derivative at an order that is one less than the total number of data in the data set D, based on the measurement times of all data in the acquired data set D, and stores the vector in the storage unit 13. More specifically, the vector generating means 12e acquires one data set D from the measurement times (t1, t2, ..., t n+1 ), a time power matrix T as shown in equation (7) is generated. Next, the vector generating means 12e generates the inverse matrix (T -1 Then, the vector generating means 12e generates an inverse matrix (T -1 A differential operator vector [a] is generated by multiplying the vector [a] by the elimination vector (step S302).

[0090] Next, the arithmetic processing means 12f divides the inner product of the differential operator vector [a] generated by the vector generating means 12e and the numerical vector [y] in which all data in the data set D are arranged in time series by the sum of the absolute values ​​of each element of the differential operator vector [a] to obtain a change in the S-parameter of the jig 50 corresponding to the magnetic permeability of the sample 85 as characteristic data. Then, the arithmetic processing means 12f stores the obtained characteristic data in the storage unit 13 (step S303).

[0091] The control unit 12 judges whether or not a data set D for which calculation of characteristic data has not been completed exists in the storage unit 13. That is, the vector generating means 12e judges whether or not the setting number j has reached the number of sets m (step S304). If the setting number j is less than the number of sets m (step S304 / No), the vector generating means 12e increments the setting number j in the storage unit 13 and reads out the unprocessed data set D from the storage unit 13 (step S301). That is, the control unit 12 repeatedly executes a series of processes from steps S301 to S303 until the setting number j reaches the number of sets m. Then, when the setting number j reaches the number of sets m (step S304 / Yes), the control unit 12 ends the calculation of the characteristic data. The control unit 12 may display the obtained characteristic data on the display unit 15. However, only one data set D may be stored in the storage unit 13, and in this case, the control unit 12 executes the calculation process of the characteristic data based on one data set D.

[0092] If we assume that the noise component is the sum of trigonometric functions up to a period (n-1) / 2 for the n+1 measurement of the S-parameter, then as shown in the following formula (9), the inner product of the differential operator vector [a] and the numerical vector [y] divided by the sum of the absolute values ​​of each element of the differential operator vector [a] is d(ω / nπ) for the amplitude P. n Here, since ω / nπ is less than 1, we use the nth power of (ω / nπ) n is a value that is infinitesimally close to 0, so even under this assumption, it is thought that the amplitude P can be approximately calculated.

[0093]

number

[0094] That is, the control unit 12 may obtain characteristic data by processing based on equation (9) on the assumption that for the number of measurements of the S parameter n+1, “the noise component is the sum of trigonometric functions up to a period (n-1) / 2.”

[0095] Next, the algorithm for removing noise components from actual measurement values ​​will be explained using a concrete example with reference to Fig. 11 and equations (3), (6), (7), (8), and (10) to (14). Fig. 11 illustrates measurement data and virtual noise components in a coordinate system with the horizontal axis representing the measurement time t and the vertical axis representing the actual measurement value (measurement value) y. In Fig. 11, the measurement data (t i ,y i ) are assumed to be (0,1.5), (1.2,-4.22), (2.7,-5.52), and (2.9,-3.88). The noise components including drift noise are expressed by the quadratic polynomial "d=2t 2 -8t-2", that is, processing using the arithmetic expression (8) is assumed. Also, assume that the amplitude P of the signal to be actually measured is 0.5. In the present embodiment 1, the amplitude P corresponds to the change in the S parameter of the jig 50, which corresponds to the magnetic permeability of the sample 85.

[0096] Under these conditions, the time vector consisting of the measurement times of the four measured values ​​is (0, 1.2, 2.7, 2.9). Therefore, the time power matrix T shown in equation (7) is embodied as in equation (10) below. Therefore, its inverse matrix is ​​as shown in equation (11) below, and the differential operator vector [a] shown in equation (6) is as shown in equation (12) below.

[0097]

number

[0098]

number

[0099]

number

[0100] In addition, the numerical vector [y] in which all the measured values ​​are arranged in chronological order can be expressed from equation (3) as shown in the following equation (13).

[0101]

number

[0102] Applying these data to equation (8) gives the following equation (14). In other words, if the inner product of the differential operator vector [a] and the numerical vector [y] is divided by the sum of the absolute values ​​of each element of the differential operator vector [a], the noise components are completely cancelled. In other words, the result of the calculation of equation (8) is 0.5, which corresponds to the amplitude P.

[0103]

number

[0104] When the number of actual measurement values ​​is n+1, the actual noise components may not be completely expressed by an n-1 degree polynomial. However, since the noise components contained in the actual measurement values ​​exhibit relatively gradual changes, by using the data calculation method in this embodiment 1, it is possible to accurately remove the noise components from the actual measurement values ​​and obtain an amplitude P that is extremely close to the true value.

[0105] As described above, the data calculation method, the data calculation program 13p, and the data calculation device 10 of the first embodiment are adapted to obtain characteristic data using the data set D in which the first data and the second data showing the properties of the sample 85 under opposing conditions are arranged alternately in a time series. That is, the control unit 12 of the data calculation device 10 obtains characteristic data by dividing the inner product of the differential operator vector [a] based on the measurement time of each data in the data set D and the numerical vector [y] in which the data are arranged, by the sum of the absolute values ​​of each element of the differential operator vector [a]. Since the differential operator vector [a] represents the time differential of each data in the data set D at an order that is one less than the total number of the data, the noise components included in each data in the data set D can be accurately removed by taking the inner product with the numerical vector [y]. Therefore, many noise components in a plurality of data showing the state change of the sample 85 under opposing conditions can be eliminated, and the magnetic permeability of the sample 85 can be accurately obtained.

[0106] More specifically, the differential operator vector [a] is a time vector (t1, t2, …, t n+1 ) is the inverse matrix T of the time power matrix T, which is a matrix in which the multiplier for each element is increased by 1 from 0 and arranged vertically for the total number of data. -1 and an erasure vector, which is a column vector that erases data of orders less than the total number of data (see formulas (6) and (7)). In other words, if the noise component [d] is an n-1 degree polynomial, when the inner product of the differential operator vector [a] and the numerical vector [y] is divided by the sum of the absolute values ​​of each element of the differential operator vector [a] as in formula (8), the calculation result becomes the amplitude P corresponding to the magnetic permeability of the sample 85 (see FIG. 9(b)).

[0107] This method can be suitably used not only when the measurement intervals of each data are constant, but also when the intervals of the measurement times of each data vary. For example, in the standard mode illustrated in FIG. 6, the measurement intervals of each data can be adjusted to be equal by adjusting the convergence conditions. On the other hand, in the high-speed mode illustrated in FIG. 7, the measurement intervals of each data become unequal. In this respect, this method performs uniform processing based on formulas (1) to (8) or formulas (1) to (7) and (9) according to the flow illustrated in FIG. 10, regardless of the measurement intervals of the plurality of first data and the plurality of second data. Therefore, regardless of the measurement intervals of the plurality of data in the data set D, the desired physical quantity can be obtained with high accuracy. That is, according to this method, when deriving desired data from a minute change in the signal, the noise mixed into the signal can be accurately reduced.

[0108] Embodiment 2 A data calculation device, a data calculation system, a data calculation program, and a data calculation method according to a second embodiment of the present invention will be described with reference to Figures 12 to 15. The overall configuration of the data calculation device and the data calculation system in this second embodiment is similar to the configuration example of the first embodiment described with reference to Figures 1 to 5, so the same reference numerals are used for each component, and the description thereof will be omitted or simplified. The data calculation device 10 executes the acquisition process of the first data and the second data in the standard mode of Figure 6 or the high-speed mode of Figure 7.

[0109] In the above-mentioned first embodiment, a method is adopted in which all data in the data set D is processed collectively. That is, in the first embodiment, when the total number of data in the data set D is n+1, processing equivalent to n-th order differentiation is directly performed on the data. The higher the order of differentiation, the more fluctuations in the higher order data can be eliminated, so a certain degree of effectiveness can be expected from the method of the first embodiment.

[0110] However, when the data to be processed are arranged in a chronological order, the more the total number of data is, the more data at both ends of the time axis are practically excluded in the inner product calculation between the differential operator vector [a] and the numerical vector [y]. In other words, when the number of data in the data set D is large, one or more data measured at an early timing (including the data measured first) and one or more data measured at a late timing (including the data measured last) are hardly reflected in the inner product calculation between the differential operator vector [a] and the numerical vector [y] in the formula (8) or (9), and only data near the center of the time axis are reflected, so that the contribution of white noise becomes large.

[0111] This point will be explained in more detail with reference to FIG. 12. FIG. 12 shows an example in which the number of data to be processed is 20, and each data is arranged along the horizontal axis corresponding to the time series. That is, the horizontal axis of FIG. 12 indicates the measurement order of each data with numbers 1 to 20. The vertical axis of FIG. 12 indicates each element of the differential operator vector [a]. For example, when 20 data in the data set D are subjected to a process equivalent to a 19th order differentiation, as shown in graph G1 (dashed line) of FIG. 12, the four data before and after (1 to 4, 17 to 20) are almost meaningless, and only a few data near the center are reflected in the calculation process.

[0112] Therefore, in the second embodiment, the data in the data set D is divided into a plurality of groups, a differential operator vector (group differential operator vector) is generated for each group, these are added to obtain a differential operator vector, and then characteristic data is obtained by formula (8) or formula (9). Therefore, the control unit 12 in the second embodiment is different from the first embodiment in that part of the function of the vector generating means 12e in particular is different.

[0113] That is, the vector generating means 12e extracts a plurality of data groups, which are combinations of a set number of data arranged in a time series, from the first data and the second data in the data set D by shifting the shift number. The set number corresponds to the total number of data in the data group, that is, the total number of first data and second data included in the data group, and is set to a natural number of 2 or more. The shift number is set to an even number equal to or less than the set number. The shift number is preferably 2, 4, or 6, and is most preferably 2. In addition, the vector generating means 12e generates a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, for each of the extracted plurality of data groups, based on the measurement time of each of the first data and the second data in the data group. Then, the vector generating means 12e generates a differential operator vector [a] by adding up the generated plurality of group differential operator vectors.

[0114] If the total number of data in the data set D is an odd number, the set number is set to an odd number, and the operation based on the group differential operator vector corresponds to even-order differentiation (see FIG. 13). If the total number of data in the data set D is an even number, the set number is set to an even number, and the operation based on the group differential operator vector corresponds to odd-order differentiation (see FIG. 14). The set number is experimentally set according to the stability of the measurement system and the magnitude of the white noise, etc. For example, if the measurement system is stable, the set number may be small, and if the white noise is small, the set number should be large. Although it depends on the number of data in the data set D, the set number is preferably 3 to 8. Other functions of the vector generating means 12e are the same as those in the example of the first embodiment.

[0115] Fig. 13 is an example showing an overview of the data calculation method according to the second embodiment. Fig. 13 illustrates a case where the number of data in the data set D is 9, the number of settings is 5, and the number of shifts is 2. In this example, the number of settings is 5, that is, the total number of data in each data group is 5, so the group differential operator vector corresponds to the fourth-order differential for the five data in the data group.

[0116] 13, the vector generation means 12e extracts three data groups, each of which combines five pieces of data arranged in a time series, from all data in the data set D, shifted by two pieces, and stores the extracted data groups in the storage unit 13. The vector generation means 12e generates a group differential operator vector, which is a differential operator vector for each of the three data groups.

[0117] Here, the group differential operator vector of the data group relating to times t1 to t5 is [a1], the group differential operator vector of the data group relating to times t3 to t7 is [a2], and the group differential operator vector of the data group relating to times t5 to t9 is [a3], and the component representations of these are shown in the following formulas (15) to (17). Note that [ ] is a substitute for the arrow symbol (→) indicating a vector in each formula.

[0118]

number

[0119]

number

[0120]

number

[0121] In this way, the group differential operator vector has "0" except for the components corresponding to the data groups among the total number of components of the data set D. Therefore, the differential operator vector [a] obtained by adding up the group differential operator vectors corresponding to each of the multiple data groups is expressed by the following formula (18).

[0122]

number

[0123] Fig. 14 is another example showing an overview of the data calculation method according to the second embodiment. Fig. 14 illustrates a case where the data set D contains 20 pieces of data, the number of settings is 6, and the number of shifts is 2. In this example, the number of settings is 6, that is, the total number of pieces of data in each data group is 6, so the group differential operator vector corresponds to the fifth-order differential for the six pieces of data in the data group.

[0124] In the example of Fig. 14, the vector generation means 12e extracts eight data groups, each of which is a combination of six data items arranged in a time series, from all data in the data set D, shifted by two, and stores the extracted data groups in the storage unit 13. The vector generation means 12e generates a group differential operator vector, which is a differential operator vector for each of the eight data groups. The vector generation means 12e then adds up the eight group differential operator vectors to obtain the differential operator vector [a]. The configuration of the group differential operator vector is the same as in the example of formulas (15) to (17), and therefore the configuration of the differential operator vector [a] is the same as in the example of formula (18).

[0125] Incidentally, graph G4 (solid line) in Fig. 12 corresponds to the example in Fig. 14. That is, by extracting multiple data groups made up of six pieces of data (preset number) arranged in chronological order from the 20 pieces of data in dataset D, shifting them by two pieces (shifting by the shift number), and performing processing taking into account the multiple extracted data groups, it is possible to make use of most of the data except for the two ends (1-20), as shown in graph G4 (solid line) in Fig. 12.

[0126] If multiple data groups each consisting of 14 pieces of data (a set number) arranged in a time series are extracted from the 20 pieces of data in the data set D with a shift of two, and processing is performed taking into account the multiple extracted data groups, more valid data will be obtained as shown in graph G2 (two-dot chain line) in FIG. 12 than in the case of graph G1. Note that graph G1 in FIG. 12 corresponds to each element of the differential operator vector [a] when processing equivalent to 19th-order differentiation is performed on the 20 pieces of data in the data set D. If multiple data groups each consisting of 10 pieces of data (a set number) arranged in a time series are extracted from the 20 pieces of data in the data set D with a shift of two, and processing is performed taking into account the multiple extracted data groups, more valid data will be obtained as shown in graph G3 (one-dot chain line) in FIG. 12 than in the case of graph G2.

[0127] Next, a flow of a process for calculating characteristic data from a data set D in the data calculation method according to the second embodiment will be described with reference to the flowchart of FIG. M A flow for obtaining individual characteristic data from the data set D corresponding to each of the steps j will be described below. The same step numbers will be used for steps similar to those in the process of FIG. 10, and descriptions thereof will be omitted or simplified.

[0128] First, the vector generating means 12e selects one data set D in the storage unit 13, and extracts a plurality of data groups, which are combinations of a set number of data arranged in a time series, from all data in the selected data set D by shifting the number of shifts (step S401). Next, the vector generating means 12e generates a group differential operator vector for each of the extracted data groups. That is, for each data group, the vector generating means 12e generates a group differential operator vector representing a time derivative at an order that is one less than the total number of data in the data group, based on the measurement time of each of all data in the data group, and stores the group differential operator vector in the storage unit 13 (step S402).

[0129] The vector generating means 12e repeatedly executes the process of step S402 until the generation of group differential operator vectors for all extracted data groups is completed (step S403 / No). When the generation process of group differential operator vectors for all data groups is completed (step S403 / Yes), the vector generating means 12e obtains a differential operator vector [a] by adding up the generated group differential operator vectors.

[0130] Next, the arithmetic processing means 12f divides the inner product of the differential operator vector [a] generated by the vector generating means 12e and the numerical vector [y] in which all data in the data set D are arranged in time series by the sum of the absolute values ​​of each element of the differential operator vector [a] to obtain a change in the S-parameter of the jig 50 corresponding to the magnetic permeability of the sample 85 as characteristic data. Then, the arithmetic processing means 12f stores the obtained characteristic data in the storage unit 13 (step S303).

[0131] The vector generating means 12e judges whether the setting number j has reached the number of sets m (step S304), and if the setting number j is less than the number of sets m (step S304 / No), it increments the setting number j in the storage unit 13 and reads out the unprocessed data set D from the storage unit 13 (step S301). That is, the control unit 12 repeatedly executes a series of processes from steps S401 to S404 and S303 until the setting number j reaches the number of sets m, and when the setting number j reaches the number of sets m (step S304 / Yes), it ends the calculation of the characteristic data. The control unit 12 may execute only the calculation process of the characteristic data based on one data set D.

[0132] The above operation has been described in the order of the step numbers in the flowchart of Fig. 15, but this flow is an example and is not limited to this. Also, the control unit 12 may obtain characteristic data by processing based on formula (9) on the assumption that "noise components are the sum of trigonometric functions up to a period (n-1) / 2" for the number n+1 of measurements of the S parameter.

[0133] As described above, the control unit 12 in the second embodiment obtains characteristic data by dividing the inner product of the differential operator vector [a] based on the measurement time of each data in the data set D and the numerical vector [y] in which the data are arranged by the sum of the absolute values ​​of each element of the differential operator vector [a]. Since the differential operator vector [a] represents the time differential of each data in the data set D at an order that is one less than the total number of the data, the noise components included in each data in the data set D can be accurately removed by taking the inner product with the numerical vector [y]. Therefore, many noise components in a plurality of data showing the state changes of the sample 85 under contradictory conditions can be eliminated, and the magnetic permeability of the sample 85 can be accurately obtained.

[0134] More specifically, the control unit 12 extracts a plurality of data groups, which are combinations of two or more data arranged in a time series, from the first data and the second data in the data set D by shifting the number of shifts, which is an even number less than the total number of data in the data group. That is, the control unit 12 extracts a plurality of data groups with overlapping data from all data in the data set D. In addition, the control unit 12 generates a group differential operator vector representing a time derivative at an order that is one less than the total number of data in the data group, for each of the extracted plurality of data groups, based on the measurement time of each of the first data and the second data in the data group. Then, the control unit 12 adds up the generated plurality of group differential operator vectors to obtain the differential operator vector [a]. In this way, the control unit 12 is configured to obtain the differential operator vector [a] by using the group differential operator vector corresponding to each of the plurality of data groups extracted by shifting the number of shifts. Therefore, even if the total number of data in the data set D is large, it is possible to widely and effectively utilize data at both ends of the time axis, and the contribution of white noise can be reduced.

[0135] <Modification> A data calculation device, a data calculation system, a data calculation program, and a data calculation method according to a modification of the second embodiment will be described with reference to Figures 12 to 14 and 16. The overall configuration of the data calculation device and the data calculation system in this modification is similar to the configuration example of the first embodiment described with reference to Figures 1 to 5, so the same reference numerals are used for each component, and the description thereof will be omitted or simplified. The data calculation device 10 executes the acquisition process of the first data and the second data in the standard mode of Figure 6 or the high-speed mode of Figure 7.

[0136] In this modification, even if the total number of data is large, the differential operator vector [a] is processed by dividing the data into a plurality of data groups so that the data at both ends of the time axis are reflected as much as possible (see Figs. 12 to 14). More specifically, this modification employs a method in which the data in the data set D is divided into a plurality of groups, each group is processed in the same manner as in the first embodiment, and characteristic data is obtained by averaging the data obtained. Therefore, the control unit 12 in this modification differs from the first embodiment in some of the functions, particularly the vector generating means 12e and the calculation processing means 12f.

[0137] That is, the vector generating means 12e extracts, from all data in the data set D, a plurality of data groups, each of which is a combination of a set number of data arranged in a time series, by shifting the number of shifts, as described above. Furthermore, for each of the plurality of data groups, the vector generating means 12e generates a group differential operator vector representing a time derivative at an order that is one less than the total number of data in the data group, based on the measurement time of each of all data in the data group. The group differential operator vector in this modification is expressed in the same way as in formula (4) or formula (6). Other functions of the vector generating means 12e are the same as those in the example of embodiment 1.

[0138] The arithmetic processing means 12f divides the inner product of the group differential operator vector generated by the vector generating means 12e and the numerical vector [y] in which all data in the data group are arranged in time series by the sum of the absolute values ​​of each element of the group differential operator vector for each of the multiple data groups to obtain group characteristic data for each data group. The group characteristic data is data indicating the characteristics of the sample 85 associated with each of the multiple data groups. The arithmetic processing means 12f then obtains the average of the group characteristic data for each of the multiple data groups as characteristic data. Other functions of the arithmetic processing means 12f are the same as those in the example of embodiment 1.

[0139] Next, a flow of a process for calculating characteristic data from a data set D in the data calculation method according to the second embodiment will be described with reference to the flowchart of FIG. M A flow for obtaining individual characteristic data from the data set D corresponding to each of the steps j will be described. The same step numbers will be used for steps similar to those in the process of FIG. 10, and descriptions thereof will be omitted or simplified.

[0140] First, the vector generating means 12e selects one data set D in the storage unit 13, and extracts multiple data groups, which are combinations of a set number of data items arranged in a time series, from all data in the selected data set D by shifting the number of shifts (step S401). Next, the control unit 12 executes a process of generating a differential operator vector [a] (step S402) and a process of generating group characteristic data (step S501) for each of the extracted data groups.

[0141] That is, the vector generating means 12e generates, for each data group, a group differential operator vector that represents a time differential at an order that is one less than the total number of data in the data group, based on the measurement time of each of all data in the data group, and stores the vector in the storage unit 13 (step S402). In addition, the calculation processing means 12f obtains group characteristic data for each data group by dividing the inner product of the group differential operator vector and a numerical vector [y] in which all data in the data group are arranged in chronological order, by the sum of the absolute values ​​of each element of the group differential operator vector (step S501).

[0142] The vector generating means 12e repeatedly executes the process of step S402 until the generation of the group differential operator vector [a] for all the extracted data groups is completed (step S502 / No). Similarly, the calculation processing means 12f repeatedly executes the process of step S501 until the vector generating means 12e has determined the group characteristic data for each of all the data groups extracted (step S502 / No).

[0143] When the calculation of the group characteristic data of all the data groups is completed (step S502 / Yes), the calculation processing means 12f adds up all the group characteristic data and divides the sum by the number of data groups to obtain characteristic data. Then, the calculation processing means 12f stores the obtained characteristic data in the storage unit 13 (step S503).

[0144] If the setting number j is less than the number of groups m (step S304 / No), the control unit 12 increments the setting number j in the storage unit 13 and repeats the series of processes of steps S401, S402, and S501 to S504 until the setting number j reaches the number of groups m. Then, when the setting number j reaches the number of groups m (step S304 / Yes), the control unit 12 ends the calculation of the characteristic data. The control unit 12 may cause the display unit 15 to display the obtained characteristic data.

[0145] The above operation has been described in the order of step numbers in the flowchart of FIG. 16, but the flow is an example and is not limited thereto. For example, the control unit 12 may perform each process for the next data group after completing the process of generating the differential operator vector [a] (step S402) and the process of generating the group characteristic data (step S501) for one data group. The control unit 12 may perform the process of generating the differential operator vector [a] (step S402) and the process of generating the group characteristic data (step S501) in parallel. The control unit 12 may start the process of generating the group characteristic data (step S501) after completing the process of generating the differential operator vector [a] (step S402) for all data groups. However, only one data set D may be stored in the storage unit 13, and in this case, the control unit 12 executes the calculation process of the characteristic data based on one data set D. The control unit 12 may obtain characteristic data by processing based on equation (9) on the assumption that for the number n+1 measurements of the S parameter, "the noise component is the sum of trigonometric functions up to a period (n-1) / 2."

[0146] As described above, the data calculation method, data calculation program 13p, and data calculation device 10 of the second embodiment are adapted to extract a plurality of data groups, which are combinations of two or more pieces of data arranged in a time series, from all data in the data set D, by shifting the number of shifts, which is an even number less than the total number of pieces of data in the data group. That is, the control unit 12 of the data calculation device 10 extracts a plurality of data groups with overlapping data from all data in the data set D. Then, for each of the extracted plurality of data groups, the control unit 12 generates a group differential operator vector representing a time derivative at an order that is one less than the total number of pieces of data in the data group, based on the measurement time of each of all data in the data group. Furthermore, for each of the plurality of data groups, the control unit 12 obtains group characteristic data by dividing the inner product of the group differential operator vector and the numerical vector [y] by the sum of the absolute values ​​of each element of the group differential operator vector, and obtains characteristic data by averaging the obtained plurality of group characteristic data.

[0147] Since the group differential operator vector represents the time differential of each data in the data group at an order one less than the total number of the data, it is possible to accurately remove noise components contained in each data in the data group by taking the inner product with the numerical vector [y] related to the data group. As described above, the control unit 12 obtains group characteristic data corresponding to each of the multiple data groups extracted by shifting the number of shifts, and obtains characteristic data based on these. Therefore, even if the total number of data in the data set D is large, it is possible to effectively utilize data from both ends of the time axis widely, and the contribution of white noise can be reduced.

[0148] Here, the above-mentioned embodiments are specific examples of the data calculation program, the data calculation device, and the data calculation method, and the technical scope of the present invention is not limited to these aspects. For example, in the above-mentioned embodiments, the first condition and the second condition are switched by the change of the magnetic field, and the S parameters obtained under each condition are used as the first data and the second data, but the present invention is not limited to this. As described above, the method in each of the above-mentioned embodiments can be suitably used not only when the measurement time of a plurality of data in the data set D is equally spaced, but also when it is not equally spaced, and therefore can be effectively used in various fields. As an example, the first condition and the second condition may be switched by the change of temperature, and the physical properties obtained under each condition may be used as the first data and the second data. As a specific example, the method of the present invention may be used when the electrical resistivity, dielectric constant, etc. of a semiconductor element at a first temperature (reference temperature) are used as the first data, and the electrical resistivity, dielectric constant, etc. of the semiconductor element at a second temperature (temperature changed from the reference temperature) are used as the second data, and the changes in these are analyzed. As another example, the first and second conditions may be set in association with the pumping action of the human or animal heart, and vital data, etc. obtained under each condition may be used as the first data and the second data. As yet another example, the first and second conditions may be set in association with the breathing (inhalation and exhalation) of a human or animal, and vital data, etc. obtained under each condition may be used as the first data and the second data. As yet another example, the first condition may be a condition in which light (including sunlight) is irradiated onto an object, and the second condition may be a condition in which light is not irradiated onto the object, and the method of the present invention may be used when analyzing changes in the physical properties of the object obtained under each condition. The methods in each of the above embodiments can also be widely used in the field of astronomy. [Explanation of symbols]

[0149] 10 data calculation device, 11 communication unit, 12 control unit, 12a input processing means, 12b output processing means, 12c magnetic field processing means, 12d acquisition processing means, 12e vector generation means, 12f calculation processing means, 13 memory unit, 13p data calculation program, 14 input unit, 15 display unit, 20 fixture, 20 signal processing device, 22b excitation unit, 25 amplifier, 30 magnetic field control device, 35 electromagnet, 50 fixture, 60 input waveguide, 61 substrate, 62 signal line, 62a input unit, 62b excitation unit, 62s short-circuit end, 63 ground line (earth line), 63a input unit, 70 output waveguide, 71 substrate, 72 signal line, 72a output unit, 72b detection unit, 72b excitation unit, 72s short-circuit end, 73 Ground line (earth line), 73a input section, 80 non-magnetic insulating layer, 85 sample, 91 coaxial cable, 91a signal line, 91b ground line, 92 coaxial cable, 92a signal line, 92b ground line, 100 data calculation system.

Claims

1. a vector generating means for generating a differential operator vector representing a time derivative at an order one less than the total number of data in a data set, based on measurement times of the first data and the second data in the data set, the first data indicating a property of a sample under a first condition and the second data indicating a property of the sample under a second condition opposite to the first condition being arranged alternately in a chronological order; and a data calculation program for causing a computer to function as a calculation processing means for obtaining characteristic data indicating the characteristics of the sample by dividing the inner product of the differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by the sum of the absolute values ​​of each element of the differential operator vector.

2. a vector generating means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a time series, from among the first data and the second data in a data set in which a plurality of first data indicating properties of a sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition are arranged alternately in a time series, by shifting the number of shifts by an even number not greater than the total number of data in the data group, for each of the extracted plurality of data groups, generating a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of the first data and the second data in the data group, and calculating a differential operator vector by adding up the generated plurality of group differential operator vectors; and a data calculation program for causing a computer to function as a calculation processing means for obtaining characteristic data indicating the characteristics of the sample by dividing the inner product of the differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by the sum of the absolute values ​​of each element of the differential operator vector.

3. a vector generating means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a time series, from among the first data and the second data in a data set in which a plurality of first data indicating a property of a sample under a first condition and a plurality of second data indicating a property of the sample under a second condition opposite to the first condition are arranged alternately in a time series, by shifting the number of shifts by an even number not greater than the total number of data in the data groups, and for each of the extracted plurality of data groups, generating a group differential operator vector representing a time derivative at an order one less than the total number of data in the data group, based on the measurement times of the first data and the second data in the data group; and a data calculation program for causing a computer to function as a calculation processing means for obtaining group characteristic data by dividing, for each of a plurality of the data groups, the inner product of the generated group differential operator vector and a numerical vector in which the first data and the second data in the data group are arranged in chronological order, by the sum of the absolute values ​​of each element of the group differential operator vector, and for obtaining characteristic data indicative of the characteristics of the sample by averaging the plurality of group characteristic data thus obtained.

4. a control unit that calculates characteristic data indicating characteristics of the sample using a data set in which a plurality of first data indicating properties of the sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition are alternately arranged in a time series; The control unit is a vector generating means for generating a differential operator vector representing a time derivative at an order that is one less than the total number of data in the data set, based on measurement times of the first data and the second data in the data set; a calculation processing means for dividing an inner product of the differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by a sum of absolute values ​​of each element of the differential operator vector to obtain the characteristic data.

5. a control unit that calculates characteristic data indicating characteristics of the sample using a data set in which a plurality of first data indicating properties of the sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition are alternately arranged in a time series; The control unit is a vector generating means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a time series from the first data and the second data in the data set, by shifting the data by an even number of shifts less than the total number of pieces of data in the data group, generating, for each of the extracted plurality of data groups, a group differential operator vector representing a time derivative at an order one less than the total number of pieces of data in the data group, based on the measurement times of the first data and the second data in the data group, and calculating a differential operator vector by adding up the generated plurality of group differential operator vectors; and a calculation processing means for dividing an inner product of the differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by a sum of absolute values ​​of each element of the differential operator vector to obtain characteristic data indicating characteristics of the sample.

6. a control unit that calculates characteristic data indicating characteristics of the sample using a data set in which a plurality of first data indicating properties of the sample under a first condition and a plurality of second data indicating properties of the sample under a second condition that is opposite to the first condition are alternately arranged in a time series; The control unit is a vector generating means for extracting a plurality of data groups, each of which is a combination of two or more pieces of data arranged in a time series from the first data and the second data in the data set, by shifting the number of shifts by an even number not greater than the total number of pieces of data in the data groups, and for each of the extracted plurality of data groups, generating a group differential operator vector representing a time derivative at an order that is one less than the total number of pieces of data in the data group, based on the measurement times of the first data and the second data in the data group; and a calculation processing means for calculating, for each of a plurality of said data groups, an inner product of the group differential operator vector generated by the vector generating means and a numerical vector in which the first data and the second data in the data group are arranged in time series, by a sum of the absolute values ​​of each element of the group differential operator vector to obtain group characteristic data, and for calculating the characteristic data by taking an average of the plurality of group characteristic data thus obtained.

7. one or more processors, a differential operator vector representing a time derivative at an order one less than the total number of data in the data set is generated based on measurement times of the first data and the second data in a data set in which a plurality of first data indicating a property of a sample under a first condition and a plurality of second data indicating a property of the sample under a second condition opposite to the first condition are alternately arranged in a time series; A data calculation method comprising: dividing an inner product of the differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by a sum of absolute values ​​of each element of the differential operator vector to obtain characteristic data indicating characteristics of the sample.

8. one or more processors, A data set in which a plurality of first data indicating a property of a sample under a first condition and a plurality of second data indicating a property of the sample under a second condition opposite to the first condition are alternately arranged in a time series is extracted from the first data and the second data in the data set, the data being a combination of two or more pieces of data arranged in a time series, by shifting the number of shifts by an even number not greater than the total number of pieces of data in the data group; For each of the plurality of data groups, a group differential operator vector is generated that represents a time derivative at an order that is one less than the total number of data in the data group, based on the measurement time of each of the first data and the second data in the data group; A differential operator vector is obtained by adding up the generated group differential operator vectors; A data calculation method comprising: dividing an inner product of the obtained differential operator vector and a numerical vector in which a plurality of the first data and a plurality of the second data are arranged in a time series by a sum of absolute values ​​of each element of the differential operator vector, thereby obtaining characteristic data indicating characteristics of the sample.

9. one or more processors, A data set in which a plurality of first data indicating a property of a sample under a first condition and a plurality of second data indicating a property of the sample under a second condition opposite to the first condition are alternately arranged in a time series is extracted from the first data and the second data in the data set, the data being a combination of two or more pieces of data arranged in a time series, by shifting the number of shifts by an even number not greater than the total number of pieces of data in the data group; For each of the plurality of data groups, generating a group differential operator vector representing a time derivative at an order that is one less than the total number of data in the data group, based on the measurement times of the first data and the second data in the data group, and dividing an inner product of the generated group differential operator vector and a numerical vector in which the first data and the second data in the data group are arranged in a time series by a sum of absolute values ​​of each element of the group differential operator vector to obtain group characteristic data; A data calculation method comprising: averaging the group characteristic data of each of the obtained plurality of data groups to obtain characteristic data indicating the characteristics of the sample.

Images