Detection and identification of change in vibrometer
By employing statistical methods to analyze conduit stiffness changes, the system accurately detects and identifies issues like erosion or corrosion in vibrometers, reducing false alarms and improving maintenance efficiency.
Patent Information
- Application Number
- JP2024150958
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-09-02
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing vibrometers struggle to accurately detect and identify changes in conduit stiffness due to factors like erosion, corrosion, or damage, leading to false alarms and inadequate maintenance procedures.
The implementation of instrument electronics that utilize statistical methods, including central tendency values and variance values, to determine stiffness changes at multiple positions within the conduit, and calculate probabilities using confidence intervals and bias deadbands to identify the nature of these changes.
This approach allows for early detection of conduit changes, reduces false alarms, and provides precise identification of conditions such as erosion, corrosion, or damage, ensuring appropriate maintenance actions are taken.
Smart Images

Figure 2024164271000001_ABST
Abstract
Description
[Technical field]
[0001] The embodiments described below relate to changes in a vibrometer, and more particularly, to detecting and identifying changes in a vibrometer. [Background technology]
[0002] Vibration meters, such as Coriolis mass flow meters, liquid densitometers, gas densitometers, liquid viscometers, gas / liquid specific gravity meters, gas / liquid relative density meters, and gas molecular weight meters, are commonly known and used to measure fluid properties. Generally, a vibration meter comprises a meter assembly and an electronics section. The material in the meter assembly may be flowing or stationary. Each type of sensor may have unique characteristics that the meter must take into account to achieve optimal performance. For example, some sensors may require the tube device to vibrate at a certain displacement level. Other meter assembly types may require special compensation algorithms.
[0003] Among other functions, the meter electronics typically includes stored sensor calibration values for the particular sensor being used. For example, the meter electronics may include stiffness measurements. The stiffness of the reference sensor represents a baseline measurement for the sensor geometry of a particular meter assembly, e.g., as measured at the factory under reference conditions or when it was last calibrated. Changes between the stiffness measured after the vibrometer is installed at the customer site and the stiffness of the reference sensor may represent physical changes in the meter assembly due to coatings, erosion, corrosion, or damage to the conduits in the meter assembly, among other causes. A meter validation or health check test can detect these changes.
[0004] The meter verification can determine if the difference between the measured stiffness and the reference stiffness is within range. For example, the comparison can determine if the measured stiffness is within the range of the reference stiffness. If the comparison indicates a change greater than or outside the range, the vibrometer can send an alarm to notify the user to investigate the fault. However, this simple comparison of a single stiffness value may not indicate the root cause of the fault. That is, the user will not know if the fault is due to erosion / corrosion, damage (e.g., freezing, overpressure, etc.), or coating. This is because the ranges are set to include all possible root causes or changes to the conduit and to prevent false alarms due to causes not attributable to changes in the conduit. An example of a false alarm is an increase in the variability of the stiffness measurement caused by a high velocity or high noise gas flow. Summary of the Invention [Problem to be solved by the invention]
[0005] If changes can be detected correctly, they can be detected at an early stage as they form. Furthermore, if changes can be detected correctly, false alarms can be minimized. If changes to the conduit can be identified, an indication of the nature of the change can be provided to the user. This can prevent downtime of the vibrometer due to false alarms and ensure that post-alarm procedures are more appropriate to the condition of the vibrometer. Thus, there is a need to detect and identify changes in a vibrometer. [Means for solving the problem]
[0006] Meter electronics for detecting and identifying changes in a vibration meter are provided. In one embodiment, the meter electronics includes an interface configured to receive a sensor signal from a meter assembly and provide information based on the sensor signal, and is communicatively coupled to the interface. and a processing system configured to use the information to determine a first stiffness change associated with a first position of the conduit of the vibrometer, determine a second stiffness change associated with a second position of the conduit of the vibrometer, and determine a condition of the conduit based on the first stiffness change and the second stiffness change.
[0007] Meter electronics for detecting and identifying changes in a vibration meter is provided. In one embodiment, the meter electronics comprises a processing system including a storage system configured to store a central tendency value of a meter validation parameter and a variance value of the meter validation parameter. The processing system is configured to retrieve the central tendency value and the variance value from the storage system and determine a probability based on the central tendency value and the variance value to detect whether the central tendency value differs from a baseline value.
[0008] A method is provided for detecting and identifying changes in a vibrometer, the method including determining a first stiffness change associated with a first position of the vibrometer of a conduit, determining a second stiffness change associated with a second position of the vibrometer of the conduit, and determining a condition of the conduit based on the first stiffness change and the second stiffness change.
[0009] A method is provided for detecting and identifying changes in a conduit of a vibrometer, which in one embodiment includes obtaining a central tendency value of a meter validation parameter and a variance value of the meter validation parameter from a storage device in the meter electronics of the vibrometer, and determining a probability based on the central tendency value and the variance value to determine whether the central tendency value differs from a baseline value.
[0010] Aspects According to one aspect, meter electronics (20) for detecting and identifying changes in a vibrometer (5) includes an interface (201) configured to receive a sensor signal (100) from a meter assembly (10) and provide information based on the sensor signal (100), and a processing system (202) communicatively coupled to the interface (201). The processing system (202) is configured to use the information to determine a first stiffness change associated with a first position of the conduit (130, 130') of the vibrometer (5), determine a second stiffness change associated with a second position of the conduit (130, 130') of the vibrometer (5), and determine a condition of the conduit (130, 130') based on the first stiffness change and the second stiffness change.
[0011] Preferably, the first location is the location of a left pickoff sensor (170l) on the conduit (130, 130') of the vibrometer (5) and the second location is the location of a right pickoff sensor (170r) on the conduit (130, 130') of the vibrometer (5).
[0012] Preferably, the first stiffness change represents a physical stiffness change of the conduit (130, 130') between the driver (180) and a first position of the conduit (130, 130'), and the second stiffness change represents a physical stiffness change of the conduit (130, 130') between the driver (180) and a second position of the conduit (130, 130').
[0013] Preferably, the meter electronics (20) is further configured to determine the stiffness symmetry of the conduits (130, 130').
[0014] Preferably, the stiffness symmetry of the conduits (130, 130') is determined based on the first stiffness variation and the second stiffness variation.
[0015] Preferably, the processing system (202) is further configured to provide an alert based on the determination of the condition of the conduit.
[0016] Preferably, the condition of the conduits (130, 130') includes at least one of erosion, corrosion, damage, and coating of the conduits (130, 130') of the vibrometer (5).
[0017] Preferably, the processing system (202) is further configured to determine a condition of the conduits (130, 130') based on at least one of a residual flexibility, damping, and mass of the conduits (130, 130').
[0018] The meter electronics (20) for detecting and identifying changes in a vibration meter (5) comprises a processing system (202) including a storage system (204) configured to store a central tendency value of a meter validation parameter and a variance value of the meter validation parameter. The processing system (202) is configured to retrieve the central tendency value and the variance value from the storage system (204) and determine a probability based on the central tendency value and the variance value to detect whether the central tendency value differs from a baseline value.
[0019] Preferably, the processing system (202) is configured to determine the probability based on the central tendency value and the variance value, and the processing system (202) is configured to calculate a t-value and to calculate the probability using the t-value.
[0020] Preferably, the processing system (202) configured to determine the probability based on the central tendency value and the variance value comprises a processing system (202) configured to calculate the degrees of freedom based on the number of instrument stiffness measurements that constitute the instrument stiffness.
[0021] Preferably, the processing system (202) configured to determine the probability based on the central tendency value and the variance value comprises a processing system (202) configured to calculate the standard error based on the standard deviation and the degrees of freedom.
[0022] Preferably, the standard error is calculated using the formula:
[0023]
number
[0024] Where: stddev pooled is the standard deviation of the instrument stiffness, n DOF is the degree of freedom. Preferably, the variance value is a pooled standard deviation including the standard deviation of the instrument stiffness measurement and the baseline instrument stiffness measurement.
[0025] Preferably, the probabilities include confidence intervals for the central tendency values.
[0026] Preferably, the confidence interval is compared to zero to detect that the central tendency value is not equal to the baseline value if the confidence interval does not include zero, and to detect that the central tendency value is equal to the baseline value if the confidence interval includes zero.
[0027] Preferably, the central tendency value is meter stiffness and the confidence interval for the meter stiffness is calculated using the formula:
[0028]
number
[0029] Where: CI is the confidence interval of the instrument stiffness, Stiffness mean is the gauge stiffness (204) retrieved from the storage system, CI range is the range of the confidence interval calculated based on the standard deviation of the instrument stiffness and the t-value . Preferably, the range of the confidence interval is calculated using the formula:
[0030]
number
[0031] Where: stddev pooled is the standard deviation of the instrument stiffness, t student,99,8 is determined from the significance level and the number of instrument stiffness measurements that make up the instrument stiffness. The t-value was calculated based on the degrees of freedom given.
[0032] Preferably, the processing system (202) is further configured to set a bias deadband, such that if the central tendency value is less than the bias deadband, the meter stiffness is not detected as differing from the central tendency value.
[0033] According to one aspect, a method for detecting and identifying changes in a vibrometer includes determining a first stiffness change associated with a first position of a conduit of the vibrometer, determining a second stiffness change associated with a second position of the conduit of the vibrometer, and determining a condition of the conduit based on the first stiffness change and the second stiffness change.
[0034] Preferably, the first location is the location of a left pickoff sensor on the conduit of the vibrometer and the second location is the location of a right pickoff sensor on the conduit of the vibrometer.
[0035] Preferably, the first stiffness change represents a physical stiffness change of the conduit between the driver and a first position of the conduit, and the second stiffness change represents a physical stiffness change of the conduit between the driver and a second position of the conduit.
[0036] Preferably, the method further comprises determining stiffness symmetry of the conduit.
[0037] Preferably, the stiffness symmetry of the conduit is determined based on the first stiffness variation and the second stiffness variation.
[0038] Preferably, the method further comprises providing an alarm based on the determination of the condition of the conduit.
[0039] Preferably, the condition of the conduit includes at least one of erosion, corrosion, damage, and coating of the conduit of the vibrometer.
[0040] According to one aspect, a method for detecting and identifying changes in a conduit of a vibrometer includes: The method includes obtaining a central tendency value of the meter validation parameter and a variance value of the meter validation parameter from a storage device in the electronic device, and determining a probability based on the central tendency value and the variance value to determine whether the central tendency value differs from a baseline value.
[0041] Preferably, determining the probability based on the central tendency value and the variance value comprises calculating a t-value and calculating the probability using the t-value.
[0042] Preferably, determining the probability based on the central tendency value and the variance value includes calculating the degrees of freedom based on the number of instrument stiffness measurements that make up the instrument stiffness.
[0043] Preferably, determining the probability based on the instrument stiffness includes calculating a standard error based on the standard deviation and the degrees of freedom.
[0044] Preferably, the standard error is calculated using the formula:
[0045]
number
[0046] Where: stddev pooled is the standard deviation of the instrument stiffness, n DOF is the degree of freedom. Preferably, the variance value is a pooled standard deviation including the standard deviation of the instrument stiffness measurement and the baseline instrument stiffness measurement.
[0047] Preferably, the probabilities include confidence intervals for the central tendency values.
[0048] Preferably, the confidence interval is compared to zero to detect that the central tendency value is not equal to the baseline value if the confidence interval does not include zero, and to detect that the central tendency value is equal to the baseline value if the confidence interval includes zero.
[0049] Preferably, the central tendency value is meter stiffness and the confidence interval for the meter stiffness is calculated using the formula:
[0050]
number
[0051] Where: CI is the confidence interval of the instrument stiffness, Stiffness mean is the gauge stiffness (204) retrieved from the storage system, CI range is the range of the confidence interval calculated based on the standard deviation of the instrument stiffness and the t-value. Preferably, the range of the confidence interval is calculated using the formula:
[0052]
number
[0053] Where: stddev pooled is the standard deviation of the instrument stiffness, t student,99,8 is determined from the significance level and the number of instrument stiffness measurements that make up the instrument stiffness. The t-value was calculated based on the degrees of freedom given.
[0054] Preferably, the method further comprises setting a bias deadband, where if the central tendency value is less than the bias deadband, the central tendency value is not detected as differing from the baseline value.
[0055] According to one aspect, a method for detecting and identifying changes in a vibrometer includes determining a first stiffness change associated with a first position of a conduit of the vibrometer, determining a second stiffness change associated with a second position of the conduit of the vibrometer, and determining a condition of the conduit based on the first stiffness change and the second stiffness change.
[0056] Preferably, the first location is the location of a left pickoff sensor on the conduit of the vibrometer and the second location is the location of a right pickoff sensor on the conduit of the vibrometer.
[0057] Preferably, the first stiffness change represents a physical stiffness change of the conduit between the driver and a first position of the conduit, and the second stiffness change represents a physical stiffness change of the conduit between the driver and a second position of the conduit.
[0058] Preferably, the method further comprises determining stiffness symmetry of the conduit.
[0059] Preferably, the stiffness symmetry of the conduit is determined based on the first stiffness variation and the second stiffness variation.
[0060] Preferably, the method further comprises providing an alarm based on the determination of the condition of the conduit.
[0061] Preferably, the condition of the conduit includes at least one of erosion, corrosion, damage, and coating of the conduit of the vibrometer.
[0062] Preferably, the method further comprises determining a condition of the conduit based on at least one of the residual flexibility, damping, and mass of the conduit.
[0063] Like reference numbers refer to like elements in all drawings. It should be understood that the drawings are not necessarily to scale. [Brief description of the drawings]
[0064] [Figure 1] FIG. 1 is a diagram showing a vibrometer 5. [Diagram 2] FIG. 2 is meter electronics 20 for detecting and identifying changes in a vibrometer, according to one embodiment. [Figure 3a] FIG. 3a is a graph 300a illustrating stiffness changes determined during multiple instrument validation runs. [Figure 3b] FIG. 3b is a graph 300b illustrating stiffness symmetry variations determined during multiple instrument validation runs. [Figure 4a] FIG. 4a is a graph 400a illustrating stiffness change data points determined during multiple instrument validation runs, with a probability distribution assigned to each data point. [Figure 4b] FIG. 4b is a graph 400b illustrating stiffness symmetry variation data points determined during multiple instrument validation runs, with a probability distribution assigned to each data point. [Figure 5a] FIG. 5a is a graph 500a illustrating stiffness change data points determined during multiple instrument validation runs, with a probability assigned to each data point. [Figure 5b] FIG. 5b is a graph 500b illustrating stiffness symmetry variation data points determined during multiple instrument validation runs, with a probability assigned to each data point. [Figure 6] FIG. 6 is a method 600 for detecting and identifying changes in a vibrometer, according to one embodiment. [Figure 7] FIG. 7 is a method 700 for detecting and identifying changes in a vibrometer, according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0065] 1-7 and the following description provide specific examples to teach those skilled in the art how to make and use the best mode of embodiments for detecting and identifying changes in a vibrometer. For the purpose of teaching the principles of the invention, some conventional aspects have been simplified or omitted. Those skilled in the art will appreciate variations from these examples that are within the scope of the description. Those skilled in the art will appreciate that the features described below can be combined in various ways to form multiple variations for detecting and identifying changes in a vibrometer. As a result, the embodiments described below are not limited to the specific examples described below, but only by the claims and their equivalents.
[0066] By utilizing statistics, changes in the vibrometer can be accurately detected. Statistics have not been used in meter electronics before because the meter electronics has limited computing power, for example, compared to a computer workstation running statistical software. The statistical methods used herein utilize data available in the registers of the processing system of the meter electronics to enable the embedded code to calculate the probability that there is no change in the vibrometer. By calculating this probability, the null hypothesis that no change has occurred can be disproved, thereby indicating that a change has occurred in the vibrometer. Since the probability is calculated by the meter electronics, this probability can be updated as meter validation is performed, even in cases where computing resources are limited. Thus, changes that would not be detected by comparing stiffness changes, for example, to a predefined limit, can be detected. In addition, accurate detection of changes can prevent false alarms.
[0067] The changes in the vibrometer may be identified by determining a condition, such as erosion, corrosion, damage, etc., of the vibrometer conduit based on a first stiffness change associated with a first position of the conduit and a second stiffness change associated with a second position of the conduit. For example, the condition may be determined based on whether the first and second stiffness changes indicate an increase or decrease in stiffness. Additionally, symmetry of the first and second stiffness changes may be used to determine the condition. In one example, if the first stiffness change indicates a decrease and the second stiffness change indicates an increase, and the stiffness symmetry is considered "right low," the determined condition may be erosion or corrosion of the vibrometer conduit.
[0068] FIG 1 illustrates a vibrometer 5. As shown in FIG 1, the vibrometer 5 includes a meter assembly 10 and meter electronics 20. The meter assembly 10 responds to the mass flow rate and density of a process material. The meter electronics 20 is connected to the meter assembly 10 via a sensor signal 100 and provides density, mass flow, and temperature information on path 26, as well as other information.
[0069] The meter assembly 10 includes a pair of manifolds 150 and 150' and a flange neck 1 The conduits 130, 130' include flanges 103, 103' having ends 10 and 110', a pair of parallel conduits 130, 130', a driver 180, a resistance temperature detector (RTD) 190, and a pair of pick-off sensors 170l and 170r. The conduits 130, 130' have two essentially straight inlet sections 131, 131' and outlet sections 134, 134' that converge towards each other at the conduit mounting blocks 120, 120'. The conduits 130, 130' bend at two symmetrical locations along their length and are essentially parallel throughout their length. The brace bars 140, 140' serve to define axes W and W' about which each conduit 130, 130' oscillates. Sections 131, 131' and 134, 134' of conduits 130, 130' are fixedly attached to conduit mounting blocks 120 and 120' which are fixedly attached to manifolds 150 and 150'. This provides a continuous, closed material path through meter assembly 10.
[0070] When flanges 103 and 103' having holes 102 and 102' are connected via inlet end 104 and outlet end 104' to a process line (not shown) carrying the process material being measured, the material enters the meter at inlet end 104 through orifice 101 in flange 103 and is directed through manifold 150 to conduit mounting block 120 having surface 121. Within manifold 150, the material is split and directed through conduits 130, 130'. Upon exiting conduits 130, 130', the process material is recombined into a single stream within block 120' having surface 121' and manifold 150' before being directed to outlet end 104' connected to the process line (not shown) by flange 103' having holes 102'.
[0071] The conduits 130, 130' are selected to have substantially the same mass distribution, moment of inertia, and Young's modulus about bending axes WW and W'-W', respectively, and are appropriately mounted in the conduit mounting blocks 120, 120'. These bending axes pass through the brace bars 140, 140'. The Young's modulus of the conduits changes with temperature and this change affects flow and An RTD 190 is attached to the conduit 130' to continuously measure the temperature of the conduit 130' for affecting the density calculation. The temperature of the conduit 130', and therefore the voltage appearing across the RTD 190 for a given current passing through it, is governed by the temperature of the material passing through the conduit 130'. The temperature dependent voltage appearing across the RTD 190 is used by meter electronics 20 in a well known manner to compensate for changes in the modulus of elasticity of the conduits 130, 130' due to changes in the conduit temperature. The RTD 190 is connected to meter electronics 20 by leads 195.
[0072] Both conduits 130, 130' are driven in opposite directions about their respective bending axes W and W', in what is referred to as the first out-of-phase bending mode of the flowmeter, by a driver 180. This driver 180 may include any one of a number of well-known devices, such as a magnet attached to conduit 130' and an opposing coil attached to conduit 130 through which an alternating current is passed to vibrate both conduits 130, 130'. An appropriate drive signal is applied to driver 180 by meter electronics 20 via leads 185.
[0073] Meter electronics 20 receives the RTD temperature signal on lead 195 and the left and right sensor signals appearing on sensor line 100, which carry left and right sensor signals 165l, 165r, respectively. Meter electronics 20 generates a drive signal appearing on lead 185 to driver 180 to vibrate conduits 130, 130'. Meter electronics 20 processes the left and right sensor signals and the RTD signal to calculate the mass flow rate and density of material passing through meter assembly 10. This information, along with other information, is applied by meter electronics 20 as a signal on path 26.
[0074] Mass Flow Measurement
[0075]
number
[0076] can be generated according to the following formula:
[0077]
number
[0078] The Δt term includes an operationally derived (i.e., measured) time delay value that includes the time delay that exists between pickoff sensor signals, for example, due to the Coriolis effect associated with the mass flow rate through the vibrometer 5. The measured Δt term ultimately determines the mass flow rate of the flow material through the vibrometer 5. The Δt0 term includes a time delay at zero flow calibration constant. The Δt0 term is typically determined at the factory and programmed into the vibrometer 5. The time delay at zero flow rate, Δt0, does not change even when flow conditions are changing. The flow calibration factor FCF is proportional to the stiffness of the flow meter.
[0079] A concern is that the conduits may change over time, and the original factory calibration may change over time as the conduits 130, 130' corrode, erode, or otherwise change. As a result, the stiffness of the conduits 130, 130' may change from an initial representative stiffness value (or original measured stiffness value) over the life of the vibrometer 5. As described below, meter validation can detect such changes in the stiffness of the conduits 130, 130'.
[0080] 2 is a meter electronics 20 for detecting and identifying changes in a vibration meter, according to one embodiment. The meter electronics 20 can include an interface 201 and a processing system 202. The meter electronics 20 receives, for example, a vibration response from the meter assembly 10. The meter electronics 20 processes the vibration response to obtain flow characteristics of a flow material flowing through the meter assembly 10.
[0081] As previously mentioned, the flow calibration factor FCF reflects the material properties and cross-sectional properties of the flow tube. The mass flow rate of the flow material through the meter is determined by multiplying the measured time delay (or phase difference / frequency) by the flow calibration factor FCF. The flow calibration factor FCF can be related to the stiffness characteristics of the meter assembly. If the stiffness characteristics of the meter assembly change, the flow calibration factor FCF also changes. Thus, the stiffness change of the flow meter will affect the flow measurements produced by the flow meter. Affects the accuracy of the set value.
[0082] The interface 201 receives the vibration response from one of the pickoff sensors 170l, 170r via the sensor signal 100 of FIG. 1. The interface 201 may perform any necessary or desired signal conditioning, such as, for example, any manner of formatting, amplification, buffering, etc. Alternatively, some or all of the signal conditioning may be performed in the processing system 202. Additionally, the interface 201 may enable communication between the meter electronics 20 and an external device. The interface 201 may be capable of any manner of electronic, optical, or wireless communication. The interface 201 may provide information based on the vibration response.
[0083] In one embodiment, the interface 201 is coupled to a digitizer (not shown) and the sensor signal comprises an analog sensor signal. The digitizer samples and digitizes the analog vibration response to generate a digital vibration response.
[0084] The processing system 202 performs the operations of the meter electronics 20 and processes the flow measurements from the meter assembly 10. The processing system 202 executes one or more processing routines, thereby processing the flow measurements to generate one or more flow characteristics. The processing system 202 is communicatively coupled to the interface 201 and is configured to receive information from the interface 201.
[0085] Processing system 202 may comprise a general purpose computer, a microprocessing system, a logic circuit, or some other general purpose or customized processing device. Additionally or alternatively, processing system 202 may be distributed among multiple processing devices. Processing system 202 may also include any type of integrated or stand-alone electronic storage medium, such as storage system 204.
[0086] The storage system 204 can store flow meter parameters and data, software routines, constant values, and variable values. In one embodiment, the storage system 204 includes routines executed by the processing system 202, such as operation routines 210 and validation 220 of the vibrometer 5. The storage system can also store statistics such as standard deviations, confidence intervals, etc.
[0087] The storage system 204 can store the baseline instrument stiffness 230. The baseline instrument stiffness 230 may be determined during manufacturing or calibration of the vibrometer 5, or during a previous recalibration. For example, the baseline instrument stiffness 230 can be determined by validation 220 before the vibrometer 5 is installed in the field. The baseline instrument stiffness 230 represents the stiffness of the conduit 130, 130' before any changes such as erosion / corrosion, damage (e.g., freezing, overpressure, etc.), coating, etc. occur. The baseline instrument stiffness 230 may be an average value of multiple baseline instrument stiffness measurements. Thus, the baseline instrument stiffness 230 may have an associated variance characteristic, as described in more detail below, and the baseline instrument stiffness measurements may vary. The greater the change in the baseline instrument stiffness measurements, the greater the variance.
[0088] The storage system 204 can store the meter stiffness 232. The meter stiffness 232 includes a stiffness value determined from a vibration response generated during operation of the vibrometer 5. The meter stiffness 232 may be generated to verify proper operation of the vibrometer 5. The meter stiffness 232 may be generated for a verification process, where the meter stiffness 232 serves the purpose of verifying proper and accurate operation of the vibrometer 5. Similar to the baseline meter stiffness 230, the meter stiffness 232 may be an average value of multiple meter stiffness measurements. Thus, the meter stiffness 232 may have an associated variance characteristic, as discussed in more detail below, where the meter stiffness measurements may vary. The greater the change in the meter stiffness measurements, the greater the variance characteristic.
[0089] The storage system 204 may store the stiffness change 234. The stiffness change 234 may be a value determined by comparing the baseline meter stiffness 230 and the meter stiffness 232. For example, the stiffness change 234 may be the difference between the baseline meter stiffness 230 and the meter stiffness 232. In this example, a negative number may indicate that the stiffness of the conduit 130, 130' has increased since being installed in the field. A positive number may indicate that the physical stiffness of the conduit 130, 130' has decreased since the baseline meter stiffness 230 was determined.
[0090] As can be appreciated, the comparison may be performed in a variety of ways. For example, stiffness change 234 may be the difference between meter stiffness 232 and baseline meter stiffness 230. Thus, an increase in stiffness results in a positive number and a decrease in stiffness results in a negative number. Additionally or alternatively, ratios may be used that use values derived from baseline meter stiffness 230 and / or meter stiffness 232 or values related thereto, such as other values such as conduit geometry, dimensions, etc.
[0091] If the meter stiffness 232 is substantially the same as the baseline meter stiffness 230, it can be determined that the vibrometer 5, and more specifically, the conduit 130, 130', may be relatively unchanged since it was manufactured, calibrated, or since the vibrometer 5 was last recalibrated. Alternatively, if the meter stiffness 232 differs significantly from the baseline meter stiffness 230, it can be determined that the conduit 130, 130' has deteriorated and may not be operating accurately and reliably, such as the conduit 130, 130' has changed due to erosion, corrosion, damage (e.g., freezing, overpressure, etc.), coating, or other conditions.
[0092] As described above, the baseline meter stiffness 230 and meter stiffness 232 are determined for both the left and right pickoff sensors 170l, 170r. That is, the baseline meter stiffness 230 and meter stiffness 232 are proportional to the stiffness of the conduits 130, 130' between the left and right pickoff sensors 170l, 170r. As a result, different conditions of the conduits 130, 130' may cause similar stiffness changes 234. For example, erosion, corrosion, and / or damage to the conduits 130, 130' may result in a similar decrease in physical stiffness, which may be indicated by a negative or "low" stiffness change 234. Thus, when relying solely on the stiffness change 234, a particular condition of the conduits 130, 130' may not be ascertained.
[0093] However, the left pickoff sensor 170l and the right pickoff sensor 170r may each have their own associated stiffness value. More specifically, as described above, the driver 180 applies a force to the conduits 130, 130' and the pickoff sensors 170l, 170r measure the resulting deflection. The amount of deflection of the conduits 130, 130' at the location of the pickoff sensors 170l, 170r is proportional to the stiffness of the conduits 130, 130' between the driver 180 and the pickoff sensors 170l, 170r.
[0094] Thus, the stiffness associated with the left pickoff sensor 170l is proportional to the physical stiffness of the conduit 130, 130' between the driver 180 and the left pickoff sensor 170l, and the stiffness associated with the right pickoff sensor 170r is proportional to the physical stiffness of the conduit 130, 130' between the driver 180 and the right pickoff sensor 170r. Thus, if there is erosion, corrosion, damage, coating, etc. between the driver 180 and, for example, the right pickoff sensor 170r, the stiffness associated with the right pickoff sensor 170r may decrease, but the stiffness associated with the left pickoff sensor 170l may not change. To track changes, the storage system 204 may also include stiffness values associated with the left and right pickoff sensors 170l, 170r.
[0095] 2, the storage system 204 includes a baseline LPO stiffness 240 proportional to the physical stiffness of the conduit 130, 130' between the driver 180 and the location of the left pickoff sensor 170l on the conduit 130, 130'. Similarly, the storage system 204 also includes a baseline RPO stiffness 250 proportional to the physical stiffness of the conduit 130, 130' between the driver 180 and the location of the right pickoff sensor 170r on the conduit 130, 130'. The baseline LPO and RPO stiffnesses 240, 250 may be measured before the vibrometer 5 is installed in the field, such as during manufacture or calibration of the vibrometer 5, or during a prior recalibration. This may be determined by evidence 220.
[0096] The storage system 204 also includes an LPO stiffness 242 and an RPO stiffness 252. The LPO stiffness 242 is proportional to the physical stiffness of the conduits 130, 130' between the driver 180 and the location of the left pickoff sensor 170l, but after the baseline LPO stiffness 240 has been determined. Similarly, the RPO stiffness 252 is proportional to the physical stiffness of the conduits 130, 130' between the driver 180 and the location of the right pickoff sensor 170r, but after the baseline RPO stiffness 250 has been determined.
[0097] 2, the storage system 204 further includes an LPO stiffness change 244 and an RPO stiffness change 254. The LPO and RPO stiffness change 244, 254 is proportional to the difference between the baseline LPO and RPO stiffness 240, 250 and the LPO and RPO stiffness 242, 252. For example, a negative LPO stiffness change 244 may indicate that the physical stiffness of the conduit 130, 130' between the driver 180 and the left pickoff sensor 170l has increased. A positive LPO stiffness change 244 may indicate that the physical stiffness of the conduit 130, 130' between the driver 180 and the left pickoff sensor 170l has decreased since the baseline LPO stiffness 240 was determined. Alternatively, the LPO and RPO stiffness change 244, 254 may be the difference between the LPO and RPO stiffness 242, 252 and the baseline LPO and RPO stiffness 240, 250. Thus, for example, a positive LPO stiffness change 244 may indicate that the physical stiffness of the conduit 130, 130' between the driver 180 and the left pickoff sensor 170l has increased since the baseline LPO stiffness 240 was determined. Although the LPO and RPO stiffness changes 244, 254 are described as being determined by difference, any values derived from or related to the baseline LPO and RPO stiffnesses 240, 250 and LPO and RPO stiffnesses 242, 252 may be used, such as ratios of stiffness values to other values such as conduit geometry, dimensions, etc. The LPO and RPO stiffness changes 244, 254 may be expressed in any suitable units, such as integers, ratios, percentages, etc.
[0098] An increase or decrease in the physical stiffness associated with the left and right pickoff sensors 170l, 170r may be indicative of an underlying condition of the conduit 130, 130' causing the physical stiffness change. For example, erosion of the inner wall of the conduit 130, 130' may decrease the physical stiffness of the conduit 130, 130'. In particular, erosion of the inner wall of the conduit 130, 130' between the left pickoff sensor 170l and the driver 180, for example, may decrease the physical stiffness of the conduit 130, 130' between the left pickoff sensor 170l and the driver 180. Conversely, an increase in stiffness may be indicative of, for example, the formation of a coating on the inner wall.
[0099] Additionally, a relative increase or decrease in the physical stiffness of the conduits 130, 130' between the driver 180 and the left pickoff sensor 170l and the physical stiffness of the conduits 130, 130' between the driver 180 and the right pickoff sensor 170r may further indicate the underlying condition of the conduits 130, 130' that causes the physical stiffness change. This relative increase or decrease in physical stiffness may be indicated by stiffness symmetry 260 of the storage system 204.
[0100] Stiffness symmetry 260 may be any suitable value indicating, for example, the relative value of LPO stiffness change 244 and RPO stiffness change 254. For example, LPO stiffness change 244 and RPO stiffness change 254 may indicate that the physical stiffness of the conduits 130, 130' associated with the left and right pickoff sensors 170l, 170r have both increased, but that, for example, the physical stiffness associated with the left pickoff sensor 170l has increased more than the physical stiffness associated with the right pickoff sensor 170r. In one example, stiffness symmetry 260 may be a function of the relative stiffness of the conduits 130, 130' associated with the left and right pickoff sensors 170l, 170r. It can be expressed as a percentage and can be determined by:
[0101]
number
[0102] Where: SMV Stiffness, LPO% is the LPO stiffness change, in this example expressed as a percentage change. 4, SMV Stiffness, RPO% is the RPO stiffness change expressed as a percentage change in this example. The answer is 4.
[0103] Stiffness change 234, LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260 may be any suitable values, such as, for example, a value directly proportional to the property being measured, an intermediate value representing the physical stiffness, a value indicating whether there has been an increase or decrease in the physical stiffness, etc. For example, LPO stiffness change 244 may be a positive or negative value proportional to the stiffness change. The processing system 202 may further process these values to generate a toggle indicator in which only an increase or decrease in the physical stiffness of the conduit 130, 130' between the driver 180 and the left pickoff sensor 170l is indicated. These values and / or toggle indicators may be used to determine an underlying change in the conduit 130, 130', as shown in the truth table below.
[0104] [Table 1]
[0105] As can be seen, a combination of LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260 can be used to distinguish between various possible changes in the conduits 130, 130'. For example, in both cases J and N, the stiffness symmetry 260 value is "low" and the RPO stiffness change 254 is "low." However, in case J, the LPO stiffness change 244 is "low," while in case N, the LPO stiffness change 244 is "high." Case J is shown as a possible erosion / corrosion of the conduits 130, 130' and case N is shown as a possible coating of the conduits 130, 130'.
[0106] Although the above table utilizes LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260 to determine the condition of the conduit 130, 130', any suitable means, such as alternative tables, logic, objects, relationships, circuits, processors, routines, etc., may be used to determine the condition within the conduit. For example, with reference to the meter electronics 20 described with reference to FIG. 2, only LPO stiffness change 244 and RPO stiffness change 254 may be utilized to determine the condition of the conduit 130, 130'. However, as can be appreciated, utilizing stiffness symmetry 260 may allow for a more specific determination of the condition of the conduit 130, 130'.
[0107] Additionally or alternatively, the actual values of LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260 may be used in place of the toggle indicators to determine the condition of the conduit. For example, the condition determined by the above table may be supplemented by a further step of determining that case J is likely to be corrosion rather than erosion, e.g., if stiffness symmetry 260 is relatively small "right low." " The rigid symmetry 260 may be due to the more uniform nature of corrosion compared to erosion, which may be more prevalent at the entrance to the conduit.
[0108] While the above discussion has been related to instrument stiffness, other instrument validation parameters may be used in addition or as an alternative. For example, residual flexibility may be compared to a baseline residual flexibility. Residual flexibility may be defined as a portion of the frequency response associated with one vibration mode at the resonant frequency of another vibration mode. For example, the frequency responses of various vibration modes (e.g., bending, torsion, etc.) may be characterized as frequency response functions (e.g., amplitude response versus frequency). The frequency response functions are typically centered on the resonant frequency of a given vibration mode, with the amplitude decreasing diagonally in proportion to the distance from the resonant frequency. For example, a first bending mode (e.g., a primary out-of-phase bending mode) having two nodes located on the brace bar may have a first bending mode resonant frequency ω1. A second bending mode having four nodes may have a second bending mode resonant frequency higher than the first bending mode resonant frequency ω1. The frequency response function of the second bending mode may overlap the first bending mode resonant frequency ω1. Therefore, the residual flexibility of the first bending mode caused by the second bending mode is part of the frequency response function of the second bending mode at the resonant frequency ω1 of the first bending mode. As can be appreciated, the occurrence of erosion, corrosion, damage, coatings, etc. can change this residual flexibility value for a given mode, as it changes the frequency response of each vibration mode. Thus, the residual flexibility can also be used to identify changes in the vibrometer.
[0109] Damping can also be used. For example, in meter verification, measured damping values can be compared to baseline damping values. Damping can be useful in detecting coatings because it is not affected by erosion or corrosion.
[0110] Similarly, the mass associated with the left or right pickoff sensor 170l, 170r can be compared to a baseline mass associated with the left or right pickoff sensor 170l, 170r. In one example, a predicted mass may be used. In one example, a predicted mass based on calibrated air and water mass values, and the measured or known density of the process fluid, may be calculated using the following formula:
[0111]
number
[0112] It can be calculated using Where: m expected is the expected mass, i.e., the mass that would be measured if no changes occurred in the vibrometer. is the mass, m factory,air is the measured mass when the vibrometer is filled with air, ρ air is the density of air, ρ water is the density value of water, ρ kwown is the density of the material being measured. Estimated mass m expected Using the formula: The deviation can be calculated,
[0113]
number
[0114] Where: m measured is the mass measured during instrument verification, m Deviation is the predicted mass m expected Mass deviation of the measured mass from m measured As can be appreciated, erosion, corrosion, damage, coatings, etc. can affect the mass of the vibrometer conduit, and therefore the predicted mass can be used to detect changes in the vibrometer by comparing the measured mass to the predicted mass.
[0115] As discussed above, the geometry of the conduit may also be considered in determining the condition of the conduit. For example, a U-shaped tube may be more susceptible to erosion than corrosion at certain locations within the conduit, as compared to, for example, a straight tube. Additionally or alternatively, some process / conduit combinations may be more susceptible to certain conditions. For example, the conduit 130, 130' may be more susceptible to damage in a cryogenic process using nitrogen, as compared to a high temperature process using a corrosive material. Thus, the LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260, or methods of using these values, may include other values, such as factors related to the geometry, structure, dimensions, process variables, etc. of the conduit.
[0116] 2, the storage system 204 may also store stiffness standard deviation 236, LPO stiffness standard deviation 246, and RPO stiffness standard deviation 256. These values may be determined from instrument stiffness measurements, including, for example, baseline instrument stiffness 230 and instrument stiffness 232. For example, stiffness standard deviation 236 may be a pooled standard deviation. Thus, stiffness standard deviation 236 is a measure of how much instrument stiffness 232 has changed, including the instrument stiffness measurements that make up baseline instrument stiffness 230. LPO stiffness standard deviation 246 and RPO stiffness standard deviation 256 may also be pooled standard deviations.
[0117] The example shown in Figure 2 uses the stiffness standard deviation, but in the instrument verification parameter data Other measures of variation and dispersion may be employed. For example, dispersion may be used instead of standard deviation. That is, stiffness standard deviation 236, LPO stiffness standard deviation 246, and RPO stiffness standard deviation 256 are dispersion values of exemplary instrument validation parameters. Additionally or alternatively, other measures of central tendency may be used instead of the mean values that may be used for baseline instrument stiffness 230 and instrument stiffness 232. Thus, baseline instrument stiffness 230 and instrument stiffness 232 are central tendency values of exemplary instrument validation parameters.
[0118] The storage device may also store other statistics such as a confidence interval 270. As described in more detail below, the confidence interval 270 may be calculated based on a t-value 272, a significance level 274, and degrees of freedom 276. The significance level 274 may be a scalar value, for example, set by validation 220. The significance level 274 may be defined as the probability of rejecting the null hypothesis when the hypothesis is actually true (e.g., detecting a change in the vibrometer when no change occurred), and is typically a small value such as 1% or 0.01. The degrees of freedom 276 are calculated, for example, from the number of samples used to determine the stiffness standard deviation 236. A bias deadband 278 is also shown, which is a scalar value that may also be set by validation 220 to ensure that a bias in the vibrometer does not induce a false flag.
[0119] The confidence interval 270 can detect small changes in the physical stiffness of the vibrometer 5 while also reducing the number of false alarms, for example compared to predetermined limits previously used in meter validation. Furthermore, the confidence interval 270 can be calculated using relatively simple mathematical operations, thereby allowing the processing system 202 to use robust statistical techniques using validation 220 with a relatively simple embedded code.
[0120] Predefined alarm limits 3a and 3b show graphs 300a, 300b illustrating stiffness change and stiffness symmetry variation determined during multiple instrument validation runs. As shown, the graphs 300a, 300b include a run number axis 310a, 310b. The run number axis 310a, 310b ranges from 0 to 600 and indicates the instrument validation run number. For example, run number "100" indicates the 100th instrument validation run out of 600 instrument validation runs. The graph 300a also includes a percent change on the stiffness axis 320a, which may be, for example, a percentage representation of the LPO stiffness change 244 and the RPO stiffness change 254. The graph 300b includes a percent stiffness difference axis 320b, which may be, for example, a percentage representation of stiffness symmetry 260. For example, a stiffness difference of 0 percent means, for example, that the LPO stiffness change 244 is equal to the RPO stiffness change 254. Additionally, the graphs 300a and 300b each show stiffness change data 330a and stiffness difference data 330b.
[0121] The stiffness change data 330a and stiffness difference data 330b are comprised of data points determined in groups of runs for various flow material / flow rate configurations where the coating is present in the conduit. More specifically, there are four groups of data that are identifiable from the stiffness difference data 330b. The first two groups may be based on high and low water flows. The latter two groups may be based on high and low air flows.
[0122] Graph 300a shown in FIG. 3a includes stiffness change data 330a comprised of data points representing stiffness change for a given meter validation run. As can be seen, stiffness change data 330a ranges from about -0.3% to about 2.0%. As can be seen, this appears to indicate that stiffness is changing. However, if an alarm limit is set at, for example, 4%, an alarm may not be generated.
[0123] Graph 300b shown in FIG. 3b illustrates stiffness differences, e.g., LPO stiffness change 244 and RPO stiffness 3 includes stiffness difference data 330b comprised of data points representing stiffness symmetry changes 254. As can be seen, stiffness difference data 330b ranges from about -0.4% to about 0.6%. As can also be seen, stiffness difference data 330 includes sporadic data points that do not follow a discernible trend. In addition, stiffness difference data 330b suggests that stiffness symmetry values may be affected by materials within the conduit.
[0124] Graphs 300a, 300b show that if the alarm limit or range is larger than the stiffness change associated with the vibrometer change, no alarm may be issued. Furthermore, if the alarm limit is smaller than the sporadic data points, false alarms may occur. The following addresses this issue by removing the limit and employing statistics that can be implemented in an embedded system.
[0125] Statistics for embed codes Statistical methods that calculate the probability of an outcome can be used to detect changes in the vibration meter, but due to their complexity cannot be implemented in the meter electronics 20. For example, P and T statistics may be used to test whether a null hypothesis is met for a given set of data. Rejecting the null hypothesis is not determining whether a condition exists in the vibration meter, but rather that the absence of a condition is false. In the case of meter validation, the null hypothesis may be defined as "the current meter validation results have the same mean as the baseline meter validation results." If this null hypothesis is disproved, it can be inferred that the current results do not have the same mean as the baseline meter validation results due to changes in the vibration meter.
[0126] As an example, in a t-test, the t-value can be calculated using the formula:
[0127]
number
[0128] Where: μ0 is some specified value,
[0129]
number
[0130] is the sample mean, s is the sample standard deviation, n is the sample size. In the context of instrument validation, μ0 is a reference instrument validation value, such as a baseline stiffness value. The instrument validation measurements are used to determine the average value of the samples to compare with the reference instrument validation value.
[0131]
number
[0132] and calculate the sample standard deviation s. The number of instrument validation measurements is the sample size n. A t-test typically also includes the degrees of freedom, which in equation [1] above is defined as n-1.
[0133] As discussed above, a t-test can be used to test the null hypothesis, which in the case of instrument validation is the sample mean
[0134]
number
[0135] The null hypothesis may be defined as whether the current instrument validation results are equal to the baseline instrument validation value. To test the null hypothesis, a P-value may be calculated using the known distribution of t-values. To test the null hypothesis, the P-value is compared to a significance level α. The significance level α is typically set to a small value, such as 0.01, 0.05, or 0.10. If the P-value is equal to or less than the significance level α, the null hypothesis is rejected in favor of the alternative hypothesis. Since the null hypothesis is defined as "the current instrument validation results have the same mean as the baseline instrument validation results," the alternative hypothesis is that the current instrument validation does not have the same mean and therefore, a change has occurred in the instrument.
[0136] However, P values are difficult to calculate with limited computational resources. For example, P values can be calculated on a computer workstation with an operating system and statistical software, but cannot be easily calculated on an embedded system. The meter electronics 20 described above may be an embedded system with limited computational resources. In addition, the ability to reject the null hypothesis in-situ or in real-time in the meter electronics can prevent the meter electronics 20 from sending false alarms while still accurately detecting changes in the conduit 130, 130', which is a significant improvement over the use of predetermined alarm limits.
[0137] For this purpose, instead of P-values, confidence intervals are used that leverage the limited computing resources of the meter electronics 20. As a result, the confidence intervals can be calculated using code embedded in the meter electronics 20. For example, the meter electronics 20 can have a current stiffness value and a stiffness standard deviation value stored in two registers. As can be seen, the above-mentioned t-values can be calculated using the current stiffness values by using the significance level α and the degrees of freedom. As an example, the significance level α may be set to 0.01, which is a 99% confidence level. The number of meter validation tests may be set as 5. Thus, the pooled degrees of freedom is determined to be 2·(5-1)=8. The two-tailed Student's t-value can be calculated from the significance level α and the pooled degrees of freedom using the Student's t-value function as follows:
[0138]
number
[0139] A pooled standard deviation of the stiffness values associated with the left and right pickoff sensors 170l, 170r may also be used. In the general case, calculating the pooled standard deviation may be complicated. However, because the meter electronics 20 stores the measured stiffness standard deviations in a register, the pooled standard deviation may simply be a stored standard deviation, such as stiffness standard deviation 236 discussed above. Calculating the Pooled Standard Error It can also be defined as follows:
[0140]
number
[0141]
number
[0142]
number
[0143] Using the standard error and t-value determined above, the confidence interval range can be calculated as follows:
[0144]
number
[0145]
number
[0146] Finally, the confidence interval can be calculated using the stiffness mean value and the range of the confidence interval, which is shown below:
[0147]
number
[0148] The confidence interval can be used to test the null hypothesis by determining if the confidence interval contains 0.0. If the confidence interval contains 0.0, the null hypothesis is not rejected and the meter validation passes. If the confidence interval does not contain 0.0, the null hypothesis is rejected and an meter validation failure may be sent.
[0149] As can be appreciated, by using confidence intervals instead of P values, where the meter electronics 20 stores stiffness values and stiffness standard deviation values, the calculations are relatively simple and can be performed using embedded code. For example, The meter electronics 20 may be capable of calculating confidence intervals for performing in-situ or real-time statistical analysis. As can also be appreciated, the confidence intervals may be used to test null hypotheses at a desired confidence level.
[0150] In addition to the confidence interval, a bias deadband can be defined around zero to account for bias in the meter stiffness measurements, which may be due to mounting, density, temperature gradients, or other conditions of the vibrometer that may affect the meter validation measurements. This bias deadband in the t-test is a value around zero where a small bias with small variations that would otherwise cause the hypothesis to be rejected by the confidence interval check does not reject the hypothesis. Thus, this bias deadband can be set to a value that reduces the number of false alarms sent by the meter electronics 20.
[0151] In the example of a confidence interval compared to zero, the bias deadband is in the range around zero, and if zero is not within the confidence interval, but part of the bias deadband is within the confidence interval, then the null hypothesis is not rejected. Mathematically, this test can be expressed as whether the mean instrument stiffness value is less than the bias deadband. Or, using the nomenclature above,
[0152]
number
[0153] where db bias is the bias dead zone, and we cannot reject the null hypothesis.
[0154] The bias deadband can be implemented alone or in combination with other deadbands. For example, the bias deadband can be implemented in combination with a drift deadband. In one example, the drift deadband is variation =db bias / t student,99,8 can be determined from, where db variation is the fluctuation deadband. The fluctuation deadband is compared with the standard deviation of the instrument stiffness to determine the null hypothesis. It can be determined whether the theory needs to be rejected. In one example, the bias deadband may be compared as described above, and the fluctuation deadband can be compared to the standard deviation of the meter stiffness as follows:
[0155]
number
[0156] If , and s <db variation If , we cannot reject the null hypothesis. can be used after the null hypothesis has been rejected by a confidence interval check, or
[0157]
number
[0158] If , and s <db variation If , the average instrument stiffness
[0159]
number
[0160] is set to zero and the meter stiffness variation is equal to the variation deadband. Thus, when a confidence interval check is performed, the null hypothesis may not be rejected due to bias in the meter stiffness measurements.
[0161] 4a and 4b show graphs 400a, 400b illustrating stiffness change and stiffness symmetry variation data points determined during multiple meter validation runs, with a probability distribution assigned to each data point. As shown, the graphs 400a, 400b include a run number axis 410a, 410b. The run number axis 410a, 410b ranges from 0 to 600 and indicates the run number of the meter validation. The graph 400a also includes a percent change in stiffness axis 420a, which is, for example, a percentage representation of the LPO stiffness change 244 and the RPO stiffness change 254. The graph 400b includes a percent stiffness difference axis 420b, which is, for example, a percentage representation of stiffness symmetry 260.
[0162] 5a and 5b are graphs 500a, 500b illustrating stiffness change and stiffness symmetry variation data points determined during multiple meter validation runs, with a probability assigned to each data point. As shown, the graphs 500a, 500b include a run number axis 510a, 510b. The run number axis 510a, 510b ranges from 0 to 140 and indicates the run number for the meter validation. The graph 500a also includes a percent change in stiffness axis 520a, which is, for example, a percentage representation of the LPO stiffness change 244 and the RPO stiffness change 254. The graph 500b includes a percent stiffness difference axis 520b, which is, for example, a percentage representation of stiffness symmetry 260.
[0163] The graphs 400a, 500a include stiffness deviation plots 430a, 530a comprised of a number of data points representing stiffness deviations, which may be stiffness changes 234 stored in the storage system 204 of the instrument stiffness. The graphs 400b, 500b include stiffness symmetry plots 430b, 530b comprised of data points representing stiffness symmetry changes. Also shown are change indication plots 440a-540b shown as exclamation point points, which indicate that the confidence interval does not include zero.
[0164] In Figures 4a-5b, change indicator plots 440a-540b are used to indicate that rejection of the null hypothesis occurs for a given data point. As discussed above, the null hypothesis is when the measured value is equal to the baseline value, but the test may be performed by probability. As shown in Figures 4a-5b, the probability is a confidence interval, but any suitable probability may be used. The confidence interval is represented by a bar associated with each data point. In the example shown in Figures 4a-5b, the bar represents a 99% confidence interval.
[0165] As can be seen, the exclamation point is associated with a data point whose confidence interval does not include the zero axis. In FIG. 5b, the zero axis of stiffness symmetry represents the null hypothesis that the measured stiffness symmetry is equal to the baseline stiffness symmetry value. That is, the zero axis represents no change in the stiffness symmetry of the vibrometer. Thus, if the confidence interval does not include the zero axis, the null hypothesis is rejected. This indicates that, for example, with at least 99% confidence, with the significance level set at 0.01, the null hypothesis is rejected and a change has occurred in the vibrometer.
[0166] As can be appreciated, various systems and methods may use the above-described LPO stiffness change 244, RPO stiffness change 254, and stiffness symmetry 260 to indicate changes in the conduit 130, 130'. An exemplary method is described in more detail below with reference to FIG.
[0167] FIG. 6 illustrates a method 600 for detecting and identifying a change in a vibrometer, according to one embodiment. As shown in FIG. 6, the method 600 begins in step 610 by determining a first stiffness change associated with a first position of a conduit of the vibrometer. The vibrometer and conduit may be one of the vibrometer 5 and conduits 130, 130′ described with reference to FIG. 1. According to this example, the first position of the conduit may be, for example, the position of the left pickoff sensor 170l on the conduit 130, although any suitable position may be used. Thus, the first stiffness change associated with the first position may be the LPO stiffness change 244, which may represent a physical stiffness change of the conduit 130 between the driver 180 and the position of the left pickoff sensor 170l, as discussed above.
[0168] The method 600 may determine a second stiffness change associated with a second position of the vibrometer conduit at step 620. Continuing with the example described above with reference to step 610, the second position of the conduit may be the position of the right pickoff sensor 170r on the conduit 130, although any suitable position may be used. Thus, the second stiffness change associated with the second position may be the RPO stiffness change 254 associated with the position of the right pickoff sensor 170r on the conduit 130, which may represent a physical stiffness change of the conduit 130 between the driver 180 and the position of the right pickoff sensor 170r, as described above.
[0169] In step 630, the method 600 determines a condition in the conduit based on the first stiffness change and the second stiffness change. In the example described above, the condition may be determined based on the LPO stiffness change 244 and the RPO stiffness change 254. The condition may be anything that affects the stiffness of the conduit, such as erosion, corrosion, damage (e.g., freezing, overpressure, etc.), coating, etc. By way of example, the first and second stiffness changes may be the LPO stiffness change 244 and the RPO stiffness change 254 designated as "low". Additionally, the stiffness symmetry 260, which may also be based on the LPO stiffness change 244 and the RPO stiffness change 254, may be "low right". The method 600 may determine that the condition of the conduit 130 is erosion / erosion, for example, using a table similar to that described above.
[0170] The method 600 may further identify, suggest or enable procedures appropriate for each determined condition of the conduit. For example, an alert may be provided for the determined condition of the conduit, and the user may proceed with further diagnostics, maintenance, service, etc. specific to that condition. In the case of conduit damage, procedures may include removing the vibration meter 5 from operation and repairing / replacing the instrument assembly 10. In the case of a coating, a procedure to reduce or remove the coating without removing the vibration meter 5 from operation may be more appropriate.
[0171] Figure 7 illustrates a method 700 for detecting and identifying changes in a vibrometer. As shown in Figure 7, method 700 begins by obtaining a central tendency value of a meter validation parameter and a variance value of the meter validation parameter from storage in the meter electronics of the vibrometer in step 710. In step 720, method 700 determines a probability based on the meter validation parameter and the variance value to determine whether the central tendency value differs from a baseline value.
[0172] In step 710, the central tendency value and the variance value may be retrieved, for example, from the storage system 204 described above with reference to FIG. 2. The storage system 204 may be a register in the processing system 202. Thus, the processing system 202 may retrieve the central tendency value and the variance value from the register and perform simple mathematical operations to determine the probability. In one example, the central tendency value may be the instrument stiffness and the variance value may be the standard deviation of the instrument stiffness.
[0173] In an example using instrument stiffness and variance values, in step 720, the processing system 202 can calculate a t value based on the number of instrument stiffness measurements that make up the instrument stiffness, and use the t value to calculate the probability. In one example, the t value can be determined from the significance level α and the degrees of freedom, as described above. The instrument stiffness may be an average instrument stiffness determined from instrument stiffness measurements taken after a baseline value, such as a baseline instrument stiffness, is determined. The baseline value may be a baseline central tendency value. Thus, the baseline instrument stiffness may be an average value of the baseline instrument stiffness measurements.
[0174] Method 700 may include additional steps, such as, for example, setting a bias deadband. As described above, if the meter stiffness, which may be a central tendency value, is less than the bias deadband, method 700 may determine that the meter stiffness and the baseline meter stiffness are not different. For example, before comparing the meter stiffness to the bias deadband, a flag may be set indicating that the null hypothesis has been rejected because the confidence interval may not include zero. However, if the meter stiffness is less than the bias deadband, the flag may be reset to indicate that the null hypothesis has not been rejected. Thus, method 700 may not send an alert.
[0175] The meter electronics 20, or other electronic devices, devices, etc. described with reference to FIG. 1, may perform the methods 600, 700, or other methods, for detecting and / or identifying changes in a vibrometer. Thus, the meter electronics 20 and the processing system 202 may be configured to receive information from the interface 201 and determine a first stiffness change associated with a first position of the conduit 130, 130' of the vibrometer 5 and determine a second stiffness change associated with a second position of the conduit 130, 130' of the vibrometer 5. With reference to the vibrometer 5 of FIG. 1, the first position may be the position of the left pickoff sensor 170l on the conduit 130, 130' of the vibrometer 5. Similarly, the second position may be the position of the right pickoff sensor 170r on the conduit 130, 130' of the vibrometer 5.
[0176] Meter electronics 20 may also be configured to determine a condition of conduits 130, 130' based on the first stiffness change and the second stiffness change. Meter electronics 20 may also be configured to determine a stiffness symmetry of conduits 130, 130', such as stiffness symmetry 260 shown in FIG. 2. Meter electronics 20 may also be configured to provide an alert based on the determination of the condition of the conduits. The alert may be provided, for example, by sending a signal, message, packet, etc. over path 26.
[0177] The meter electronics 20, and in particular the processing system 202, can also obtain the meter stiffness and the standard deviation of the meter stiffness from storage within the meter electronics 20. The meter electronics 20 or the processing system 202 can determine a probability based on the meter stiffness and the standard deviation of the meter stiffness to determine whether the meter stiffness differs from the baseline meter stiffness.
[0178] The above description provides a meter electronics 20 and method 600, 700 that can detect and identify a change in the vibrometer 5. The change can be identified by detecting a condition of the conduit 130, 130' within the vibrometer 5 based on a first stiffness change associated with a first position of the conduit and a second stiffness change associated with a second position of the conduit. These and other steps can be performed by the meter electronics 20, the processing system 202 within the meter electronics 20, and / or the method 600, or other electronic devices, systems, and / or methods.
[0179] The change can be detected by using statistics in a particular way so that the probability can be determined with limited computational resources. For example, the probability can be a confidence interval around the instrument stiffness where if zero is within the confidence interval, the null hypothesis is rejected. Furthermore, the instrument stiffness To ensure that bias in the stiffness measurement does not induce false alarms, the meter electronics 20 can compare the meter stiffness to a bias deadband. Thus, the probability that it can be continuously updated, as opposed to a limit that does not change, can accurately detect changes in the vibration meter 5 without causing false alarms.
[0180] Although the above discussion refers to a vibrometer 5 shown in FIG. 1, any suitable vibrometer may be used. For example, a vibrometer with two or more drivers and two or more pick-off sensors may be used. Thus, in an exemplary vibrometer having two pick-off sensors and two drivers, three or more stiffness changes may be determined. In this example, the stiffness change between each of the drivers and each of the pick-off sensors may be determined. Similarly, the symmetry between the stiffness changes between the two drivers and the two sensors may also be determined.
[0181] The detailed description of the above embodiments is not an exhaustive description of all embodiments contemplated by the inventors to be within the scope of the present description. Indeed, one of ordinary skill in the art will recognize that certain elements of the above-described embodiments can be combined or deleted in various ways to create further embodiments, and such further embodiments will fall within the scope and teachings of the present description. It will also be apparent to one of ordinary skill in the art that the above embodiments can be combined in whole or in part to create additional embodiments within the scope and teachings of the present description.
[0182] Thus, while specific embodiments have been described herein for illustrative purposes, various equivalent modifications are possible within the scope of this description, as those skilled in the art will recognize. The teachings provided herein may be applied to other identifications of changes in a vibrometer, as well as the embodiments described above and shown in the accompanying figures. The scope of the above-described embodiments should therefore be determined from the appended claims.
Claims
1. Meter electronics (20) for detecting and identifying changes in a vibration meter (5), comprising: an interface (201) configured to receive a sensor signal (100) from an instrument assembly (10) and provide information based on said sensor signal (100); a processing system (202) communicatively coupled to the interface (201) for using the information to: determining a first stiffness change associated with a first position of the conduit (130, 130') of the vibrometer (5); determining a second stiffness change associated with a second position of the conduit (130, 130') of the vibrometer (5); and a processing system (202) configured to determine a condition of the conduit (130, 130') based on the first stiffness change and the second stiffness change.
2. the first position is a position of a left pickoff sensor (170l) on the conduit (130, 130') of the vibrometer (5); The meter electronics (20) of claim 1, wherein the second location is a location of a right pickoff sensor (170r) on the conduit (130, 130') of the vibrometer (5).
3. the first stiffness change represents a physical stiffness change of the conduit (130, 130') between a driver (180) and the first position of the conduit (130, 130'); The meter electronics (20) of claim 1 or 2, wherein the second stiffness change represents a physical stiffness change of the conduit (130, 130') between the driver (180) and the second position of the conduit (130, 130').
4. The meter electronics (20) of any one of claims 1 to 3, wherein the meter electronics (20) is further configured to determine stiffness symmetry of the conduits (130, 130').
5. The meter electronics of claim 4, wherein the stiffness symmetry of the conduits is determined based on the first stiffness change and the second stiffness change.
6. The meter electronics (20) of any one of claims 1 to 5, wherein the processing system (202) is further configured to provide an alarm based on the determination of the condition of the conduit.
7. The meter electronics (20) of any one of claims 1 to 6, wherein the condition of the conduit (130, 130') comprises at least one of erosion, corrosion, damage, and coating of the conduit (130, 130') of the vibrometer (5).
8. 8. The meter electronics of claim 1, wherein the processing system is further configured to determine the condition of the conduit based on at least one of a residual flexibility, damping, and mass of the conduit.
9. Meter electronics (20) for detecting and identifying changes in a vibration meter (5), comprising: A processing system (202) including a storage system (204) configured to store a central tendency value of a meter validation parameter and a variance value of said meter validation parameter, Obtaining the central tendency value and the variance value from the storage system (204); meter electronics (20), comprising a processing system (202) configured to determine a probability based on the central tendency value and the variance value to detect whether the central tendency value differs from a baseline value.
10. 10. The meter electronics of claim 9, wherein the processing system configured to determine a probability based on the central tendency value and the variance value comprises the processing system configured to calculate a t-value and to calculate the probability using the t-value.
11. 11. The meter electronics of claim 9 or 10, wherein the processing system configured to determine the probability based on the central tendency value and the variance value comprises the processing system configured to calculate degrees of freedom based on a number of the meter stiffness measurements that constitute the meter stiffness.
12. 12. The meter electronics of claim 11, wherein the processing system configured to determine the probability based on the central tendency value and the variance value comprises the processing system configured to calculate a standard error based on a standard deviation and the degrees of freedom.
13. The standard error is calculated using the following formula: [0027] Where: stddev pooled is the standard deviation of the meter stiffness, n DOF The meter electronics of claim 12, wherein: is the degree of freedom.
14. 14. The meter electronics (20) of any one of claims 9 to 13, wherein the variance value is a pooled standard deviation including standard deviations of meter stiffness measurements and baseline meter stiffness measurements.
15. The meter electronics (20) of any one of claims 9 to 14, wherein the probability comprises a confidence interval for the central tendency value.
16. The confidence interval is compared to zero; detecting that the central tendency value is not equal to the baseline value if the confidence interval does not include zero; 16. The meter electronics of claim 15, further comprising: detecting that the central tendency value is equal to the baseline value if the confidence interval includes zero.
17. The central tendency value is meter stiffness, and the confidence interval for the meter stiffness is calculated using the following formula: [0028] Where: CI is the confidence interval of the instrument stiffness; Stiffness mean is the gauge stiffness obtained from the storage system (204), CI range is the range of the confidence interval calculated based on the standard deviation and t-value of the instrument stiffness. The meter electronics (20) of claim 16.
18. The range of the confidence interval is calculated using the formula: [0029] Where: stddev pooled is the standard deviation of the meter stiffness, t student,99,8 is determined from the significance level and the number of instrument stiffness measurements that make up the instrument stiffness.
20. The meter electronics of claim 17, wherein the t value is calculated based on a set of degrees of freedom.
19. 19. The meter electronics of claim 9, wherein the processing system is further configured to set a bias deadband, and wherein if the central tendency value is less than the bias deadband, the meter stiffness is not detected as differing from the central tendency value.
20. 1. A method for detecting and identifying changes in a vibrometer, comprising: determining a first stiffness change associated with a first position of the conduit of the vibrometer; determining a second stiffness change associated with a second position of the conduit on the vibrometer; and determining a condition of the conduit based on the first stiffness change and the second stiffness change.
21. the first location is a location of a left pickoff sensor on the conduit of the vibrometer; 21. The method of claim 20, wherein the second location is a location of a right pickoff sensor on the conduit of the vibrometer.
22. the first stiffness change represents a physical stiffness change of the conduit between a driver and the first position of the conduit; 22. The method of claim 20 or 21, wherein the second stiffness change represents a physical stiffness change of the conduit between the driver and the second position of the conduit.
23. 23. The method of any one of claims 20 to 22, further comprising determining stiffness symmetry of the conduit.
24. The method of claim 23 , wherein stiffness symmetry of the conduit is determined based on the first stiffness change and the second stiffness change.
25. 25. The method of any one of claims 20 to 24, further comprising providing an alert based on the determination of the condition of the conduit.
26. 26. The method of any one of claims 20 to 25, wherein the condition of the conduit comprises at least one of erosion, corrosion, damage, and coating of the conduit of the vibrometer.
27. 1. A method for detecting and identifying changes in a conduit of the vibrometer, comprising: obtaining a central tendency value of a meter validation parameter and a variance value of the meter validation parameter from a storage device in the meter electronics of the vibration meter; determining a probability based on the central tendency value and the variance value to determine whether the central tendency value differs from a baseline value.
28. 28. The method of claim 27, wherein determining the probability based on the central tendency value and the variance value comprises: calculating a t-value; and calculating the probability using the t-value.
29. 29. The method of claim 27 or 28, wherein determining the probability based on the central tendency value and the variance value comprises calculating degrees of freedom based on a number of instrument stiffness measurements that make up the instrument stiffness.
30. 30. The method of claim 29, wherein determining the probability based on the instrument stiffness comprises calculating a standard error based on a standard deviation and the degrees of freedom.
31. The standard error is calculated using the following formula: [0030] Where: stddev pooled is the standard deviation of the meter stiffness, n DOF The method of claim 30, wherein is the degree of freedom.
32. 32. The method of any one of claims 27 to 31, wherein the variance value is a pooled standard deviation including the standard deviation of the instrument stiffness measurement and the baseline instrument stiffness measurement.
33. 33. The method of any one of claims 27 to 32, wherein the probability comprises a confidence interval for the central tendency value.
34. The confidence interval is compared to zero; detecting that the central tendency value is not equal to the baseline value if the confidence interval does not include zero; 34. The method of claim 33, detecting that the central tendency value is equal to the baseline value if the confidence interval includes zero.
35. The central tendency value is meter stiffness, and the confidence interval for the meter stiffness is calculated using the following formula: [0031] Where: CI is the confidence interval of the instrument stiffness; Stiffness mean is the gauge stiffness obtained from the storage system (204), CI range is the range of the confidence interval calculated based on the standard deviation of the instrument stiffness and the t-value The method of claim 34.
36. The range of the confidence interval is calculated using the formula: [0032] Where: stddev pooled is the standard deviation of the meter stiffness, t student,99,8 is determined from the significance level and the number of instrument stiffness measurements that make up the instrument stiffness.
36. The method of claim 35, wherein the t-value is calculated based on a set of degrees of freedom.
37. 37. The method of any one of claims 27 to 36, further comprising setting a bias deadband, wherein if the central tendency value is less than the bias deadband, the central tendency value is not detected as differing from the baseline value.
38. 1. A method for detecting and identifying changes in a vibrometer, comprising: determining a first stiffness change associated with a first position of the conduit of the vibrometer; determining a second stiffness change associated with a second position of the conduit on the vibrometer; and determining a condition of the conduit based on the first stiffness change and the second stiffness change.
39. the first location is a location of a left pickoff sensor on the conduit of the vibrometer; the second location being a location of a right pickoff sensor on the conduit of the vibrometer; 39. The method of claim 38.
40. the first stiffness change represents a physical stiffness change of the conduit between a driver and the first position of the conduit; 40. The method of claim 38 or 39, wherein the second stiffness change represents a physical stiffness change of the conduit between the driver and the second position of the conduit.
41. 41. The method of any one of claims 38 to 40, further comprising determining stiffness symmetry of the conduit.
42. 42. The method of any one of claims 38 to 41, wherein the stiffness symmetry of the conduit is determined based on the first stiffness change and the second stiffness change.
43. 43. The method of any one of claims 38 to 42, further comprising providing an alert based on the determination of the condition of the conduit.
44. 44. The method of any one of claims 38 to 43, wherein the condition of the conduit comprises at least one of erosion, corrosion, damage, and coating of the conduit of the vibrometer.
45. 45. The method of any one of claims 38 to 44, further comprising determining the condition of the conduit based on at least one of a residual flexibility, damping, and mass of the conduit.
Citation Information
Patent Citations
Method and device for monitoring anomaly in gas turbine
JP2007192138A
Flow meter electronics and method for verification diagnostics of flow meters
JP2009509163A
Vibrating flow meter and zero-check method
JP2014522972A
