Design of a crude oil ultrasonic flowmeter

0 Introduction
At present, a difficult problem faced in the oil field crude oil production is the online measurement of the produced crude oil. The main reason is that the composition of crude oil is very complex. Crude oil contains oil, water, gas and other impurities. It is a multi-phase complex fluid, and the single well crude oil flows intermittently, so the general flow meter cannot meet the requirements. This paper designs a metering system based on ultrasonic correlation flow calculation, which solves the problem of non-contact online measurement of crude oil.

1 Principle of ultrasonic correlation flowmeter
The correlation method uses correlation technology to measure fluid flow. The measurement accuracy is independent of the sound velocity in the fluid, and the measurement accuracy is high. It is suitable for multi-phase flow and the measurement of fluids with large interference. When the fluid flows in the pipeline, if it contains other impurities, there are various random disturbances inside it, which generates flow signals related to the flow conditions and has certain statistical characteristics. The structure of the correlation method flowmeter is shown in Figure 1. A, A' and B, B' are two sets of ultrasonic transmitting and receiving transducers, and L is the distance between the upstream transducer and the downstream transducer. When the ultrasonic signal passes through the pipeline, it will be modulated by the noise in the fluid. The modulated ultrasonic signal contains the information of the fluid velocity field. The ultrasonic signal is analyzed to extract the flow signals A(t) and B(t) related to the flow condition, and A(t) and B(t) are correlated to obtain the correlation function RAB(τ):

The time displacement corresponding to the peak of the function is τ, which is the time for the fluid to be transferred from the upstream transducer to the downstream transducer, that is, the transfer time (also called transit time) in the system. The flow rate calculation formula is:

According to the flow rate, the flow rate is obtained:

Where: D is the diameter of the pipeline.

2 Structure of single-well crude oil ultrasonic correlation flow measurement system
As shown in Figure 1, the single-well flow measurement system is mainly composed of three parts: gas-liquid separation preprocessing, ultrasonic detection, and signal processing. The following analyzes these three parts respectively.

2.1 Gas-liquid separation pretreatment part
The structure of the pretreatment part is shown in Figure 2. The functions of the pretreatment part are first to separate gas from liquid (in addition to the mixture of oil and liquid, the crude oil extracted by the pumping unit also contains gas and other impurities. The gas will cause a large error in the measurement of oil and liquid, so gas-liquid separation must be performed when measuring the oil and liquid); second, to solve the full pipe measurement during intermittent flow (the crude oil extracted by the pumping unit each time is not equal, and it flows intermittently, which may cause the oil in the pipeline to be not full, and will also cause a large measurement error). For these reasons, pretreatment is required before measuring the oil and liquid. After pretreatment, gas-liquid separation and full pipe of oil and liquid passing through the metering oil pipe are achieved to reduce the metering error. The working principle is: crude oil enters the oil storage tank from the oil inlet through the sand settling tank, and oil and gas are separated in the oil storage tank. The separated gas is output from the upper valve (gas outlet valve) of the oil storage tank through the gas pipeline. When the oil in the oil storage tank reaches a certain height, the float floats up to open the lower valve (oil outlet valve), and the upper valve seals the gas port to build up pressure. Under the action of pressure, the oil in the oil storage tank flows to the oil outlet through the measuring pipeline. When the oil in the oil storage tank drops to a certain level, the float sinks to seal the lower valve and open the upper valve. Repeating this process completes gas-liquid separation. [page]

2.2 Ultrasonic detection part
The detection part is mainly composed of two pairs of ultrasonic sensors. The detection of ultrasonic sensors is completed by transmitting and receiving ultrasonic energy. The core is the transducer (converting ultrasonic energy into electrical energy or converting electrical energy into ultrasonic energy. Reversible transducers refer to transducers that convert two forms of energy into each other with equal efficiency). Common transducers include piezoelectric crystal oscillators, magnetostrictive oscillators, etc. Ultrasonic waves used for related flow measurement generally have two forms: sine waves and pulse waves. Pulse ultrasonic and sine wave related flow meters both integrate the velocity information of the flow field cross section to obtain the flow velocity. This design uses a piezoelectric crystal ultrasonic sensor with a center frequency of 200 kHz. In order to overcome the influence of standing waves, ultrasonic waves use a phase-locked loop pulse signal generator.
2.3 Signal processing part
The signal processing part is mainly composed of an ultrasonic receiving transducer and a DSP.
The signal conditioning circuit consists of a receiving transducer, a three-stage amplifier circuit, a filter circuit, and an envelope detection circuit. The preamplifier is composed of MAX410 instrument amplifier module, the secondary amplifier and the final amplifier are composed of INA128 precision low-power instrument amplifier; the filter circuit is a bandpass filter composed of MAX275 analog integrated filter, with a center frequency of 200 kHz, and a low-pass filter composed of TL14, which mainly extracts the signal after detection. The envelope detection circuit is composed of a diode and a capacitor to form a peak detector.
The other part is a data acquisition and processing circuit composed of a DSP module. This circuit uses TI's TMS320F2812 DSP chip. In the current process control field, it is the most advanced DSP microprocessor. Compared with traditional single-chip microcomputers, it has outstanding performances such as powerful functions, rich resources, and low power consumption. It has perfect performance and the best integrated peripheral interface. It integrates flash memory, high-speed A/D converter, high-performance CAN module, etc. During measurement, the upstream and downstream transmitting transducers emit high-frequency ultrasonic waves. When ultrasonic waves propagate in the fluid, the flow signal will produce amplitude, phase and frequency modulation on the ultrasonic waves. The high-frequency modulated signal received by the receiving transducer is demodulated after filtering and amplification to obtain the flow signal, which is sent to the A/D converter for data acquisition. The collected information is sent to the DSP for related processing to obtain the flow rate of the fluid. 3 System Program Design The software system includes DSP initialization, calculation module, flow display, interrupt processing module and other parts. The main program flow chart is shown in Figure 3. After the main program completes initialization, it enters a loop program to process the sampled data, respond to external A/D interrupt requests, serial communication interrupt requests and timer interrupt requests at any time, and also judge whether the flow display timing has arrived at any time. The main program responds to the above interrupt requests and calls each corresponding processing program to complete data acquisition and processing.

Initialization is to set up the working environment of DSP on the one hand, and to prepare for the subsequent signal processing on the other hand. The system initialization program includes the internal initialization that affects the operation of DSP chip CPU and the peripheral initialization that affects the operation of each peripheral, as well as the initialization of peripheral programmable devices (such as A/D, D/A, etc.). Specifically, it includes the following functions: setting the clock generator, setting the timer, initializing each status register, opening the interrupt, etc.
The interrupt processing module includes three interrupts: the timer interrupt processing module is used to start the A/D converter and control the sampling frequency, the serial communication interrupt processing module is used to communicate with the host computer, and the A/D interrupt processing module is used to read the A/D converter sampling data. Its flow chart is shown in Figure 4.
The display module refreshes the instrument regularly to display the instantaneous flow value and the cumulative flow value.
The system processing process is: set the timing cycle, the timer generates an interrupt, this interrupt starts the A/D converter, after the conversion is completed, the A/D converter requests the DSP to read the data interrupt, the DSP responds to the A/D converter interrupt request, calls the A/D interrupt processing module, reads the sampled data, and sends it to the data buffer. Since the fluid flows intermittently, after receiving the N-point data of the upstream and downstream signals, the DSP performs Fourier analysis on the data to determine whether the fluid flows. If it flows, the calculation program is called to perform correlation operations on the sampled data, find the peak of the correlation function, determine the transit time T, and obtain the instantaneous flow value and the cumulative flow value according to the instrument parameters and temperature compensation, and store the results in the data storage unit for display by the display instrument.
In the correlation flow measurement, one of the key issues is the calculation method of the correlation function, which requires the acquisition of a large number of random modulation signals, the correlation integral operation and the peak search of the correlation function to be completed at high speed and accurately. The algorithms of the correlation function mainly include the polarity coincidence method and the zero-point crossing method. In order to improve the calculation speed, this system uses the correlation operation in the frequency domain. After the input data is transformed by FFT, the correlation operation in the frequency domain can be obtained. Then, the correlation result in the time domain can be obtained by IFFT, which can be used for peak search.

4 Conclusion
Based on the analysis of the single well working conditions and the principle of correlation flow measurement in the oil field, a device suitable for single well crude oil metering is designed. After field testing, good results have been achieved, and its error is less than 2%. However, there are still several problems: First, the signal fluctuates greatly, mainly because the crude oil contains gas and impurities, which causes a large signal difference, and the detection circuit needs to add an AGC circuit. Second, it is difficult to adjust the correction coefficient. Different wells have different water contents, and the viscosity of the oil varies greatly. At the same time, the fluidity of the oil varies greatly at different temperatures, so the correction coefficient must be adjusted many times in different environments, which brings inconvenience to use. Third, the error is relatively large when the flow rate is low. These are all aspects that need to be improved in future research.