Håkon Viumdal

Ultrasonic level sensors for flowmetering of non-Newtonian fluids in open Venturi channels: Using data fusion based on Artificial Neural Network and Support Vector Machines

In drilling operations related to oil & gas or geothermal applications, the improved monitoring and control of the flowrate of drilling fluid is important. This will help in reducing cost as well as improving system and Health, Safety and Environmental (HSE) performances. A relatively accurate and low-cost flow monitoring system functioning as a supervisory unit for the drilling fluid in the return path would be useful for this purpose. Inclusion of appropriate sensors and modifying the geometry of an already existing open channel in the transport of drilling fluids are possible approaches for estimating the flowrate of the drilling fluid. Forming a Venturi flume in the already existing open channel structure of the transporting conduit for the drilling fluid offers some interesting possibilities. Using a set of three ultrasonic level meters for determining the levels at various points in the open channel and fusing the data from other sensors in the test loop, the flow rate in a Venturi channel is successfully estimated. Two different empirical models using Artificial Neural Network (ANN) and Support Vector Machine (SVM) are used as alternatives to the mass balance approach. For the flowrate of drilling fluid in the range of (250-550) kg/min, the performances of ANN and SVM models are much better than that of the mechanistic model. The sampling rate for SVM is about 90 times more than that of the mechanistic model. Whereas, the sampling rate of ANN is about 100 times more than that of the SVM model. The Mean Absolute Percentage Error (MAPE) for both empirical models is less than 2%.

Enhancing signal to noise ratio by fine-tuning tapers of cladded/uncladded buffer rods in ultrasonic time domain reflectometry in smelters

Buffer rods (BR) as waveguides in ultrasonic time domain reflectometry (TDR) can somewhat extend the range of industrial applications of ultrasonics. Level, temperature and flow measurements involving elevated temperatures, corrosive fluids and generally harsh environments are some of the applications in which conventional ultrasonic transducers cannot be used directly in contact with the media. In such cases, BRs with some design modifications can make ultrasonic TDR measurements possible with limited success. This paper deals with TDR in conjunction with distance measurements in extremely hot fluids, using conventional ultrasonic transducers in combination with BRs. When using BRs in the ultrasonic measurement systems in extreme temperatures, problems associated with size and the material of the buffer, have to be addressed. The resonant frequency of the transducer and the relative size of the transducer with respect to the diameter of BR are also important parameters influencing the signal to noise ratio (SNR) of the signal processing system used in the ultrasonic TDR. This paper gives an overview of design aspects related to the BRs with special emphasis on tapers and cladding used on BRs. As protective cladding, zirconium oxide-yttrium oxide composite was used, with its proven thermal stability in withstanding temperatures in rocket and jet engines up to 1650 °C. In general a BR should guide the signals through to the medium and from and back to the transducer without excessive attenuation and at the same time not exacerbate the noise in the measurement system. The SNR is the decisive performance indicator to consider in the design of BR based ultrasonic TDR, along with appropriate transducer, with suitable size and operating frequency. This work presents and analyses results from extensive experiments related to fine-tuning both geometry of and signals in cladded/uncladded BRs used in high temperature ultrasonic TDR with focus on overall performance based on measured values of SNR.

Soft Sensing of Non-Newtonian Fluid Flow in Open Venturi Channel Using an Array of Ultrasonic Level Sensors—AI Models and Their Validations

In oil and gas and geothermal installations, open channels followed by sieves for removal of drill cuttings, are used to monitor the quality and quantity of the drilling fluids. Drilling fluid flow rate is difficult to measure due to the varying flow conditions (e.g., wavy, turbulent and irregular) and the presence of drilling cuttings and gas bubbles. Inclusion of a Venturi section in the open channel and an array of ultrasonic level sensors above it at locations in the vicinity of and above the Venturi constriction gives the varying levels of the drilling fluid in the channel. The time series of the levels from this array of ultrasonic level sensors are used to estimate the drilling fluid flow rate, which is compared with Coriolis meter measurements. Fuzzy logic, neural networks and support vector regression algorithms applied to the data from temporal and spatial ultrasonic level measurements of the drilling fluid in the open channel give estimates of its flow rate with sufficient reliability, repeatability and uncertainty, providing a novel soft sensing of an important process variable. Simulations, cross-validations and experimental results show that feedforward neural networks with the Bayesian regularization learning algorithm provide the best flow rate estimates. Finally, the benefits of using this soft sensing technique combined with Venturi constriction in open channels are discussed.

Dependency of signal to noise ratio on transducer diameter and buffer diameter in guided wave time domain reflectometry

The use of waveguides or buffer rods (BR) is well known in performing Non-Destructive Testing (NDT) with ultrasonic or generally in high-temperature ultrasonic time domain reflectometry (TDR). However, determining optimal geometrical shapes of these buffers and parameters of the ultrasonic transducers are not straightforward. In this paper the signal to noise ratio (SNR) achieved with different combinations of buffer diameters and transducer diameters are considered based mainly on experiments and on some dedicated simulations. Experimental results show large variations of the SNR as the ratio between the transducer and the buffer rod diameter varies. Corresponding simulated results are less influenced by such variations in the ratio of diameters considered, showing possible limitations of these simulations.

Rheological characterization of non-newtonian drilling fluids with non-invasive ultrasonic interrogation

The drilling process is generally costly and time consuming and prone to serious hazards. Cost-efficiency and enhanced safety measures are vital for any drilling operation. Recent studies indicate that poor reliability in the drilling process resulted in as much as 30% loss of production time. Improved sensor technology with process automation can improve process performance and safety. During drilling operations, along with the drillstring, a drilling fluid, commonly very dense and viscous fluid, is circulated in a closed flow-loop. The drilling fluid, non-Newtonian in its rheological behavior, serves three main objectives: keeping the bottom-hole pressure at an acceptable level, lubricating the drill bit and facilitating the removal of cuttings and debris from downhole. These three goals have to be kept in balance and are achieved by adjusting the density (ρ), viscosity (μ) and the flow-rate (qv) of the drilling fluid. These three drilling process parameters need to be continuously monitored for optimizing process performance and securing safety. The cuttings in the drilling fluids make it especially challenging when conventional in-line sensor systems are used due to the unavoidable erosion and maintenance costs. Non-invasive ultrasonic measurement techniques can be part of a robust and easily implementable control and monitoring system. In this work ultrasonic properties of different drilling fluids are studied. Propagational properties of different samples of drilling fluids are studied with focus on attenuation and frequency characteristics in transmission mode. Experimental results using different sets of ultrasonic transducers with different frequencies, confirm the high attenuation of ultrasonic pulses. A model is proposed to estimate the attenuation and viscosity of the drilling fluid based on ultrasonic and rheological parameters. This study presents results from ultrasonic interrogation of non-Newtonian fluids with focus on their rheological properties.

Estimating viscosity of non-Newtonian fluids using support vector regression method: Rheological parameters of drilling fluids using data fusion

In contrast to flow studies involving water in hydraulics and hydrodynamics falling in the realm of Newtonian fluids, the flow behavior of food, oil and polymers fall into the category of non-Newtonian fluids, where the rheological parameters viscosity, density, flow velocity, Reynolds number are all interdependent. Because of these interdependencies, in addition, the viscosity is dependent on the dimension and geometry of the conduit used for transporting the medium. Many models exist for different conditions relating the shear rate to shear stress for the estimation of viscosity. For engineering applications involving drilling fluids as encountered in oil & gas industries or geothermal applications, a knowledge of viscosity is important. This paper presents the estimation of viscosity using Support Vector Regression (SVR) method. In an earlier study, without resorting to analytical techniques involving a plethora of equations, the viscosity was estimated, using Artificial Neural Networks (ANN). In this paper, measurements performed in an open Venturi channel with the non-Newtonian fluid flow are used to estimate the viscosity using ANN and SVR techniques.

Improving ultrasonic multi-level measurements with wavelets - results from separators and smelters

Usage of ultrasonic level measurements is encountered in different industrial applications such as gas, oil, water and sand in separators, liquid-solid in sedimentation processes, molten metals, etc. When the reflected signal level from multi-layers is low and swamped in noise with very poor signal to noise ratio (SNR), getting good enough signals for level estimation and interface detection will be difficult. This will be even more difficult in flowing mediums with multi-levels or in ultrasonic interrogations of molten metals. To enhance the signal detection algorithms, wavelets have been applied with some success. The objective of this paper is to present wavelet based algorithms for the detection of reflected signals from multi-layers when the SNR is very low. Ultrasonic time domain reflectometry (UTDR) is performed using buffer rods for signal transmission and wavelets for signal analysis. Suitable mother wavelet selection, threshold and thresholding algorithms are proposed and successfully implemented. Discrete wavelet Transform (DWT) based algorithms for signal denoising gives better signal trains in the UTDR studies performed in multilayered systems including molten metals. A system consisting of three layers is studied using ultrasonic transmission via a buffer rod. Useful denoising properties are achieved using Sym7 mother wavelet, Rigrsure threshold, and soft thresholding for analyzing the reflected signals in multi-layered systems. A significant improvement is observed in the case of UTDR based studies involving molten metals too.