Method for correcting flow measurement values in an ultrasonic flowmeter, and ultrasonic flowmeter of this type

Abstract

The invention relates to a method (100) for correcting measurement values of an average flow velocity of a medium flowing through a pipe which are determined by means of an ultrasonic flowmeter (1), wherein: a plurality of ultrasonic transducers (10) are arranged on or in the pipe (50) or a measurement channel (60) integrated into the pipe and span an arrangement of at least three signal paths; an electronic unit (20) operates the ultrasonic transducers, evaluates measurement signals from the ultrasonic transducers, and outputs corrected measurement values of the average flow velocity; the electronic unit, in each case for a time period, - calculates a local flow velocity of the medium by means of the measurement signals from ultrasonic transducers for each signal path, - calculates at least one, in particular dimensionless, functional relationship between local flow velocities from calculated local flow velocities, - calculates a basic measurement value for the average flow velocity from at least one determined local flow velocity; the basic measurement value is corrected by means of a correction factor determined by comparison and is output.

Description

[0001] Method for correcting flow measurement values ​​in an ultrasonic flow meter and such an ultrasonic flow meter

[0002] Ultrasonic flow meters measure the flow rate of media through pipelines by sending ultrasonic signals at an angle to the flow direction, both against and in the direction of flow. From the measured transit times or transit time differences, an averaged local measurement of a flow velocity-dependent quantity (e.g., flow velocity) can be determined along the signal path. However, this measurement is only partially representative, as local flow velocities depend, among other things, on the distance of the signal path from a pipe axis and the flow profile.

[0003] US20190186967A1 discloses a flow measurement system with an ultrasonic flow meter as the flow meter and transmitter electronics configured to determine exactly one dimensionless quantity PF (“profile factor”) from four locally determined flow velocities and, depending on the PF, to determine a correction quantity MF (“meter factor”) to correct the measured value of the flow velocity of the medium.

[0004] DE19717940A1 discloses a method for determining the vortex intensity of the medium to be monitored for an ultrasonic flow meter. In this method, a ratio between the angular momentum and longitudinal momentum of the medium is used to determine the vortex intensity, which is then used to correct the measured flow rate.

[0005] EP0804717B1 aims to provide an ultrasonic flowmeter with improved measurement accuracy. This is achieved by determining a current Reynolds number, for which an operating flow profile is first recorded as a function of several measured values ​​of the medium's flow velocity, in order to account for the influence of inlet effects. It has been shown that such methods assume that a flow profile is already essentially formed and does not deviate significantly from rotational symmetry.

[0006] DE102013106108A1 discloses a method for determining a compensated flow rate in which, taking into account a distance of the measuring setup to the disturbance, the type of disturbance and optionally the Reynolds number, the correction factor to be applied is determined and used for determining the flow rate.

[0007] A disadvantage of the aforementioned state of the art is that, in addition to the determined local measured values ​​of the flow velocity-dependent quantity, further information is always necessary to identify the flow profile in order to determine a clear correction value suitable for the present situation.

[0008] The object of the invention is to provide a remedy. This object is achieved by the method according to independent claim 1, as well as by the method according to claim 11 and the ultrasonic flow meter according to independent claim 13.

[0009] The inventive method for determining a correction value for correcting a measured value of a flow velocity-dependent measured quantity of a medium flowing along a measuring channel, determined by means of an ultrasonic flow meter, wherein the ultrasonic flow meter comprises several ultrasonic transducers that span at least three signal paths running through the measuring channel, wherein the signal paths each have an axial extension in the longitudinal direction of the measuring channel, comprises the following method steps:

[0010] - Determine one local measurement value of the flow velocity-dependent measured quantity for each of the at least three signal paths;

[0011] - Determine from the at least three local measured values ​​an n-tuple with n > 2 of, in particular dimensionless, functional relationships, wherein at least two of the, in particular dimensionless, functional relationships differ at least in that at least one local measured value is included in the respective determination, which is not included in the correspondingly other, in particular dimensionless, functional relationship; and

[0012] - Determining the correction value depending on the n-tuple.

[0013] The flow velocity-dependent measured quantity can be, for example, a determined transit time of an ultrasound signal, a difference in transit times of two different ultrasound signals, a flow velocity derived from it, a volume flow rate derived from it, or a mass flow rate derived from it.

[0014] The measuring channel describes the volume or space through which the fluid flows. The shape of the measuring channel is usually determined by the shape of the pipeline or measuring tube. The measuring channel can be cylindrical, cubic, or any other shape.

[0015] The ultrasonic transducers can be arranged directly or indirectly on an outer surface of a pipeline (e.g., WO 2019 206 507 A1) or in contact with the medium (e.g., DE 102014 105 840 A1) in an opening of the measuring tube. The measured value of the flow velocity-dependent quantity to be corrected can, for example, be a local measurement from at least three local measurements of the flow velocity-dependent quantity or a base value determined from at least two local measurements.

[0016] The at least two, in particular dimensionless, quantities differ at least in that at least one local measurement is included in the determination of the respective, in particular dimensionless, quantity, which is not included in the determination of the corresponding other, in particular dimensionless, quantity. This means that one of the at least two, in particular dimensionless, quantities depends on a local measurement that is not included in the determination of the other of the at least two, in particular dimensionless, quantities. Further dimensionless quantities can also be included in the n-tuple, which do not differ by the local measurements included in the respective determination, but rather by the mathematical function itself used to determine the, in particular dimensionless, quantity.

[0017] The measuring channel or pipeline is not limited to circular cross-sections. The method according to the invention can be carried out with virtually all cross-sectional shapes. For example, a cross-section can follow a polygon, particularly a regular one. A cross-section can also follow an irregular polygon, such as a rectangle.

[0018] After a disturbance, the flow profile of a flowing medium requires a certain flow path in a straight section of pipe to reach a fully developed flow state.

[0019] The formation of functional relationships, especially dimensionless ones, allows a statement to be made about a flow profile independently of an absolute flow velocity.

[0020] One embodiment provides that when determining the correction value, the n-tuple is compared with a data set comprising an m-tuple of, in particular dimensionless, functional reference relationships, which is linked to a correction value, where m = n can hold.

[0021] A comparison of the n-tuple of, in particular, dimensionless, functional relationships with the m-tuple of, in particular, dimensionless, functional reference relationships means searching for the best fit between, in particular, dimensionless, functional relationships and, in particular, dimensionless, functional reference relationships. This can, for example, involve searching for the nearest reference relationship within a most similar group of reference relationships and / or interpolating surrounding reference relationships. This can be done, for example, via multivariate regression analysis or categorization.

[0022] One implementation provides that the m-tuple and / or the n-tuple is linked to one of the following variables:

[0023] - Type of upstream flow disturbance;

[0024] - Distance of the flow disturbance to the ultrasonic flow meter;

[0025] - Orientation of the flow disturbance towards the ultrasonic flow meter; and / or

[0026] - Reynolds number.

[0027] Flow disturbances in a pipeline include, for example, bends of all conceivable variations, valves, grids, changes in the cross-section of the pipeline such as expansions or reductions.

[0028] In one embodiment, the variable "type of upstream flow disturbance" includes the variables distance and orientation of the flow disturbance with respect to the arrangement of the signal paths.

[0029] The distance can be measured, for example, relative to a center point of all signal paths. As a starting point for a distance measurement, for example, a first cross-section of the measuring tube or pipeline in a straight section without geometric interference can be selected.

[0030] One embodiment provides that a look-up table, a calculation of a Pareto front, or a pattern recognition algorithm, in particular using a neural network, is applied when determining the correction value.

[0031] One embodiment provides that if the n-tuple deviates from the m-tuple of the, in particular dimensionless, functional reference relationships, the correction value is determined by one of the following procedures:

[0032] Searching for the nearest data set of a most similar group of, in particular dimensionless, functional reference relationships;

[0033] Interpolation of surrounding, especially dimensionless, functional reference contexts;

[0034] Applying a regression analysis procedure using linear regression; or

[0035] Applying a regression analysis method using nth-order polynomials, with n > 2. One embodiment provides that the determination of the correction value incorporates an m-dimensional correction function, particularly one stored in the ultrasonic flowmeter, which depends on the n-tuple, where m > 2, and in particular m = n.

[0036] The correction function can, for example, be an approximation function derived from a fit function or an interpolation function with data points as support points. The at least two functional relationships, particularly dimensionless ones, are input variables, i.e., arguments, of the correction function when determining the correction factor.

[0037] One embodiment provides that the m-dimensional correction function depends on

[0038] - the type of upstream flow disturbance,

[0039] - the distance of the flow disturbance to the ultrasonic flow meter,

[0040] - the orientation of the flow disturbance to the ultrasonic flow meter, and / or

[0041] - the Reynolds number, is selected.

[0042] In one embodiment, the electronic unit receives information on at least one of the following variables: the Reynolds number of the medium, the type and / or distance and / or orientation of the flow disturbance. This can be done wirelessly, for example, via a handheld device or a WAN (Wide Area Network, a cloud). Alternatively, the information can also be provided to the operator via a display.

[0043] This allows the comparison to be limited to a portion of the data set. Particularly when dealing with a small number of signal paths, this helps avoid ambiguities during the comparison.

[0044] One embodiment provides that a pattern recognition algorithm based on a neural network is used to determine the correction value, wherein the neural network is trained such that n-tuples with n > 2 from measured and / or simulated local values ​​of the flow velocity-dependent measured quantity for different disturbances and for different medium velocities (optionally different distances / orientations) in conjunction with a measured and / or simulated correction value are provided to the neural network, wherein the neural network is configured to create a neural data set from the provided n-tuples, which comprises fc-tuples with k > 2, each of which is assigned a theoretical correction value, and to compare this neural data set with the provided n-tuples.

[0045] One embodiment provides that at least half of a number of the signal paths are N-traverse signal paths, with N being an odd number, and / or where at least half of a number of the signal paths are N-traverse signal paths, with N being an even number.

[0046] One embodiment provides that at least two of the signal paths are chosen such that, in the presence of a rotationally symmetric or rotationally symmetric flow profile in the flowing medium, different local measured values ​​of the flow velocity-dependent measured quantity are measured.

[0047] One embodiment provides that the functional relationships, especially those without dimensions, are each determined from at least one of the following functional relationships:

[0048] - Ratio of local measured values ​​of the flow velocity-dependent measured quantity with respect to different signal paths;

[0049] - Ratio of sums, differences or linear combinations of local measured values ​​of the flow velocity-dependent measured quantity with respect to multiple signal paths;

[0050] - Products, sums, logarithms or powers of the local measured values ​​of the flow velocity-dependent measured quantity;

[0051] - Result of an exponential function of local measured values ​​of the flow velocity-dependent quantity or the quantities listed above.

[0052] A specialist is able to create dimensionless functional relationships, if necessary, depending on their interests.

[0053] One embodiment provides that the type of upstream flow disturbance, the distance of the flow disturbance to the ultrasonic flow meter, the orientation of the flow disturbance relative to the ultrasonic flow meter, and / or the Reynolds number are not included in the calculation of the correction value, and / or that only the n-tuple is included in the calculation of the correction value. Thus, the electronic unit does not need to be configured to first determine a Reynolds number via a physical and / or mathematical approach and then calculate the correction value from that.

[0054] The method according to the invention for determining a corrected measured value of a flow velocity-dependent measured quantity comprises the following process steps:

[0055] - Determining a correction value using a method of the preceding claims; and

[0056] - Determining the corrected measured value of the flow velocity-dependent quantity as a function of the determined correction value and a determined baseline measured value of the flow velocity-dependent quantity; and

[0057] - Optional: Output of the corrected measured value of the flow velocity-dependent measured quantity.

[0058] The base measurement value can be, for example, a single local measurement value from at least three local measurements, or a weighted sum of several local measurements. A calibration value, particularly one that is stored, can also be included in the determination of the flow velocity-dependent measurement value.

[0059] One embodiment provides that the at least three signal paths comprise a first, second, third and fourth signal path, wherein the first and third signal paths each form a 1-traverse signal path, and the second and fourth signal paths each form a 2-traverse signal path.

[0060] One embodiment provides that the first and third signal paths each intersect a longitudinal axis of the measurement channel, while the second and fourth signal paths each pass through the measurement channel off-center.

[0061] The measurement channel is typically shaped such that all cross-sections of the measurement channel have a center point that is intersected by the longitudinal axis of the measurement channel. An off-center signal path does not intersect any center point of the measurement channel's cross-sections.

[0062] One design stipulates that the procedure includes the following procedural steps:

[0063] - Determine a local measurement value of the flow velocity-dependent measured quantity for each of the at least three signal paths; - Determine a base value as a function of the at least three local measurement values ​​of the flow velocity-dependent measured quantity, wherein, in addition to the at least three local measurement values, at least one weighting value is also included in the determination of the base value, and the base value is included in the determination of the corrected measurement value of the flow velocity-dependent measured quantity.

[0064] The procedural steps can be exchanged between each other, provided there are no causal reasons against it.

[0065] The ultrasonic flowmeter according to the invention for determining a measured value of a flow velocity-dependent quantity of a flowing medium in a measuring channel comprises: several ultrasonic transducers which are configured to transmit and / or receive ultrasonic signals along at least three signal paths through the measuring channel; an electronic unit which is configured to operate the ultrasonic transducers, to provide measurement signals from the ultrasonic transducers and to provide measured values ​​of the flow velocity-dependent quantity which are corrected by means of a correction value, the correction factor of which is determined by means of a method according to the invention.

[0066] The ultrasonic flow meter can be or include an inline meter, a clamp-on meter, or a Lamb wave meter.

[0067] One embodiment provides that the multiple ultrasonic transducers form at least four pairs of ultrasonic transducers, wherein the respective ultrasonic transducers of one of the four pairs are connected to each other via a first, second, third, or fourth signal path. The first and third signal paths each form a single-traverse signal path, while the second and fourth signal paths each form a two-traverse signal path. The first and third signal paths each intersect a longitudinal axis of the measurement channel, and the second and fourth signal paths each traverse the measurement channel off-center. Signal paths with an even number of traverses exhibit inherent crossflow compensation. Compensation for crossflows results in the loss of information that can no longer be used. This can be disadvantageous for the implementation of the method according to the invention.

[0068] In one embodiment, at least one pair of 1-traverse signal paths is provided, wherein the signal paths of the pair run in a plane and wherein two straight lines running along the signal paths enclose an angle α of at least 30 degrees enclosing the longitudinal axis of the measuring channel, wherein signal paths of at least one pair of the at least one pair have a relative distance of at most 0.8 r, and in particular at most 0.5 r and preferably at most 0.2 r from the longitudinal axis of the measuring channel, with r being the length of a line between the longitudinal axis of the measuring channel and a wall section of the measuring channel, wherein the line passes through a point of the signal path with the smallest distance of the signal path to the longitudinal axis of the measuring channel.

[0069] In this way, ultrasound signals which run along the signal paths of the pair can be used for the computational compensation of cross-flow influences on the flow measurement, but also the cross-flow influences can be taken into account for the method according to the invention.

[0070] In one embodiment, signal paths of different relative distances r to the longitudinal axis of the measuring channel have a relative difference of at least 0.1 r and in particular at least 0.2 r and at most 0.8 r and in particular at most 0.5 r, with r being the length of a line between the longitudinal axis of the measuring channel and a wall section of the measuring channel, wherein the line passes through a point of the signal path with the smallest distance of the signal path to the longitudinal axis of the measuring channel.

[0071] This provides a sufficiently good scan of the flow profile.

[0072] In one embodiment, at least n / 2 different projections exist in a projection of the signal paths onto a cross-section of the measuring channel, wherein at least n / 4 of the signal paths do not pass through the longitudinal axis of the measuring channel, where n is a number of signal paths.

[0073] This provides a sufficiently good sample of the flow profile. If n / 2 or n / 4 is not a whole number, the rounded-up value applies.

[0074] In one embodiment, at least two of the signal paths are configured to sample different local flow velocities when a rotationally symmetric or rotationally symmetric flow profile is present.

[0075] This ensures that even with an almost fully developed flow profile, sufficiently diverse information is perceived.

[0076] In one configuration, the data set is created experimentally or using computer-based methods such as computational fluid dynamics simulation.

[0077] The invention will be described below using exemplary embodiments.

[0078] Fig. 1a sketches a longitudinal section through one embodiment of the ultrasonic flow meter;

[0079] Fig. 1b sketches a cross-section through an embodiment of the ultrasonic flow meter of Fig. 1a;

[0080] Fig. 2a shows the ratio of two determined local medium velocities as a function of the Reynolds number;

[0081] Fig. 2b shows the error of an ultrasonic flow meter as a function of the Reynolds number;

[0082] Fig. 3a shows the ratio of two determined local medium velocities as a function of the Reynolds number for different measurement path combinations;

[0083] Fig. 3b shows the remaining error after correction using the method according to the invention as a function of the error resulting from uncorrected measured values; and

[0084] Fig. 4 shows a process flow of an embodiment of the method according to the invention.

[0085] Fig. 1a sketches a longitudinal section through an embodiment of the ultrasonic flowmeter 1, in which ultrasonic transducers 10 are arranged on a pipe 50 with a measuring channel 60. Alternatively, the ultrasonic transducers 10 can also be arranged in openings provided for this purpose in a measuring tube and projecting into the measuring channel, or at least in contact with the medium. In the embodiment shown (Figs. 1a and 1b), eight ultrasonic transducers 10 are provided. A different number of ultrasonic transducers 10 can also be provided to generate at least three measuring paths A, B, C. The ultrasonic transducers 10 are configured to emit ultrasonic signals into a medium flowing along the measuring channel 60 and also to detect ultrasonic signals traveling through the medium along one of the at least three signal paths A, B, C. More than three signal paths A, B, C can also be provided. The embodiment shown in Fig.1a and 1b contain four pairs of ultrasonic transducers UW1, UW2, UW3, and UW4, which together form four signal paths A, B, C, and D. An electronic unit 20 is configured to provide operating signals to the ultrasonic transducers 10, to evaluate and / or correct measurement signals from the ultrasonic transducers 10, and to output measured values ​​of a corrected flow velocity-dependent measurement quantity. The ultrasonic flowmeter 1 can, for example, operate according to the transit-time difference principle.

[0086] Ultrasonic transducers 10 can be arranged in pairs as shown here. Signal paths A, B, C through the medium can have an even number of crossheads (Fig. 1b), for example 2, or an odd number of crossheads, as shown here (Fig. 1a), for example 1. A signal path with one crosshead describes a straight path of the ultrasonic signal without reflections of the ultrasonic signal from the pipe wall or measuring tube wall. If one assumes that an imaginary cross-sectional plane – in which the longitudinal axis of the measuring channel can also lie – divides the pipe or measuring tube into two sides, then the ultrasonic transducers 10 forming a pair are located on opposite sides of the pipe or measuring tube. A signal path with two crossheads describes a path of the ultrasonic signal that is reflected exactly once from the pipe wall or measuring tube wall.In this case, the two ultrasonic transducers forming a pair are usually arranged on the same side of the pipeline or measuring tube.

[0087] Fig. 1b sketches a cross-section through an embodiment of the ultrasonic flowmeter 10 of Fig. 1a. As shown in Fig. 1a, the signal paths A, C of the at least three signal paths A, B, C can run through a longitudinal axis 70 of the measuring channel. Alternatively, the signal paths B, D of the at least three signal paths A, B, C, D can run off-center, see Fig. 1b. By means of signal paths that have an odd number of crossheads and run off-center (relative to the center of the cross-section), transverse components of a flow, such as a rotational flow component in the medium, can be detected, for example. With signal paths with an even number of crossheads, such transverse components are compensated. The number of reflections N Reflexof the ultrasound signal along the signal path results from the relationship N Reflex = N - 1 with the number of traverses N.

[0088] The measuring channel or pipeline shown here is schematic; for example, edges may be rounded. A wall thickness can be selected by a specialist as required.

[0089] Figure 2a shows the ratio of two measured local medium velocities as a function of the Reynolds number for an ultrasonic flowmeter with two measuring paths. If a value for the ratio of two local measured values, e.g., the medium velocity, is known, the Reynolds number for the ultrasonic flowmeter and the medium in the measuring channel can be determined if the functional relationship between the ratio of the medium velocities and the Reynolds number is known or stored. Figure 2b shows the error of an ultrasonic flowmeter as a function of the Reynolds number. Once the Reynolds number has been determined and the functional relationship between the error and the Reynolds number is known or stored, the applicable correction value can be determined.

[0090] Figure 3a shows the ratio of two measured local medium velocities as a function of the Reynolds number for different measurement path combinations of an ultrasonic flowmeter. It can be seen that if only the ratio of the medium velocities is known, a unique Reynolds number cannot be determined. If the Reynolds number cannot be uniquely determined, then a unique correction value cannot be determined either.

[0091] Fig. 3b shows the remaining error after correction using the inventive method (see Fig. 4) as a function of the error resulting from uncorrected measurements. It can be seen that the measurement error remaining after correction is predominantly between ±1%. Individual outliers result in a residual measurement error of over ±2%. The results demonstrate that the inventive method can reduce the measurement error to a level of ±2% that is acceptable to the customer.

[0092] Fig. 4 shows a process flow of an embodiment of the inventive method 100.

[0093] An ultrasonic flowmeter with at least three signal paths (see Figs. 1a and 1b) can be used to apply an exemplary method 100 according to the invention for determining a correction factor or for determining a corrected flow velocity-dependent measured quantity of the medium as outlined in Fig. 4. The signal paths each have an axial extension in the longitudinal direction of the measuring channel. This means that the signal paths do not extend exclusively in a cross-section (as disclosed, for example, in EP 0 804 717 B1), but extend at least section by section in the longitudinal direction of the measuring channel.

[0094] The method (100) for determining a correction value for correcting a measured value of a flow velocity-dependent quantity of a medium flowing along a measuring channel, as determined by an ultrasonic flow meter, comprises the following method steps:

[0095] I. Determine one local measurement of the flow velocity-dependent quantity for each of the at least three signal paths. This will result in at least three local measurements of the flow velocity-dependent quantity. These local measurements can be, for example, local transit-time differences or velocities derived therefrom. II. From the at least three local measurements, determine an n-tuple with n > 2 of different, in particular dimensionless, functional relationships. At least two of the, in particular dimensionless, functional relationships differ at least in that at least one local measurement is used in the determination of one of the at least two, in particular dimensionless, functional relationships and not in the corresponding other, in particular dimensionless, functional relationship.

[0096] The functional relationships, especially dimensionless ones, can each be determined from at least one of the following functional relationships:

[0097] - Ratio of local measured values ​​of the flow velocity-dependent measured quantity with respect to different signal paths;

[0098] - Ratio of sums, differences or linear combinations of local measured values ​​of the flow velocity-dependent measured quantity with respect to multiple signal paths;

[0099] - Products, sums, logarithms or powers of the local measured values ​​of the flow velocity-dependent measured quantity;

[0100] - Result of an exponential function of local measured values ​​of the flow velocity-dependent quantity or the quantities listed above.

[0101] III. Determining the correction value depending on the n-tuple.

[0102] Various methods can be used for this. According to one initial approach, determining the correction value involves comparing the n-tuple with a dataset comprising an m-tuple of, in particular, dimensionless, functional reference relationships. The m-tuple is linked to a correction value. A lookup table, a Pareto front calculation, or a pattern recognition algorithm, especially using a neural network, can be employed to determine the correction value.

[0103] The m-tuple and / or the n-tuple can be linked to one of the following variables:

[0104] - Type of upstream flow disturbance;

[0105] - Distance of the flow disturbance to the ultrasonic flow meter;

[0106] - Orientation of the flow disturbance towards the ultrasonic flow meter; and / or

[0107] - Reynolds number. This means that these variables must be known and possibly stored in the ultrasonic flow meter.

[0108] Alternatively, the mT tuple and / or the nT tuple may not be linked to any of the variables mentioned above. In this case, a correction value is determined using the n-tuple without knowledge of the type of flow disturbance, the distance, the orientation, or the Reynolds number. Only the n-tuple is then used to calculate the correction value.

[0109] If there is a discrepancy between the n-tuple and the m-tuple, one of the following procedural steps can also be performed.

[0110] Searching for the nearest data set of a most similar group of, in particular dimensionless, functional reference relationships;

[0111] Interpolation of surrounding, especially dimensionless, functional reference contexts;

[0112] Applying a regression analysis procedure using linear regression; or

[0113] Applying a regression analysis procedure using nth-order polynomials, with n > 2.

[0114] Alternatively, according to a second embodiment, an rn-dimensional correction function with m > 2, which is dependent on the n-tuple and is stored in the ultrasonic flowmeter, can be included in the determination of the correction value.

[0115] Here, the m-dimensional correction function can also be compared to the above, depending on the

[0116] - Type of upstream flow disturbance,

[0117] - Distance of the flow disturbance to the ultrasonic flow meter,

[0118] - Orientation of the flow disturbance towards the ultrasonic flow meter, and / or

[0119] - Reynolds number, be selected, or depend on it.

[0120] Alternatively, exactly one correction function can be provided that does not depend on the previously mentioned variables.

[0121] Furthermore, according to a third embodiment, a pattern recognition algorithm based on a neural network can be used to determine the correction value. The neural network can be trained such that n-tuples with n > 2, derived from measured and / or simulated local values ​​of the flow velocity-dependent measured quantity for different disturbances and medium velocities (optionally different distances / orientations), are provided to the neural network in conjunction with a measured and / or simulated correction value. During training, the neural network can be configured to create a neural dataset based on the provided n-tuple. This dataset comprises fc-tuples with k > 2, each assigned a theoretical correction value, and is then compared with the provided n-tuples.

[0122] Further procedural steps may be provided to determine the corrected measured value of a flow velocity-dependent measured quantity.

[0123] IV. Determine one local measurement value of the flow velocity-dependent measured quantity for each of the at least three signal paths;

[0124] V. Determining a base value based on at least three local measured values ​​of the flow velocity-dependent measured quantity,

[0125] VI. Determining the corrected measured value of the flow velocity-dependent measured quantity as a function of the determined correction value and a determined baseline measured value of the flow velocity-dependent measured quantity.

[0126] Optionally, the output of the corrected measured value of the flow velocity-dependent measured quantity can also be provided.

[0127] Optionally, in addition to the at least three local measured values, at least one weighting value can also be included in the determination of the base value, and the base value can be included in the determination of the corrected measured value of the flow velocity-dependent measured quantity.

[0128] Reference symbol list

[0129] 1 ultrasonic flow meter

[0130] 10 ultrasound transducers

[0131] A, B, C, D Signal path 20 Electronic unit

[0132] 50 Pipeline

[0133] 60 measuring channels

[0134] 70 Measuring channel longitudinal axis

[0135] 100 Procedure I First Procedure Step

[0136] II second procedural step

[0137] III third procedural step

[0138] IV fourth procedural step

[0139] VI fifth procedural step

Claims

Patent claims 1. Method (100) for determining a correction value for correcting a measured value of a flow velocity-dependent quantity of a medium flowing along a measuring channel (60) determined by means of an ultrasonic flow meter (1), wherein the ultrasonic flow meter (1) comprises several ultrasonic transducers (10) spanning at least three signal paths (A, B, C) running through the measuring channel (60), wherein the signal paths (A, B, C) each have a partial axial extension in the longitudinal direction of the measuring channel (60), comprising: - Determine one local measurement value of the flow velocity-dependent measured quantity for each of the at least three signal paths (A, B, C); - Determine from the at least three local measured values ​​an n-tuple with n > 2 of, in particular dimensionless, functional relationships, wherein at least two of the, in particular dimensionless, functional relationships differ at least in that at least one local measured value is included in the respective determination, which is not included in the correspondingly other, in particular dimensionless, functional relationship; and - Determining the correction value depending on the n-tuple.

2. Method according to claim 1, wherein, when determining the correction value, the n-tuple is compared with a data set comprising an m-tuple of, in particular dimensionless, functional reference relationships, which m-tuple is linked to a correction value.

3. Method according to claim 2, where the m-tuple and / or the n-tuple is linked to one of the following variables: - Type of upstream flow disturbance; - Distance of the flow disturbance to the ultrasonic flow meter (1); - Orientation of the flow disturbance towards the ultrasonic flow meter (1); and / or - Reynolds number.

4. Method according to claim 1 or 2, wherein a look-up table, a calculation of a Pareto front or a pattern recognition algorithm, in particular by means of a neural network, is used to determine the correction value.

5. Method according to one of the preceding claims, wherein, in the event of a deviation of the n-tuple from the m-tuple of the, in particular dimensionless, functional reference relationships, the correction value is determined by one of the following procedures: Searching for the nearest data set of a most similar group of, in particular dimensionless, functional reference relationships; Interpolation of surrounding, especially dimensionless, functional reference contexts; Applying a regression analysis procedure using linear regression; or Applying a regression analysis procedure using nth-order polynomials, with n > 2.

6. Method according to claim 1, wherein the determination of the correction value incorporates an m-dimensional correction function, in particular one stored in the ultrasonic flowmeter (1), which is dependent on the n-tuple, where m > 2.

7. The method of claim 6, wherein the m-dimensional correction function depends on the - Type of upstream flow disturbance, - Distance of the flow disturbance to the ultrasonic flow meter (1), - Orientation of the flow disturbance towards the ultrasonic flow meter (1), and / or - Reynolds number, selected.

8. Method according to claim 1, wherein a pattern recognition algorithm based on a neural network is used to determine the correction value, wherein the neural network is trained such that n-tuples with n > 2 from measured and / or simulated local values ​​of the flow velocity-dependent measured quantity for different disturbances and for different medium velocities in conjunction with a measured and / or simulated correction value are provided to the neural network, wherein the neural network is configured to create a neural data set from the provided n-tuples, which comprises fc-tuples with k > 2, each of which is assigned a theoretical correction value, and to compare this neural data set with the provided n-tuples.

9. Method according to one of the preceding claims, wherein the functional relationships, in particular dimensionless ones, are each determined from at least one of the following functional relationships: - Ratio of local measured values ​​of the flow velocity-dependent measured quantity with respect to different signal paths (A, B, C); - Ratio of sums, differences or linear combinations of local measured values ​​of the flow velocity-dependent measured quantity with respect to multiple signal paths (A, B, C); - Products, sums, logarithms or powers of the local measured values ​​of the flow velocity-dependent measured quantity; - Result of an exponential function of local measured values ​​of the flow velocity-dependent quantity or the quantities listed above.

10. Method according to any of the preceding claims, except claims 3 and 7, wherein the type of upstream flow disturbance, the distance of the flow disturbance to the ultrasonic flow meter (1), the orientation of the flow disturbance to the ultrasonic flow meter (1) and / or the Reynolds number do not take place in the determination of the correction value, and / or wherein only the n-tuple takes place in the determination of the correction value.

11. Method for determining a corrected measured value of a flow velocity-dependent measured quantity, comprising the following method steps: - Determining a correction value using a method of the preceding claims; and - Determining the corrected measured value of the flow velocity-dependent quantity as a function of the determined correction value and a determined baseline measured value of the flow velocity-dependent quantity; and - Optional: Output of the corrected measured value of the flow velocity-dependent measured quantity.

12. A method according to one of claim 11, comprising the method steps: - Determine one local measurement value of the flow velocity-dependent measured quantity for each of the at least three signal paths (A, B, C); - Determining a base value as a function of at least three local measured values ​​of the flow velocity-dependent measured quantity, wherein, in addition to the at least three local measured values, at least one weighting value is also included in the determination of the base value, and the base value is included in the determination of the corrected measured value of the flow velocity-dependent measured quantity.

13. Ultrasonic flowmeter (1) for determining a measured value of a flow velocity-dependent quantity of a flowing medium in a measuring channel (60), comprising: several ultrasonic transducers (11) which are configured to transmit and / or receive ultrasonic signals along at least three signal paths through the measuring channel (60); an electronic unit (20) which is configured to operate the ultrasonic transducers (11), to provide measuring signals from the ultrasonic transducers and to provide measured values ​​of the flow velocity-dependent quantity which are corrected by means of a correction value, the correction factor of which is determined by a method according to one of claims 1 to 10.

14. Ultrasonic flowmeter (1) according to claim 13, wherein the multiple ultrasonic transducers (11) form at least four pairs of ultrasonic transducers (UWP1 , UWP2, UWP3, UWP4), wherein the respective ultrasonic transducers of one of the four pairs of ultrasonic transducers (UWP1 , UWP2, UWP3, UWP4) are connected to each other via a first, second, third or fourth signal path (A, B, C, D), wherein the first and third signal paths (A, C) each form a 1-traverse signal path, wherein the second and fourth signal paths (B, D) each form a 2-traverse signal path, wherein the first and third signal paths (A, C) each intersect a longitudinal axis (70) of the measuring channel (60), and wherein the second and fourth signal paths (B, D) each pass through the measuring channel (60) off-center.

Images