Method and device for measuring the flow rate of oil-gas two-phase flow in the oil return pipeline of an aeroengine

Abstract

The present invention discloses a method and device for measuring the flow rate of oil-gas two-phase flow in the oil return pipeline of an aeroengine. Based on a large number of capacitance data reflecting the distribution of oil-gas two-phase at different temperatures collected by a high-temperature ECT sensor, the method uses data mining technology and machine learning algorithms to propose a new method for temperature compensation of void fraction measurement, establishes a void fraction measurement model under different flow patterns, and selects the corresponding void fraction measurement model and two-phase flow rate calculation correlation formula according to the flow pattern discrimination result, realizing the measurement of the flow rate of oil-gas two-phase flow in the oil return pipeline of an aeroengine. The proposed method for temperature compensation of void fraction measurement reduces the measurement error of void fraction caused by temperature changes, improves the measurement accuracy of void fraction and flow rate of oil-gas two-phase flow in the oil return pipeline of an aeroengine, and provides a new idea and method for measuring the parameters of the oil return pipeline of an aeroengine.

Description

Technical Field

[0001] The present invention relates to the technical field of multiphase flow measurement, and particularly to a method and device for measuring the oil-gas two-phase flow rate in the oil return pipeline of an aeroengine. Background Art

[0002] An aeroengine is a device with a complex structure and high precision. During its operation, the lubrication system provides lubrication and cooling for components such as gears and bearings of the aeroengine, removes foreign impurities such as abrasive particles, and plays roles such as rust prevention, cleaning, sealing, and buffering. The oil return system returns the lubricating oil discharged from the oil outlet holes to the lubricating oil tank to ensure the balance of the quality and temperature of the lubricating oil in the bearings, and plays an important role in maintaining the normal operation of the engine. In order to better analyze the flow state of the lubricating oil in the oil return pipeline and design the oil return pipeline more reasonably, it is necessary to measure the relevant flow parameters in the oil return pipeline.

[0003] The void fraction, also known as the cross-sectional gas holdup, refers to the ratio of the gas phase area to the total cross-sectional area on the cross-section of two-phase flow. This parameter is one of the important flow parameters for two-phase flow analysis and measurement of other related parameters.

[0004] Flow rate is mainly characterized by two forms: mass and volume. The mass flow rate refers to the total mass of the gas-liquid two-phase fluid flowing through the pipeline cross-section per unit time (the sum of the gas phase mass flow rate and the liquid phase mass flow rate), and the volume flow rate is the total volume of the two-phase fluid flowing through the pipeline cross-section per unit time.

[0005] An Electrical Capacitance Tomography (ECT) sensor can, based on the different dielectric constant distributions caused by different medium distributions in the measured field, invert the medium distribution in the field by measuring the capacitance values and sensitivity matrix distributions in the field, providing an effective way to solve the problem of measuring the void fraction of two-phase flow.

[0006] There is a complex and variable lubricating oil-air two-phase mixed flow in the oil return pipeline of an aeroengine, and the diameter of the oil return pipeline is small. When the engine is working, the temperature of the oil return pipeline is relatively high and there are continuous temperature changes. The electrical properties of the lubricating oil in the oil return pipeline change with temperature, and existing ECT sensors usually work in a room temperature environment and cannot adapt to the changes in the electrical properties of the measured medium caused by temperature changes.

[0007] When using a high-temperature ECT sensor to measure the flow rate of two-phase flow, usually the image reconstruction of the pipeline cross-section is first performed, then the void fraction is calculated based on the reconstructed image, and finally the two-phase flow rate is obtained according to the flow rate calculation correlation. Among them, the calculation complexity of image reconstruction is relatively high, time-consuming and laborious, affecting the real-time performance of void fraction measurement, and thus resulting in high measurement complexity and being time-consuming and laborious for the flow rate.

[0008] At present, there is no applicable measurement method for the oil-gas two-phase flow in the return oil pipeline of an aero-engine. It is necessary to seek a suitable measurement method according to the actual working conditions to meet the real-time measurement of the void fraction and flow rate under the condition that the temperature of the two-phase fluid inside the pipeline is constantly changing, and the measuring device can withstand a relatively high temperature for a long time. Summary of the Invention

[0009] Aiming at the deficiencies of the existing technology, the present invention provides a measurement method and device for the oil-gas two-phase flow in the return oil pipeline of an aero-engine. This method can reduce the influence of temperature on the measurement result error, enable the device to adapt to the temperature change of the oil-gas two-phase flow, measure the flow rate of the oil-gas two-phase flow in real time, and improve its measurement accuracy.

[0010] The object of the present invention is achieved by the following technical solutions:

[0011] A measurement method for the oil-gas two-phase flow in the return oil pipeline of an aero-engine includes a model construction stage and an actual measurement stage; in the model construction stage, the ECT sensor collects the capacitance signals of the measured cross-section of the return oil pipeline to form a data set, analyzes the data sets under three different flow patterns to obtain the void fraction measurement model with temperature compensation under the corresponding flow patterns, then determines the parameters in the different two-phase flow rate calculation correlation formulas under the three flow patterns, and analyzes and calculates the errors of different correlation formulas under different flow patterns; during actual measurement, it is not necessary to collect the temperature information of the return oil pipeline, only the capacitance signals between two non-adjacent electrodes of the ECT sensor need to be collected in real time, the flow pattern is judged according to the capacitance signals, and then the capacitance is input into the void fraction measurement model with temperature compensation under the corresponding flow pattern to obtain the void fraction, and a flow rate calculation correlation formula with the smallest error under the corresponding flow pattern is selected to obtain the two-phase flow rate.

[0012] In the above technical solution, further, analyzing the data sets under three different flow patterns to obtain the void fraction measurement model with temperature compensation under the corresponding flow pattern specifically includes: using the kernel principal component analysis method to extract features from the data sets under the three flow patterns of bubble flow, annular flow and stratified flow to obtain the data after dimensionality reduction, and then using support vector regression for the data after dimensionality reduction to obtain the void fraction measurement model with temperature compensation under the corresponding flow pattern.

[0013] The construction method of the void fraction measurement model with temperature compensation is as follows:

[0014] (1) Data collection: Collect the capacitance data under different temperatures, different flow patterns and different void fractions to form a training set;

[0015] (2) Feature extraction: Extract the features of the data to be dimensionally reduced through the kernel principal component analysis (KPCA) method to obtain the features after dimensionality reduction. The specific steps are as follows:

[0016] i Combine the training set data of the three flow patterns at the same temperature to form the matrix \(C\) to be dimensionally reduced. N*20 = [c 1 , c 2 , …, c N T , where \(c i \) represents each group of capacitance values. By using the Gaussian kernel function, the kernel matrix \(K\) is calculated, and after centering processing, \(H\) is obtained:

[0017]

[0018]

[0019] where \(I\) is the identity matrix;

[0020] ii Calculate the eigenvalues and eigenvectors of the centered kernel matrix \(H\):

[0021] \(H\mu=\lambda\mu\)

[0022] iii Set the cumulative contribution rate \(A\), and select the top \(r\) eigenvalues \(\lambda j \((j = 1, …, r)\) and the corresponding eigenvectors \(\mu j \((j = 1, …, r)\) from largest to smallest:

[0023]

[0024] iv Calculate the dimensionally reduced matrix \(C p \):

[0025]

[0026] (3) Model establishment: Using the extracted principal component eigenvalues, adopt the machine learning algorithm - support vector regression SVR to establish the void fraction measurement models under different flow patterns

[0027] \(\alpha(C p ) = u T C p + b

[0028] where \(C p \) is the principal component feature after the dimensional reduction of the original capacitance data by KPCA, and \(u T \) and \(b\) are the regression coefficient and the residual respectively.

[0029] An oil-gas two-phase flow flow measurement device for an aeroengine oil return pipeline, comprising the following three parts:

[0030] High-temperature ECT sensor: A multi-electrode high-temperature array capacitance sensor made of FPC material;

[0031] ​Venturi tube flowmeter: A differential pressure measuring port is set in the contraction section of the Venturi tube and a pressure measuring point is designed in the rear diffuser section, which is used to measure the two-phase medium flowing through the pipeline;

[0032] Data acquisition unit: used to alternately excite and detect the electrodes of the high-temperature ECT sensor to obtain the array capacitance value, and transmit the capacitance value to the host computer; collect and transmit the output signals of pressure and differential pressure measurements; the host computer identifies the flow type and obtains the flow data based on the void ratio measurement model with temperature compensation and the flow calculation correlation formula obtained by the above method.

[0033] The beneficial effects of the present invention are as follows:

[0034] The method for measuring the flow rate of oil-gas two-phase flow in the oil return pipeline of an aircraft engine of the present invention adopts a high-temperature ECT sensor to measure the capacitance signal, directly establishes a porosity and flow measurement model through the measured capacitance signal, avoids the complex and time-consuming image reconstruction process to directly obtain the porosity and flow of the two-phase fluid, and effectively reduces the computational complexity and the influence of temperature on the porosity measurement. The established porosity model is based on a large amount of capacitance data reflecting the distribution of the oil-gas two-phase at different temperatures, and can mine the model change information caused by temperature changes, and realizes the temperature compensation of the porosity measurement from the data-driven perspective. The method proposed by the present invention greatly reduces the porosity measurement error caused by temperature changes. According to research, the flow measurement method and device of the present invention have a maximum porosity measurement error of no more than 10% in the temperature range of 20°C to 180°C, and a relative error of flow measurement within 10% in the temperature range of 20°C to 140°C, which improves the porosity and flow measurement accuracy compared with the existing method, and is particularly suitable for the porosity and flow measurement of the oil-gas two-phase flow in the oil return pipeline of an aircraft engine. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the flow measurement device of the present invention;

[0036] Figure 2 Schematic diagram of the experimental device used in the model building stage of the present invention;

[0037] Figure 3 It is a schematic diagram of the flow measurement method of the present invention;

[0038] Figure 4 This is a physical picture of the FPC electrode array of the present invention;

[0039] Figure 5 This is a schematic diagram of the void ratio measurement model of the present invention;

[0040] Figure 6 This is the comparison chart of the maximum absolute error before and after temperature compensation;

[0041] Figure 7 It is a comparison chart of the maximum measurement errors of the Chisholm correlation under three flow patterns; Specific implementation manners

[0042] The present invention will be described in detail below according to the accompanying drawings and preferred embodiments. The purpose and effects of the present invention will become more apparent. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0043] The method and device for measuring the gas-liquid two-phase flow rate of the oil return pipeline of an aeroengine of the present invention include a model construction stage and an actual measurement stage; in the model construction stage, first, a high-temperature ECT sensor is used to collect capacitance signals of the measured cross-section of the oil return pipeline to form a data set. The kernel principal component analysis method is used to extract features from the data sets under various flow patterns to obtain the dimension-reduced data. Then, support vector regression is performed on the dimension-reduced data to obtain a void fraction measurement model with temperature compensation under the corresponding flow pattern. Then, the parameters in the quality gas content formula are determined according to the void fraction. Furthermore, according to the differential pressure signal measured by the Venturi tube and the quality gas content calculated by the quality gas content formula with determined parameters, the parameters in the different two-phase flow rate calculation correlations (James, Collins, and Chisholm flow rate calculation correlations) under the three flow patterns are determined, and the errors of the three correlations under different flow patterns are analyzed and calculated; in actual measurement, it is not necessary to collect the temperature information of the oil return pipeline. Only the capacitance signals between two non-adjacent electrodes of the high-temperature ECT sensor need to be collected in real time. First, the flow pattern is determined according to the capacitance signals (the flow pattern can be determined by using the clustering method to extract features for classification, or any method reported in the literature can be used, such as Meng Zhenzhen. Research on a new method for measuring the flow rate of gas-water two-phase flow [D]. Zhejiang University, 2011.). Then, the capacitance is input into the void fraction measurement model of the corresponding flow pattern to obtain the void fraction. Combining with the measured differential pressure signal, the flow rate calculation correlation with the smallest error under this flow pattern is selected to obtain the two-phase flow rate. Using this method can greatly improve the measurement accuracy of the void fraction compared with the existing methods, thereby improving the measurement accuracy of the flow rate. In addition, since the relationship between the void fraction and the capacitance value is non-linear, if a linear feature extraction method is used, there will be a large error. In the present invention, the method combining KPCA and SVR is used to construct a void fraction measurement model with temperature compensation, which can well solve the problem of non-linear feature extraction, extract the target information to the greatest extent, and the combination of these two methods is simple to implement and suitable for the industrial field of rapid measurement.

[0044] The void fraction and flow rate measurement model can be applicable to any measured cross-section of the oil return pipeline. However, for ECT sensors with different parameters, model training needs to be carried out separately.

[0045] As Figure 1 shown, the flow rate measurement device of the present invention mainly includes the following parts:

[0046] (1) High-temperature ECT sensor: Aiming at the difficult conditions of high temperatures (20 °C to 180 °C) in the oil return pipeline, an 8-electrode high-temperature array capacitive sensor was designed. The high-temperature ECT sensor consists of three parts: an insulating pipeline part, an 8-electrode array part, and a shielding cover. Among them, the 8-electrode array part uses FPC material, which can ensure that the electrodes do not deform or are difficult to fix under high-temperature conditions;

[0047] (2) Venturi tube flowmeter: The Venturi tube is provided with a differential pressure measurement port in the contraction section and a pressure measurement point in the post-diffuser section, and the two-phase medium flowing through the pipeline is measured through the corresponding transmitters respectively; The measurement principle of the Venturi tube flowmeter is based on the continuous flow equation and Bernoulli equation, so it can only be used for the flow measurement of single-phase flow. To apply it to the measurement of oil-gas two-phase flow, it is necessary to combine the flow parameters of multiphase flow such as flow pattern and void fraction to correct the parameters of the flow calculation correlation formula to achieve a higher-precision measurement of multiphase flow.

[0048] (3) Data acquisition unit: This part is mainly responsible for realizing two major functions: alternately exciting and detecting the electrodes of the high-temperature ECT sensor to obtain the array capacitance value and transmitting the capacitance value to the upper computer; collecting and transmitting the output signals of the pressure and differential pressure transmitters; The upper computer discriminates the flow pattern and obtains the final flow data according to the void fraction measurement model and flow calculation correlation formula constructed in the model construction stage.

[0049] In the model construction stage, it needs to be realized by using an experimental device. As Figure 2 shown, the oil return pipeline condition is simulated through a mixer. Before mixing, both oil and gas are single-phase, and their respective flows can be directly measured by the corresponding flowmeters. A temperature sensor is also set at the post-diffuser section of the Venturi tube in this device. In actual measurement, this setting is not required. The flow calculated according to the flow calculation correlation formula is compared with the result shown by the flowmeter before mixing to obtain the corresponding error. Select the flow calculation correlation formula with the smallest error under the corresponding flow pattern for subsequent measurement.

[0050] As Figure 3 shown, the flow measurement method of the present invention specifically includes the following steps:

[0051] S1: Fill the high-temperature ECT sensor with lubricating oil, insert hollow polytetrafluoroethylene (PTFE) tubes with different diameters into the high-temperature ECT sensor to simulate typical flow patterns. Since the dielectric constant of polytetrafluoroethylene is close to that of oil, the gas inside the hollow tube can simulate the bubbles in the cross-section. Therefore, by inserting hollow polytetrafluoroethylene tubes with different diameters into the high-temperature ECT sensor filled with lubricating oil inside, the typical flow patterns of oil-gas two-phase flow under different void fractions in the oil return pipeline can be simulated, and the capacitance signals between non-adjacent pairs of electrodes of the high-temperature ECT sensor at different void fractions from 20°C to 140°C are measured to form an offline data set. When the high-temperature ECT sensor has 8 electrodes, there are 20 capacitance signals between non-adjacent pairs of electrodes measured, so the obtained offline data set is 20-dimensional.

[0052] S2: Obtain the flow pattern according to the flow pattern discrimination, and select the void fraction measurement model with temperature compensation under the corresponding flow pattern to obtain the void fraction;

[0053] The void fraction measurement model with temperature compensation is specifically:

[0054] (1) Data acquisition: Collect capacitance data at different flow patterns and different void fractions at 20°C, 40°C, 60°C, 80°C, 100°C, 120°C, 140°C, 160°C, and 180°C to form a training set;

[0055] (2) Feature extraction: Extract features from the data to be dimension-reduced through the KPCA method to obtain the dimension-reduced features. The specific steps are as follows:

[0056] i. Combine the training set data of the three flow patterns at the same temperature to form the matrix C to be dimension-reduced N*20 =[c 1 ,c 2 ,…,c N T , where c i represents each group of capacitance values. By using the Gaussian kernel function, the kernel matrix K is calculated, and after centering processing, H is obtained:

[0057]

[0058]

[0059] where I is the identity matrix;

[0060] ii. Calculate the eigenvalues and eigenvectors of the centered kernel matrix H:

[0061] Hμ = λμ

[0062] iii. Set the cumulative contribution rate to 95%, and select the top r eigenvalues λ from largest to smallest j ​(j = 1, …, r) and the corresponding eigenvector μ j (j = 1, …, r):

[0063]

[0064] iv Calculate the dimensionality reduction matrix C p :

[0065]

[0066] (3) Model establishment: Using the extracted principal component eigenvalues, conventional machine learning algorithms - Support Vector Regression (SVR) are adopted to establish void fraction measurement models for different flow patterns (bubble flow, annular flow, and stratified flow) respectively

[0067] α(C p ) = u T C p + b

[0068] where C p is the principal component feature after the original capacitance data is reduced in dimension by KPCA, and u T and b are the regression coefficient and the residual respectively

[0069] S3: Input the void fraction and the differential pressure signal measured by the Venturi tube into the flow rate measurement model together to obtain the two-phase flow rate

[0070] The specific flow rate measurement model is as follows

[0071] (1) Determine the parameters in the quality gas content formula according to the void fraction measured by the void fraction measurement model under the corresponding flow pattern

[0072] (2) Determine the parameters in the different two-phase flow rate calculation correlations (James, Collins, and Chisholm flow rate calculation correlations) for the three flow patterns according to the differential pressure signal and the quality gas content, and analyze and calculate the errors of the three correlations under different flow patterns

[0073] (3) Obtain the flow pattern through flow pattern discrimination, select the flow rate calculation correlation with the smallest error under this flow pattern, and calculate the two-phase flow rate

[0074] The void fraction and flow rate measurement method of the present invention can be used for various oil return pipelines. To further verify the void fraction and flow rate measurement method proposed by the present invention, taking the oil return pipeline of an aeroengine as an example, corresponding experimental studies were carried out. During the experiment, the medium temperature was set, and the lubricating oil was heated to this temperature by a constant temperature heating oil tank. The gear pump and air compressor were adjusted to achieve the corresponding flow pattern and void fraction. Under the stable state of this working condition, the capacitance signal collected by the high-temperature ECT sensor and the pressure drop signal of the Venturi tube were recorded, and the separated phase flow rates of the gas-liquid two-phase were recorded as reference values. Among them, since only a small section of the oil in the tube needs to be heated to change the temperature during void fraction measurement, but the whole device needs to be pumped to flow during flow rate measurement, and it is difficult to achieve too high a temperature when heating all the oil. Therefore, in the flow rate measurement verification stage of the present invention, the temperature was only set up to a maximum of 140 °, while the maximum temperature in the void fraction measurement stage was up to 180 °. That is, the set temperatures for the flow rate verification experiment were 20 °C, 40 °C, 60 °C, 80 °C, 100 °C, 120 °C, and 140 °C respectively.

[0075] The void fraction measurement and flow rate measurement models are established based on the capacitance measurement values at different temperatures. The data is divided into a training set and a test set. Among them, the training set data is used for the establishment of the void fraction and flow rate measurement models, and the test set data is used for the verification of the measurement results.

[0076] At 20 °C to 180 °C, the void fraction measurement method of the present invention was used to process the training set data, and the maximum absolute error change of the void fraction obtained is as Figure 6 shown. Using the void fraction measurement method proposed by the present invention can greatly reduce the influence of temperature on measurement, and compared with the capacitance normalization method (CN113340951A) proposed by us before (i.e., Figure 6 the full-tube calibration in it), the void fraction measurement accuracy can be further improved. In the temperature range of 20 °C to 180 °C, the maximum measurement error of the void fraction by the method of the present invention does not exceed 10%. Compared with the existing methods, the method of the present invention can greatly improve the void fraction measurement accuracy, thereby correspondingly improving the flow rate measurement accuracy. In addition, different flow rate correlation empirical formulas and different flow patterns will all affect the calculation of the two-phase flow rate. The present invention can determine the correlation formula with the smallest error under the corresponding flow pattern by combining the judgment of the flow pattern and the selection of the flow rate correlation, thereby further reducing the flow rate error and improving the accuracy.

[0077] Under the experimental working conditions of the present invention (the quality gas content range is 0 to 0.064, the flow rate range is 0 kg / min to 19.37 kg / min, and the temperature range is 20 °C to 140 °C), the measurement effect of the Chisholm flow rate calculation correlation formula is the best, and the error is within 10%. Therefore, the optimized Chisholm model under this working condition can be used for the measurement of the oil-gas two-phase flow rate in the oil return pipeline.

[0078] Those of ordinary skill in the art can understand that the above are only preferred examples of the invention and are not used to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent replacements for some of the technical features. Any modifications, equivalent replacements, etc. made within the spirit and principles of the invention shall be included within the protection scope of the invention.

Claims

1. A method for measuring the oil-gas two-phase flow rate in the oil return pipeline of an aeroengine, characterized in that, it includes a model construction stage and an actual measurement stage; in the model construction stage, a data set is composed of capacitance signals collected by an ECT sensor at the measured cross-section of the oil return pipeline. The data sets under three flow patterns are analyzed to obtain the void fraction measurement models with temperature compensation under the corresponding flow patterns. Then, the parameters in the different two-phase flow rate calculation correlation formulas under the three flow patterns are determined, and the errors of different correlation formulas under different flow patterns are analyzed and calculated; during actual measurement, it is not necessary to collect the temperature information of the oil return pipeline. Only the capacitance signals between two non-adjacent electrodes of the ECT sensor need to be collected in real time. The flow pattern is judged according to the capacitance signals, and then the capacitance is input into the void fraction measurement model with temperature compensation under the corresponding flow pattern to obtain the void fraction. The flow rate calculation correlation formula with the smallest error under the corresponding flow pattern is selected to obtain the two-phase flow rate; The analysis of the data sets under the three flow patterns to obtain the void fraction measurement models with temperature compensation under the corresponding flow patterns is specifically as follows: collect the capacitance data at different temperatures and different void fractions under the three flow patterns to form a data set; the three flow patterns are bubbly flow, annular flow and stratified flow; the kernel principal component analysis method is used to extract the features of the data sets under the three flow patterns to obtain the data after dimensionality reduction. Then, support vector regression is performed on the data after dimensionality reduction to obtain the void fraction measurement models with temperature compensation under the corresponding three flow patterns.

2. The method for measuring the oil-gas two-phase flow rate in the oil return pipeline of an aeroengine according to claim 1, characterized in that, the construction method of the void fraction measurement model with temperature compensation is as follows: (1) Data collection: collect the capacitance data at different temperatures, three flow patterns and different void fractions to form a data set; (2) Feature extraction: perform feature extraction on the data to be dimensionally reduced through the kernel principal component analysis KPCA method to obtain the features after dimensionality reduction. The specific steps are as follows: i composes the training set data of the three flow patterns at the same temperature into the matrix to be dimensionally reduced , where represents the capacitance value of each group. By using the Gaussian kernel function, the kernel matrix K is calculated, and H is obtained after centering processing: ， ， wherein , I is the identity matrix; ii Calculate the eigenvalues and eigenvectors of the centralized kernel matrix H: ； iii Set the cumulative contribution rate A, and select the top r eigenvalues from largest to smallest and the corresponding eigenvectors : ； iv Calculate the dimensionality reduction matrix : ； (3) Model establishment: use the extracted principal component eigenvalues and adopt the machine learning algorithm - support vector regression SVR to establish the void fraction measurement models under the three flow patterns respectively ； Among them are the principal component features after the original capacitance data is reduced in dimension by KPCA, and are the regression coefficient and the residual respectively.

Images