Flow rate measuring method, program, information processing apparatus, and error measuring method

Abstract

To provide a flow rate measuring method or the like capable of suitably estimating a flow rate of a fluid flowing through a pipe.SOLUTION: The flow rate measuring method causes a computer to execute processing of acquiring a measured value of a flow rate of a fluid flowing through a pipe measured by an ultrasonic flowmeter 2 and data indicating a flow condition of the fluid, inputting the acquired measured value and the data indicating the flow condition to an error prediction model for outputting an error between the measured value and an actual flow rate in a case where the measured value and the data indicating the flow condition are input to output the error, and estimating the actual flow rate based on the output error and the measured value.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a flow rate measurement method, a program, an information processing device, and an error measurement method. [Background technology]

[0002] Clamp-on ultrasonic flowmeters are used as a non-destructive method for measuring the flow rate of a fluid (e.g., steam flow) flowing through a pipe. For example, Patent Document 1 discloses a flow pattern discrimination device that discriminates the flow pattern of a fluid flowing through a pipe based on the standard deviation of the amplitude of the received signal obtained by receiving ultrasonic waves transmitted into the pipe by the clamp-on ultrasonic flowmeter. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7246634 Summary of the Invention [Problem to be solved by the invention]

[0004] In one aspect, an object is to provide a flow rate measurement method and the like that can suitably estimate the flow rate and the like of a fluid flowing inside a pipe. [Means for solving the problem]

[0005] In one aspect, the flow rate measurement method acquires a measurement value measured by an ultrasonic flow meter of the flow rate of a fluid flowing through a pipe and data indicating the flow state of the fluid, inputs the acquired measurement value and the data indicating the flow state into an error prediction model that outputs the error between the measurement value and the actual flow rate when the measurement value and the data indicating the flow state are input, and outputs the error, and a computer executes a process to estimate the actual flow rate based on the output error and the measurement value. [Effects of the Invention]

[0006] In one aspect, the flow rate and the like of a fluid flowing through a pipe can be suitably estimated. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is an explanatory diagram illustrating an example of the configuration of a flow rate estimation system. [Figure 2] FIG. 1 is a block diagram illustrating an example of the configuration of an estimation device. [Figure 3] FIG. 1 is a diagram illustrating an overview of an embodiment. [Figure 4] FIG. 1 is an explanatory diagram of a guided wave. [Figure 5] 10 is a flowchart showing the steps of a process for generating a flow state prediction model. [Figure 6] 10 is a flowchart showing the procedure of a process for generating an error prediction model. [Figure 7] 10 is a flowchart showing the procedure of a flow rate estimation process. DETAILED DESCRIPTION OF THE INVENTION

[0008] The present invention will be described in detail below with reference to the drawings showing embodiments thereof. (Embodiment) 1 is an explanatory diagram showing an example of the configuration of a flow rate estimation system. In this embodiment, a flow rate estimation system that estimates the flow rate of a fluid (e.g., steam flow) flowing in a pipe will be described. The flow rate estimation system includes an estimation device 1 and an ultrasonic flowmeter 2.

[0009] The ultrasonic flowmeter 2 is a so-called clamp-on type ultrasonic flowmeter and includes a first ultrasonic transducer 21 and a second ultrasonic transducer 22. The first ultrasonic transducer 21 is installed upstream of the flow of fluid flowing through a pipe, and the second ultrasonic transducer 22 is installed downstream of the first ultrasonic transducer 21. The first ultrasonic transducer 21 and the second ultrasonic transducer 22 each include a transmitting unit that transmits ultrasonic waves and a receiving unit that receives ultrasonic waves, and are configured to transmit ultrasonic waves into the pipe and receive ultrasonic waves transmitted from the other. The first ultrasonic transducer 21 and the second ultrasonic transducer 22 convert the ultrasonic waves received by the receiving unit into reception signals (voltage signals) and output the same.

[0010] Because the fluid flowing in the pipe flows in a fixed direction (leftward in FIG. 1), a time difference occurs between the propagation time t1 of the ultrasonic wave transmitted from the first ultrasonic transducer 21 to the second ultrasonic transducer 22 and the propagation time t2 of the ultrasonic wave transmitted from the second ultrasonic transducer 22 to the first ultrasonic transducer 21. Based on this time difference Δt = t2 - t1, the ultrasonic flowmeter 2 measures (calculates) the flow rate of the fluid flowing in the pipe with reference to the diameter D of the pipe, the incident angle θ of the ultrasonic waves, etc. The method of measuring the flow rate in a clamp-on ultrasonic flowmeter is well known, and therefore a detailed description thereof will be omitted in this embodiment.

[0011] The estimation device 1 is an information processing device capable of various information processing, such as a personal computer or a server computer. In this embodiment, the estimation device 1 predicts an error between the measurement value of the flow rate measured by the ultrasonic flowmeter 2 and the actual flow rate, and estimates the actual flow rate based on the predicted error. As will be described in detail later, the estimation device 1 estimates the actual flow rate using two models: a flow state prediction model 51 for predicting the flow state (flow pattern) of the fluid flowing through the pipe, and an error prediction model 52 for predicting the error between the measurement value of the flow rate and the actual flow rate (see FIG. 3). The estimation device 1 first inputs waveform data of the ultrasonic waves received by the ultrasonic flowmeter 2 into the flow state prediction model 51 to calculate a score (data) indicating the flow state of the fluid. Next, the estimation device 1 inputs the calculated score and the measurement value of the flow rate measured by the ultrasonic flowmeter 2 into the error prediction model 52 to calculate the error between the measurement value and the actual flow rate. The estimation device 1 then estimates the actual flow rate based on the calculated error and the measurement value of the flow rate.

[0012] In this embodiment, the estimation device 1 is described as being a separate device from the ultrasonic flowmeter 2, but the estimation device 1 and the ultrasonic flowmeter 2 may be configured as an integrated device.

[0013] 2 is a block diagram showing an example of the configuration of the estimation device 1. The estimation device 1 includes a control unit 11, a main memory unit 12, a communication unit 13, a display unit 14, an input unit 15, and an auxiliary memory unit 16. The control unit 11 has one or more arithmetic processing devices such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), a GPU (Graphics Processing Unit), etc., and performs various information processing, control processing, etc. by reading and executing programs stored in the auxiliary storage unit 16. The main storage unit 12 is a temporary storage area such as an SRAM (Static Random Access Memory) or a DRAM (Dynamic Random Access Memory), and temporarily stores data necessary for the control unit 11 to execute arithmetic processing. The communication unit 13 is a communication module for performing communication-related processing, and transmits and receives information to and from the outside. The display unit 14 is a display screen such as a liquid crystal display, and displays images. The input unit 15 is an operation interface such as a keyboard or a mouse, and accepts operation input.

[0014] The auxiliary storage unit 16 is a non-volatile storage area such as a large-capacity memory or a hard disk, and stores programs (program products) and other data required for the control unit 11 to execute processing. The auxiliary storage unit 16 also stores a flow state prediction model 51 and an error prediction model 52. The flow state prediction model 51 is a machine learning model that has learned predetermined training data, and is a model that outputs a score (data) indicating the flow state of the fluid when waveform data of ultrasonic waves transmitted into the pipe is input. The error prediction model 52 is a multiple regression model constructed from predetermined training data, and is a model that outputs the error between the measurement value and the actual flow rate when a measurement value of the flow rate of the fluid flowing through the pipe measured by the ultrasonic flowmeter 2 and a score (data) indicating the flow state of the fluid are input.

[0015] The auxiliary storage unit 16 may be an external storage device connected to the estimation device 1. The estimation device 1 may also be a multi-computer consisting of multiple computers, or may be a virtual machine virtually constructed by software.

[0016] Furthermore, in this embodiment, the estimation device 1 is not limited to the above configuration, and may include, for example, an input unit that accepts operation input, a display unit that displays images, etc. Furthermore, the estimation device 1 may include a reading unit that reads a portable storage medium 1a such as a CD (Compact Disk)-ROM or a DVD (Digital Versatile Disc)-ROM, and may read and execute a program from the portable storage medium 1a.

[0017] 3 is a diagram showing an outline of the embodiment, and the outline of the embodiment will be described with reference to FIG.

[0018] 3 conceptually illustrates the processing flow from estimating the actual flow rate of the fluid flowing through the pipe based on the waveform (upper left) of the ultrasonic waves received by the ultrasonic flowmeter 2. As already described, the estimation device 1 performs processing to estimate the actual flow rate using two models, the flow state prediction model 51 and the error prediction model 52.

[0019] The flow state prediction model 51 is a machine learning model that has learned predetermined training data, and is a model that outputs a score indicating the flow state of a fluid when waveform data of an ultrasonic wave transmitted into a pipe through which a fluid flows is input. In this embodiment, an SVM (Support Vector Machine) is used as the flow state prediction model 51.

[0020] In this embodiment, the flow state prediction model 51 is described as being an SVM, but the flow state prediction model 51 may be another machine learning model such as a neural network. In particular, when the flow state prediction model 51 is a neural network, the received waveform of the ultrasonic wave received by the ultrasonic flowmeter 2 can be used as an input as is to calculate the score.

[0021] The "waveform data" input to the flow state prediction model 51 is a parameter (feature amount) that represents the characteristics of the received ultrasonic waveform, and is a statistical index value such as the average amplitude of the received waveform, standard deviation, skewness, kurtosis, etc. As described above, the received waveform may be used as is for input as the "waveform data."

[0022] The "score indicating the flow state" output from the flow state prediction model 51 is a score that expresses the flow state (flow pattern) of the fluid flowing inside the pipe with a value between 0 and 1, for example. In this embodiment, two types of flow patterns, "wavy flow" and "annular mist flow", are handled as the flow states to be predicted, and "wavy flow" is expressed as "0" and "annular mist flow" as "1".

[0023] In this embodiment, the problem handled by the flow state prediction model 51 is described as a regression problem, and the flow state is expressed by a score of 0 to 1, but the problem handled by the flow state prediction model 51 may be a classification problem, and the flow state may be classified into classes using two values, 0 or 1. In other words, the output of the flow state prediction model 51 may be data indicating the flow state, and the output is not limited to continuous numerical values ​​(scores) from 0 to 1.

[0024] In addition, in this embodiment, the two types of flow patterns to be predicted are "wavy flow" and "annular mist flow", but three or more types of flow patterns may also be predicted.

[0025] The estimation device 1 learns training data in which training waveform data is associated with correct scores indicating the flow state of a fluid, thereby generating a flow state prediction model 51. Using the generated flow state prediction model 51, the estimation device 1 outputs a score indicating the flow state of a fluid flowing in a pipe to be predicted.

[0026] Specifically, the estimation device 1 acquires multiple received waveforms obtained by transmitting ultrasonic waves into the pipe multiple times (for example, hundreds to thousands of times) from the ultrasonic flowmeter 2. The estimation device 1 calculates an average waveform from the multiple received waveforms. The estimation device 1 calculates index values ​​such as the average amplitude, standard deviation, skewness, and kurtosis of the calculated average waveform as waveform data to be used as input (feature values) to the flow state prediction model 51.

[0027] The estimation device 1 extracts a guided wave region corresponding to the received portion of the guided wave that has propagated within the wall of the pipe from the waveform data calculated above, and uses the extracted region as an input to the flow state prediction model 51.

[0028] Figure 4 is an explanatory diagram of guided waves. Guided waves are a type of ultrasonic wave used in the field of non-destructive testing, and are waves that propagate longitudinally along the boundary surface of a plate or cylindrical pipe. The cross-sectional view of a pipe shown on the left side of Figure 4 illustrates the propagation of ultrasonic waves within the flow path of the pipe, as well as the propagation of guided waves within the pipe wall.

[0029] As shown on the right side of Figure 4, in the received waveform of the ultrasonic waves, a portion corresponding to the guided wave is observed a certain time after the start of ultrasonic transmission. According to the research of the inventors of this application, the characteristics of the flow state are expressed in this guided wave.

[0030] Therefore, the estimation device 1 extracts a guided wave region from the ultrasonic waveform data and uses it as input to the flow state prediction model 51. Specifically, the estimation device 1 identifies a region where the standard deviation (index value) of the amplitude of the received waveform rises as the guided wave region, and extracts waveform data of the region.

[0031] Returning to Figure 3, the explanation will continue. The estimation device 1 inputs the waveform data extracted above into a flow state prediction model 51, and outputs a score indicating the flow state of the fluid flowing inside the pipe. The estimation device 1 uses the score output from the flow state prediction model 51 as input to a subsequent error prediction model 52, and outputs (calculates) the error between the measurement value of the ultrasonic flowmeter 2 and the actual flow rate.

[0032] The error prediction model 52 is a multiple regression model constructed from predetermined training data, and when the measurement value of the flow rate of the fluid flowing through the pipe measured by the ultrasonic flowmeter 2 and a score indicating the flow state of the fluid are input, the model outputs the error between the measurement value and the actual flow rate.

[0033] In this embodiment, the error prediction model 52 is described as being a multiple regression model, but the error prediction model 52 is not limited to being a multiple regression model and may be, for example, a machine learning model such as a neural network.

[0034] The estimation device 1 generates an error prediction model 52 using training data in which the ground truth value of the error between the measurement value and the actual flow rate is associated with the measurement value of the flow rate obtained by the ultrasonic flowmeter 2 and a score indicating the flow state of the fluid. The estimation device 1 inputs the measurement value of the flow rate of the fluid flowing in the pipe to be predicted, measured by the ultrasonic flowmeter 2, and the score output from the flow state prediction model 51 into the generated error prediction model 52, and outputs the error between the measurement value and the actual flow rate.

[0035] The estimation device 1 estimates the actual flow rate based on the error output from the error prediction model 52 and the measurement value of the flow rate measured by the ultrasonic flowmeter 2. That is, the estimation device 1 corrects the measurement value of the ultrasonic flowmeter 2 based on the error output from the error prediction model 52. For example, the estimation device 1 displays the estimated value of the actual flow rate on the display unit 14 and presents it to the user.

[0036] As described above, the estimation device 1 predicts the error between the measurement value and the actual flow rate by inputting not only the flow rate measurement value obtained by the ultrasonic flowmeter 2 but also the score indicating the flow state predicted by the flow state prediction model 51 into the error prediction model 52. According to research by the inventors of the present application, by providing the flow state prediction result obtained by the flow state prediction model 51 to the error prediction model 52, the accuracy of error prediction can be improved.

[0037] 5 is a flowchart showing the procedure of the process for generating the flow state prediction model 51. The process for generating the flow state prediction model 51 by machine learning will be described with reference to FIG. The control unit 11 of the estimation device 1 acquires training data for generating the flow state prediction model 51 (step S11). The training data is data in which waveform data of ultrasonic waves transmitted into the pipe is associated with correct scores (data) indicating the flow state of the fluid flowing through the pipe. The "waveform data" is index values ​​such as the mean, standard deviation, skewness, and kurtosis related to the amplitude of the received waveform of ultrasonic waves, but the received waveform may be treated as waveform data as is.

[0038] Based on the training data, the control unit 11 generates a flow state prediction model 51 that outputs a score indicating the flow state of the fluid flowing through the pipe when ultrasonic waveform data is input (step S12). Specifically, as described above, the control unit 11 generates an SVM as the flow state prediction model 51. The control unit 11 then ends the series of processes.

[0039] 6 is a flowchart showing the procedure of the process for generating the error prediction model 52. The process for generating the error prediction model 52 will be described with reference to FIG. The control unit 11 of the estimation device 1 acquires training data for generating the error prediction model 52 (step S31). The training data is data in which the flow rate of the fluid flowing through the pipe is measured by the ultrasonic flowmeter 2, and the correct value of the error between the measurement value and the actual flow rate is associated with a score (data) indicating the flow state of the fluid.

[0040] Based on the training data, the control unit 11 generates an error prediction model 52 that outputs the error between the measurement value and the actual flow rate when the measurement value of the flow rate and the score indicating the flow state are input (step S32). Specifically, as described above, the control unit 11 generates a multiple regression model as the error prediction model 52. The control unit 11 then ends the series of processes.

[0041] 7 is a flowchart showing the procedure of the flow rate estimation process, and the processing content when estimating the actual flow rate flowing through the pipe will be described with reference to FIG. The control unit 11 of the estimation device 1 acquires from the ultrasonic flowmeter 2 the received waveform of ultrasonic waves transmitted into a pipe through which a fluid flows (step S51). For example, the control unit 11 acquires multiple received waveforms obtained by transmitting ultrasonic waves multiple times into the pipe. The control unit 11 also acquires a measured value of the flow rate of the fluid flowing through the pipe from the ultrasonic flowmeter 2 (step S52).

[0042] The control unit 11 calculates an average waveform of the multiple received waveforms acquired in step S51, and calculates index values ​​such as the mean, standard deviation, skewness, and kurtosis related to the amplitude of the average waveform as waveform data (step S53).The control unit 11 extracts waveform data of a guided wave region corresponding to the received portion of the guided wave that propagated inside the wall of the pipe based on the calculated index value (standard deviation) (step S54).

[0043] The control unit 11 inputs the extracted waveform data into the flow state prediction model 51, thereby outputting a score (data) indicating the flow state of the fluid flowing through the pipe (step S55). The control unit 11 inputs the output score and the measurement value of the flow rate measured by the ultrasonic flowmeter 2 into the error prediction model 52, thereby outputting the error between the measurement value and the actual flow rate (step S56). The control unit 11 estimates the actual flow rate based on the output error and the measurement value of the flow rate (step S57). The control unit 11 ends the series of processes.

[0044] In the above, the generation (learning) of the flow state prediction model 51 and the error prediction model 52 and the estimation of the flow rate based on the models are performed by the same device (estimation device 1), but both processes may be performed by separate devices.

[0045] As described above, according to this embodiment, it is possible to suitably estimate the error between the flow rate measured by the ultrasonic flowmeter 2 and the actual flow rate, the flow rate of the fluid flowing in the pipe, and the like.

[0046] The embodiments disclosed herein are to be considered as illustrative in all respects and not restrictive. The scope of the present invention is defined by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims.

[0047] The matters described in each embodiment can be combined with each other. Furthermore, the independent claims and dependent claims described in the claims can be combined with each other in any combination, regardless of the reference format. Furthermore, although the claims use a format in which a claim references two or more other claims (multi-claim format), this is not limited to this format. A format in which multiple claims (multi-multi-claims) that reference at least one other multiple claim may also be used. [Explanation of symbols]

[0048] 1. Estimation device (information processing device) 11 Control section 12 Main memory 13 Communications Department 14 Display section 15 Input section 16 Auxiliary storage 51 Flow state prediction model 52 Error prediction model

Claims

1. A measurement value of the flow rate of a fluid flowing through a pipe measured by an ultrasonic flowmeter and data indicating the flow state of the fluid are acquired; inputting the acquired measurement value and the data indicating the flow state into an error prediction model that outputs an error between the measurement value and an actual flow rate when the measurement value and the data indicating the flow state are input, and outputting the error; The actual flow rate is estimated based on the output error and the measurement value. A flow measurement method in which processing is performed by a computer.

2. Acquiring waveform data of the ultrasonic waves obtained by receiving the ultrasonic waves transmitted into the pipe; inputting the acquired waveform data into a flow state prediction model that has been trained to output data indicating the flow state when the waveform data is input, thereby outputting data indicating the flow state; The output data indicating the flow state and the measurement value are input to the error prediction model to output the error. The method of claim 1 .

3. extracting a guided wave region from the waveform data, the guided wave region corresponding to a received portion of the guided wave propagating within the wall of the pipe; The extracted waveform data in the guided wave region is input to the flow state prediction model, and data indicating the flow state is output. The method of claim 2 .

4. Acquire the received waveform of the ultrasonic wave; calculating a predetermined index value as the waveform data from the received waveform; The calculated waveform data is input to the flow state prediction model, and data indicating the flow state is output. The method of claim 2 .

5. Acquiring the received waveforms obtained by transmitting the ultrasonic waves a plurality of times; Calculating an average waveform from the received waveforms multiple times; The index value is calculated from the average waveform. The method of claim 4 .

6. A measurement value of the flow rate of a fluid flowing through a pipe measured by an ultrasonic flowmeter and data indicating the flow state of the fluid are acquired; inputting the acquired measurement value and the data indicating the flow state into an error prediction model that outputs an error between the measurement value and an actual flow rate when the measurement value and the data indicating the flow state are input, and outputting the error; The actual flow rate is estimated based on the output error and the measurement value. A program that causes a computer to perform a process.

7. An information processing device including a control unit, The control unit A measurement value of the flow rate of a fluid flowing through a pipe measured by an ultrasonic flowmeter and data indicating the flow state of the fluid are acquired; inputting the acquired measurement value and the data indicating the flow state into an error prediction model that outputs an error between the measurement value and an actual flow rate when the measurement value and the data indicating the flow state are input, and outputting the error; The actual flow rate is estimated based on the output error and the measurement value. Information processing device.

8. Acquire training data in which a measurement value of a flow rate of a fluid flowing through a pipe measured by an ultrasonic flowmeter and data indicating a flow state of the fluid correspond to a correct value of an error between the measurement value and an actual flow rate; generating an error prediction model that outputs the error when the measurement value and the data indicating the flow state are input based on the training data; Acquire the measurement value related to the pipe to be predicted and data indicating the flow state; The acquired measurement value and the data indicating the flow state are input to the error prediction model, and the error is output. An error measurement method in which processing is performed by a computer.

Images