Prediction processing device, prediction processing system, prediction processing method and prediction processing program

Abstract

To provide a prediction processing device, a prediction processing system, a prediction processing method and a prediction processing program, which quickly and accurately predict information indicating a moisture content of gas.SOLUTION: A prediction processing device includes initial arithmetic means that determines a measurement time constant as a time constant indicating a speed of a temporal change in a measured value of information indicating a moisture content of gas on the basis of a temperature of the gas to be measured, and determines as an initial value a measured value of information indicating a moisture amount at a rising point relating to a measured humidity change of the gas. The prediction processing device includes prediction means that determines as a reference value a measured value of information indicating a moisture content when a set time has elapsed from the rising, and uses the initial value, the reference value and the measurement time constant to determine a prediction value of a value to which the measured value of the information indicating the moisture content converges.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a prediction processing device, a prediction processing system, a prediction processing method, and a prediction processing program that predict information indicating the moisture content of a gas such as air. [Background technology]

[0002] Known information indicating the moisture content of gas includes dew point temperature, absolute humidity, and water vapor pressure. These can be easily converted into each other, and for example, dew point can be measured using an optical method that optically detects the state of condensation on a mirror surface, or a calculation method that calculates based on temperature and relative humidity (see, for example, Patent Document 1). The dew point temperature detector in Patent Document 1 is configured to apply the detected temperature and detected humidity to a calculation formula to determine the dew point. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2008-281376 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the optical method requires cooling of the mirror surface, and the conventional calculation method requires stabilizing the sensor state, and either method requires a certain waiting time for the sensor's actual measurement value to converge. Therefore, there is a need to speed up the measurement of information indicating the moisture content of gas.

[0005] The present invention has been made to solve the above-mentioned problems, and aims to provide a prediction processing device, a prediction processing system, a prediction processing method, and a prediction processing program that accurately predict information indicating the moisture content of gas in a short period of time. [Means for solving the problem]

[0006] A prediction processing device according to one embodiment of the present invention comprises an initial calculation means for calculating a measurement time constant, which is a time constant representing the rate of change over time of the actual measured value of information indicating the moisture content of the gas, based on the temperature of the gas to be measured, and calculating the actual measured value of the information indicating the moisture content at the time of rise related to the actual measured humidity change of the gas as an initial value; and a prediction means for calculating the actual measured value of the information indicating the moisture content after a set time has elapsed from the rise as a reference value, and using the initial value, reference value, and measurement time constant to calculate a predicted value to which the actual measured value of the information indicating the moisture content will converge.

[0007] A prediction processing system according to one aspect of the present invention includes a measurement processing unit that measures the temperature and relative humidity of a gas to be measured, and the prediction processing device described above.

[0008] A prediction processing program according to one embodiment of the present invention causes a computer to function as an initial calculation means that calculates a measurement time constant, which is a time constant that represents the rate of change over time of information indicating the moisture content of the gas, based on the temperature of the gas to be measured, and calculates, as an initial value, the actual measured value of information indicating the moisture content at the start of the actual humidity change of the gas, and a prediction means that calculates, as a reference value, the actual measured value of information indicating the moisture content after a set time has elapsed since the start of the change, and uses the initial value, the reference value, and the measurement time constant to calculate a predicted value to which the actual measured value of information indicating the moisture content will converge.

[0009] A prediction processing method according to one embodiment of the present invention uses one or more processors to calculate a measurement time constant, which is a time constant that represents the rate of change over time of information indicating the moisture content of the gas, based on the temperature of the gas to be measured; calculates an actual measured value of the information indicating the moisture content at the start of the actual measured humidity change of the gas as an initial value; calculates an actual measured value of the information indicating the moisture content after a set time has elapsed since the start of the change as a reference value; and calculates a predicted value to which the actual measured value of the information indicating the moisture content will converge using the initial value, the reference value, and the measurement time constant. [Effects of the Invention]

[0010] The present invention is configured to calculate a predicted value to which the actual measured value of information indicating the moisture content will converge using a time constant obtained by applying the time constant to the time constant correspondence data of the gas temperature, an initial value indicating the moisture content at the time of rise related to the actual measured humidity change of the gas, and a reference value indicating the moisture content after a set time has elapsed since the rise. Therefore, the information indicating the moisture content of the gas can be predicted accurately in a short time without waiting for the time required for the actual measured value of the sensor to converge. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a configuration diagram illustrating a prediction processing system according to a first embodiment of the present invention. [Figure 2] FIG. 2 is a perspective view showing an example of the appearance of the measurement processing unit of FIG. 1. [Figure 3] FIG. 3 is a plan view of the measurement processing unit of FIG. 2. [Figure 4] FIG. 4 is a circuit diagram showing a simplified example of a circuit related to driving the heater portion of FIGS. 1 to 3. [Figure 5] 10 is a graph showing the change over time in output y when τ is changed in equation (5) described below. [Figure 6] 1 is a graph showing the relationship between the change over time in the measured dew-point temperature of the first gas in Table 1 and a model curve according to equation (6) described below. [Figure 7] 1 is a graph showing the relationship between the change over time in the measured values ​​of the dew-point temperature of the second gas in Table 1 and a model curve according to equation (6) described below. [Figure 8] 2 is a graph showing the change over time in the measured dew point temperature when a predetermined gas is passed through a flow path of the flow path member of FIG. 1. [Figure 9] FIG. 3 is a schematic diagram illustrating the sensor unit of FIG. 2. [Figure 10] FIG. 3 is a schematic diagram illustrating a sensor unit on which a flow path member is mounted as in FIG. 2. [Figure 11] 10 is a graph showing the change over time in the measured dew-point temperature when humidity-control gas is sprayed from the front and side of the detection unit in FIG. 9. [Figure 12]11 is a graph showing the change over time in the measured dew point temperature when humidity-controlling gas is blown from one opening and a side of the flow path member 12 in FIG. 10. [Figure 13] 11 is a graph showing the change over time in the measured dew-point temperature when a water vapor source is brought close to the sensor unit of FIG. 9 and the sensor unit of FIG. 10. [Figure 14] 11 is a graph showing the change over time in the measured values ​​of the dew-point temperature when water vapor is sprayed onto the sensor unit of FIG. 9 and the sensor unit of FIG. 10. [Figure 15] 10 is a graph showing the change in dew point temperature over time when heating by a heater unit is performed and when heating is not performed under predetermined conditions. [Figure 16] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a temperature of 40% and a model curve. [Figure 17] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a temperature of 50% and a model curve. [Figure 18] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a temperature of 60% and a model curve. [Figure 19] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a temperature of 70% and a model curve. [Figure 20] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a model curve. [Figure 21] 1 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 36° C. and a model curve. [Figure 22] 10 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 20° C. and a humidity of 90% when the set temperature of the heater unit is changed in stages. [Figure 23] 10 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 40° C. and a humidity of 90% when the set temperature of the heater unit is changed in stages. [Figure 24] 10 is a graph showing the change over time in the measured dew point temperature of a gas at a temperature of 50° C. and a humidity of 90% when the set temperature of the heater unit is changed in stages. [Figure 25] 10 is a graph showing the change over time in the measured dew point temperatures of a plurality of gases with different humidities at a temperature of 20° C. when the set temperature of the heater unit is 28° C. [Figure 26] 10 is a graph showing the change over time in the measured dew point temperatures of a plurality of gases with different humidities at a temperature of 20°C when the set temperature of the heater unit is 31°C. [Figure 27] 10 is a graph showing the difference in the change over time in the measured dew point temperature of gas depending on whether or not one end of the flow path member is closed with a sheet member. [Figure 28] 10 is a graph showing the change over time in the measured dew point temperature of a gas when the flow rate of the gas passing through a flow path is changed in stages. [Figure 29] 10 is a graph for examining the prediction accuracy of moisture content data when the prediction time is set to τ or τ / 2 after the rise of the measured humidity change of the gas. [Figure 30] 1 is a flowchart illustrating an example of the flow of operations of a prediction processing method according to the first embodiment of the present invention. [Figure 31] FIG. 10 is a configuration diagram illustrating an example of a prediction processing system according to a modified example of the first embodiment of the present invention. [Figure 32] 10 is a flowchart illustrating an example of the flow of operations in a prediction processing method according to a modified example of the first embodiment of the present invention. [Figure 33] FIG. 10 is a configuration diagram illustrating a prediction processing system according to a second embodiment of the present invention. [Figure 34] 10 is a flowchart illustrating an example of the flow of operations of a prediction processing method according to a second embodiment of the present invention. [Figure 35] FIG. 10 is a configuration diagram illustrating a prediction processing system according to a third embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Embodiment 1 The measurement processing unit, prediction processing system, prediction processing method, and prediction processing program according to the first embodiment will be described with reference to Figures 1 to 30. In some figures, some reference numerals and components are omitted to avoid complication.

[0013] First, the overall configuration and functional configuration of a prediction processing system 100 in the first embodiment will be described with reference to Fig. 1. As shown in Fig. 1, the prediction processing system 100 is composed of a measurement processing unit 10 and a prediction processing device 50. The measurement processing unit 10 and the prediction processing device 50 are connected via a network N such as the Internet so as to be able to communicate with each other via wired or wireless communication.

[0014] The measurement processing unit 10 includes a flow path member 12, a heater section 15, a switching section 16, a light-emitting section 17, a sensor section 20, and a measurement processing device 30. The flow path member 12 is cylindrical (pipe-shaped) with open ends and is arranged so that the gas to be measured flows in through one opening and flows out through the other opening. The flow path member 12 is made of a metal with good thermal conductivity, such as aluminum or copper. The heater section 15 heats the flow path member 12, thereby increasing the temperature of the gas in the flow path O of the flow path member 12. The switching section 16 is, for example, a switch, and switches the heater section 15 between on and off. The light-emitting section 17 is, for example, an LED, and indicates the state of the heater section 15. The light-emitting section 17 is configured to emit light when the switching section 16 is in an on state, that is, when electricity is applied to the heater section 15.

[0015] The sensor unit 20 is a sensing device for detecting the environment in the flow path O, and has a detection unit 21 that detects information related to temperature and relative humidity. Hereinafter, relative humidity will be simply referred to as humidity, and is distinguished from absolute humidity. The detection unit 21 is disposed in the flow path O of the flow path member 12, and detects information related to the temperature and humidity of gas that is retained in or passing through the flow path O. The measurement processing device 30 has a function of measuring the temperature and humidity of the gas to be measured based on detection data from the detection unit 21 disposed in the flow path O of the flow path member 12. The measurement processing device 30 has a communication unit 31, an information processing unit 32, and a memory unit 33.

[0016] The communication unit 31 is an interface through which the information processing unit 32 communicates with external devices via wired or wireless communication. The storage unit 33 stores various information in addition to the operating program of the information processing unit 32. The storage unit 33 is configured by RAM (Random Access Memory), ROM (Read Only Memory), flash memory, eMMC (embedded Multi Media Card), SSD (Solid State Drive), HDD (Hard Disk Drive), or the like.

[0017] The information processing unit 32 measures the temperature and humidity based on the data detected by the detection unit 21 and transmits the measured values ​​to the prediction processing device 50. The information processing unit 32 can be configured by a computing device such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or an MPU (Micro-Processing Unit), and software that cooperates with the computing device to realize the above functions. Note that the sensor unit 20 may have an information processing unit that measures the temperature and humidity based on the data detected by the detection unit 21. In this case, the information processing unit 32 transmits the measured values ​​acquired from the sensor unit 20 via the communication unit 31 to the prediction processing device 50.

[0018] The prediction processing device 50 predicts a convergence value of information indicating the moisture content of the gas in the flow path O based on the measurement values ​​transmitted from the information processing unit 32. The information indicating the moisture content of the gas includes the dew point temperature, absolute humidity, water vapor pressure, etc. derived from the temperature and humidity of the gas, and is hereinafter also referred to as "moisture content data." The convergence value of the moisture content data is the value to which the moisture content data converges (saturates) after a certain time has passed since the rise of the measurement value. The prediction processing device 50 is configured by a PC (Personal Computer), such as a desktop PC, notebook PC, tablet PC, or smartphone.

[0019] The prediction processing device 50 includes a communication unit 51, a prediction processing unit 52, a storage unit 53, an operation unit 54, and a display unit 55. The communication unit 51 is an interface for the prediction processing unit 52 to communicate with external devices via wired or wireless communication. The storage unit 53 stores various information, including the operation program of the prediction processing unit 52, including the prediction processing program 53p. The storage unit 53 stores, for example, time constant correspondence data that associates the temperature of a given gas with a time constant τ that represents the rate of change in information indicating the moisture content of the gas. The time constant correspondence data may be, for example, table information that associates multiple gas temperature ranges with multiple time constants τ, graph information that associates gas temperatures with time constants τ, or a function that determines the time constant τ by substituting the gas temperature (a function for deriving the time constant τ using the gas temperature as a variable). The storage unit 53 is configured with RAM, ROM, flash memory, eMMC, SSD, HDD, or the like.

[0020] The operation unit 54 accepts an input operation by the user and outputs an operation signal corresponding to the content of the accepted operation to the prediction processing unit 52. The display unit 55 is formed of, for example, a liquid crystal display (LCD), and displays various information in response to instructions from the prediction processing unit 52. Instead of the operation unit 54 and the display unit 55, the prediction processing device 50 may have a touch panel including a display panel that displays characters, images, etc., and detection means that is stacked on the display panel and detects touch operations.

[0021] The prediction processing unit 52 predicts a convergence value of the moisture content data based on the temperature and humidity of the gas. The prediction processing unit 52 has an acquisition processing means 52a, an initial calculation means 52b, a prediction means 52c, and an output processing means 52d. The acquisition processing means 52a acquires the measurement values ​​transmitted from the information processing unit 32, i.e., the information on the temperature and humidity in the flow path O, and stores the information in the memory unit 53.

[0022] The initial calculation means 52b applies the gas temperature or the set temperature of the heater unit 15 to the time constant correspondence data to find the time constant τ. The time constant found by the initial calculation means 52b corresponds to the "measured time constant." When the measurement target is air blown out from an air conditioner, the gas temperature applied to the time constant correspondence data is the set temperature of the air conditioner or the set temperature plus a predetermined value. When the measurement target is the exhaled air of a person or animal, the gas temperature applied to the time constant correspondence data is, for example, the general body temperature of the person or animal or the actual measured body temperature of the person. The set temperature of the heater unit 15 will be described later.

[0023] The initial calculation means 52b also detects a rise in the measurement value. For example, the initial calculation means 52b may calculate the dew-point temperature from the measurement value over time and detect the rise in the dew-point temperature as the rise in the measurement value. The initial calculation means 52b in the first embodiment has a function of calculating at least one of the dew-point temperature, absolute humidity, and water vapor pressure of the gas using a known calculation method. However, the initial calculation means 52b may also detect the rise in humidity among the measurement values ​​as the rise in the measurement value. In this case, the initial calculation means 52b does not need to sequentially calculate the dew-point temperature, absolute humidity, or water vapor pressure using the measurement value. Furthermore, the initial calculation means 52b calculates the actual moisture content data at the time of the rise in the measurement value as the initial value Yb. Here, the change in the actual dew-point temperature over time depends on the change in the actual relative humidity over time. Therefore, the rise in the measurement value is also referred to as a "rise related to the change in the actual humidity," and hereinafter, simply referred to as a "rise" to avoid complication.

[0024] The prediction means 52c predicts the time t SThe actual measured value of the moisture content data after t S ) is calculated as the set time t S is the waiting time from the start of the process to the calculation of the predicted value. S is set based on the time constant τ, for example. S is better from the viewpoint of shortening the prediction time, but from the viewpoint of prediction accuracy, it is good to set it to τ / 2 or more, preferably 2τ / 3 or more, and more preferably τ or more. This point will be described in detail later. The prediction means 52c has a function of determining at least one of the dew point temperature, absolute humidity, and water vapor pressure of the gas by a well-known calculation method or the like. That is, the prediction means 52c determines the reference value Y(t S ), calculation based on equation (1) is not performed.

[0025] The prediction means 52c calculates the initial value Yb and the reference value Y(t S ), and the time constant τ, to obtain a predicted value to which the measured value of the moisture content data will converge. More specifically, the prediction means 52c uses the initial value Yb, the reference value Y(t S ), and the time constant τ, a prediction coefficient A corresponding to the difference between the true value of the moisture content data and the initial value Yb is calculated, and the calculated prediction coefficient A is added to the initial value Yb to calculate a predicted value. The prediction means 52c of the first embodiment calculates a predicted value by calculation based on the following equation (1) using time t as a variable.

[0026]

number

[0027] That is, the prediction means 52c calculates the initial value Yb and the reference value Y(t S ), and the time constant τ are applied to the following equation (2) to obtain the prediction coefficient A. Note that equation (2) is obtained by rearranging equation (1) with respect to the prediction coefficient A.

[0028]

number

[0029] Here, the parenthesized part of the second term on the right side of equation (1) can be regarded as "1" when a sufficient amount of time has passed since the rise, as shown in equation (3) below.

[0030]

number

[0031] Therefore, the predicted value corresponding to the moisture content data when a sufficient amount of time has passed since the rise is the sum of the initial value Yb and the prediction coefficient A, as shown in the following equation (4). In other words, the prediction means 52c calculates the predicted value by adding the initial value Yb and the prediction coefficient A. However, the prediction means 52c may also perform the calculations based on equations (1) and (2) all at once.

[0032]

number

[0033] The output processing means 52d displays various information on the display unit 55. For example, the output processing means 52d displays on the display unit 55 information including the predicted value obtained by the prediction means 52c.

[0034] The prediction processing unit 52 is composed of one or more processors. In other words, the prediction processing unit 52 can be composed of a calculation device such as a CPU, GPU, or MPU, and software (including a prediction processing program 53p) ​​that cooperates with the calculation device to realize the various functions described above or below. Some of the various functions in the prediction processing unit 52 may be realized by hardware. The prediction processing program 53p is a program that causes a computer to function as an acquisition processing means 52a, an initial calculation means 52b, a prediction means 52c, and an output processing means 52d. The storage unit 53 corresponds to a computer-readable recording medium on which the prediction processing program 53p is recorded.

[0035] Next, a specific configuration example of the measurement processing unit 10 will be described with reference to Figures 2 and 3. Figures 2 and 3 show x, y, and z axes to clarify the positional relationship between the flow of the gas to be measured and the measurement processing unit 10. The measurement processing unit 10 is disposed so that the gas to be measured flows in the flow path member 12 in the direction of the outline arrow shown in Figures 2 and 3 (positive direction of the y axis). Hereinafter, the direction in which the gas to be measured flows in the flow path member 12 will be referred to as the "flow path direction."

[0036] The measurement processing unit 10 illustrated in Figures 2 and 3 has a rectangular parallelepiped outer shell and a hollow housing 11 having two opposing side walls each provided with holes 11h that face each other. A flow path member 12 is disposed such that both ends correspond to the holes 11h. In Figures 2 and 3, the end of the flow path member 12 on the gas inlet side protrudes from the housing 11, and the end on the gas outlet side is substantially flush with the surface of the housing 11. The degree to which both ends of the flow path member 12 protrude from the housing 11 is not limited to the example in Figures 2 and 3, and can be adjusted as desired.

[0037] The measurement processing unit 10 has a moisture-permeable sheet member 13 provided at one opening of the flow path member 12. The sheet member 13 is configured and arranged to function as a check valve that prevents gas from flowing in the opposite direction to the flow path direction within the flow path member 12. The sheet member 13 in each figure is attached so that the degree of opening to the outside increases as the flow rate of gas flowing toward the sheet member 13 within the flow path O increases. Because the measurement processing unit 10 has the sheet member 13, it is possible to suppress the effects of external disturbances on the detection unit 21 and improve the stability of the detection data.

[0038] The sheet member 13 may be arranged so as to completely block one opening of the flow path member 12, but in that case, resistance may increase and moisture permeability may decrease. For this reason, it is desirable that a part of the sheet member 13 of the measurement processing unit 10 be attached to the housing 11 so that it functions as a check valve. The attachment position of the sheet member 13 to the housing 11 is not limited to the examples in Figures 2 and 3, and the sheet member 13 may be attached to various positions, such as a position on the x-axis positive side, a position on the z-axis positive side, or a position on the z-axis negative side of one opening of the flow path member 12.

[0039] 2 and 3 illustrate a heater section 15 having a wire-shaped heat generating portion. The heat generating portion is wound around the outer periphery of the flow path member 12. The number of windings and winding density of the heat generating portion can be adjusted as appropriate. However, the heat generating portion of the heater section 15 is not limited to the examples shown in the figures, and any shape is possible as long as it can heat the flow path member 12 and the gas in the flow path O. The sensor section 20 and the measurement processing device 30 are disposed inside the housing 11. Note that the configurations of the switching section 16 and the light emitting section 17 are merely exemplary. A circuit for driving the heater section 15 is configured, for example, as shown in the circuit diagram of FIG. 4. That is, the measurement processing unit 10 of the first embodiment is configured so that when the switching section 16 is turned on, power is supplied from the power supply device 18 to the heater section 15 and the light emitting section 17.

[0040] [Time required for the mathematical model and graph to converge] The properties of the mathematical model of equation (1) will be explained with reference to Figure 5. Here, we will explain an example of a method for predicting the value of a function that changes over time using the following equation (5), which excludes the initial value Yb on the right side of equation (1).

[0041]

number

[0042] If a mathematical model for prediction is determined for each sensor, it is possible to predict the convergence value from the transient sensor response. Therefore, we wondered whether the mathematical model of equation (5), which is a function of time t and output y, might be suitable for measurements based on temperature and humidity sensors. In equation (5), the convergence value of output y is A. In equation (5), A is a constant and τ is a time constant. The time constant τ is an index that represents the speed of change over time in output y, which corresponds to moisture content data.

[0043] Figure 5 is a graph showing the change in output y when τ in equation (5) is changed from 1, 2, 3, to 4. The output y in each graph roughly converges (saturates) when the time t is 5τ. To be precise, it is estimated that it takes at least 5τ from the rise time to reach 99% of the converged value. In other words, the time constant τ is a characteristic quantity that determines the time it takes for each graph to converge to a constant value.

[0044] [Comparison of actual measurement results with the model curve related to equation (1)] Based on the above, we have constructed a mathematical model to predict the change over time in the measured moisture content data. Table 1 shows the gas to be measured, with a dew point temperature of 14°C. dp Air (first gas) at a temperature of 25°C and a dew point temperature of 19°C dp The figures show the values ​​of each data when air (second gas) at a temperature of 22° C. was used and flowed through the flow path O of the flow path member 12. When measuring the dew point temperature, the heater unit 15 was in an off state.

[0045] [Table 1]

[0046] In Table 1, the temperature and humidity of each humidity-controlled gas are listed in the temperature and humidity column. D , initial value Yb, prediction coefficient A, and time constant τ correspond to the following equation (6). The time constant τ is uniquely determined based on the temperature of each humidity-controlled gas, and is approximately 5 seconds for each controlled gas. Note that equation (6) is essentially the same as equation (1).

[0047]

number

[0048] The time-dependent changes in the measured dew point temperature from the temperature and humidity of each humidity-controlled gas were calculated and graphed. When the mathematical model of Equation (1) was fitted to the graph, a very good fit was obtained, as shown in Figures 6 and 7. Hereinafter, the time-dependent changes in the measured values ​​will also be referred to as the actual changes. Figure 6 is a graph showing the relationship between the measured changes in the dew point temperature of the first gas (temperature 25°C, humidity 50% RH) and the model curve of Equation (6). Figure 7 is a graph showing the relationship between the measured changes in the dew point temperature of the second gas (temperature 22°C, humidity 90% RH) and the model curve of Equation (6).

[0049] Between 0 and 30 seconds, no humidity-controlling gas is flowing through the flow path O, so the measurement processing unit 10 measures the dew-point temperature in the atmosphere. Between 30 and 90 seconds, each humidity-controlling gas is flowing through the flow path O, so the measurement processing unit 10 measures the dew-point temperature of each humidity-controlling gas. During the period when humidity-controlling gas is flowing through the flow path O, the model curve according to equation (6) can be applied to the rising portion of the graph to obtain fitting parameters for the mathematical model.

[0050] [Setting time t S and prediction accuracy) Based on FIG. 8 and Table 2, the above-mentioned set time t S The relationship between the time constant τ and prediction accuracy will be explained. FIG. 8 is a graph showing the measured changes in dew point temperature from 5 seconds after the start of operation and from 40 seconds after the start of operation when a humidity-controlled gas with a temperature of 20°C and a humidity of 80% RH is passed through the flow path O of the flow path member 12 at a constant flow rate. Table 2 is a table corresponding to FIG. 8 and includes multiple waiting periods from the start of operation and predicted values ​​corresponding to each waiting period. In the example of FIG. 8 and Table 2, the time constant τ is set to 4.5.

[0051] [Table 2]

[0052] The predicted times in Table 2 are the times when the predicted values ​​are calculated, and are 1, 2, 3, 4, and 5 seconds after the start-up time. The waiting time is set to t S Here, the dew point temperature is used as the moisture content data.

[0053] For example, when the waiting time is 5 seconds, adding the prediction coefficient A and the initial value Yb results in a predicted value of 16.709°C. dp The prediction coefficient A is calculated by the initial value Yb at the time of rising and the set time t S , and the set time t S The reference value Y(t S ) can be calculated by applying the above formula (2). On the other hand, when the waiting time is 40 seconds, the dew point temperature has almost converged, and the measured value of the dew point temperature at this time is 16.745°C. dp The difference from the predicted value is only 0.036℃. dp In this way, the prediction processing system 100 can accurately obtain a predicted value of the dew-point temperature even in about 1 / 8 of the time required for the actual measured values ​​of the dew-point temperature to converge.

[0054] More specifically, the predicted value when the waiting time is 2 seconds is 16.348°C. dp The difference from the predicted value is 0.397℃. dp Therefore, it can be seen that a certain degree of prediction accuracy is guaranteed even at this point. Furthermore, the predicted value when the waiting time is 3 seconds is 16.666℃. dp , the predicted value when the waiting time is 4 seconds is 16.709℃ dp In both cases, the error is 0.1°C. dp It fits in below.

[0055] Here, comparing the waiting time and time constant τ in Table 2, 1 second = 0.22τ, 2 seconds = 0.44τ, 3 seconds = 0.67τ (≒2τ / 3), 4 seconds = 0.89τ, and 5 seconds = 1.10τ. That is, in the prediction processing system 100, the set time t SIt is estimated that if the set time t is set to 2 / 3 times or more of τ, the moisture content data can be predicted with high accuracy. S It seems that the prediction accuracy will improve if the time is increased, but there is no significant difference in the error after 2τ / 3. If the time is increased unnecessarily, the advantage of shortening the prediction time will be reduced. Therefore, the setting time t S The upper limit of can be considered to be about 2τ.

[0056] [Advantages of providing the flow path member 12] Using the sensor unit 20 alone (see Figure 9) and the sensor unit 20 (sensor unit 25: see Figure 10) with the flow path member 12 attached so as to cover the detection unit 21, we investigated the extent to which the flow path member 12 mitigates the influence of the surrounding environment on the moisture content data.

[0057] (Gas spraying) Fig. 11 is a graph showing the change over time in the measured dew-point temperature when the humidity-conditioning gas is sprayed from the front side of the detection unit 21 in Fig. 9 (see solid line) and when the humidity-conditioning gas is sprayed from the side (see dashed line). Fig. 12 is a graph showing the change over time in the measured dew-point temperature when the humidity-conditioning gas is sprayed from one opening of the flow path member 12 in Fig. 10 (see solid line) and when the humidity-conditioning gas is sprayed from the side of the flow path member 12 (see dashed line). Note that with regard to the direction of gas spraying onto the detection unit 21, the solid line graph in Fig. 11 corresponds to the dashed line graph in Fig. 12, and the dashed line graph in Fig. 11 corresponds to the solid line graph in Fig. 12.

[0058] 11 shows that, unless the detection unit 21 is covered with the flow path member 12, there is not much difference in the measured change in dew point temperature between when the humidity-conditioning gas is sprayed from the front side of the detection unit 21 and when it is sprayed from the side. In addition, although not shown, it was confirmed that even when the humidity-conditioning gas is sprayed from the back side of the detection unit 21, the dew point temperature rises to about half of the level when it is sprayed from the front side.

[0059] On the other hand, from Figure 12 it can be seen that in the case of sensor unit 25, the dew point temperature changes along the model curve only when the humidity-conditioning gas is sprayed in a direction along the flow path O of the flow path member 12. In other words, with sensor unit 25, even when the humidity-conditioning gas is sprayed from the side of the flow path member 12, there is almost no change in the dew point temperature. In this way, by using sensor unit 25, it is possible to accurately capture changes in the moisture content data without being easily affected by external disturbances such as changes in the surrounding airflow, and therefore it is possible to accurately predict the convergence value of the moisture content data.

[0060] (Proximity of steam source) In addition, the influence of a water vapor generation source was examined by bringing it close to the configurations shown in Figures 9 and 10. In Figure 13, the dashed line shows the measured change in dew point temperature when a water vapor generation source is brought close to sensor unit 20 without flow path member 12 (state shown in Figure 9), and the solid line shows the measured change in dew point temperature when a water vapor generation source is brought close to sensor unit 20 without flow path member 12 (state shown in Figure 10). A human finger was brought close to sensor unit 20 as a water vapor generation source.

[0061] As shown in Figure 13, with the sensor unit 20 alone, the influence of humidity fluctuations due to changes in the distance from the water vapor generation source is significantly reflected in the measured change in dew point temperature. On the other hand, with the sensor unit 25, the detection unit 21 is protected by the flow path member 12, so the influence of humidity fluctuations is greatly mitigated. Comparing the two graphs in Figure 13, it can be seen that the influence of the water vapor generation source is greatly suppressed by the flow path member 12. In other words, if the detection unit 21 is covered by the flow path member 12, it becomes less susceptible to changes in the moisture content of the surrounding environment, and therefore it is possible to directly measure only the temperature and humidity of the flowing gas, thereby enabling more accurate prediction of moisture content data.

[0062] (Steam spraying) Furthermore, water vapor was sprayed onto the configurations shown in Figures 9 and 10 to investigate the effect. In Figure 14, the dashed line shows the measured change in dew point temperature when water vapor is sprayed onto the sensor unit 20 without the flow path member 12 (as shown in Figure 9), and the solid line shows the measured change in dew point temperature when water vapor is sprayed onto the sensor unit 20 without the flow path member 12 (as shown in Figure 10). An ultrasonic humidifier was used to spray the water vapor.

[0063] When measuring water vapor emitted from an ultrasonic humidifier, as shown in Figure 14, the sensor unit 25 (with flow path member) showed a more stable rise in the dew point temperature and the dew point temperature converged more quickly. Looking at the convergence value on the graph, the dew point temperature was 15.5°C, which is equivalent to 85% RH when converted to humidity. This is presumably a reflection of the dew point temperature of the air remaining in the flow path O. On the other hand, with the sensor unit 20 alone, not only was the rise unstable, but the dew point temperature continued to rise and become unstable even after the spraying of water vapor had stopped.

[0064] Furthermore, when the mathematical model according to Equation (1) is applied to the rising portion of the sensor response in sensor unit 25, the graph of dew point temperature and the model curve according to Equation (1) match very well. In other words, since the rising portion of the sensor response in sensor unit 25 is stable, it is easy to model the graph, and the accuracy of the parameters used to predict moisture content data can be improved.

[0065] [Advantages of the heater unit 15] 15 to 21, the stabilization of moisture content data by heating the flow path member 12 will be described. In the case of gas with a dew point temperature higher than room temperature, condensation occurs in the flow path O, and it is presumed that the occurrence of this condensation affects the stability of the moisture content data. Therefore, an experiment was conducted to determine whether or not heating the flow path member 12 results in a difference in the measured dew point temperature of the gas passing through the flow path O of the flow path member 12.

[0066] The experiment shown in Fig. 15 was carried out in an environment of room temperature 22°C by flowing humidity-controlled gas with a dew-point temperature of 30°C through the flow path O of the flow path member 12. In Fig. 15, the solid line indicates the measured change in dew-point temperature when the flow path member 12 is heated by the heater unit 15 (set temperature 39°C), and the dashed line indicates the measured change in dew-point temperature when the flow path member 12 is not heated.

[0067] Looking at each graph, in all cases, the dew point temperature rises almost simultaneously with the gas inflow, and rises to a certain extent. In the graph with heating, the temperature converges at 30°C and drops sharply once the gas inflow ends. When measurements are performed repeatedly within a short period of time, a short recovery time for the dew point temperature to return to its original value is desirable. In this regard, the measurement processing unit 10 is suitable for use in such cases because the recovery time is shortened by heating the flow path member 12. On the other hand, in the graph without heating, condensation occurs near the detection unit 21, which not only makes it impossible to accurately measure the dew point temperature but also causes fluctuations in the dew point temperature. Such fluctuations in the sensor output affect the modeling of the sensor response (estimation of numerical parameters), which will be described later. Furthermore, in the graph without heating, the sensor output remains high due to condensation after the gas inflow ends.

[0068] (Modeling of sensor response) While heating the flow path member 12 with the heater section 15, a gas with a dew point temperature higher than room temperature was passed through the flow path O to measure the actual change in dew point temperature and attempt to model the sensor response. Figures 16 to 21 are graphs showing the actual change in dew point temperature of multiple gases with the same temperature but different humidity, and the modeled curves.

[0069] Figure 16 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 40% RH, as well as a model curve. Figure 17 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 50% RH, as well as a model curve. Figure 18 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 60% RH, as well as a model curve. Figure 19 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 70% RH, as well as a model curve. Figure 20 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 80% RH, as well as a model curve. Figure 21 shows the measured change in dew point temperature for air at a temperature of 36°C and a humidity of 90% RH, as well as a model curve.

[0070] By heating the flow path member 12, the dew point temperature can be measured stably, as shown in the graphs of each figure, and the sensor response can be modeled appropriately, as shown by the gray dashed model curves. Table 3 lists the numerical parameters used in the modeling of Figures 16 to 21. These values ​​can be read from the graphs. It can be seen that the predicted value (Yb+A) corresponds to the convergence value of the measured changes in dew point temperature in Figures 16 to 21. Using the above procedure, it is possible to estimate the value of the time constant τ used to predict moisture content data using equations (2) and (4).

[0071] [Table 3]

[0072] (Consider the appropriate temperature setting range for the heater unit 15) Fig. 22 is a graph showing the measured change in dew point temperature of a gas at a temperature of 20°C and a humidity of 90% RH when the set temperature (hereinafter also referred to as the heating temperature) of the heater unit 15 is changed in stages. In the experiment shown in Fig. 22, the heating temperature was set to 22°C, 25°C, 28°C, and 31°C. As a result, as shown in Fig. 22, an overshoot occurred only when the heating temperature was 31°C.

[0073] Fig. 23 is a graph showing the measured changes in dew point temperature of a gas at a temperature of 40°C and a humidity of 90% RH when the heating temperature is changed in stages. In the experiment shown in Fig. 23, the heating temperature was set to 40°C, 45°C, 50°C, 55°C, 60°C, and 65°C. As shown in Fig. 23, no overshoot occurred when the heating temperature was 50°C or lower. In other words, overshoot occurred when the heating temperature was set to 55°C or higher.

[0074] Fig. 24 is a graph showing the measured changes in dew point temperature of a gas at a temperature of 50°C and a humidity of 90% RH when the heating temperature is changed in stages. In the experiment shown in Fig. 24, the heating temperature was set to 50°C, 55°C, 60°C, 65°C, 40°C, and 70°C. As shown in Fig. 24, no overshoot occurred when the heating temperature was 60°C or lower. In other words, overshoot occurred when the heating temperature was set to 65°C or higher.

[0075] Figure 25 is a graph showing the measured changes in dew point temperature for multiple gases with different humidities at a temperature of 20°C when the heating temperature is set to 28°C. As shown in Figure 25, when the heating temperature is set to 8°C higher than the gas temperature, no overshoot occurs, regardless of the gas humidity.

[0076] Figure 26 is a graph showing the measured changes in dew point temperature for multiple gases with different humidities at a temperature of 20°C when the heating temperature is set to 31°C. As shown in Figure 26, when the heating temperature is set to 11°C higher than the gas temperature, overshoot occurs regardless of the gas humidity, although to different degrees.

[0077] Taking into account all the information that can be gleaned from Figures 22 to 26, it can be inferred that overshoot will not occur if the difference between the set temperature of the heater unit 15 and the gas temperature is set to 10°C or less. If overshoot occurs in the sensor response, it becomes more difficult to apply the mathematical model according to Equation (1), which affects the accuracy of calculation of the predicted value. In other words, if the sensor response jumps out to the upper side before the moisture content data converges, it will affect the mathematical modeling. Therefore, it is recommended that the set temperature of the heater unit 15 be set to within +10°C of the gas temperature, that is, so that the difference between the set temperature of the heater unit 15 and the gas temperature is 10°C or less.

[0078] [Effect of closing one end of flow path O] 27, the advantages obtained when one end of the flow path member 12 is closed with the sheet member 13 will be described. 3 A flowing humidity-control gas with a dew-point temperature of 19.3°C at a flow rate of 1 / min (sccm) was used. In Fig. 27, the solid line shows the measured change in dew-point temperature when one opening of the flow path member 12 is blocked with the sheet member 13, and the dashed line shows the measured change in dew-point temperature when both ends of the flow path member 12 are open.

[0079] In the experiment shown in Figure 27, the atmospheric dew-point temperature was measured from 0 to 30 seconds, and the humidity-controlled gas was flowed through the flow path O from 30 to 90 seconds. First, we focus on the measurement portion (baseline) of the atmospheric dew-point temperature from 0 to 30 seconds. When both ends of the flow path member 12 are open, fluctuations of approximately 0.5°C in the dew-point temperature appear. This is presumably due to outside air entering the flow path O when the measurement processing unit 10 is moved. On the other hand, when one end of the outlet side of the flow path member 12 is closed with the sheet member 13, which functions as a check valve, the baseline fluctuations are small. When the baseline is stabilized, the rising portion of the moisture content data becomes clear, enabling more accurate prediction of the dew-point temperature.

[0080] [Gas flow rate and ventilation time] Figure 28 shows a gas with a dew point temperature of about 19°C, with a flow rate of 100 cm 3 / min, 200cm 3 / min, 1000cm 3 28 shows the actual change in dew point temperature when the flow rate was changed from 1 / min to 1 / min and the air was passed through flow path O. In the experiment shown in FIG. 28, the dew point temperature of the outside air was about 10°C. The sensor response time was evaluated based on these results, with reference to Table 4.

[0081] [Table 4]

[0082] For each gas flow rate, the time required for the dew point temperature to change by 90% from the initial value to the convergence value is t 90 and the time it takes for the value to change by 63% from the rising point to the convergence point, t 63 As a result of calculating the two types of flow rate, t 90 is about 12 seconds, and t 63 The time required for measurement was about 5 seconds. That is, the measured change in the dew point temperature of the gas had a unique shape regardless of the flow rate. This is because the amount of gas passing through the flow path O of the flow path member 12 (hereinafter also referred to as the inside of the pipe) was small compared to the flow rate of the gas being measured per unit time.

[0083] The fourth column of Table 4 shows the calculation results indicating the ventilation time of the gas inside the pipe for each flow rate. Since the sampling frequency is 10 Hz, the sampling time is 0.1 seconds. In the method of the first embodiment, which predicts the sensor measurement results using the sensor's time constant τ, it is desirable to keep the ventilation time of the gas inside the pipe as short as possible. From this perspective, it is recommended that the flow path member 12 have an inner diameter of approximately 3 mm to 6 mm and a length in the flow path direction of approximately 35 mm to 65 mm. When designing the flow path member 12, it is recommended to keep in mind that if the length in the flow path direction is too short, it will be susceptible to external disturbances, and if it is too long, it will take up space and be inconvenient to carry.

[0084] An example of the process for predicting the convergence value of moisture content data using the mathematical model according to Equation (1) is shown below, with reference to the actual measurement results shown in Figure 29. In the experiment shown in Figure 29, the gas used for measurement was a gas with a temperature of 22°C and a dew point temperature of 19.2°C. The time constant τ is 4.82 seconds. If the total change from the rise to the convergence value is A, then if the function follows the prediction model, there will be an increase of approximately 63% of the total change (5.79 / 9.1 ≒ 0.63) at t = 4.82 seconds.

[0085] Estimating the amount of change from the graph in Figure 29 gives 0.63A = 5.79, and A = 9.2. In other words, the predicted dew point temperature is 19.3°C, which is a 9.2°C increase from the initial value of 10.1°C. Since the actual convergence value is 19.2°C, the prediction is highly accurate.

[0086] On the other hand, at t = 2.41 seconds (corresponding to 0.5τ), there will be an increase of approximately 39% (3.13 / 9.1 ≒ 0.34) of the total change. Estimating the amount of change from the graph in Figure 29 gives 0.39A = 3.13, and A = 8.0. In other words, the predicted dew point temperature is 18.1°C, an 8.0°C increase from the initial value of 10.1°C. If the time from the rising edge to the predicted time is short, an error will be introduced in the detection of the rising portion of the graph, which will likely result in an inaccurate predicted dew point temperature. Therefore, designing a flow path that minimizes the effects of disturbances is important for accurate prediction of the dew point temperature.

[0087] Let's look again at equation (5) and Figure 5. The approximate value of output y when the same amount of time as the time constant τ has passed since the rise is 0.63A (1 - 0.3678794... ≒ 0.63). Therefore, if the mathematical model of equation (5) matches well with the change in the measured value over time, then equation (7) below can be used to calculate the prediction coefficient A, rather than equation (2). In other words, when the time constant τ has passed since the rise, the convergence value can be found using calculations based on equations (4) and (7).

[0088]

number

[0089] Similarly, when the time constant τ has elapsed since the rise, the approximate value of the output y is 0.39 A (1 - 0.6065306 ... ≒ 0.39). Therefore, if the mathematical model of equation (5) matches well with the change in the measured value over time, the following equation (8) may be used to calculate the prediction coefficient A. However, from the viewpoint of prediction accuracy, equation (7) is more suitable.

[0090]

number

[0091] Next, the flow of operations related to the prediction processing method of the first embodiment will be described with reference to the flowchart of FIG.

[0092] First, the user (operator) checks the outside air temperature and the gas temperature (step S101) and determines whether the temperature of the gas to be measured is higher than the outside air temperature. The outside air temperature here refers to the temperature around the measurement processing unit 10. The user may check the outside air temperature by visually checking a separately provided thermometer. The gas temperature is, for example, the set temperature of an air conditioner (step S102).

[0093] If the gas temperature is higher than the outside air temperature (step S102 / Yes), the user operates the switching unit 16 to start heating the flow path member 12 with the heater unit 15. At this time, the user determines the set temperature of the heater unit 15 so that the value obtained by subtracting the gas temperature from the set temperature of the heater unit 15 is 10°C or less and is higher than the gas temperature. Then, the user registers the set temperature of the heater unit 15 in the prediction processing device 50 (step S103). On the other hand, if the gas temperature is lower than the outside air temperature (step S102 / No), the user keeps the switching unit 16 in the off state and registers the gas temperature in the prediction processing device 50 (step S104).

[0094] When the set temperature of the heater unit 15 is registered (step S102 / Yes, step S103), the prediction processing unit 52 applies the set temperature of the heater unit 15 to the time constant correspondence data to determine the time constant τ (step S105). On the other hand, when the gas temperature is registered (step S102 / No, step S104), the prediction processing unit 52 applies the gas temperature to the time constant correspondence data to determine the time constant τ (step S106).

[0095] Then, the prediction processing unit 52 starts acquiring the temperature and humidity inside the flow path O. The prediction processing unit 52 of the first embodiment starts calculating the dew-point temperature based on the acquired temperature and humidity inside the flow path O (step S107). The prediction processing unit 52 waits until it detects a rise in the measurement value (step S108 / No), and when it detects a rise in the measurement value (step S108 / Yes), it calculates the dew-point temperature Yb at the time of the rise (step S109).

[0096] The prediction processing unit 52 performs the process for a set time t S (Step S110 / No) S When time has passed (step S110 / Yes), the dew point temperature Y(t S ) (step S111). Next, the prediction processing unit 52 calculates the dew-point temperature Yb, the dew-point temperature Y(t S ), and the time constant τ, to obtain a prediction coefficient A from equation (2) (step S112), and to obtain a predicted value (Yb+A) by adding the dew-point temperature Yb and the prediction coefficient A (step S113). Then, the prediction processing unit 52 causes the display unit 55 to display information including the predicted value (Yb+A) (step S114).

[0097] The above-mentioned operational flow has been described in the order of the step numbers shown in FIG. 30, but the order may be changed as appropriate, or multiple steps may be combined into one process, as long as the predicted value (Yb+A) can ultimately be obtained. For example, the prediction processing unit 52 may perform the process of step S107 before or simultaneously with the calculation of the time constant τ (step S105 or step S106). The prediction processing unit 52 may omit the process of obtaining the prediction coefficient A (processing of step S112) and directly obtain the predicted value (Yb+A). While FIG. 30 illustrates the dew-point temperature as moisture content data, the moisture content data is not limited to this, and may also be absolute humidity, water vapor pressure, or the like.

[0098] As described above, the prediction processing device 50 of the first embodiment is configured to calculate a predicted value to which the actual moisture content data will converge, using the time constant τ calculated by applying the gas temperature to the time constant correspondence data, the initial value Yb indicating the moisture content at the time of rise related to the measured humidity change of the gas, and the reference value Y(tk) indicating the moisture content when a set time has elapsed since the rise. Therefore, the moisture content data of the gas can be predicted accurately in a short time, without waiting for the time it takes for the actual sensor value to converge.

[0099] In the first embodiment, a prediction process using the mathematical model of Equation (1) is adopted. That is, the prediction means 52c predicts the initial value Yb, the reference value Y(t S ), and time constant τ (measurement time constant), to determine a prediction coefficient A corresponding to the difference between the true value of the moisture content data and the initial value Yb. The prediction means 52c then adds the determined prediction coefficient A to the initial value Yb to determine a predicted value. Here, equation (1) corresponds to a prediction function in which the initial value Yb is a constant term and the elapsed time t from the rise is a variable, and the sum of the initial value Yb and the prediction coefficient A corresponding to the difference between the true value of the moisture content data and the initial value Yb becomes a convergence value. In other words, the prediction means 52c adds the initial value Yb, the reference value Y(t S ), and time constant τ (measurement time constant) are applied to obtain the predicted value.

[0100] The prediction processing system 100 of the first embodiment has a measurement processing unit 10 that measures the temperature and relative humidity of the gas to be measured, and a prediction processing device 50, and can quickly and accurately predict moisture content data through cooperation between the measurement processing unit 10 and the prediction processing device 50. The prediction processing device 50 has a display unit 55 that displays information including the predicted value, allowing the user to quickly visually confirm the derived predicted value.

[0101] In the measurement processing unit 10 of the first embodiment, the detection unit 21 is disposed within the flow path O of the cylindrical flow path member 12, thereby suppressing the influence of disturbances around the detection unit 21. Therefore, the detection data relating to the temperature and humidity within the flow path O is stabilized, making it possible to achieve stable and highly accurate measurement of moisture content data, which is information indicating the moisture content of the gas. In the first embodiment, since the measurement value by the information processing unit 32 is stable, the moisture content data of the gas within the flow path O is also stable, allowing the prediction processing device 50 to accurately predict the convergence value.

[0102] <Modification> An example of the overall configuration of a prediction processing system according to a modified example will be described with reference to Fig. 31. The same components as those in the main part of the above-described first embodiment will be assigned the same reference numerals, and their description will be omitted or simplified.

[0103] The prediction processing system 100 according to this modification is characterized in that the measurement processing device 30 controls the heater section 15. That is, the information processing section 32 according to this modification has a measurement processing means 32a and a heating control means 32h. The temperature sensor 60 in FIG. 31 is configured to include, for example, a thermistor, and measures the outside air temperature, which is the temperature around the measurement processing unit 10. The temperature sensor 60 may be an internal component of the prediction processing system 100, or may be an external component.

[0104] The measurement processing means 32a measures the temperature and humidity based on the detection data from the detection unit 21, and transmits the measurement values ​​to the prediction processing device 50. When the sensor unit 20 has an information processing unit that measures the temperature and humidity based on the detection data from the detection unit 21, the measurement processing means 32a transmits the measurement values ​​acquired from the sensor unit 20 via the communication unit 31 to the prediction processing device 50.

[0105] The heating control means 32h acquires the outside air temperature from the temperature sensor 60 and also acquires the temperature of the gas to be measured, that is, the assumed temperature of the gas in the flow path O (estimated gas temperature), from the outside. For example, the temperature of the gas to be measured may be configured to be transmitted from the prediction processing device 50 to the measurement processing device 30. Alternatively, the measurement processing device 30 may be provided with a function for setting the gas temperature, and the heating control means 32h may acquire the gas temperature set and registered by the user. Alternatively, if the air blown out from an air conditioner is the object to be measured, the heating control means 32h may acquire the gas temperature (the set temperature of the air conditioner) from the air conditioner or its remote control.

[0106] The heating control means 32h has a function of determining whether the gas temperature is higher than the outside air temperature. If the gas temperature is higher than the outside air temperature, the heating control means 32h calculates an adjustment temperature based on the gas temperature. The heating control means 32h then adds the calculated adjustment temperature to the gas temperature to calculate a set temperature (heating temperature) for the heater unit 15, and causes the heater unit 15 to generate heat at the calculated heating temperature. The heating control means 32h also transmits the calculated set temperature to the prediction processing device 50.

[0107] Here, the adjustment temperature is a temperature to be added to the gas temperature to determine the heating temperature. The adjustment temperature may be a constant set to be higher than 0°C and equal to or lower than 10°C. Alternatively, adjustment data correlating the gas temperature with the adjustment temperature may be stored in the memory unit 33 or the like, and the heating control means 32h may be configured to determine the adjustment temperature by applying the gas temperature to the adjustment data. The adjustment data may be table information correlating multiple gas temperature ranges with multiple adjustment temperatures, or graph information correlating the gas temperature with the adjustment temperature, or a function into which the adjustment temperature is determined by substituting the gas temperature (a function for deriving the adjustment temperature with the gas temperature as a variable). In any case, the adjustment temperature should be higher than 0°C and equal to or lower than 10°C.

[0108] If the gas temperature is equal to or lower than the outside air temperature, the heating control means 32h transmits the gas temperature to the prediction processing device 50. If the prediction processing device 50 has acquired the gas temperature in advance, the heating control means 32h may transmit a non-heating signal to the prediction processing device 50 indicating that heating by the heater unit 15 will not be performed.

[0109] As described above, since the measurement processing device 30 determines the set temperature (including no heating) for the heater unit 15, the prediction processing device 50 calculates the time constant τ based on the information transmitted from the measurement processing device 30. That is, the acquisition processing means 52a stores the heating temperature transmitted from the measurement processing device 30 in the memory unit 53. When the gas temperature is transmitted from the measurement processing device 30, the acquisition processing means 52a stores the gas temperature in the memory unit 53. Then, the initial calculation means 52b applies the gas temperature or the heating temperature to the time constant correspondence data to calculate the time constant τ. The other configurations are the same as those of the main part of the first embodiment described above.

[0110] Next, the flow of operations related to the prediction processing method of this modified example will be described with reference to the flowchart of Fig. 32. Steps that are the same as those in Fig. 30 are given the same reference numerals, and their description will be omitted.

[0111] First, the information processing unit 32 acquires the outside air temperature and the temperature of the gas to be measured (step S201) and determines whether the gas temperature is higher than the outside air temperature (step S202). If the gas temperature is higher than the outside air temperature (step S202 / Yes), the information processing unit 32 calculates an adjustment temperature based on the gas temperature, calculates a heating temperature by adding the adjustment temperature to the gas temperature, and causes the heater unit 15 to generate heat at the calculated heating temperature to heat the flow path member 12. Then, the information processing unit 32 transmits the heating temperature to the prediction processing device 50 (step S203). On the other hand, if the gas temperature is equal to or lower than the outside air temperature (step S202 / No), the information processing unit 32 transmits a gas temperature or non-heating signal to the prediction processing device 50 (step S204).

[0112] When the information processing unit 32 transmits a heating temperature (step S202 / Yes, step S203), the prediction processing unit 52 applies the heating temperature to the time constant correspondence data to determine the time constant τ (step S205). On the other hand, when the information processing unit 32 transmits a gas temperature or a non-heating signal (step S202 / No, step S204), the prediction processing unit 52 applies the gas temperature to the time constant correspondence data to determine the time constant τ (step S206). The series of processes from step S107 to step S114 are the same as those described with reference to FIG. 30.

[0113] As described above, when the estimated gas temperature, which is the temperature of the gas in the flow path O, is higher than the outside air temperature, the information processing unit 32 in this modification controls the heater unit 15 so that the difference obtained by subtracting the estimated gas temperature from the heating temperature is 10°C or less. Therefore, it is possible to suppress the occurrence of condensation in the flow path O of the flow path member 12 and avoid the occurrence of overshoot, thereby enabling stable measurement of moisture content data without requiring user operation. Other effects and the like are the same as those in the main part of the first embodiment.

[0114] Embodiment 2 An example of the overall configuration of a prediction processing system 200 according to the second embodiment of the present invention will be described with reference to Fig. 33. The same components as those in the first embodiment will be denoted by the same reference numerals, and their description will be omitted or simplified.

[0115] The prediction processing system 200 of the second embodiment is characterized in that it employs processing based on machine learning, and the prediction processing device 150 has a prediction processing unit 152 equipped with prediction means 152c. More specifically, the initial calculation means 52b calculates a measurement time constant τ, which is a time constant representing the rate of change over time in the measured value of the moisture content data of the gas, based on the temperature of the gas to be measured. a The actual measured value of the moisture content data at the time of rising relating to the actual measured humidity change of the gas is calculated as the initial value Yb. The memory unit 53 stores the temperature of an arbitrary gas and a learning time constant τ n and the actual measured value of the moisture content data at the time of rising relating to the actual measured humidity change of the gas, and the set time t S The prediction model 53m generated by machine learning based on the actual measured value of the moisture content data of the gas after the time has elapsed and the convergence value of the moisture content data of the gas is stored. The convergence value of the moisture content data of the gas becomes the label in the machine learning. Then, the prediction means 152c calculates the initial value Yb, the reference value Y(t S ), and measurement time constant τ a is input into the prediction model 53m to obtain a predicted value to which the actual measured value of the moisture content data for the gas to be measured will converge.

[0116] The prediction model 53m may be generated outside the prediction processing device 150 and then stored in the storage unit 53. However, in the second embodiment, the learning processing means 152e of the prediction processing unit 152 generates the prediction model 53m. That is, the learning processing means 152e generates the prediction model 53m by calculating the temperature of an arbitrary gas and the learning time constant τ associated with the gas. n and the actual measured value of the moisture content data at the time of rising relating to the actual measured humidity change of the gas, and the set time t S By machine learning based on the actual measured value of the moisture content data of the gas after the time has elapsed and the convergence value of the moisture content data of the gas, the initial value Yb and the reference value Y(t S ), and measurement time constant τ aA prediction model 53m that outputs a predicted value corresponding to the input data is constructed. The learning processing means 152e in the second embodiment is configured to generate the prediction model 53m by supervised learning using a DNN (Deep Neural Network). However, the learning processing means 152e may also generate the prediction model 53m by other machine learning techniques. Other configurations of the prediction processing system 200 are similar to those of the prediction processing system 100 in the first embodiment.

[0117] Next, the flow of operations related to the prediction processing method of the second embodiment will be described with reference to the flowchart of Fig. 34. The same steps as those in Fig. 30 are denoted by the same reference numerals, and their description will be omitted. Note that, in steps S105 and S106 in Fig. 34, the learning time constant τ n To distinguish between the measured time constant τ a That is, the only difference from the prediction processing method of the first embodiment is step S301. Specifically, the prediction processing unit 52 calculates the dew-point temperature Yb, the dew-point temperature Y(t S ), and measurement time constant τ a is input to the prediction model 53m, and a predicted value is calculated to determine the value to which the actually measured value of the moisture content data related to the gas to be measured will converge (step S301).

[0118] As described above, the prediction processing system 200 of the second embodiment uses the prediction model 53m based on machine learning to calculate predicted values. Therefore, even when it is difficult to apply a mathematical model to the actual changes in moisture content data of the gas being measured, the predicted value of moisture content data can be calculated quickly and accurately. Furthermore, the measurement processing unit 10 of the second embodiment has the detection unit 21 disposed within the flow path O of the cylindrical flow path member 12, thereby suppressing the influence of disturbances around the detection unit 21. Therefore, the detection data related to the temperature and humidity within the flow path O is stabilized, enabling stable and highly accurate measurement of moisture content data, which is information indicating the moisture content of the gas. Other effects are similar to those of the first embodiment. The prediction processing program 53p of the second embodiment is a program for causing a computer to function as the acquisition processing means 52a, initial calculation means 52b, prediction means 152c, output processing means 52d, and learning processing means 152e. The configuration of the modified example can also be applied to the configuration of the second embodiment. That is, the information processing unit 32 may include a heating control means 32h.

[0119] Embodiment 3 An example configuration of a prediction processing system 300 according to the third embodiment of the present invention will be described with reference to Fig. 35. The same components as those in the first and second embodiments are denoted by the same reference numerals, and their description will be omitted or simplified.

[0120] The prediction processing system 300 includes a measurement processing unit 210 and a prediction processing device 250. The prediction processing system 300 of the third embodiment is characterized in that the measurement processing device 230 of the measurement processing unit 210 predicts the convergence value of the moisture content data. More specifically, the measurement processing device 230 includes a communication unit 31, an information processing unit 232, and a storage unit 33.

[0121] The information processing unit 232 has a measurement processing means 32a, an initial calculation means 232b, and a prediction means 232c. The storage unit 33 stores various information in addition to the operation programs of the information processing unit 232, including a prediction processing program 33p. The prediction processing program 33p is a program for causing a computer to function as at least the initial calculation means 232b and the prediction means 232c.

[0122] The initial calculation means 232b functions in the same way as the initial calculation means 52b in the first embodiment. That is, the initial calculation means 232b applies the gas temperature or the set temperature of the heater unit 15 to the time constant correspondence data to obtain the time constant τ (measurement time constant). The initial calculation means 232b also detects the rise of the measurement value. Furthermore, the initial calculation means 232b obtains the actual measured value of the moisture content data at the time of the rise of the measurement value as the initial value Yb.

[0123] The prediction means 232c functions in the same manner as the prediction means 52c in the first embodiment. That is, the prediction means 232c predicts whether the predetermined time t S The actual measured value of the moisture content data after t S ) is calculated as the initial value Yb and the reference value Y(t S ), and the time constant τ, to obtain a predicted value to which the measured value of the moisture content data will converge. Then, prediction means 232c transmits the obtained predicted value to prediction processing device 250. Therefore, prediction processing unit 252 of prediction processing device 250 obtains the predicted value transmitted from measurement processing device 230 by acquisition processing means 52a, and stores it in memory unit 33. Measurement processing device 230 of this third embodiment corresponds to the prediction processing device according to the claims of this application.

[0124] The operational flow of the prediction processing method of the third embodiment is the same as the flow of the flowchart in Fig. 30. Specifically, in the third embodiment, the processes of steps S105 to S113 are executed by the information processing unit 232. However, it is not necessary for the information processing unit 232 to execute all of the processes of steps S105 to S113. For example, the processes of steps S105 to S111 may be executed by the information processing unit 232, and the processes of steps S112 to S113 may be executed by the prediction processing unit 252.

[0125] As described above, in the configuration of the third embodiment, the measurement processing unit 210 can perform all processes up to the calculation of the predicted value. Therefore, the prediction processing device 250 manages the predicted value transmitted from the measurement processing device 230 and displays the predicted value on the display unit 55 as necessary. Here, the measurement processing device 230 may have a display unit, such as a liquid crystal panel, that displays various information. With such a configuration, the measurement processing device 230 can display the calculated predicted value on the display unit. However, the measurement processing device 230 may also have an operation unit that accepts user operations or a touch panel. Furthermore, in the measurement processing unit 210 of the third embodiment, the detection unit 21 is disposed within the flow path O of the cylindrical flow path member 12, thereby suppressing the influence of disturbances around the detection unit 21. Therefore, the detection data related to the temperature and humidity within the flow path O is stabilized, enabling stable and highly accurate measurement of moisture content data, which is information indicating the moisture content of the gas. In other words, since the measurement value by the information processing unit 232 is stable, the moisture content data of the gas within the flow path O is also stable, allowing the measurement processing device 230 to accurately predict the convergence value. Other effects are the same as those of the first and second embodiments.

[0126] The configuration of the modified example can also be applied to the configuration of the third embodiment. That is, the information processing unit 232 may include a heating control means 32h. Furthermore, the measurement processing device 230 of the third embodiment may also incorporate processing based on machine learning, similar to the prediction processing device 150 of the second embodiment. That is, the prediction means 232c calculates the dew-point temperature Yb, the dew-point temperature Y(t S ), and measurement time constant τ amay be used as an input for the prediction model to obtain a predicted value to which the actual measured value of the moisture content data for the gas to be measured will converge. S ), and measurement time constant τ a The prediction model may include a learning processing means for constructing the prediction model and outputting a predicted value corresponding to the temperature of an arbitrary gas and a learning time constant τ associated with the gas. n and the actual measured value of the moisture content data at the time of rising relating to the actual measured humidity change of the gas, and the set time t S The moisture content data of the gas after the lapse of time is generated by machine learning based on the actual measured value of the moisture content data of the gas and the converged value of the moisture content data of the gas. The measurement processing device 230 having such a configuration also corresponds to the prediction processing device according to the claims of this application.

[0127] The above-described embodiments are merely examples of the measurement processing unit and prediction processing system, and the technical scope of the present invention is not limited to these embodiments. For example, although FIGS. 2 and 3 illustrate a rectangular parallelepiped housing 11, the housing 11 may have various shapes, such as a cuboid, a prolate spheroid, or a sphere. The flow path member 12 may be disposed outside the housing 11. In this case, the detection unit 21 may be disposed outside the housing 11 by passing through a hole provided in the housing 11 and a hole in the flow path member 12, eliminating the need for the pair of holes 11h in the housing 11. The measurement processing units 10 and 210 may not include the light-emitting unit 17, but the presence of the light-emitting unit 17 allows the state of the heater unit 15 to be grasped at a glance. The measurement processing units 10 and 210 may not include the sheet member 13, but the presence of the sheet member 13 can suppress the influence of external disturbances on the detection unit 21 and improve the stability of the detection data. The measurement processing units 10 and 210 do not necessarily have to include the heater section 15. However, if the heater section 15 is included, it is possible to suppress the occurrence of condensation inside the flow path member 12 and stabilize the change over time in the measured value of the moisture content data, even in an environment where the outside air temperature is likely to be lower than the temperature of the gas being measured, such as in cold countries or during the winter in Japan. [Explanation of symbols]

[0128] 10, 210 measurement processing unit, 11 housing, 11h hole, 12 flow path member, 13 sheet member, 15 heater section, 16 switching section, 17 light emitting section, 18 power supply device, 20 sensor section, 21 detection section, 25 sensor unit, 30, 230 measurement processing device, 31, 51 communication section, 32, 232 information processing section, 32a measurement processing means, 32h heating control means, 33, 53 memory section, 33p, 53p prediction processing program, 50, 150, 250 prediction processing device, 52, 152, 252 prediction processing section, 52a acquisition processing means, 52b, 232b initial calculation means, 52c, 152c, 232c prediction means, 52d output processing means, 53m prediction model, 54 operation section, 55 display section, 60 Temperature sensor, 100, 200, 300 prediction processing system, 152e learning processing means, A prediction coefficient, O flow path, Y reference value (dew point temperature), Yb initial value (dew point temperature), t S Setting time, τ time constant, τa measurement time constant, τn learning time constant.

Claims

1. an initial calculation means for calculating a measurement time constant, which is a time constant representing the rate of change over time of the measured value of information indicating the moisture content of the gas, based on the temperature of the gas to be measured, and for calculating an initial value of the measured value of the information indicating the moisture content at the time of the rise of the measured humidity change of the gas; a prediction means for determining an actual measured value of the information indicating the moisture content after a set time has elapsed since the rise as a reference value, and for determining a predicted value to which the actual measured value of the information indicating the moisture content will converge using the initial value, the reference value, and the measurement time constant.

2. The prediction means 2. The prediction processing device according to claim 1, wherein a prediction coefficient corresponding to a difference between a true value of the information indicating the moisture content and the initial value is calculated using the initial value, the reference value, and the measurement time constant, and the calculated prediction coefficient is added to the initial value to calculate the predicted value.

3. a storage unit that stores a prediction function in which the initial value is a constant term and the elapsed time from the rising time is a variable, and the sum of the initial value and a prediction coefficient corresponding to the difference between the true value of the information indicating the moisture content and the initial value is a convergence value; The prediction means The prediction processing device according to claim 1 , wherein the predicted value is obtained by applying the initial value, the reference value, and the time constant to the prediction function.

4. a memory unit that stores a prediction model generated by machine learning based on the temperature of an arbitrary gas, a learning time constant that is the time constant associated with the gas, an actual measurement value of information indicating the moisture content at the time of rise related to the actual humidity change of the gas, an actual measurement value of information indicating the moisture content of the gas after a set time has elapsed since the rise, and a convergence value of the moisture content data of the gas, The prediction means The prediction processing device according to claim 1 , wherein the initial value, the reference value, and the measurement time constant are input to the prediction model to obtain the predicted value.

5. a measurement processing unit that measures the temperature and relative humidity of the gas to be measured; A prediction processing system comprising: the prediction processing device according to any one of claims 1 to 4.

6. an initial calculation means for calculating a measurement time constant, which is a time constant representing the rate of change over time of information indicating the moisture content of the gas, based on the temperature of the gas to be measured, and for calculating an actual measured value of the information indicating the moisture content at the time of rise of the actual measured humidity change of the gas as an initial value; and a prediction processing program for causing a computer to function as a prediction means for determining an actual measured value of the information indicating the moisture content when a set time has elapsed since the start-up as a reference value, and for determining a predicted value to which the actual measured value of the information indicating the moisture content will converge using the initial value, the reference value, and the measurement time constant.

7. by one or more processors, A measurement time constant is calculated based on the temperature of the gas to be measured, the measurement time constant being a time constant that represents the rate of change over time in information indicating the moisture content of the gas; The actual measured value of the information indicating the moisture content at the time of rising relating to the actual measured humidity change of the gas is obtained as an initial value; an actual measurement value of the information indicating the moisture content when a set time has elapsed since the start of the test is obtained as a reference value; A prediction processing method for obtaining a predicted value to which an actually measured value of the information indicating the moisture content will converge using the initial value, the reference value, and the measurement time constant.

Images