From disturbance to measurement: Application of Coriolis meter for two-phase flow with gas bubbles

Highlights

• Explanation of Multi-Frequency Technology.

• Detection of entrained gas.

• Pattern Identification of Gas Bubbles.

Abstract

Entrained gas has been regarded as disturbance to measurements based on Coriolis meters, since measurement accuracy can be degraded because of this disturbance. Recent research from Endress + Hauser has discovered that different types of gas bubbles, namely free bubbles and suspended bubbles, have various impact on the meter measurement performance. It is important to understand the error mechanism for different effects, namely bubble effect and resonator effect, which are introduced by different bubble types, and to take the corresponding measures to cope with the effects. It is also crucial to identify the bubble pattern in the measuring tube of a Coriolis meter to make a diagnosis and reduce the negative influence of the disturbance accordingly. For free bubbles that typically cause inhomogeneity of a medium, the fluctuation of the resonance frequency of the measuring tube in a Coriolis meter is directly correlated to the existence of this type of bubbles, since this medium under a flowing condition causes density fluctuation to the meter as gas density is typically much lower than that of a liquid. For homogenous suspended bubbles that lead to a significantly increased compressibility of a medium, the innovative Multi-Frequency Technology in Promass Q sensor offers the means to qualitatively detect the existence of this type of bubbles and quantitatively calculate the volume fraction of the gas phase, based on its ability to derive the speed of sound in a medium containing such bubbles. Identification of the type of bubbles helps not only for crediting the measurement reliability, but also for obtaining more detailed medium properties, and in turn a better process insight, with which a process optimization can be enabled to improve the quality of production.

Introduction

In recent decades, Coriolis Mass Flowmeters (CMFs) have been widely used in industry for mass flow and density measurements. The measuring technique has reached a high degree of acceptance and new fields of applications emerge every year. Together with this high acceptance, CMFs are utilized as a multivariable sensor with not only mass flow and density, but also temperature and viscosity measurements [1]. There is a trend to use those additional measured parameters, for example density and viscosity, for monitoring product quality.

Fig. 1 shows a typical CMF, which consists of two parallel measuring tubes, a housing that protects the inner part as well as other components adhering to the measuring tube such as a driver for exciting the tube and sensors for sensing the tube motion. The measuring tube, which in commercial designs can be of various shapes, is the core element of a CMF. In order to be energy efficient, the tube is continuously excited at its natural frequency. This measured natural frequency is a function of the medium density in the tube, therefore, forms the basis of the density measurement. The induced tube vibration generates an angular velocity. Together with the mass flow of a medium inside the tube, Coriolis forces are generated, which causes an anti-symmetrically distortion of the tube. The magnitude of this distortion sensed by the sensors is directly proportional to the mass flow rate and forms the basis of the mass flow measurement.

Similar to many other measuring principles, it is known that accuracy of a CMF can be affected by the existence of entrained gas in a liquid flow. A number of research activities have been carried out in the past to understand the error mechanisms of Coriolis metering under two-phase conditions, which is summarized in some PhD theses, such as in Ref. [2] and in Ref. [3], as well as in some review papers, such as in Ref. [4] and in Ref. [5]. In many industrial applications dealing with liquids, where gas is not intended to be put in and consequently the gas void fraction (GVF) in a CMF is relatively low, the gas phase is usually conveyed in the form of bubbles. Therefore, bubbly flows have been causing one of the biggest challenges for Coriolis metering. In the opinion of Reizner [6], approximately 90–95% of non-hardware CMF problems were due to entrained air issues. Early in the 1980s, various experiments and research activities were already attempted to elucidate the influence of bubbles on CMF. Grumski et al. and Cascetta et al. tested different CMFs with different two-phase flows, analysed these phenomena based on the results of the experiments and gave suggestions for improving the reliability of the meters [7,8]. Hemp and Sultan theoretically addressed this problem related to bubbly flow and proposed the bubble theory to explain the observed errors [9,10]. Since then, more and more CMF manufacturers have become involved in this issue by enterprise R&D or cooperation with other research institutes. Bubble effect was studied intensively in Ref. [3] and renamed as decoupling effect. A digital transmitter with the capability of handling entrained gas flow was proposed by Henry et al. [11], and a method of correcting mass flow errors caused by two-phase flow was developed with the help of neural network [12] based on this digital transmitter. In addition to the bubble effect, another effect, which is named as resonator effect with a significant impact on Coriolis metering, was first disclosed and analysed analytically and numerically in two conference papers [13,14] during the PhD study of [2]. Based on this finding, the resonator effect was further developed theoretically by Hemp et al. [15] and renamed as compressibility effect. However, compressibility is only one of the two major factors that dominate the corresponding effect for Coriolis metering. The other important factor is mass. That is the reason why a single-phase gas does not cause a similar order of effect as a two-phase gas-liquid fluid does, although its compressibility is higher than the latter. Therefore, in this paper, the name resonator effect is still preferred, since the frequency of a resonator consists of the contribution from both compressibility and mass. In order to compensate the resonator effect, a two-mode compensation scheme was proposed in Ref. [2], which formed the basis for Multi-Frequency Technology (MFT) developed later [16].

Very often, bubbles bring a negative effect rather than a benefit to the process industries: they cause additional pressure losses; they can damage the pumps, propellers and impellers when cavitation takes place; more importantly, they cause increased uncertainties to process measurements employing different measuring technologies. However, it has been found that in many industrial applications, especially in Food industry and Life Science, gas is injected into liquid products intentionally, for example for the sake of product flavour such as in cream cheese, or as a special feature such as in shampoo.

According to the effects on Coriolis metering, gas bubbles in liquid flows are classified as “free bubble” and “suspended bubble” that lead to “bubble effect” and “resonator effect”, respectively. The primary goal of the investigation is to reduce or eliminate the negative impacts of the above two effects on the measurement. As the best practice, the bubble effect due to the existence of free bubbles should be suppressed by eliminating the free bubbles with the help of the process optimization, e.g. using gas eliminator or holding tank, since free bubbles are much easier to separate, compared with suspended bubbles. Typically, free bubbles cannot be maintained as a product feature, as they will escape soon from the product anyway after they are put in. The resonator effect, which is caused by suspended bubbles that are difficult to separate from liquid, can be compensated by MFT. The theoretical background of the two effects and MFT is given in the next section of this paper.

The second goal of the investigation is that for applications, where free bubbles and suspended bubble can both exist, or at least one of them can, it is important to detect the existence of entrained gas and identify the type of entrained bubbles that has relevance to the measurement reliability. Furthermore, very often entrained gas bubbles affect product quality in an adverse way, for example in chemical industry when a glue is produced. However, sometimes homogeneous suspended micro-bubbles are wanted as a product feature, for example in food industry when cream cheese is produced, as previously mentioned. Exactly for this product, big free bubbles are unwanted and regarded as being disadvantageous for product quality. In the meanwhile, the existence of big “free bubbles” also indicates a less optimal manufacturing process, in which the injected gas is not well mixed into the cream cheese for generating homogeneous micro-bubbles. Therefore, the detection of gas bubbles and the identification of the bubble types are crucial for product quality control and optimization of production process.