A crude oil water cut detection method and system based on multi-sensor data fusion

Abstract

The application discloses a crude oil water content detection method and system based on multi-sensor data fusion, which comprises the following steps: collecting and preprocessing crude oil water content information data sets according to a capacitive sensor, a microwave sensor, a conductivity sensor and a pressure sensor; obtaining the best order and degree by performing autocorrelation and partial autocorrelation analysis on the preprocessed data sequence; calculating the information data by using an ARIMA time series water content residual model to obtain an initial prediction value; and calculating the water content residual data set; taking the water content residual data set as the input of a LightGBM model, calculating the information data by using the LightGBM model to obtain a further prediction value; and adding the further prediction value and the initial prediction value to obtain the final crude oil water content. The ARIMA-LightGBM hybrid model is used for fusing the measurement data of the four sensors, the range of crude oil water content detection is expanded, and the precision is improved.

Description

Technical Field

[0001] This invention relates to the field of crude oil water content detection technology, specifically to a crude oil water content detection method and system based on multi-sensor data fusion. Background Technology

[0002] Petroleum is an important energy source and a high-quality chemical raw material, and a vital strategic resource related to national livelihood. During crude oil extraction, to determine the crude oil production of each well and understand changes in formation oil and gas content and geological structure, it is necessary to continuously measure the volume ratio of each component in the produced fluid and provide real-time measurement data to optimize production parameters and improve recovery rates. Water cut is a crucial monitoring parameter in oil well production technology, playing a significant role in improving the economic efficiency of oil extraction.

[0003] In the 1980s, various online methods for measuring crude oil water content were put into use in oil fields, mainly including the capacitance method, conductivity method, microwave method, density method, and gamma-ray method. The capacitance method is accurate when the water content is less than 30%; the conductivity method can achieve an accuracy of ±3% under water content conditions of 50%-100%; microwaves have strong penetrating power and high measurement accuracy, with an accuracy within 2% when the water content is higher than 50% or lower than 20%; the density method has a measurement range of 0-100%, offering a wide range of water content measurement. However, due to limitations in technology and processes, the accuracy, stability, and reliability of these commonly used crude oil water content detection methods still have many problems.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] The purpose of this invention is to provide a crude oil water content detection method and system based on multi-sensor data fusion, which can achieve high-precision and full-range water content measurement of crude oil, and provide important water content reference value for oil well production.

[0006] This invention is achieved through the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting the water content of crude oil based on multi-sensor data fusion, the method comprising:

[0008] Crude oil water content data set (X1, X2, X3, X4) was collected using capacitive sensors, microwave sensors, conductivity sensors, and pressure sensors. The data set was then preprocessed to obtain a water content preprocessed data sequence.

[0009] By performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data series, the optimal hierarchy and order are obtained; based on the optimal hierarchy and order, an ARIMA time series moisture content residual model is constructed; the ARIMA time series moisture content residual model is used to calculate the information data to obtain the initial predicted value; and the moisture content residual dataset is calculated based on the pre-processed moisture content data series.

[0010] The residual data of water content is used as input to the LightGBM model to construct and train the LightGBM model for information data calculation, thereby obtaining the residual prediction sequence and further prediction values. The further prediction values ​​are then added to the initial prediction values ​​to obtain the final crude oil water content.

[0011] Furthermore, the capacitance sensor, microwave sensor, conductivity sensor, and pressure sensor correspond to the capacitance method, microwave method, conductivity method, and density method, respectively, of the traditional crude oil water content measurement methods.

[0012] Furthermore, by performing autocorrelation and partial autocorrelation analyses on the preprocessed moisture content data series, the optimal hierarchy and order were obtained, including:

[0013] The stationarity of the moisture content preprocessed data series was determined using the unit root test (ADF) method, and a stationary time series was obtained.

[0014] By performing autocorrelation and partial autocorrelation analysis on stationary time series, the optimal hierarchy and order can be obtained.

[0015] Furthermore, the formulas for autocorrelation and partial autocorrelation analysis are as follows:

[0016]

[0017]

[0018] Where ρi is the autocorrelation coefficient, Cov(·) is the autocovariance function, Var(·) is the variance function, and α ij It is a partial correlation coefficient.

[0019] Furthermore, the expression for the ARIMA time series moisture content residual model is as follows:

[0020]

[0021] Where Ψt is the stationary sequence after differencing, μ is the constant term, p is the autoregressive order, q is the moving average term, and ε is the constant term. t It is wandering white noise. It is an AR autoregressive model. It is a moving average (MA) model.

[0022] Furthermore, the Bayesian criterion (BIC) coefficients are used to construct the magnitude of the ARIMA time series moisture content residual model.

[0023] Further, the LightGBM model is trained, including:

[0024] The LightGBM model trains a basic learner f(x). The next learner will learn more from the data samples where the previous learner's prediction bias is large. The iterative training yields the crude oil water content output, which is used as the residual prediction sequence.

[0025] Furthermore, the LightGBM model is represented as a weighted sum function of multi-base learners, and the expression used to learn and train the LightGBM model is:

[0026]

[0027] in, Let f be the weight of the nth basic learner, x be the input data, and f be the weight of the nth basic learner. n (x) is the output of the nth base learner, F n (x) represents the total output of the LightGBM model.

[0028] Furthermore, training the LightGBM model also includes:

[0029] The GOSS and EFB algorithms were used to discretize and extract features from the training samples of the moisture content residual dataset.

[0030] Secondly, the present invention provides a crude oil water content detection system based on multi-sensor data fusion, which uses the aforementioned crude oil water content detection method based on multi-sensor data fusion; the system includes:

[0031] The acquisition and preprocessing unit is used to collect crude oil water content information dataset (X1, X2, X3, X4) based on capacitive sensors, microwave sensors, conductivity sensors and pressure sensors, and to preprocess the information dataset to obtain a water content preprocessed data sequence.

[0032] The initial prediction calculation unit is used to obtain the optimal hierarchy and order by performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data series; based on the optimal hierarchy and order, an ARIMA time series moisture content residual model is constructed; the initial prediction value is obtained by calculating the information data using the ARIMA time series moisture content residual model; and the moisture content residual dataset is calculated based on the pre-processed moisture content data series.

[0033] The final crude oil water content calculation unit is used to take the water content residual dataset as input to the LightGBM model, construct and train the LightGBM model to perform information data calculation, obtain the residual prediction sequence and further prediction values; and add the further prediction values ​​to the initial prediction values ​​to obtain the final crude oil water content.

[0034] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0035] This invention discloses a method and system for detecting crude oil water content based on multi-sensor data fusion. The invention fuses measurement datasets from capacitive, conductive, microwave, and pressure sensors, enabling accurate measurement of crude oil water content across the entire range. The data fusion employs a hybrid model combining the RIMA time-series water content residual model and the LightBGM model, which optimizes and rapidly processes measurement data from the four sensors. This improves the reliability and accuracy of the crude oil water content measurement model, expands the detection range, and enhances precision, better meeting the current needs of crude oil extraction for crude oil water content measurement technology. Attached Figure Description

[0036] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0037] Figure 1 This is a flowchart of a crude oil water content detection method based on multi-sensor data fusion according to the present invention.

[0038] Figure 2 This is a detailed flowchart of a crude oil water content detection method based on multi-sensor data fusion according to the present invention;

[0039] Figure 3 This is a flowchart of the ARIMA time series moisture content residual model of the present invention;

[0040] Figure 4 This is a flowchart of the LightBGM model algorithm of the present invention;

[0041] Figure 5 The images show the test results of the capacitance method, density method, microwave method, conductivity method, and the ARIMA-Light fusion method of this invention.

[0042] Figure 6 This is a structural block diagram of a crude oil water content detection system based on multi-sensor data fusion according to the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0044] Existing online methods for measuring crude oil water content mainly include capacitance, conductivity, microwave, density, and gamma-ray methods. The capacitance method is accurate when the water content is less than 30%; the conductivity method achieves an accuracy of ±3% under water content conditions of 50%-100%; microwaves have strong penetrating power and high accuracy, with an accuracy within 2% when the water content is above 50% or below 20%; the density method has a wide measurement range of 0-100%. However, due to limitations in technology and processes, the accuracy, stability, and reliability of these commonly used crude oil water content detection methods still have many problems. Therefore, this invention, considering the characteristics of crude oil production in my country, proposes combining multiple sensors to measure crude oil water content and improve detection accuracy.

[0045] Multi-sensor data fusion is a novel data processing method. Information fusion involves comprehensively processing information from multiple sensors or sources to arrive at a more accurate and reliable description. For example, the accuracy of capacitance measurement decreases significantly when the moisture content is above 30%, and the accuracy of conductivity measurement decreases significantly when the moisture content is below 30%. Multi-sensor data fusion combines these two methods, expanding the measurement range and improving accuracy. Multi-sensor data fusion methods can currently be divided into two categories: probabilistic and statistical methods and artificial intelligence methods. Probabilistic and statistical methods include estimation theory, Bayesian methods, and decision statistics theory; artificial intelligence methods include evidence theory, logical reasoning, and artificial neural networks.

[0046] In recent years, with the continuous research of many scholars, it has been found that sometimes single models are not thorough enough in mining information about the research object and their effects are not stable enough. At this time, research has turned to hybrid models. The Autoregressive Model of Differential Integrated Moving Average (ARIMA) is a model that transforms a non-stationary time series into a stationary time series, and then regresses the dependent variable only on its lagged values ​​and the present value and lagged values ​​of the random error term. LightGBM is a distributed gradient boosting (GBDT) framework based on decision tree algorithm, open-sourced by the Distributed Machine Learning Toolkit Team of Microsoft Research Asia. It supports highly efficient parallel training and has the advantages of faster training speed, lower memory consumption, better accuracy, and support for distributed processing of massive amounts of data. This invention proposes an ARIMA-LightGBM hybrid model, which has a good fusion processing effect on data generated by multiple sensors in traditional crude oil water content measurement. The novel multi-sensor data fusion method described provides ideas for new sensors required for crude oil water content measurement, which has important social and economic significance.

[0047] This invention discloses a crude oil water content detection method based on multi-sensor data fusion. First, the crude oil water content is measured using capacitance, conductivity, microwave, and density methods. Then, the measurement results of these four methods are fused using an ARIMA-LightGBM hybrid model to obtain the crude oil water content. Specifically, the ARIMA time-series model is used to extract linear features from the data, while the LightGBM regression model is used to extract nonlinear features. The hybrid model, combining both methods, possesses both linear and nonlinear processing capabilities. In traditional crude oil water content measurement methods, capacitance is suitable for low water content crude oil, conductivity and microwave methods are suitable for high water content crude oil, and density is suitable for crude oil with a wide range of water content but lacks accuracy. This invention utilizes the ARIMA-LightGBM hybrid model to fuse measurement data from four sensors, expanding the detection range and improving accuracy for crude oil water content.

[0048] Example 1

[0049] like Figure 1 and Figure 2 As shown, this invention discloses a method for detecting crude oil water content based on multi-sensor data fusion, the method comprising:

[0050] Step 1: Collect crude oil water content data set (X1, X2, X...) using capacitive sensors, microwave sensors, conductivity sensors, and pressure sensors. 3, X4), and preprocess the information dataset to obtain the moisture content preprocessed data sequence;

[0051] In step 1, the four types of sensors (capacitive sensor, microwave sensor, conductivity sensor, and pressure sensor) correspond to the traditional crude oil water content measurement methods: capacitance method, microwave method, conductivity method, and density method, respectively.

[0052] This approach takes into account the measurement range by combining the capacitance method (suitable for low water content), the conductivity method (suitable for high water content), and the density method (suitable for the entire range but with lower accuracy) to improve the accuracy of crude oil water content across the entire range. Furthermore, the measurement methods combine contact measurement methods such as capacitance, conductivity, and density with non-contact measurement methods such as microwave to compensate for the impact of salinity on accuracy.

[0053] In this embodiment, crude oil water content samples were prepared by adding emulsifier to diesel oil and tap water to form 11 crude oil samples with water content of 0%, 10%, 20%, ... 100%. 120 ml of each water content level was prepared for testing at various temperatures.

[0054] The above four sensor data acquisition methods are all known methods, and the acquired datasets are all voltage data. The capacitance sensor changes capacitance based on the different water content being measured, which in turn changes the voltage value after detection. The microwave sensor can replace the relatively difficult measurement of microwave power by measuring the voltage amplitude. The magnitude of the voltage sensed by the conductivity sensor is inversely proportional to the conductivity of the fluid flowing through the sensor. The pressure sensor can reflect the density of crude oil and thus the water content by calculating the pressure difference.

[0055] The preprocessing steps for collected data include outlier handling and missing value handling.

[0056] Specifically, the water content of crude oil samples was tested using known methods, including capacitance, density, microwave, and conductivity methods. The test results for 30 samples are as follows: Figure 5 As shown.

[0057] Step 2: By performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data series, the optimal hierarchy and order are obtained; based on the optimal hierarchy and order, an ARIMA time series moisture content residual model is constructed; the ARIMA time series moisture content residual model is used to calculate the information data to obtain the initial predicted value; and the moisture content residual dataset is calculated based on the pre-processed moisture content data series.

[0058] Specifically, the expression for the ARIMA time series moisture content residual model is as follows:

[0059]

[0060] Where Ψt is the stationary sequence after differencing, μ is the constant term, p is the autoregressive order, q is the moving average term, and ε is the constant term. t It is wandering white noise. It is an AR autoregressive model. It is a moving average (MA) model.

[0061] Specifically, in step 2, the optimal hierarchy and order are obtained by performing autocorrelation and partial autocorrelation analysis on the moisture content preprocessing data sequence, including:

[0062] Step A: Determine the stationarity of the pre-processed moisture content data series using the unit root test (ADF) method to obtain a stationary time series. Specifically, the stationarity of the pre-processed moisture content data series is determined by the unit root test (ADF). If the obtained statistic is significantly lower than the critical statistic at three confidence levels (1%, 5%, 10%), then the original series is rejected as stationary, and the series is stationary through finite-order difference operations.

[0063] Step B involves performing autocorrelation and partial autocorrelation analysis on the stationary time series to obtain the optimal hierarchy and order.

[0064] The optimal order and rank are determined by calculating the sample autocorrelation function and partial autocorrelation function based on the observed values ​​of the moisture content sample.

[0065] Specifically, the formulas for autocorrelation and partial autocorrelation analysis in step 2 are as follows:

[0066]

[0067]

[0068] Where ρi is the autocorrelation coefficient, Cov(·) is the autocovariance function, Var(·) is the variance function, and α ij It is a partial correlation coefficient.

[0069] Specifically, when training the ARIMA time series water content residual model, increasing the values ​​of parameters p and q increases the model's complexity and leads to overfitting. Therefore, this invention uses Bayes' criterion (BIC) coefficients to construct the magnitude of the ARIMA time series water content residual model. The BIC measurement standard formula (4) is expressed as follows:

[0070] BIC=kln(n)-2ln(L) (4)

[0071] Specifically, the flowchart for constructing the ARIMA time series moisture content residual model in step 2 is as follows: Figure 3 As shown, the white noise test was performed on the moisture content residual data to determine that the established model can effectively extract information from the original data.

[0072] Step 3: Use the water content residual dataset as input to the LightGBM model, construct and train the LightGBM model to perform information data calculation, obtain the residual prediction sequence and further prediction values; and add the further prediction values ​​to the initial prediction values ​​to obtain the final crude oil water content.

[0073] Specifically, step 3 involves training the LightGBM model, including:

[0074] The LightGBM model trains a basic learner f(x). The next learner will learn more from the data samples where the previous learner's prediction bias is large. The iterative training yields the crude oil water content output, which is used as the residual prediction sequence.

[0075] In this embodiment, the algorithm flowchart of the LightBGM model is as follows: Figure 4 As shown, the LightGBM model is represented as a weighted sum function of multiple base learners, and the expression used to learn and train the LightGBM model is:

[0076]

[0077] in, Let f be the weight of the nth basic learner, x be the input data, and f be the weight of the nth basic learner. n (x) is the output of the nth base learner, F n (x) represents the total output of the LightGBM model.

[0078] In this embodiment, training the LightGBM model further includes:

[0079] A gradient-based one-sided sampling algorithm (GOSS algorithm) and an exclusive feature binding algorithm (EFB) are used to discretize and extract features from the training samples of the water content residual dataset. Specifically, the water content training data is first sorted according to its gradient, and a predetermined proportion is set to retain data samples with gradients higher than this proportion. Data samples with gradients lower than this proportion are not discarded directly, but are sampled proportionally from the samples with smaller gradients. When calculating information gain, the dataset with smaller gradients is multiplied by a coefficient to amplify the information. Specifically, the EFB algorithm extracts features from the water content training samples by adding an offset constant to different features within the bound feature clusters, so that the values ​​of different features are distributed in different value spaces.

[0080] The parameters determine the quality of the trained model. Appropriate settings include the maximum depth of the decision tree, the minimum number of records per leaf, and the learner weight reduction factor. Specifically, the node with the highest benefit is selected in the decision tree tree growth strategy, and a limit is set on the maximum leaf depth to prevent overfitting caused by excessively deep leaves.

[0081] By employing a histogram-based partitioning point search algorithm, feature values ​​are discretized into several value spaces, reducing the algorithmic complexity of finding the optimal split point, effectively lowering computational and storage costs, and enabling faster model training.

[0082] The final crude oil water content (i.e., the final crude oil water content) is the sum of the initial prediction from the ARIMA model and the prediction from the LightGBM model. The test results for 30 crude oil samples are as follows: Figure 5 As shown, the prediction results are better than those of the four traditional moisture content methods.

[0083] This invention fuses measurement datasets from capacitive, conductive, microwave, and pressure sensors to achieve accurate measurement of crude oil water content across the entire range. The data fusion employs a hybrid model combining the RIMA time-series water content residual model and the LightBGM model, which optimizes and rapidly processes measurement data from the four sensors. This improves the reliability and accuracy of the crude oil water content measurement model, better meeting the current needs of crude oil extraction for crude oil water content measurement technology.

[0084] Example 2

[0085] like Figure 6 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a crude oil water content detection system based on multi-sensor data fusion. This system uses a crude oil water content detection method based on multi-sensor data fusion from Embodiment 1. The system includes:

[0086] The acquisition and preprocessing unit is used to collect crude oil water content information dataset (X1, X2, X3, X4) based on capacitive sensors, microwave sensors, conductivity sensors and pressure sensors, and to preprocess the information dataset to obtain a water content preprocessed data sequence.

[0087] The initial prediction calculation unit is used to obtain the optimal hierarchy and order by performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data series; based on the optimal hierarchy and order, an ARIMA time series moisture content residual model is constructed; the initial prediction value is obtained by calculating the information data using the ARIMA time series moisture content residual model; and the moisture content residual dataset is calculated based on the pre-processed moisture content data series.

[0088] The final crude oil water content calculation unit is used to take the water content residual dataset as input to the LightGBM model, construct and train the LightGBM model to perform information data calculation, obtain the residual prediction sequence and further prediction values; and add the further prediction values ​​to the initial prediction values ​​to obtain the final crude oil water content.

[0089] The execution process of each unit can be carried out according to the steps of the crude oil water content detection method based on multi-sensor data fusion in Example 1, and will not be described in detail in this example.

[0090] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0094] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting crude oil water content based on multi-sensor data fusion, characterized in that, The method includes: Crude oil water content data sets are collected using capacitive sensors, microwave sensors, conductivity sensors, and pressure sensors. The data sets are then preprocessed to obtain a water content preprocessed data sequence. By performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data sequence, the optimal hierarchy and order are obtained; based on the optimal hierarchy and order, an ARIMA time series moisture content residual model is constructed; the ARIMA time series moisture content residual model is used to calculate the information data to obtain the initial predicted value; and the moisture content residual dataset is calculated based on the pre-processed moisture content data sequence. The water content residual dataset is used as input to the LightGBM model to construct and train the LightGBM model for information data calculation, thereby obtaining the residual prediction sequence and further prediction values; and the further prediction values ​​are added to the initial prediction values ​​to obtain the final crude oil water content. The capacitance sensor, microwave sensor, conductivity sensor, and pressure sensor correspond to the capacitance method, microwave method, conductivity method, and density method, respectively, of the traditional crude oil water content measurement methods. Training the LightGBM model includes: The LightGBM model trains a basic learner f(x), and the next learner will learn more from the data samples where the previous learner's prediction bias is large. The iterative training yields the crude oil water content output and serves as the residual prediction sequence. Training the LightGBM model also includes: The GOSS algorithm and EFB algorithm are used to discretize and extract features from the training samples of the moisture content residual dataset.

2. The crude oil water content detection method based on multi-sensor data fusion according to claim 1, characterized in that, By performing autocorrelation and partial autocorrelation analyses on the preprocessed moisture content data series, the optimal hierarchy and order were obtained, including: The stationarity of the moisture content preprocessed data series was determined using the unit root test method to obtain a stationary time series; By performing autocorrelation and partial autocorrelation analysis on the stationary time series, the optimal hierarchy and order are obtained.

3. A method for detecting crude oil water content based on multi-sensor data fusion according to claim 1 or 2, characterized in that, The formulas for the autocorrelation and partial autocorrelation analyses are as follows: ； ； Where ρi is the autocorrelation coefficient, Cov(·) is the autocovariance function, Var(·) is the variance function, and α ij It is a partial correlation coefficient.

4. The crude oil water content detection method based on multi-sensor data fusion according to claim 1, characterized in that, The expression for the ARIMA time series moisture content residual model is as follows: ； Where Ψt is the stationary sequence after differencing, μ is the constant term, p is the autoregressive order, q is the moving average term, and ε is the constant term. t It is wandering white noise. It is an AR autoregressive model. It is a moving average (MA) model.

5. The crude oil water content detection method based on multi-sensor data fusion according to claim 4, characterized in that, The size of the ARIMA time series moisture content residual model is determined by the Bayesian criterion coefficients.

6. The crude oil water content detection method based on multi-sensor data fusion according to claim 1, characterized in that, The LightGBM model is represented as a weighted sum function of multi-base learners, and the expression used to learn and train the LightGBM model is: ； in, n is the weight of the nth basic learner, x is the input data, and f is the weight of the nth basic learner. n (x) is the output of the nth base learner, F n (x) represents the total output of the LightGBM model.

7. A crude oil water content detection system based on multi-sensor data fusion, characterized in that, The system uses a crude oil water content detection method based on multi-sensor data fusion as described in any one of claims 1 to 6; the system includes: The acquisition and preprocessing unit is used to collect crude oil water content information datasets based on capacitive sensors, microwave sensors, conductivity sensors and pressure sensors, and to preprocess the information datasets to obtain a water content preprocessed data sequence. The initial prediction calculation unit is used to obtain the optimal hierarchy and order by performing autocorrelation and partial autocorrelation analysis on the pre-processed moisture content data sequence; construct an ARIMA time series moisture content residual model based on the optimal hierarchy and order; calculate the initial prediction value using the ARIMA time series moisture content residual model; and calculate the moisture content residual dataset based on the pre-processed moisture content data sequence. The final crude oil water content calculation unit is used to take the water content residual dataset as input to the LightGBM model, construct and train the LightGBM model to perform information data calculation, obtain the residual prediction sequence and further prediction values; and add the further prediction values ​​to the initial prediction values ​​to obtain the final crude oil water content.

Images