PIPE FLOW SOFTWARE ... Calculate ... Simulate ... Communicate
| Fluid Flow (Home) | | | About Us | | | Navigate this site | | | Contact Us |
PIPE FLOW SOFTWARE ... Calculate ... Simulate ... Communicate |
| Fluid Flow (Home) | | | About Us | | | Navigate this site | | | Contact Us |
|
SoftwareSelector:
|  |
COMPARISON OF MODEL RESULTS WITH FIELD MEASUREMENTS The purpose of this Application Note is to provide some guidance when simulating existing pipe systems and making comparisons between model results and direct measurements in the field. Although the note is written in terms of our FluidFlow3 software, it equally appllies to AFT software. For both liquids and gases pipe friction losses are based on an expanded form of the Darcy -Weisbach equation. The friction factor term in the Darcy equation f is found from the Haaland equation which gives better than 2% agreement with the well-known Colebrook-White equation. The absolute roughness of the pipe, k, is used in the Haaland equation when determining f. FluidFlow3 detects laminar flow. A modified form of the Darcy equation is used for compressible flow to account for velocity and density changes as the fluid expands. The use of the Darcy equation implies that the fluid is Newtonian with a coefficient of dynamic viscosity that changes only with variation in temperature (so all aspects of this note may not necessarily apply to slurry flow simulations). FluidFlow3 provides an accurate solution to the in-built equations if this is the case. However, in the field there may be many circumstances that will affect flows and pressures that FluidFlow3 cannot predict. For instance …
Pipe Scaling: FluidFlow3 allows the internal diameter and roughness of pipes to be varied within a model. This can be achieved by globally updating groups of pipes so it is very easy, for instance, to provide a 10% scale to all pipes and determine how this affects the model. Cavitation: FluidFlow3 cannot model the effect of cavitation on pumps although it does calculate and display the NPSH available at the inlet to any pump in the system and flag a warning if insufficient NPSH is detected. Vapourisation: FluidFlow3's calculated values of flows and pressures at every node in the system should be inspected to see if negative pressures occur which may have an influence on flow that FluidFlow3 cannot predict. Remember, before actual vapourisation occurs, low pressures may bring air or gas out of solution. This can significantly influence the performance of a pump in reality if the vapour release is on the suction side; an effect which the software cannot simulate. Poor Pipework Design or Layout: High points in pipe systems can cause problems that a software simulation will not necessarily detect. In the simplest case, air or vapour becomes trapped at the high point reducing the cross sectional area of flow in the pipe. Air relief valves should be installed to exhaust the trapped gas. If the pipe centreline at a high point rises above the hydraulic grade line this will result in localised negative pressures. Air can come out of solution at these points long before absolute zero pressure occurs. Suction pipework can be particularly problematical. Negative pressures in the suction line will not only cause air to come out of solution, but will draw air in via leaks in the pipework. High points (even high points arising from the use of concentric reducers) can localise air blockages. Air in a suction system can become trapped within the eye of the impeller of a centrifugal pump causing a lowering of performance. Equipment Deterioration: Site measurements/inspection required. FluidFlow3 allows a fixed head drop to be entered at any location in the model so such head losses can be tested, perhaps to accord with site pressure measurements. Control Valves Operating out of Recommended Range: FluidFlow3 detects and warns of two conditions: (a) if the valve is operating outside user-specified recommended operational limits, and (b) if the valve is operating outside the limits of manufacturer's test data. Equipment Head Loss Coefficients: Site-specific data required. FluidFlow3 has a wide range of options for simulating the head loss across junctions, fittings and equipment including the methods of Crane and Idelchick. FluidFlow3 allows fittings data to be entered in the Crane fashion (K') or in the more traditional way of K = u2/2g. If the K method is used, the values of K supplied for a 100mm fitting will not necessarily be correct for a 200mm fitting. So, in any simulation, adopting the correct head loss coefficient for a fitting can be difficult. Is it certain that a fitting is actually performing in accordance with manufacturer's data? Over time the internals of the fitting may have deteriorated, causing a higher or lower pressure drop than the coefficient would indicate. Finally, most coefficients are provided for fittings tested in isolation from other line equipment. How closely spaced fittings interact with each other is often uncertain. Site Data: Obtaining accurate site data with which to compare the model simulation may be difficult. Are site-metering devices accurate; have they been recently calibrated? Is the actual pump or fan performance compatible with the curve supplied - without witness testing a pump could legitimately be operating within 5% of the manufacturer's curve - with wear the discrepancy could be much higher; is the specified impeller actually installed? Have extensions to the system been added and not documented; could there be other demands on the system which have not been detailed? existing pipe network. The computer simulation allows the engineer to identify problem areas by comparing model values with site-measured values and then "calibrating" the model to agree with site conditions. SUMMARY When comparing FluidFlow3's results to field results it should be kept in mind that FluidFlow3 solves a set of equations based on input data. Circumstances in the field may be different. Although FluidFlow3 comes with a database of fittings, the only accurate source of head loss data for a particular fitting is the manufacturer or actual test data. Existing systems have often been upgraded and extended. Documentation of these past changes often does not exist. FluidFlow3's flowsheet representation of the pipe network with its choice of colours for different pipes provides an excellent schematic of the pipe network in the absence of proper drawings. The working model with its ability to provide instantaneous "what if" calculations is an effective tool for developing an understanding of how the system operates, perhaps in discussion with field operators and maintenance staff. This type of discussion can lead to a calibration of the model to close agreement with site conditions. This possibility of tuning a FluidFlow3 network to achieve agreement with field results should be considered, say by varying pipe roughness values, partially closing valves, simulating leaks, etc. Having tuned or calibrated the model it would then be reasonable to accept that percentage changes in flows and pressures resulting from changes in the calibrated model would be very similar to the percentage changes in the existing system the field. When comparing FluidFlow3's results with site you should check the following: " Adequate NPSH available
at all pumps. *Crane - Flow of Fluids Through
Valves, Fittings and
|
| Copyright | Disclaimer | Conditions of Use of Accutech Software |