Data processing device, data processing method, and data processing program

Abstract

To provide a technique capable of detecting fluid leakage with reduced influence of environmental vibration.SOLUTION: A data processing device for determining whether fluid is leaking from a pipeline network on the basis of a waveform of a vibration intensity measured from the fluid pipeline network comprises a memory which stores software programs and a processor which executes the software programs. The processor: acquires waveform data of a plurality of vibration intensities measured from the pipeline network during a different time; calculates an evaluation value for evaluating how appropriate the plurality of waveform data is used to determine whether fluid is leaking; selects waveform data used to determine whether fluid is leaking from the plurality of waveform data from among the plurality of waveform data on the basis of the evaluation value; extracts periodic characteristics from an autocorrelation coefficient of a vibration intensity of the selected waveform data; and determines whether fluid is leaking from the pipeline network on the basis of a periodic characteristic relationship between the selected waveform data.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a technique for processing data of vibration sensors for determining the presence or absence of leakage in each pipeline in a pipeline network of a fluid in an infrastructure or a factory.

Background Art

[0002] In Patent Document 1, as a technique for detecting leakage in a pipeline network as an infrastructure or a pipeline network installed in a factory, a detection system using a vibration sensor has been proposed. In Patent Document 1, a method for determining the presence or absence of leakage by using the sound pressure values of data acquired at different times has been proposed.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Pipeline networks of infrastructure and within factories may be set up in locations where various environmental vibrations occur, such as vibrations caused by the running of automobiles and trains or the operation of large machinery. The ways in which environmental vibrations occur are diverse, including those that continue for a long time within a day and those that occur intermittently at high frequencies for several days.

[0005] The method of Patent Document 1 determines water leakage by using data measured intermittently multiple times within a day. In this method, in locations where environmental vibrations that continue for a long time within a day or where environmental vibrations that occur frequently continue for several days, there is room for improvement in the possibility of misjudging the presence or absence of water leakage due to environmental vibrations. Also, the ways in which environmental vibrations occur are diverse depending on the location and time, and it is not easy to find and adjust the optimal measurement time for each location of the pipeline network.

[0006] The purpose of this disclosure includes providing a technology that enables the detection of fluid leaks while reducing the effects of environmental vibrations. [Means for solving the problem]

[0007] A data processing device according to one aspect of the present disclosure is a data processing device for determining whether or not fluid is leaking from a pipeline network based on the waveform of vibration intensity measured from the pipeline network, comprising: a memory for storing a software program; and a processor for executing the software program, wherein the processor acquires a plurality of waveform data of vibration intensity measured from the pipeline network at different times, calculates an evaluation value for evaluating how suitable or unsuitable the plurality of waveform data are for use in determining whether or not fluid is leaking, selects waveform data from the plurality of waveform data to be used for determining whether or not fluid is leaking based on the evaluation value, extracts periodic characteristics from the autocorrelation coefficient of the vibration intensity of the selected waveform data, and determines whether or not fluid is leaking from the pipeline network based on the relationship of periodic characteristics among the selected waveform data. [Effects of the Invention]

[0008] According to one aspect of this disclosure, it becomes possible to detect fluid leaks while reducing the effects of environmental vibrations. [Brief explanation of the drawing]

[0009] [Figure 1] This diagram shows the vibration intensity measured over a six-day period by vibration sensors installed in areas along main roads where no water leaks have occurred, broken down by time of day. [Figure 2] This diagram shows the intensity of vibrations measured over six days by vibration sensors installed in areas of a busy commercial district where no water leaks have occurred, broken down by time of day. [Figure 3] This figure shows an example of a flowchart for data processing for water leak detection in Example 1. [Figure 4] This figure shows an example of a flowchart for data selection processing based on evaluation values ​​for each measured waveform. [Figure 5] This figure shows an example of a flowchart for the process of acquiring measurement waveforms. [Figure 6] This figure shows an example of a flowchart for the waveform sorting process. [Figure 7] This figure shows an example flowchart for sorting N measurement waveforms. [Figure 8] This figure shows an example of a list of measurement waveform data before sorting. [Figure 9] This figure shows an example of a sorted list of measurement waveform data. [Figure 10] This diagram illustrates the list of measurement waveform data for selecting W items. [Figure 11] This figure shows an example of a flowchart for the evaluation value calculation process. [Figure 12] This figure shows another example of a flowchart for the evaluation value calculation process. [Figure 13] This figure shows yet another example of a flowchart for the evaluation value calculation process. [Figure 14] This is a schematic diagram showing the frequency characteristics of vibration intensity measured in areas with heavy vehicle traffic, such as along main roads. [Figure 15] This is a schematic diagram showing the frequency characteristics of vibration intensity due to water leakage vibrations acquired in a quiet environment with low environmental vibration. [Figure 16] This graph shows the time evolution of the autocorrelation coefficient calculated from the measured waveform including the water leakage signal. [Figure 17] This is a schematic diagram illustrating the measurement times when the time of acquisition of the measurement waveform is changed daily. [Figure 18] This figure shows an example flowchart for leak detection processing based on autocorrelation peak position data. [Figure 19] This is a block diagram showing an example of the hardware configuration of a data processing unit. [Figure 20] This is a schematic diagram showing an example of data stored in the secondary memory. [Figure 21] This block diagram shows another example of the hardware configuration of a data processing unit. [Figure 22] It is a diagram showing an example of an input screen for setting parameters in a data processing device. [Figure 23] It is a diagram showing another example of an input screen for setting parameters in a data processing device. [Figure 24] It is a diagram showing an example of a flowchart of data selection processing based on evaluation values for each measurement waveform in Example 2. [Figure 25] It is a block diagram showing an example of the hardware configuration of the data processing device in Example 2. [Figure 26] It is a block diagram showing an example of the hardware configuration of the data processing device added in Example 2. [Figure 27] It is a schematic diagram showing an example of data stored in the auxiliary storage device of the data processing device added in Example 2. [Figure 28] It is a diagram showing an example of a flowchart of data selection processing based on evaluation values for each measurement waveform in Example 3. [Figure 29] It is a diagram showing an example of a flowchart of data selection processing based on evaluation values for each measurement waveform in Example 4. [Figure 30] It is a diagram showing an example of a flowchart of evaluation value calculation processing in Example 4. [Figure 31] It is a diagram showing an example of the configuration of an evaluation value calculation model. [Figure 32] It is a diagram showing another example of the configuration of an evaluation value calculation model. [Figure 33] It is a diagram showing another example of a flowchart of evaluation value calculation processing in Example 4. [Figure 34] It is a diagram showing another example of the configuration of an evaluation value calculation model. [Figure 35] It is a diagram showing yet another example of the configuration of an evaluation value calculation model. [Figure 36] It is a diagram showing an example of environmental noise. [Figure 37] It is a schematic diagram showing an example of data stored in the auxiliary storage device in Example 4. [Figure 38]This figure shows an example of a flowchart for data selection processing based on evaluation values ​​for each measurement waveform in Example 5. [Figure 39] This figure shows an example of the data selection condition input screen in Example 5. [Figure 40] This figure shows another example of the data selection condition input screen in Example 5. [Modes for carrying out the invention]

[0010] Embodiments of the present invention will be described below with reference to the drawings. The embodiments are illustrative examples for explaining the present invention, and have been omitted and simplified as appropriate for clarity of explanation. The present invention can also be carried out in various other forms. Unless otherwise specified, each component may be singular or plural.

[0011] The position, size, shape, and extent of each component shown in the drawings may not represent the actual position, size, shape, and extent of the object, in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, and extent of the object disclosed in the drawings.

[0012] Examples of various types of information may be described using terms such as "table," "list," and "queue," but these types of information may also be represented by other data structures. For example, various types of information such as "XX table," "XX list," and "XX queue" may be referred to as "XX information." When describing identification information, terms such as "identification information," "identifier," "name," "ID," and "number" are used, and these terms are interchangeable.

[0013] When there are multiple components with the same or similar function, they may be described using the same symbol but with different subscripts. Furthermore, when it is not necessary to distinguish between these multiple components, the subscripts may be omitted in the description.

[0014] In the examples, the processes performed by executing a program may be described. Here, the computer executes the program using a processor (e.g., CPU, GPU) and performs the processing defined in the program using memory resources (e.g., memory) and interface devices (e.g., communication ports). Therefore, the main entity performing the processing by executing the program may be the processor. Similarly, the main entity performing the processing by executing the program may be a controller, device, system, computer, or node having a processor. The main entity performing the processing by executing the program may be an arithmetic unit, and may include dedicated circuits that perform specific processing. Here, dedicated circuits include, for example, FPGAs (Field Programmable Gate Arrays), ASICs (Application Specific Integrated Circuits), CPLDs (Complex Programmable Logic Devices), etc.

[0015] The program may be installed on the computer from the program source. The program source may be, for example, a program distribution server or a storage medium readable by the computer. If the program source is a program distribution server, the program distribution server includes a processor and storage resources for storing the program to be distributed, and the processor of the program distribution server may distribute the program to other computers. In addition, in the embodiment, two or more programs may be implemented as one program, or one program may be implemented as two or more programs. [Examples]

[0016] Here, the process of determining whether or not fluid is leaking from a pipeline using data acquired by vibration sensors installed at various points in a pipeline network where fluid flows through infrastructure such as factories is referred to as "data processing for leak detection." As an example, the pipeline network is assumed to be a water pipeline network, and the fluid flowing through the network is assumed to be water. The data processing for leak detection in this embodiment facilitates the accurate detection of vibrations that occur when water leaks from a water pipeline network. Vibration sensors can be installed at various points in the pipeline network, such as control valves, and can detect vibrations caused by water (fluid) leakage (for example, based on characteristics related to vibration intensity at a predetermined frequency, detection time, and vibration stationarity). The vibration sensors can then output information such as vibration intensity at a predetermined frequency and a indication of vibration to a leak monitoring device to which they are connected. The vibrations detected by the vibration sensors are not limited to vibrations transmitted through pipes, but also include vibrations originating from sound transmitted through air and liquids. A leak monitoring device is a device that performs data processing for leak detection to detect fluid leakage from a pipeline network. Hereafter, a leak monitoring device may be referred to as a data processing device.

[0017] The premise for this embodiment is that turbulence at the leak outlet is a cause of leakage vibration. A leak outlet is a hole in the pipeline from which water leaks out. Hereafter, the leak outlet will also be simply referred to as the leak hole. In high-pressure piping, localized, steep pressure fluctuations occur near minute leak holes. At this time, cavitation occurs near the leak hole, where minute bubbles are repeatedly generated and collapsed due to the pressure fluctuations in the water. The periodic generation and collapse of bubbles caused by cavitation generates a unique impact sound in the vicinity. The generated impact sound travels through the pipe and the fluid inside the pipe. By measuring this periodic impact sound with a sensor terminal, it is possible to detect the occurrence of a water leak. To accurately identify the occurrence of a water leak, it is necessary to accurately extract the periodic components observed when a water leak occurs. Generally, vibrations due to water leaks occur continuously regardless of the time of day, so if vibrations with the same periodic characteristics can be captured at different time periods, it is possible to determine that there is a water leak.

[0018] In this example, an autocorrelation function is used to understand the periodic characteristics of the signal itself. The autocorrelation function is a function that obtains the correlation between a given signal p(t) and a signal p(t+τ) obtained by shifting that signal in time, and is defined by equation (1) below.

[0019]

number

[0020] On the other hand, some environmental vibrations originating from ambient noise, not from water leaks, exhibit periodicity and occur intermittently over long periods. Measuring these types of environmental vibrations with a vibration sensor can lead to the false detection of a water leak. Furthermore, the magnitude of leakage vibrations caused by a water leak decreases as you move away from the source of the leak. Therefore, to detect a water leak at a distance from the source, it is necessary to make a judgment based on measurement waveforms taken during quiet periods when ambient vibrations are low and minute leakage vibrations appear relatively large relative to the ambient vibrations. From this perspective, it is also important to use an autocorrelation function to determine whether there are vibrations with periodic characteristics specific to water leaks, while suppressing the influence of ambient noise.

[0021] Figure 1 shows the vibration intensity measured over six days by vibration sensors installed in areas along main roads where no water leaks have occurred, broken down by time of day. Figure 2 shows the vibration intensity measured over six days by vibration sensors installed in areas in busy commercial districts where no water leaks have occurred, broken down by time of day. Figures 1 and 2 show the vibration intensity due to environmental vibrations in each environment. As can be seen from Figures 1 and 2, the vibration intensity can vary by more than 100 times depending on the time of day. When the time of day when such environmental vibrations occur overlaps with the measurement time, the environmental vibrations affect the peak information of the autocorrelation function, making it impossible to correctly determine whether a water leak has occurred. As shown in Figure 1, environmental vibrations are generally stronger during the day. However, in some locations, as shown in Figure 2, environmental vibrations occur randomly throughout the day and night. Therefore, it is not desirable to uniquely determine the time of day and measure the vibration intensity for various locations.

[0022] Therefore, in this embodiment, data from time periods with minimal environmental vibration is used to determine the presence or absence of water leakage vibration. The system includes a process to select data from multiple intermittently measured vibration waveforms that are suitable for obtaining peak information of the autocorrelation function according to an arbitrary evaluation function. The data processing for water leakage detection in this embodiment is described below. <Processing Procedure> Figure 3 shows an example of a data processing flowchart for water leak detection in Example 1.

[0023] As shown in Figure 3, in data processing for leak detection, the data processing device, upon starting processing (Step 4), first performs data selection processing based on evaluation values ​​for each measurement waveform (Step 2), then performs leak detection processing based on autocorrelation peak position data (Step 3), and finally terminates processing (Step 5). Here, the evaluation value for each measurement waveform is a value used to evaluate how suitable or unsuitable the measurement waveform is for determining the presence or absence of a leak. The data selection processing based on evaluation values ​​for each measurement waveform calculates an evaluation value for each measurement waveform, selects the measurement waveform to be used for leak detection based on the calculated evaluation value, calculates the autocorrelation coefficient of the selected measurement waveform, and extracts the peak of the autocorrelation coefficient. Details of the data selection processing based on evaluation values ​​for each measurement waveform will be described later. Furthermore, the leak detection processing based on autocorrelation peak position data is a process that determines the presence or absence of a leak based on the position of the peak of the autocorrelation coefficient calculated by the data selection processing based on evaluation values ​​for each measurement waveform. The position of the peak is represented by its position on the time axis, i.e., by time. Details of the leak detection process based on autocorrelation peak position data will be described later.

[0024] Figure 4 shows an example of a flowchart for data selection processing based on evaluation values ​​for each measured waveform.

[0025] As shown in Figure 4, when the data processing device starts processing, in the data selection process based on the evaluation value for each measurement waveform in step 2, it performs measurement waveform acquisition processing (step 41), measurement waveform selection processing (step 42), autocorrelation coefficient calculation processing (step 43), peak extraction processing (step 44), data saving processing (step 45), and determines whether the number of extracted autocorrelation peak positions has reached a predetermined number (step 46). If the number of extracted autocorrelation peak positions has not reached a predetermined number in step 46, the data processing device returns to the measurement waveform acquisition processing in step 41. If the number of extracted autocorrelation peak positions has reached a predetermined number, the data processing device outputs autocorrelation peak position data 47 indicating the extracted autocorrelation peak positions.

[0026] The waveform acquisition process is the process of acquiring waveform data measured by the vibration sensor. The waveform selection process is the process of calculating an evaluation value for the measured waveforms and selecting waveforms based on that evaluation value. The autocorrelation coefficient calculation process is the process of calculating the autocorrelation coefficient of the selected waveforms. The peak extraction process is the process of extracting the peak of the autocorrelation coefficient of the selected waveforms. The data saving process is the process of saving data indicating the position of the peak of the autocorrelation coefficient for each measured waveform. Details of the waveform acquisition process, waveform selection process, autocorrelation coefficient calculation process, and peak extraction process will be described later.

[0027] Figure 5 shows an example of a flowchart for the process of acquiring measurement waveforms.

[0028] As shown in Figure 5, in the measurement waveform acquisition process in step 41, the data processing device takes setting parameters 121, which consist of measurement time D, measurement duration T, measurement interval X, number of measurements N, and sorting number W, as input, and when measurement time D arrives, acquires the measurement waveform for T seconds (step 124), updates the next measurement time D to measurement time D+X (step 125), and determines whether the number of times the measurement waveform has been acquired is N or more (step 126). If the number of times the measurement waveform has been acquired is not N or more, the data processing device returns to step 124, and if the number of times the measurement waveform has been acquired is N or more, outputs data containing N measurement waveforms and the set sorting number W (step 123).

[0029] Figure 6 shows an example of a flowchart for the waveform sorting process.

[0030] As shown in Figure 6, in the measurement waveform selection process in step 42, the data processing device takes N measurement waveforms and a set selection number W as input, performs sorting of the N measurement waveforms (step 133), performs selection of W measurement waveforms (step 134), and outputs W measurement waveforms (step 132). The sorting of N measurement waveforms is a process that calculates an evaluation value for each of the N measurement waveforms and rearranges the N measurement waveforms based on the evaluation values. The selection of W measurement waveforms is a process that takes the data output by the sorting of N measurement waveforms as input and selects W measurement waveforms from the top of the N measurement waveforms to be used for determining the presence or absence of water leakage. Details of the sorting of N measurement waveforms will be described later.

[0031] Figure 7 shows an example of a flowchart for sorting N measurement waveforms.

[0032] As shown in Figure 7, in the sorting process of N measurement waveforms in step 133, the data processing device takes N measurement waveforms and a set sorting number W as input, and selects the Y-th measurement waveform of the N measurement waveforms as a variable Y with an initial value of 1 (step 142). It then extracts a specific frequency band from the selected waveform using a bandpass filter (step 143), calculates a predetermined evaluation value S from the extracted frequency band data by executing an evaluation value calculation process (step 144), increments Y to Y+1 (step 145), and determines whether the newly incremented Y is greater than or equal to the number of measurement waveforms N (step 146). If Y is less than N, the data processing device returns to step 142. If Y is greater than or equal to N, the data processing device sorts the measurement waveform data list, which will be described later, according to the evaluation value S (step 147), and outputs data 141 containing the N measurement waveforms, the sorting number W, and the measurement waveform data list sorted by the evaluation value S. This data 141 becomes the input for the W measurement waveform selection process described above. Here, the evaluation value S may be an evaluation value indicating the intensity of environmental vibration, or an evaluation value corresponding to the likelihood of leakage vibration. The evaluation value calculation process, i.e., the method for calculating the evaluation value S, will be described later. The sorting number W may be a fixed value for each measurement, or it may be changed for each measurement. Alternatively, the sorting number W may be set to change dynamically according to the value of the number of measurements N, such as setting the sorting number W to half of the number of measurements N (rounded down or up if N is odd).

[0033] If the evaluation value S is an evaluation value indicating the intensity of environmental vibration, the data processing device, in step 147, rearranges the measurement waveform data list so that the evaluation value S is in ascending order, so that it can preferentially select measurement waveforms suitable for determining the presence or absence of water leakage, i.e., measurement waveforms with a small evaluation value S. If the evaluation value S is an evaluation value corresponding to the likelihood of water leakage vibration, the data processing device, in step 147, rearranges the measurement waveform data list so that the evaluation value S is in descending order, so that it can preferentially select measurement waveforms suitable for determining the presence or absence of water leakage, i.e., measurement waveforms with a large evaluation value S. Therefore, step 147, which rearranges the N measurement waveform data lists according to the evaluation value S, may be performed according to either ascending or descending order of the evaluation value S.

[0034] Furthermore, if an error occurs when acquiring the measurement waveform and the evaluation value S becomes extremely small or large, the measurement waveform should not be selected. Therefore, in step 147, the data processing device may rearrange the measurement waveform data list so that the deviation or variance of the evaluation value S from the mean or median is in ascending or descending order.

[0035] Figure 8 shows an example of a measurement waveform data list before sorting. The measurement waveform data list is recorded information associated with the measurement waveform, such as measurement conditions and sensor status, and the types of information recorded are not limited to this. Figure 9 shows an example of a measurement waveform data list after sorting. In the measurement waveform data list 151 shown in Figure 8, the data name 152, measurement time D153, measurement duration T154, index E 155, evaluation value S 156, and valid data identifier V157 are recorded for each measurement waveform data. In Figure 8, as an example, the measurement interval X is 6 hours, the number of measurements N is 4, the number of selections W is 2, and the measurement duration T is 10 seconds, but the values ​​that can be set are not limited to this. The valid data identifier V157 is recorded information used to determine whether or not it is used for determining the presence or absence of water leakage. Figure 9 shows a measurement waveform data list 161 in which the measurement waveform data in the measurement waveform data list 151 of Figure 8 has been sorted so that the evaluation value S is in ascending order. The method used by the data processing device to sort the measured waveform data may be any method, such as bubble sort, quicksort, or heapsort.

[0036] Figure 10 is a diagram illustrating the process of selecting W measurement waveforms. Referring to Figure 10, the measurement waveform list 171 for selecting W waveforms contains W data selected from the beginning of the measurement waveform data list 161 sorted by the evaluation value S in Figure 9, with the valid data identifier V157 set to 1. For data that has not been selected, the valid data identifier V157 is set to 0. For example, if the valid data identifier V157 is 1, it is used to determine whether or not there is a water leak, and if it is not 1, it is not used to determine whether or not there is a water leak. Here, as an example, W=2. The measurement waveforms corresponding to the selected W measurement waveform list 171 are used to determine whether or not there is a water leak.

[0037] Figure 11 shows an example of a flowchart for the evaluation value calculation process.

[0038] As shown in Figure 11, in the evaluation value calculation process in step 144, the data processing device takes a single measurement waveform 181 from which a predetermined bandwidth has been extracted by a bandpass filter in step 143 as input, acquires the maximum amplitude within that measurement waveform as the evaluation value S (step 182), and outputs the evaluation value S 183. This process of using the maximum amplitude as the evaluation value S can be easily implemented, and by prioritizing the selection of measurement waveforms with small evaluation values ​​S, it is possible to avoid measurement waveforms that have a high vibration intensity and are likely to contain environmental vibrations. However, the evaluation value calculation process in this embodiment is not limited to that shown in Figure 11.

[0039] Figure 12 shows another example of a flowchart for the evaluation value calculation process.

[0040] In the example shown in Figure 12, in the evaluation value calculation process in step 144, the data processing device first takes a single measurement waveform 181, from which a predetermined bandwidth has been extracted by a bandpass filter in step 143, as input and performs offset correction processing on the measurement waveform (step 191). The offset correction processing on the measurement waveform is a process that applies an offset to the measurement waveform so that the average value of the vibration intensity of the measurement waveform at the time the measurement waveform was taken becomes zero. Subsequently, the data processing device calculates the sum of the squares of the vibration intensities of the measurement waveform after the offset correction processing as the evaluation value S 183 (step 192). If the measurement waveform contains constant environmental vibration, this evaluation value S represents the energy of the constant environmental vibration. Therefore, by selecting measurement waveforms from those with smaller evaluation values ​​S, it becomes possible to use measurement waveforms with less influence from environmental vibration to determine whether or not there is water leakage.

[0041] The evaluation value calculation process shown in Figures 11 and 12 calculates the evaluation value S from the vibration intensity on the time-series axis of the measured waveform, but the evaluation value calculation process is not limited to these methods. As another example, the evaluation value S may be calculated from the vibration intensity on the frequency axis of the measured waveform.

[0042] Figure 13 shows yet another example of a flowchart for the evaluation value calculation process.

[0043] In the example shown in Figure 13, in the evaluation value calculation process in step 144, the data processing device first takes a single measurement waveform 181, from which a predetermined bandwidth has been extracted by a bandpass filter in step 143, as input, and converts it into a waveform on the frequency axis by performing a fast Fourier transform on the measurement waveform (step 201). Then, it integrates the vibration intensity with frequency as the integration variable and obtains it as an evaluation value S (step 202).

[0044] Figure 14 is a schematic diagram showing the frequency characteristics of vibration intensity measured in areas with heavy vehicle traffic, such as along main roads. Figure 15 is a schematic diagram showing the frequency characteristics of vibration intensity due to water leakage vibration acquired in a quiet environment with low environmental vibration.

[0045] Vibrations caused by vehicle traffic are often distributed on the low-frequency side, below 400 Hz, as shown in Figure 14. On the other hand, water leakage vibrations, as shown in Figure 15, can be distributed around 500 Hz with a bell-shaped peak. When water leakage occurs in areas with heavy vehicle traffic, such as along main roads, it is conceivable that vibration intensities similar to those shown in Figure 14 and Figure 15 will be measured. In such cases, after extracting frequency components below 400 Hz using a bandpass filter in step 143 of Figure 7, an evaluation value S corresponding to the energy of the environmental vibration originating from steady-state vehicle traffic included in the measured waveform is calculated by integrating the vibration intensities for each frequency in step 202 of Figure 13. By selecting a measurement waveform with a small evaluation value S, it becomes possible to determine the presence or absence of water leakage using a suitable measurement waveform.

[0046] Returning to Figure 4, in the autocorrelation coefficient calculation process of step 43, the data processing device calculates the autocorrelation coefficient for each of the W selected measurement waveforms. The autocorrelation coefficient can be calculated using the autocorrelation function shown in equation (1).

[0047] Figure 16 is a graph showing the time evolution of the autocorrelation coefficient calculated for a measured waveform containing a water leakage signal. In the graph of Figure 16, the horizontal axis represents time, and the unit corresponds to the number of data sampling points of the sensor; the vertical axis represents the autocorrelation coefficient. The graph of Figure 16 shows the autocorrelation coefficients of multiple measured waveforms superimposed. As shown in Figure 16, the autocorrelation coefficient of water leakage vibration has multiple peaks over time, so the data processing device acquires a peak position information set representing the positions of these multiple peaks. Furthermore, in the case of water leakage vibration, the autocorrelation coefficient is the same for measured waveforms measured at any given time, so the peak positions in the peak position information set acquired from multiple measured waveforms will be the same.

[0048] Returning to Figure 4, in the peak extraction process of step 44, the data processing device appropriately sets a time width sufficient to extract multiple peak positions according to the sampling interval, and extracts multiple peaks from the waveform measured within that time width. The peak position information set indicating the extracted multiple peaks is stored in a non-volatile memory element or a volatile memory element by the data storage process of step 45. The data processing device repeats the process from the measurement waveform acquisition process of step 41 to the data storage process of step 45 until it is determined in step 46 that a predetermined number of peak position information sets have been acquired. When it is determined that a predetermined number of peak position information sets have been acquired, the data processing device outputs those predetermined number of peak position information sets as autocorrelation peak position data 47.

[0049] Furthermore, in the autocorrelation coefficient calculation process, the measurement waveform selected in the measurement waveform selection process may be divided into multiple parts along the time axis, and multiple autocorrelation coefficients may be calculated for each divided waveform. Also, in the measurement waveform acquisition process, in order to avoid the influence of noise sources that intermittently generate environmental vibrations, the measurement waveform may be acquired at the same time every day according to a predetermined measurement interval, as shown in Figure 5, or the measurement waveform may be acquired at different times each day.

[0050] Figure 17 is a schematic diagram showing the measurement times when the time of acquisition of measurement waveforms is changed daily. Referring to Figure 17, the time of acquisition of measurement waveforms is indicated by the word "Measurement". Within that, the time when the measurement waveform selected based on the evaluation value S was acquired is indicated by hatching. Furthermore, within that, the time when the measurement waveform in which a water leak was determined to be present was acquired is indicated by an underline. As shown in Figure 17, measurement waveforms may be acquired according to a schedule of measurement times set to different times each day. In that case, the number of measurements N is not fixed, and a schedule may be set in advance so that the number of measurements N changes daily as shown in Figure 5, or the number of measurements N and the measurement time D may be dynamically changed based on past water leak determination results, such as increasing the number of times measurement waveforms are acquired the day after a water leak is determined to be present.

[0051] Figure 18 shows an example of a flowchart for leak detection processing based on autocorrelation peak position data.

[0052] As shown in Figure 18, the data processing device takes the autocorrelation peak position data 47 as input, performs result comparison processing (step 51), performs leak detection processing (step 52), determines whether or not it is necessary to send the data to the host device based on the results of the leak detection processing (step 53), returns to step 41 if it is not necessary, and sends the data of the leak detection processing results to the host device if it is necessary (step 54), and terminates the process (step 5).

[0053] The result comparison process compares the positions of multiple peaks, indicated by the peak position information set included in the autocorrelation peak position data 47, between multiple measured waveforms and calculates the degree of agreement, which represents the degree to which the peak positions of the measured waveforms match. The leak detection process determines whether or not there is a leak based on the degree of agreement obtained from the result comparison process.

[0054] The details of the result comparison process in step 51, the water leak detection process in step 52, and the branching process in step 53 are described below.

[0055] In the result comparison process, the data processing device may, for example, compare the positions of multiple peaks in the measured waveforms and calculate the degree of agreement.

[0056] For example, a 4-second measurement waveform is divided into 16 segments along the time axis, and the autocorrelation coefficient is calculated for each of the 16 generated 250ms waveforms. From the 16 obtained autocorrelation coefficients, 6 autocorrelation coefficients with low noise are selected. Then, 6 peak position information sets are extracted from these 6 autocorrelation coefficients. If the above is repeated 15 times every hour with different time periods, 6 peak position information sets will be obtained from each of the 15 measurement waveforms, resulting in a total of 6 × 15 = 90 peak position information sets being extracted. Pairs of peak position information sets are created from these 90 peak position information sets by exhaustive combinations, and the multiple peak positions included in each pair's peak position information set are compared. The ratio of the number of matching peak positions to the total number of peak positions is defined as the degree of agreement, and pairs with a degree of agreement above a predetermined threshold are extracted. The ratio of pairs with a degree of agreement above the threshold to the total number of pairs is defined as the occurrence rate, and this occurrence rate may be used in the leak detection process in step 52 to determine whether or not there is a leak. This matching rate corresponds to the degree of matching mentioned above.

[0057] In the leak detection process, if the occurrence rate obtained from the result comparison process exceeds a predetermined threshold, it is determined that there is a leak.

[0058] However, the result comparison process and leak detection process are not limited to the examples described above. As another example, processing utilizing graph networks may be used. For example, the relationships between sets of peak position information obtained from each measurement waveform may be understood using a graph network. Here, a graph is a figure represented by multiple points and edges connecting them, and a graph in which points and edges are each given physical meaning is called a graph network. In this embodiment, the measurement waveforms to be compared can be represented as points, and the degree of agreement between the peak positions of the two measurement waveforms can be given meaning as edges.

[0059] For example, a graph network could be created by connecting points with edges where the degree of peak position agreement exceeds a threshold of 80%, and this graph network could be used to determine whether or not there is a water leak. Measurement waveforms acquired at a site where a water leak is occurring have a high degree of peak position agreement with each other, so the measurement waveforms connected by edges form a large cluster on the graph network described above. On the other hand, measurement waveforms acquired at a site where no water leak is occurring have independent peak positions and a low degree of agreement with each other, so the measurement waveforms form multiple small clusters on the graph network. From this, the presence or absence of a water leak could be determined based on information about the number of clusters or the size of the clusters.

[0060] In the branching process of step 53, the data processing unit determines whether or not it is necessary to send the resulting data to the higher-level device based on the result of the water leak detection process. For example, the data processing unit may decide that it is necessary to send data only if a water leak is detected. Alternatively, the data processing unit may send data to the higher-level device at regular intervals, regardless of whether or not there is a water leak, so that the higher-level device can confirm the status of its own device, which is a sensor terminal. The higher-level device can confirm that the data processing unit is functioning normally by receiving data from it at regular intervals.

[0061] <Device configuration> The following describes the apparatus configuration for realizing this embodiment using drawings.

[0062] Figure 19 is a block diagram showing an example of the hardware configuration of a data processing unit.

[0063] Referring to Figure 19, the data processing unit 251 consists of an input device 255, an output device 256, a network interface (I / F) 257, a microcontroller 261, a measurement module 262, a secondary memory device 259, and a battery 263, which are connected to each other via a bus 258.

[0064] The microcontroller 261 has non-volatile memory elements in addition to volatile memory elements such as RAM (Random Access Memory). Programs to be processed are stored in the non-volatile memory elements in advance, and the processor reads them out as needed to control the various parts of the sub-memory 259 and the measurement module 262, perform data communication, and execute calculations.

[0065] The secondary storage device 259 is a device that stores programs and data, etc., and has a non-volatile storage element such as an HDD (Hard Disk Drive), SSD (Solid State Drive), or flash memory.

[0066] The input device 255 is a device that accepts user input such as switches and / or buttons, and acquires information entered through user input.

[0067] The output device 256 is a device that outputs information, such as an LED (Light Emitting Diode) and / or a speaker, and for example, the result of the water leak detection is presented to the user by the lighting of the LED.

[0068] The measurement module 262 is a vibration sensor. The data processing device 251 may also include a temperature sensor or a humidity sensor in addition to the vibration sensor.

[0069] The battery 263 is composed of batteries and other components, and can operate the data processing unit 251 without receiving power from an external source.

[0070] A network (NW) is a communication network, and it can be a wired network or a wireless network. Furthermore, a network (NW) can be a global network like the internet, or a local area network (LAN).

[0071] The network interface 257 is an interface that allows the data processing device 251 to send and receive data with the cloud server 252 via the network NW. The data processing device 251 can send and receive data with the cloud server 252 connected to the network NW using the network interface 257. The network interface 257 can receive information input from the cloud server 252, thus functioning as an input device. Furthermore, the network interface 257 can send data to the cloud server 252 via the network NW, thus functioning as an output device.

[0072] Figure 20 is a schematic diagram showing an example of data stored in the sub-memory. As shown in Figure 20, the sub-memory 259 stores, for example, measurement waveforms 281, a list of measurement waveform data 171 for selecting W waveforms, and autocorrelation peak position data 292. The measurement waveforms 281 are data of all past measurement waveforms acquired by the measurement waveform acquisition process in step 41 in Figure 4. The list of measurement waveform data 171 for selecting W waveforms is a list of measurement waveform data referenced when selecting W measurement waveforms in Figure 10, and includes, for example, data name 152, measurement time D153, measurement duration T154, index E 155, evaluation value S 156, and valid data identifier V157.

[0073] Furthermore, the autocorrelation peak position data 292 represents the temporal position of the peak in the graph showing the time change of the autocorrelation coefficient shown in Figure 16.

[0074] The data processing device 251 in this embodiment may be connectable to a user terminal 272 such as a personal computer or tablet terminal.

[0075] Figure 21 is a block diagram showing another example of the hardware configuration of a data processing device. The data processing device 251 in Figure 21 differs from that in Figure 19 in that it includes a control communication module 271. The data processing device 251 in Figure 21 communicates with an external user terminal 272 via a wired or wireless communication line through the control communication module 271, and can retrieve data stored in the sub-memory 259 to the user terminal 272, or rewrite the processing program written inside the microcontroller 261 from the user terminal 272.

[0076] <Output screen> The data processing device 251 can be operated in various ways from the cloud server 252 or the user terminal 272, which are higher-level devices. The following describes the screens displayed on the cloud server 252 or the user terminal 272 when a user performs various operations on the data processing device 251 from the cloud server 252 or the user terminal 272, using the diagrams below.

[0077] Figure 22 shows an example of an input screen for setting parameters in a data processing device. The data sorting count condition input screen 371 shown in Figure 22 is displayed on the cloud server 252 or the user terminal 272. The data sorting count condition input screen 371 is a screen for setting the sorting count W, which is included in the setting parameter 121 given to the measurement waveform acquisition process shown in Figure 5, for each sensor terminal. When a user specifies the sensor terminal for which to set the sorting count W in the input box 373, enters the sorting count W to be set in the input box 374, and presses the send button 375, the sorting count W information is sent to the specified sensor terminal and set. Other various setting parameters may also be set from an input screen similar to the sorting count W.

[0078] The method for setting various parameters, such as the sorting quantity W, is not limited to the method of inputting information into the GUI (Graphical User Interface) screen as exemplified here; other methods may also be used. For example, setting parameters may be implemented using a CUI (Character User Interface) by entering commands.

[0079] Furthermore, it may be possible to set multiple parameters from a single input screen. Figure 23 shows another example of an input screen for setting parameters in a data processing device. As shown in Figure 23, the parameters that can be set and transmitted externally are not limited to the number of data selections W, but may also be predetermined parameters such as the number of data measurements N.

[0080] Thus, in this embodiment, the water leak detection system consists of a single sensor terminal configured with a data processing device 251, and a monitoring system implemented on a cloud server 252 or a user terminal 272, which has a calculation unit (not shown) for managing measurement conditions and a display unit (not shown) for displaying whether or not there is a water leak. The sensor terminal also includes the steps of: measuring vibration at least once a day using a vibration sensor under measurement conditions specified by the calculation unit; selecting measurement waveforms based on evaluation values ​​calculated according to an arbitrary function; calculating the autocorrelation function of the measurement waveforms; extracting the peak position of the autocorrelation function; saving the peak position; comparing the peak positions among multiple data; determining whether there is a water leak based on the comparison results; and transmitting the determination result.

[0081] <Effects> As described above, this embodiment allows for water leak detection within a single sensor terminal while avoiding environmental vibrations that occur for long periods or intermittently throughout the day, thereby reducing false detections due to environmental vibrations or missed water leak vibrations, and also reducing the need to unnecessarily transmit data to higher-level devices. Consequently, long-term operation is possible while suppressing battery consumption and communication costs. Furthermore, in the event of a water leak, data communication is possible only irregularly when a water leak is detected, allowing for rapid notification of the water leak on the display unit within the monitoring system and enabling real-time status monitoring.

[0082] Furthermore, this embodiment provides high accuracy in detecting water leaks. Conventionally, for example, water leaks are detected by counting the number of times the signal level of a single measurement exceeds a predetermined threshold. In this case, if the threshold is not set appropriately, there is a risk of a high rate of false positives. Also, accurate detection is difficult in environments with a lot of noise. On the other hand, the water leak detection method in this embodiment compares multiple measurement results with each other. As a result, even if the signal level and peak position differ considerably depending on the measurement site, if they have similar characteristics, a water leak can be accurately detected, which has the advantage of being highly robust to the measurement environment.

[0083] In other words, according to this embodiment, it is possible to provide a water leak detection method, a water leak detection system, and a sensor terminal used therefor, which can perform highly reliable water leak detection without being limited by the amount or frequency of data transmission.

[0084] In the above explanation, it was stated that the sensor terminal acquires the measurement waveform, calculates the peak information of the autocorrelation function, performs a leak detection, and transmits the detection result to the monitoring system via data communication. However, the sensor terminal may acquire the measurement waveform, perform the calculation of the peak information of the autocorrelation function, transmit that peak information to the monitoring system, and have the monitoring system perform the leak detection. Alternatively, the sensor terminal may acquire the measurement waveform, transmit that measurement waveform to the monitoring system, and have the monitoring system perform the calculation of the peak information of the autocorrelation function and the leak detection.

[0085] Furthermore, in this embodiment, the data processing device 251 determines the presence or absence of water leakage based on a measurement waveform selected using one of several different evaluation values ​​S explained with reference to Figures 11, 12, and 13, but it is not limited to this. As another example, the data processing device 251 may determine the presence or absence of water leakage based on a measurement waveform selected using each of the several different evaluation values ​​S, and then make a final decision on the presence or absence of water leakage based on the results of those determinations. For example, if it is determined that there is water leakage based on a measurement waveform selected by any of the evaluation values ​​S, an overall judgment of water leakage being present may be made. [Examples]

[0086] Example 2 differs from Example 1 in its data selection process based on evaluation values ​​for each measured waveform. Otherwise, Example 2 has basically the same configuration and operation as Example 1. The following mainly describes the differences between Example 2 and Example 1.

[0087] Figure 24 shows an example of a flowchart for data selection processing based on evaluation values ​​for each measurement waveform in Example 2.

[0088] Unlike Example 1, in Example 2, the data selection process based on evaluation values ​​for each measured waveform is performed by the data processing device 251. After executing the peak extraction process in step 44, the data processing device 251 does not save the peak position data of the autocorrelation coefficient of the measured waveform within its own device, but instead sends it to the cloud server 252 (step 61). As a result, the cloud server 252 accumulates the peak position data for each measured waveform. Then, in step 46, when the number of extracted autocorrelation peak positions reaches a predetermined number, the data processing device retrieves the peak position data of the autocorrelation coefficient for each measured waveform that has been accumulated up to that point from the cloud server 252 (step 62).

[0089] <Effects> In Example 1, the data processing device 251 is equipped with a secondary memory device 259, as shown in Figure 19, and stores autocorrelation peak position data 292 in the secondary memory device 259, as shown in Figure 20. On the other hand, in Modification 1, the cloud server 252 stores the autocorrelation peak position data 292 instead of the secondary memory device 259, so the secondary memory device 259 is unnecessary or the amount of data recorded in the secondary memory device 259 can be reduced.

[0090] Figure 25 is a block diagram showing an example of the hardware configuration of the data processing device of Example 2. The data processing device 251 of Example 2 shown in Figure 25 differs from that of Example 1 shown in Figure 19 in that it does not have a secondary storage device 259.

[0091] Since the data processing device 251 is a sensor terminal that is placed in a confined space such as a water control valve, miniaturization is required. In this regard, by adopting the configuration and processing of Embodiment 2, as shown in Figure 25, the sub-memory device 259 is unnecessary, or the amount of data recorded in the sub-memory device 259 is reduced, making it possible to miniaturize the data processing device 251.

[0092] In addition, in Example 2, a separate data processing device 250 may be provided, in addition to the data processing device 251 and the cloud server 252, to acquire data from the cloud server 252 and perform the series of data processing for water leak detection shown in Figure 3.

[0093] Figure 26 is a block diagram showing an example of the hardware configuration of the data processing device added in Example 2. Referring to Figure 26, the data processing device 250 differs from the data processing device 251 in that it does not have a measurement module 262 and a battery 263, and has a processor 253 and main memory 254 instead of a microcontroller 261.

[0094] Figure 27 is a schematic diagram showing an example of data stored in the sub-memory of the data processing device added in Embodiment 2. Referring to Figure 27, the sub-memory 259 of the data processing device 250 stores the processing program 311. The processor 253 reads the various data and processing program 311 stored in the sub-memory 259 into the main memory 254 and executes the processing defined in the processing program 311. [Examples]

[0095] Example 3 differs from both Example 1 and Example 2 in its data selection process based on evaluation values ​​for each measured waveform. Otherwise, Example 3 has basically the same configuration and operation as Example 1. The differences between Example 3 and Example 1 will be mainly described below.

[0096] Figure 28 shows an example of a flowchart for data selection processing based on evaluation values ​​for each measurement waveform in Example 3.

[0097] Unlike Example 1, in Example 3, the data selection process based on the evaluation value for each measurement waveform is performed by the data processing device 251. After the data acquisition process in step 41 is performed, the data storage process in step 45 is performed, and in step 46, it is determined whether the number of autocorrelation peak positions has reached a predetermined number. After the number of autocorrelation peak positions has reached a predetermined number, the measurement waveform selection process in step 42 is performed.

[0098] <Effects> As a result, the data processing device 251 acquires a predetermined number of measurement waveforms, stores the data of these measurement waveforms, and then refers to this data to perform the necessary measurement waveform selection process, autocorrelation coefficient calculation, and peak extraction process. Since the measurement waveform selection process in step 42 and the peak extraction process in step 44 can be completed in a single process, there is an effect of shortening the processing time compared to Example 1. [Examples]

[0099] Example 4 differs from Example 1 in its data selection process based on evaluation values ​​for each measured waveform. Otherwise, Example 4 has basically the same configuration and operation as Example 1. The differences between Example 4 and Example 1 will be explained below.

[0100] Figure 29 shows an example of a flowchart for data selection processing based on evaluation values ​​for each measurement waveform in Example 4.

[0101] Comparing Figure 29 with Figure 4 of Example 1, we can see that in Example 4, unlike Example 1, the data sorting process based on the evaluation value for each measurement waveform is performed by the data processing device 251, which calculates the evaluation value S using the evaluation value calculation model 101 provided as an external input. More specifically, in the evaluation value calculation process included in the data sorting process based on the evaluation value for each measurement waveform, the data processing device 251 calculates the evaluation value S using the evaluation value calculation model 101 provided as an external input. Referring to Figure 29, the data processing device 251 then performs the calculation of the autocorrelation coefficient in step 43 of Figure 4, the peak extraction process in step 44, the data storage process in step 45, and the judgment in step 46, which correspond to step 71.

[0102] <Processing Procedure> Figure 30 shows an example of a flowchart for the evaluation value calculation process in Example 4. Comparing Figure 30 with Figure 11 of Example 1, in Example 4, the data processing device 251 applies the calculation formula given from the evaluation value calculation model 101 to the time-series characteristics of the vibration shown in the measured waveform 181 (step 232) to calculate the evaluation value S.

[0103] Figure 31 shows an example of the configuration of an evaluation value calculation model. The evaluation value calculation model 101 in Figure 31 consists of an evaluation value calculation model definition 241 and an evaluation value calculation model coefficient table 242. In the example in Figure 31, the evaluation value calculation model definition 241 constructs a polynomial definition formula having multinomial components, with vibration intensity data X(T1) to X(Tm) at predetermined time points in the measured waveform as variables and predetermined constants α0 to αN as coefficients. The coefficients α0 to αN for each order are given in the evaluation value calculation model coefficient table 242. By adopting an evaluation value calculation model defined in this configuration, it becomes possible to construct a more complex evaluation function for the time-series data of measured vibration intensity, and it becomes possible to flexibly adjust the definition and coefficients of the evaluation value calculation model according to the installation location of the vibration sensor or the result of determining whether or not there is water leakage, thereby enabling it to respond to various measurement environments with different characteristics of environmental vibration.

[0104] However, the evaluation value calculation model 101 in Example 4 is not limited to a configuration in which the vibration intensity data X(T1) to X(Tm) of the measured waveform over time are used as variables, as illustrated in Figure 31. Figure 32 shows another example of the configuration of the evaluation value calculation model. The evaluation value calculation model 101 in Figure 32 uses the time derivative of the vibration intensity data X(T1) to X(Tm) as a variable.

[0105] Furthermore, the evaluation value calculation process in Example 4 is not limited to applying a calculation formula to the time-series characteristics of the measured waveform 181, as illustrated in Figure 13.

[0106] Figure 33 shows another example of the flowchart for the evaluation value calculation process in Example 4. In the example in Figure 33, the data processing device 251 performs a Fast Fourier Transform (FFT) on the measurement waveform 181 to convert it into a distribution of vibration intensity for each frequency (step 201), and calculates the evaluation value S by applying a calculation formula to the vibration intensity on the frequency axis (step 211).

[0107] Figure 34 shows another example of the configuration of the evaluation value calculation model. The evaluation value calculation model 101 in Figure 34 is an example that can be applied to the evaluation value calculation process in Figure 33. The evaluation value calculation model 101 in Figure 34 consists of an evaluation value calculation model definition 221 and an evaluation value calculation model coefficient table 222 corresponding to the frequency axis variables.

[0108] In the example shown in Figure 34, the evaluation value calculation model definition 221 constructs a definition formula with multi-order components for the vibration intensity data X(F1) to X(Fm) for each frequency of the measured waveform. The coefficients α0 to αN for each order are given in the evaluation value calculation model coefficient table 222. By adopting an evaluation value calculation model defined in this way, it becomes possible to construct a more complex evaluation function for the data obtained by transforming the measured vibration intensity waveform using FFT, and it becomes possible to flexibly adjust the definition and coefficients of the evaluation value calculation model according to the installation location of the vibration sensor or the result of determining whether or not there is water leakage, thereby enabling adaptation to various measurement environments with different characteristics of environmental vibration.

[0109] Furthermore, even in this case, the evaluation value calculation model 101 is not limited to a sum of polynomials as shown in Figure 34. Figure 35 shows yet another example of the configuration of the evaluation value calculation model. The evaluation value calculation model definition 351 shown in Figure 35 is defined by a MAX function that extracts the maximum value from among multiple variables obtained by frequency differentiation of vibration intensity data X(F1) to X(Fm). Thus, the evaluation value calculation model may be defined using a statistical function such as the MAX function, or the frequency derivatives of vibration intensity data X(F1) to X(Fm) for each frequency of the measured waveform may be used as variables.

[0110] <Effects> Figure 36 shows an example of environmental noise. Environmental vibrations generated from pumps or motors are vibrations of specific frequencies, such as 50 Hz harmonics. As shown in Figure 36, this type of environmental vibration has a steep peak in a specific frequency band. In such cases, it is effective to use the derivative of the vibration intensity in the frequency direction, as shown in Figure 35, as the evaluation value S, as this makes it easy to check the amount of fluctuation in the vibration intensity in the frequency direction.

[0111] <Device configuration> The data processing device 251 of Example 4 has basically the same configuration as the data processing device 251 of Example 1 shown in Figure 19 or Figure 21.

[0112] Figure 37 is a schematic diagram showing an example of data stored in the secondary memory in Embodiment 4. Unlike Embodiment 1, the data processing device 251 in Embodiment 4 has the evaluation value calculation model 101 pre-stored in the secondary memory 259, as shown in Figure 37. The evaluation value calculation model 101 may be provided to the data processing device 251 with the configuration shown in Figure 19 from the cloud server 252 via the network I / F 257 and updated as needed. Alternatively, the evaluation value calculation model 101 may be provided to the data processing device 251 with the configuration shown in Figure 21 from the user terminal 272 via the control communication module 271 and updated as needed. [Examples]

[0113] Example 5 differs from Example 1 in its operation of data selection processing based on evaluation values ​​for each measured waveform. Otherwise, Example 5 has basically the same configuration and operation as Example 1. The following mainly describes the differences between Example 5 and Example 1. <Processing Procedure> Figure 38 shows an example of a flowchart for data selection processing based on evaluation values ​​for each measurement waveform in Example 5.

[0114] As shown in Figure 38, in the data selection process based on the evaluation value for each measurement waveform in Example 5, unlike in Example 1, the data processing device 251 executes the measurement waveform acquisition process in step 41. If it was determined that there was a water leak in the previous step 3, it executes the measurement waveform selection process in step 42. If it was determined that there was no water leak, it does not execute the measurement waveform selection process in step 42. After that, the data processing device 251 proceeds to a series of processes from steps 43 to 46 (step 71).

[0115] In Example 1, even if the vibration intensity waveform is measured multiple times a day, if the waveform selection process is performed after acquiring the measured waveforms, only some of the measured waveforms will be used to determine whether or not there is a water leak. Therefore, it will take time from the time a water leak occurs until a water leak is detected, which raises concerns about the impact on immediacy. In this example, however, under normal circumstances, the water leak is detected without performing a waveform selection process, and the waveform selection process is started only after a water leak is detected, so the time until a water leak is detected is not prolonged. Furthermore, after a water leak is detected, the waveform selection process makes it possible to more accurately confirm whether the water leak is caused by environmental vibrations or by vibrations caused by an actual water leak.

[0116] <Output screen> In this embodiment, the data processing device 251 can be configured in various ways from a higher-level device, such as a cloud server 252 or a user terminal 272.

[0117] Figure 39 shows an example of the data selection condition input screen in Example 5.

[0118] Referring to Figure 39, the data selection condition input screen 371 is displayed. The data selection condition input screen 371 is a screen for inputting various settings related to data selection. The data selection condition input screen 371 includes an area 391 representing the sensor placement location, an area 392 showing the trend of past water leakage detection results, an input box 373 where the sensor model number can be entered, an input box 374 where the number of data to be selected can be entered, and a send button 375. The input boxes 373, 374, and send button 375 are the same as those in Embodiment 1 shown in Figure 22.

[0119] Area 391 displays the water pipeline 395 and sensor installation locations 396 in a schematic map. Area 392 displays a graph showing the trend of past leak detection results. The user can zoom in and out using the zoom in button 393 and zoom out button 394 to specify sensors at any sensor placement location. The user can also check the information related to the sensor installation locations displayed in Area 391 and the trend of past leak detection results displayed in Area 392 on the screen and determine the data selection number W. The user then enters the determined data selection number W into the input box 374 and presses the send button 375, at which point the selection number W information is sent to the sensor terminal and set.

[0120] Figure 40 shows another example of the data selection condition input screen in Example 5. In addition to the one in Figure 39, the data selection condition input screen 371 in Figure 40 includes a region 401 that shows the changes in the vibration evaluation value S for each time period and each day. The user can use the evaluation value S displayed in region 401 as a clue to determine the number of data to select W.

[0121] Although various embodiments of the present invention have been described above, the present invention is not limited to the embodiments and modifications described above, but includes various modifications and equivalent configurations within the spirit of the attached claims. For example, the embodiments and modifications described above are explained in detail to make the present invention easier to understand, and the present invention is not necessarily limited to having all the configurations described. Furthermore, a part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Furthermore, a configuration of another embodiment may be added to the configuration of one embodiment. Furthermore, a part of the configuration of each embodiment may be added, deleted, or replaced with other configurations.

[0122] Furthermore, the above-described embodiments include the following items. However, the items included in the above-described embodiments are not limited to those shown below.

[0123] (Item 1) A data processing device for determining whether or not fluid is leaking from a pipeline network based on the waveform of vibration intensity measured from the pipeline network comprises a memory for storing a software program and a processor for executing the software program. The processor acquires multiple waveform data of vibration intensity measured from the pipeline network at different times, calculates an evaluation value for evaluating how suitable or unsuitable the multiple waveform data are for determining whether or not fluid is leaking, selects waveform data from the multiple waveform data to be used for determining whether or not fluid is leaking based on the evaluation value, extracts periodic characteristics from the autocorrelation coefficient of the vibration intensity of the selected waveform data, and determines whether or not fluid is leaking from the pipeline network based on the relationship of periodic characteristics among the selected waveform data. According to this, an evaluation value is calculated for the measurement data regarding its suitability for determining whether or not fluid is leaking, and based on that evaluation value, waveform data is selected to determine whether or not fluid is leaking, thus enabling fluid leakage detection with reduced influence of environmental vibration.

[0124] (Item 2) In the data processing device described in item 1, the processor creates a list of the measured waveform data sorted in ascending or descending order of the evaluation value, and selects the waveform data to be used to determine the presence or absence of fluid leakage from the top of the list. This makes it easy to select waveform data because the waveform data to be used to determine the presence or absence of fluid leakage is selected from a list sorted based on the evaluation value.

[0125] (Item 3) In the data processing device described in item 1, the processor selects a predetermined number of waveform data from the measured plurality of waveform data to be used for determining the presence or absence of fluid leakage. This makes the process of selecting waveform data easier because the number of waveform data to be used for determining the presence or absence of fluid leakage is predetermined.

[0126] (Item 4) In the data processing device described in item 1, the processor divides the selected waveform data into multiple parts along the time axis, calculates an autocorrelation function of vibration intensity for each of the divided waveforms, and extracts periodic characteristics from the autocorrelation function. As a result, the waveform data selected as suitable for determining the presence or absence of fluid leakage is divided into multiple parts along the time axis, and periodic characteristics are extracted from the autocorrelation function of each part. Therefore, it is possible to determine the presence or absence of leakage from the pipeline network based on the relationship between the periodic characteristics of a sufficient number of waveforms suitable for determining the presence or absence of fluid leakage.

[0127] (Item 5) In the data processing device described in item 1, the time at which the waveform data is measured differs from day to day. As a result, since the waveform data is measured at different times each day, if there is ambient vibration occurring at the same time every day, it becomes possible to acquire waveforms without ambient vibration on different days and select those waveforms without ambient vibration based on the evaluation value.

[0128] (Item 6) In the data processing device described in item 1, the relationship of periodic characteristics between the waveform data is an index based on the degree of agreement of the temporal positions of the vibration intensity peaks shown in the waveform data. According to this, since fluid leakage is determined based on the degree of temporal agreement of the vibration intensity peak positions, the determination can be easily made by utilizing the properties of vibrations caused by fluid leakage from the pipeline network.

[0129] (Item 7) In the data processing device described in item 6, the processor takes the data of the periodic characteristics of the autocorrelation coefficient of the vibration intensity of the waveform data as points, connects points whose degree of agreement exceeds a predetermined threshold with edges to create a graph network, and uses the graph network to determine whether or not there is leakage of the fluid from the pipeline network. As a result, since the presence or absence of fluid leakage is determined by the graph network, the presence or absence of leakage can be easily determined from complex relationships.

[0130] (Item 8) In the data processing device described in item 1, the processor determines, based on past results of determining whether or not there is a water leak, whether to select waveform data from among the acquired waveform data to be used to determine whether or not there is a fluid leak, or to use all of the acquired waveform data to determine whether or not there is a fluid leak. This makes it possible to achieve monitoring that combines rapid determination and highly accurate determination.

[0131] (Item 9) In the data processing device described in item 1, the processor displays the temporal changes of the evaluation value in a graph. This allows the user to recognize which time period of data is suitable, as the temporal changes of the evaluation value, which indicates whether the vibration waveform data is suitable for determining the presence or absence of fluid leakage, are displayed in a graph.

[0132] (Item 10) In the data processing device described in item 1, the evaluation value is a value calculated based on at least one of the maximum amplitude of vibration intensity, vibration intensity in a specific frequency band, the integral value of vibration intensity, and the time derivative of vibration intensity. This allows for the clever exclusion of the effects of environmental vibrations depending on the environment.

[0133] (Item 11) In the data processing device described in item 10, the evaluation value is the integral value of the amplitude intensity with frequency as the integration variable. This makes it possible to select a waveform based on the difference in frequency axis characteristics between environmental vibration and vibration due to leakage. [Explanation of Symbols]

[0134] 250...Data processing unit, 251...Data processing unit, 252...Cloud server, 253...Processor, 254...Main memory, 255...Input device, 256...Output device, 257...Network interface, 258...Bus, 259...Secondary memory, 261...Microcontroller, 262...Measurement module, 263...Battery, 271...Control communication module, 272...User terminal

Claims

1. A data processing device for determining whether or not fluid is leaking from a pipeline network based on the waveform of vibration intensity measured from the pipeline network, Memory for storing software programs, The system includes a processor that executes the aforementioned software program, The aforementioned processor, Multiple vibration intensity waveform data measured at different times are obtained from the aforementioned pipeline network. For multiple waveform data, an evaluation value is calculated to assess how suitable or unsuitable they are for determining the presence or absence of fluid leakage, based on the vibration intensity in a specific frequency band. Based on the evaluation value, a waveform data suitable for extracting periodic characteristics from the autocorrelation coefficient of the vibration intensity is selected from among the multiple waveform data. The periodic characteristics are extracted from the autocorrelation coefficient of the vibration intensity of the selected waveform data. Based on the relationship of the periodic characteristics between the selected waveform data, it is determined whether or not there is leakage of the fluid from the pipeline network. Data processing device.

2. The processor creates a list by sorting the measured waveform data in ascending or descending order of evaluation value, and selects the waveform data from the top of the list to be used in determining whether or not there is fluid leakage. The data processing device according to claim 1.

3. The processor selects a predetermined number of waveform data from the measured plurality of waveform data to be used for determining the presence or absence of fluid leakage. The data processing device according to claim 1.

4. The processor divides the selected waveform data into multiple parts along the time axis, calculates the autocorrelation function of the vibration intensity for each of the divided waveforms, and extracts the periodic characteristics from the autocorrelation function. The data processing device according to claim 1.

5. The time at which the waveform data is measured varies from day to day. The data processing device according to claim 1.

6. The relationship between the periodic characteristics of the waveform data is an index based on the degree of agreement of the temporal positions of the vibration intensity peaks shown in the waveform data. The data processing device according to claim 1.

7. The processor takes the data of the periodic characteristics of the autocorrelation coefficient of the vibration intensity of the waveform data as points, connects the points whose degree of agreement exceeds a predetermined threshold with edges to create a graph network, and uses the graph network to determine whether or not there is leakage of the fluid from the pipeline network. The data processing apparatus according to claim 6.

8. The processor determines, based on past results of determining whether or not there is a water leak, whether to select waveform data from among the acquired waveform data to be used to determine whether or not there is a fluid leak, or to use all of the acquired waveform data to determine whether or not there is a fluid leak. The data processing method according to claim 1.

9. The processor displays the temporal changes of the evaluation value using a graph. The data processing device according to claim 1.

10. The aforementioned evaluation value is a value calculated based on at least one of the following: the maximum amplitude of vibration intensity, the vibration intensity in a specific frequency band, the integral value of vibration intensity, and the time derivative of vibration intensity. The data processing device according to claim 1.

11. The aforementioned evaluation value is the integral value of the amplitude intensity with frequency as the integration variable. The data processing device according to claim 10.

12. The data processing apparatus according to claim 8, wherein the processor determines whether or not there is a leak without performing the selection of waveform data to be used to determine whether or not there is a leak based on the evaluation value, and if it is determined that there is a leak in the determination, the next time the processor performs the selection of waveform data to be used to determine whether or not there is a leak based on the evaluation value and then determines whether or not there is a leak.

13. A data processing method for determining whether or not a fluid is leaking from a pipeline network based on the waveform of vibration intensity measured from the pipeline network, A data processing device having memory and a processor, Multiple vibration intensity waveform data measured at different times are obtained from the aforementioned pipeline network. For multiple waveform data, an evaluation value is calculated to assess how suitable or unsuitable they are for determining the presence or absence of fluid leakage, based on the vibration intensity in a specific frequency band. Based on the evaluation value, a waveform data suitable for extracting periodic characteristics from the autocorrelation coefficient of the vibration intensity is selected from among the multiple waveform data. The periodic characteristics are extracted from the autocorrelation coefficient of the vibration intensity of the selected waveform data. Based on the relationship of the periodic characteristics between the selected waveform data, it is determined whether or not there is leakage of the fluid from the pipeline network. Data processing method.

14. A data processing program for determining whether or not a fluid is leaking from a pipeline network based on the waveform of vibration intensity measured from the pipeline network, Multiple vibration intensity waveform data measured at different times are obtained from the aforementioned pipeline network. For multiple waveform data, an evaluation value is calculated to assess how suitable or unsuitable they are for determining the presence or absence of fluid leakage, based on the vibration intensity in a specific frequency band. Based on the evaluation value, a waveform data suitable for extracting periodic characteristics from the autocorrelation coefficient of the vibration intensity is selected from among the multiple waveform data. The periodic characteristics are extracted from the autocorrelation coefficient of the vibration intensity of the selected waveform data. The determination of whether or not there is leakage of the fluid from the pipeline network is made based on the relationship of the periodic characteristics between the selected waveform data. A data processing program to be executed by a data processing device having memory and a processor.

Images