 |
|
||||||||||||
|
 Inside this issue
 Flow Conditioning Multiple piping configurations present metering and calibration challenges to the flow metering engineer. For inferential meters to provide accurate results, their output signals should represent the average fluid velocity occurring over the complete area of the meter tube. Rotating-type inferential meters, such as turbine meters, are also particularly sensitive to rotational swirl within the meter tube. The function of the flow conditioner is to minimize swirl and to provide uniform fluid velocity within the meter tube area at the point of measurement. Research programs, in both the United States and Europe, confirm that a variety of piping configurations and fittings generate disturbances with unknown flow characteristics. Various classes of flow conditioners and lengths of straight pipe are used in combination to isolate the flowmeters from piping-induced disturbances. Tube-bundle conditioners including AGA, ASME and ISO are in wide use. However, it has been determined that they do not necessarily achieve fully-developed velocity profiles in certain piping lengths. There are several new designs incorporating perforated plates which are designed to eliminate some problems associated with tube-bundle flow conditioners; but relatively few designs have been found acceptable for widespread custody transfer use. One system design that has been widely accepted was developed by Savant Measurement Corporation. Their GallagherTM Flow Conditioner (GFC) has been found to be beneficial in numerous piping configurations, especially in rectifying swirl and flow profile effects. The GFC uses anti-swirl and profile correcting devices in-line to minimize piping-induced disturbances within the flowmeter. Independent test results from both international and domestic research laboratories demonstrate that the GFC generated pseudo-fully developed flow profiles for orifice meters, turbine meters, vortex meters, ultrasonic meters and similar inferential flow devices. A well-engineered flow conditioning system ensures that uncertainty levels are reduced in those installations that typically experience non-ideal flow conditions. Communicating with Micro Motion Mass Meters The Omni flow computer can be configured to accept mass or volume pulses from a Micro Motion Coriolis meter. It can also communicate via Modbus to the device obtaining variables such as fluid density and MM transducer alarm states. The flow computer is equipped with special firmware code to perform integrated mass or volume proving. The communication link between the Micro Motion RFT transmitter and the flow computer is via the Omni peer-to-peer link. It is possible to have multiple Micro Motion meters connected to multiple flow computers as shown in the figure. A 'Modicon Compatible' device Some adjustments to the peer-to-peer entries are needed when communicating with devices that require 'Modicon Compatible' to be selected for the peer-to- peer port [serial port #2]. 1) All database point addresses (whether source or destination) referring to the foreign Modicon compatible device should be entered as one less than the point address listed. This is needed because the Modicon device automatically adds one to the address received over the data link and subtracts one from the address before transmitting. References to database point addresses within the Omni flow computer master still use the normal point address as shown in the Omni documentation. 2) The number of points entry becomes the number of 16-bit registers to transfer, rather than the number of data variables.
Peer-to-Peer Transactions Multiple peer-to-peer transactions can be used to read data, such as the flowing density of the fluid and integer registers from the MM meter containing packed alarm status bits. These packed alarms are stored in a special alarm register within the flow computer, enabling them to be time and date tagged, logged and printed, just as though they were flow computer alarms. More transactions may be needed depending upon what additional data is required and how many meter runs are being used. Calibration Tips Because of the temperature sensitivity and bit resolutions of the flow computer A/D and D/A converters, and the high accuracy requirements, it is important that the following procedures are followed when calibrating flow computer I/O circuits: 1) Adjust the power supply to give 5.05-5.10 volts at the backplane test points. 2) All final calibrations must be performed using the actual set of modules that will be installed during normal use. 3) Before calibrating, eliminate temperature gradient errors by closing the flow computer housing and allowing at least 20 minutes for temperature stabilization to occur. Ensure that the unit is not in a high air draft area. Make adjustments, such as jumper repositioning, quickly. Wherever possible, keep the unit closed to retain internal heat. Board replacements will require that sufficient time be allowed to achieve temperature stability. 4) Observe all temperature stability requirements of any test equipment used in the calibration process. EPROMs - The Next Generation If you have not already updated your liquid or gas applications, the time may be now for you to do so. The EPROM upgrade will allow users to increase their application capabilities such as delivering 12-point meter linearization and the ability to archive up to 500 alarms. If your system is not in compliance with API and Year 2000 (Y2K), this EPROM will upgrade your system and offer some of the following features:
Calculation of Gas Net Volume and Energy Configuring any flow computer for gas calculations requires a sound knowledge of AGA and ISO gas measurement standards. The following describes the ease with which key data can be entered into an Omni. This entry is located in the 'Factor Setup' menu. Setting this entry to '0' ensures that 'gas density at base conditions' is calculated using AGA-8 (Method (6)). Entering the 'density of air at base conditions', assuming a valid 'gas relative density (SG)' is available, will override the AGA-8 calculation of 'gas density at base conditions'. In this case, 'gas density at base conditions' is calculated using either Method (3), (4) or (5). Gas Relative Density (SG) In the 'Fluid Analysis Data' menu, gas relative density (SG) can be found. For each active product one (1) entry is required. It is mandatory that this field contain a valid value of 'SG' for all AGA-8 'gross' calculation methods, except for 1985 method #4. The data in this field can be manually entered or automatically overwritten by a live 4-20mA input of 'SG', if it exists. This entry also serves as the gas chromatograph (GC) 'SG' override, if a GC is providing SG and a GC failure occurs. Entering a negative (-) value in this field will force the flow computer to calculate 'gas density at base conditions' using AGA-8. (Method (6)). Entering the SG, assuming a non-zero 'Density of Air @ Base Conditions' is entered, will override the AGA-8 calculation of 'gas density at base conditions'. In this case, 'gas density at base conditions' is calculated using either Method (3), (4) or (5). When an AGA-8 detailed method is selected and a GC is used to provide SG, this entry field is ignored unless a GC failure occurs and the 'GC Fail Code' entry is set to 'Use Override on GC Failure'. Gas Heating Value (HV) This entry is located in the 'Fluid Analysis Data' menu. One entry per active product is required. It is mandatory that this field contain a valid value for a 'gas heating value (HV)' for AGA-8 'gross' calculation method #1 and also AGA-8 1985 methods #2 and #4. The data in this field can be manually entered or automatically overwritten by a live 4-20mA input of HV, if it exists. This entry also serves as the GC 'HV' override if a GC is providing HV and a GC failure occurs. Entering a negative (-) value in this field will force the flow computer to use a 'calculated HV' using either AGA-5, GPA 2172 or ISO 6976 ( Method (10)). Entering a positive (+) value into the HV entry will override the AGA-5, GPA 2172 or ISO 6976 calculation of HV. When an AGA-8 detailed method is selected and a GC is used to provide HV, this entry field is ignored unless a GC failure occurs and the 'GC Fail Code' entry is set to 'Use Override on GC Failure'. Can you believe A/D Accuracy Claims? In hydrocarbon electronic flow measurement (EFM), the differences in measurement accuracy obtained using Analog to Digital Converters (A/Ds) of various bit resolutions (for example: 14-bit verses 16-bit) may have little relevance when several practical issues are considered. (Note that A/D resolution is not a factor when using digital process transmitters in the 'digital mode' with flow computers.) A/D Least Significant Bit (LSB) Uncertainty: An A/D will convert an analog signal into '2n' discrete steps or digital values (where 'n' is equal to the bit resolution of the A/D). A 14-bit A/D provides 16,384 discrete digital values, whereas the 16-bit A/D produces 65,536 discrete digital values. Not all of these values are available to measure analog input signal ranges such as 1-5 volts. In most cases, the A/D must be able to measure signals which are 5% or more overrange (eg., 5.25 volts instead of 5 volts). Some of the A/Ds range of values are also used to measure the elevated zero (1 volt), which is common in process measurement. The result is only 76% of the total number of discrete A/D values can be used to measure the signal between 1-5 volts. When calculated, the 14- and 16-bit A/Ds will have approximately 12,500 and 49,800 discrete digital values respectively. Therefore, the percent difference in reading, produced by a change in the Least Significant Bit 'LSB', of a 14-bit is 0.008% [(1/12500)*100], and a 16-bit is 0.002%[(1/49800)*100]. Sampling Rate: When process conditions are changing rapidly, the step change between readings taken by a slow sampling A/D system can be much larger than the precision of the analog transmitter or flow computer A/D. The timing associated with the signal sampling and its averaging method is as significant as A/D resolution. Ambient Temperature Effects on A/Ds: The assumed benefit of a higher-resolution A/D will disappear if significant drift occurs when the A/D is exposed to variations in ambient temperature. If the A/D is not properly compensated to minimize 'zero offset' and 'span' errors, then the incremental improvement of the A/D resolution would be negated, due to changes in ambient temperature. For a meaningful improvement in accuracy, the higher bit resolution must be accompanied by a matching improvement in A/D temperature stability and long-term circuit stability. The uncertainty of a process transmitter reading (+/-0.075%)** is not impacted significantly by the additional uncertainty contributed by a LSB change using either a 14-bit A/D(0.008 %) or a 16-bit A/D(0.002 %). The added uncertainty of the measured process variable caused by calibration equipment and operational procedures must also be considered. The equipment used to calibrate process transmitters is assumed to be not sensitive to temperature effects and will be calibrated against a traceable standard. Additional assumptions include, that the equipment will be accurate to a magnitude greater than the process transmitter. Practical Effects of A/D Uncertainties on Measurement: To illustrate how the various uncertainties (A/Ds verses a transmitter) can impact the actual results of a measurement, a flow computer was setup with four meter runs, all using the same natural gas composition, orifice size, pipe diameter, etc. The object of the test was to compare the calculated flowrate and gas densities for: (1) an error free meter run (NoError), (2) a meter run with simulated Dp and pressure transmitter errors (TX Error), (3) a meter run with simulated Dp and pressure errors caused by a LSB change of a 14-bit A/D (14BitLSB), and (4) a meter run with simulated Dp and pressure errors caused by a LSB change of a 16-bit A/D (16BitLSB). Override inputs were used for every parameter, including orifice Dp, static pressure, and temperature. Results from meter run #1 represent error-free results based on AGA-8 flowing density and calculated flowrates, for a Dp of 100 inches, static pressure of 750 psig, and a temperature of 75 deg.F. The other three meter runs portray the equivalent results obtained when the input Dp and static pressure overrides were adjusted slightly to simulate the uncertainty added by smart Dp and Sp transmitters (+/-0.075%), a LSB change of a 14-bit A/D (0.008%), and a LSB change of a 16-bit A/D (0.002%). Conclusion: Even when using premium smart transmitters, the measurement error caused by differing A/D resolutions was insignificant (less than 1/20th) compared to the error introduced by the differential and static pressure transmitters. The impact of differing A/D bit resolutions on liquid measurement will have even less of an impact. This result is due to the API rounding rules and because liquid density is relatively insensitive to changes in temperature and pressure. A/D stability: When exposed to the same ambient temperature change, a 16-bit A/D will drift by 4 times as many LSB discrete digital steps in value as a 14-bit A/D. For this reason, it takes a bigger temperature change to cause the LSB to change in a 14-bit A/D. Ultimately, A/Ds of differing bit resolutions, but with the same stability and temperature drift performance, will have the same impact on the measurement uncertainty. |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
| | | | | | | | | | | |
 |
©2003 Omni Flow Computers, Inc. All rights reserved |
