Following feedwater flow

29 June 2001



A highly accurate method of measuring feedwater flow allows plant operators to monitor any loss in power output arising from fouling of the feedwater venturis.


Reactor power levels are typically monitored by secondary calorimetric calculations that depend on the accurate measurement of feedwater flow. Most plants measure feedwater flow with calibrated venturis. These are subject to chemical fouling and other mechanical problems. If the loss in accuracy of the feedwater flow measurement overstates the actual flow rate, the result is a direct loss in power generated by the plant.

The Crossflow ultrasonic flowmeter improves the accuracy, reliability and maintainability of feedwater flow measurements by providing a continuous correction to the venturi-measured flow rates. Crossflow can be used as a “stand alone system” or with direct input to plant computer thermal power calculations to provide the operator with continuous high accuracy thermal power indications. A number of plants using the system have reported the recovery of between 5 and 15MWe.

Venturi fouling

From the above equation, the accuracy of the calculation of plant output is most sensitive to the feedwater flow parameter. Actual in-plant experience has shown that the feedwater flow measurement is the least reliable parameter over time, due to the difficulties in maintaining the venturi systems. This problem is clearly recognised by the US Nuclear Regulatory Commission (NRC) in the 10 CFR 50 Appendix K requirements to include a 2% power measurement uncertainty factor in safety analyses.

Venturis are subject to a variety of problems in service. These can cause a significant degradation in the accuracy of the instrumentation, as well as increase the cost of routine maintenance.

•Instrument drift. Calibration of the differential pressure transmitter can drift over the operating cycle.

•Feedwater pipe erosion. The carbon steel pipe around the upstream sensing port boss can be subject to corrosion. This causes the sensor to measure a dynamic rather than a static pressure, which reduces the differential pressure and creates a lower flow reading that can lead to an overpower condition.

•Cracked sensing tube. The high-pressure static fluid behind the venturi leaks into the sensing tube causing a lower differential pressure that leads to non-conservative flow and an overpower condition.

•Bypass flow problems. Because the venturis are often not fully welded in place there can be leakage through the stagnant region behind the venturi.

•Initial calibration problems. Venturis are calibrated in the laboratory under nearly perfect conditions, but often are not installed as carefully as needed to create the same “perfect” flow conditions as in the lab on which the calibration is based.

•Fouling. The primary issue with fouling is the increase in surface roughness – caused by the corrosion products – which impacts the boundary layer and tends to indicate a higher flow than actual. A secondary issue is the reduced cross-sectional area.

The fouling phenomenon is the most typical problem. Over time, as the venturi becomes fouled, the pressure drop increases giving a false indication that feedwater flow has increased. Because feedwater flow is directly proportional to thermal power level, this causes reactor power levels to be calculated higher than actual. Reactor power levels are limited by the plant’s operating licence. Every time the calorimetrics are re-done during an operating cycle, the thermal power output is lowered to match the apparent feedwater flow measured by the venturis, causing the plant to operate at a power level lower than planned. Because feedwater flow appears to remain constant at 100% the loss in output is often unnoticed. The losses can range from 0.5% to 3.0%.

The error in feedwater flow can be identified and measured through the use of ultrasonic instruments, chemical tracers or through careful review of standard plant parameters. Once the lost generation is recognised, correction and recovery is quite simple.

Crossflow technology

Unlike older technology employed in the transit-time ultrasonic flowmeters, the Crossflow meter relies on the modulating characteristics of the turbulence within the flow stream to determine the velocity of the fluid. Using Crossflow, a high frequency signal is transmitted through the fluid using strap-on transducers. As the signal passes through the fluid, it is modulated by the flow eddies within the fluid, imparting a unique pattern to the ultrasonic signal.

The eddies travel at the velocity of the fluid and maintain their unique pattern for a finite period.

A second set of strap-on transducers is installed downstream. As the flow eddies pass through the second set of transducers, the same modulating characteristics of the eddies are again imparted to the downstream signal. By knowing the distance between the two sets of transducers and the time that it takes for the eddies to pass between the sets, it is possible to calculate the fluid velocity and the volumetric flow rate. A mathematical process called cross-correlation is used to determine the difference in the time that it takes for the eddies to pass between the two transducer sets.

The advantage of the cross-correlation method is the system’s lack of sensitivity to changes in fluid temperature and synchronous noise that are inherent to the transit-time flowmeters. This results from using a high frequency signal as the carrier for the lower modulating signal that is used to determine the fluid velocity. Analysis of the Crossflow meter calibrated in the operating range for feedwater systems shows an effective two sigma error of 0.5% or better.

Validation of technology

The technology was verified on different pipe diameters (from 3” to 30”), pipe materials (carbon, stainless steel, plastic) and pipe wall thickness (from 0.25” to 2”). Tests were carried out in the Alden Research Laboratory, NIST Flow Laboratory, Electricité de France laboratory, Ontario Hydro Laboratory, Ontario Hydro High Temperature Flow Laboratory, National Research Council of Canada and the Stern Laboratory. In all these tests the difference between Crossflow readings and laboratory instrumentation was within uncertainty of the laboratory instrumentation. Besides laboratory tests the technology was verified under real plant conditions at the Diablo Canyon, Shearon Harris and Angra plants.

NRC review

In 1999 Westinghouse and the Advanced Measurement and Analysis Group (AMAG) submitted a topical report on the system to the NRC. The report was submitted in support of US utility applications for 1% power uprates, by means of improved feedwater flow measurement accuracy and reduced power measurement uncertainty. During the review process, the NRC probed deeply into the cross-correlation methodology in order to understand and assess this technique. The NRC concluded that the Crossflow system and methodology provides feedwater flow measurement accuracy and stability, which is adequate to allow reduction of the standard power measurement uncertainty factor of 2%. The externally mounted, bolt-on meter’s accuracy was shown to be similar to older, intrusive, spool-piece-mounted transit time meters, which have been used in the past.

The following main issues were raised during the review, mostly because this was the first application of the technology.

Crossflow history

Cross-correlation technology dates back to work performed by Canadian General Electric in the 1970s, research and development by Ontario Power Generation (formerly Ontario Hydro) through the 1980s and further development by AMAG since the early 1990s. AMAG was founded and is presently led by individuals who played key roles in development of cross- correlation technology for Ontario Power Generation and Canadian General Electric. AMAG has further advanced the development and application technology in recent years. Although the technology has been understood for years, the required data processing necessary to perform the statistical averaging in an accurate, timely and cost-effective manner was not commercially viable until the past decade. Today’s enhanced computing power together with AMAG’s advancements in ultrasonic cross-correlation application and sophisticated acoustical design, have allowed AMAG and Westinghouse to significantly evolve, validate and verify the technology.

Theoretical basis

The theory of flow measurement in a pipe using ultrasonic cross-correlation technology for single-phase flow originated in the 1970s as an empirical relationship. It was developed into a theoretical relationship and verified by AMAG in the 1990s.

Upstream disturbance

All flow measurement devices, including venturis, clamp-on transit time, chordal multi-path transit time and cross-correlation, are affected by upstream disturbance. To provide accurate flow measurement, the effect of this upstream disturbance must be accounted for and established by installation criteria that have to be met to achieve the specified accuracy. For example, the specific flow meter has to be installed at a certain distance downstream of the disturbance.

Acoustical noise

Since the 1970s, Canadian General Electric, Ontario Power Generation and AMAG have investigated the influence of acoustical noise on cross-correlation feedwater flow measurements in detail. During the last five years AMAG has performed a comprehensive analysis of this effect. This resulted in a new system design and methodology that has reduced the effect of acoustical noise to a minimum to achieve the specified accuracy.

Percentage of flow stream

The ultrasonic beam interacts with the turbulence in the central region of the pipe. The distribution of the turbulence in this region is uniquely related to the Reynolds number of feedwater flow. Thus, by knowing the Reynolds number, which is calculated by the Crossflow meter, and the velocity of the eddies within the pipe, the Crossflow meter is capable of providing an accurate measurement of the feedwater flow.

Data acquisition time

Crossflow is designed to accurately measure flow as an average of a number of readings, which is appropriate in feedwater flow applications. Due to the statistical nature of the technology, the typical data acquisition time for 0.1% uncertainty on feedwater flow measurement is from 10-15 minutes to 1-2 hours. Therefore, making comparisons of instantaneous readings with laboratory test results can be misleading. Using techniques commonly deployed in determining flow measurement accuracy yields total flow measurement uncertainty of equal to or better than 0.5%. This puts Crossflow at a comparable accuracy level of intrusive spool-piece systems and makes it the most accurate non-intrusive flow meter available.

Venturi correction and power calculation

Since Crossflow is not susceptable to problems associated with venturi fouling it can be used as a direct input to plant computer thermal power calculations to provide the operator with continuous high accuracy thermal power indication. Furthermore, Crossflow can be compared (in real-time) with venturi output and thereby used to alert plant staff to venturi fouling or de-fouling events. The additional accuracy gained combined with continuous online monitoring permits maximum reactor output while providing an extra measure of safeguard against exceeding rated thermal power.



Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.