System for detecting technological parameters in petroleum production process

Abstract

The invention provides a petroleum production process technological parameter detection system, which comprises a data acquisition module for acquiring various technological parameters including but not limited to temperature, pressure, flow, concentration and vibration parameters in a petroleum production process by using a sensor; a data processing unit; a user terminal module; compared with the prior art, the optimization control module has the following beneficial effects that a multi-sensor data fusion technology is adopted, the problem that the monitoring precision of a single sensor is insufficient is solved, and more accurate real-time monitoring data is provided. Through the analysis and optimization functions of the intelligent data processing unit, automatic process control is supported, and production parameters can be automatically adjusted when abnormity occurs, so that the production efficiency and safety are improved, the system integrates the multi-dimensional data preprocessing, analysis, fusion and optimization functions, and an omnibearing process parameter monitoring closed loop is formed.

Description

Technical Field

[0001] The invention relates to data monitoring and processing in the oil production process and is a detection system for real-time monitoring and analysis of oil production process parameters. Background Art

[0002] The oil production process involves multiple complex process links, such as crude oil extraction, transportation, storage, etc., each of which requires precise parameter control and real-time monitoring. At present, traditional oil production process parameter detection systems mostly rely on single sensors and monitoring equipment, which have problems such as low detection accuracy, poor real-time performance, and inability to perform multi-parameter fusion analysis, making it difficult to achieve intelligent and precise management. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention aims to provide a system for detecting process parameters of an oil production process to solve the problems raised in the above-mentioned background technology.

[0004] The present invention is implemented by the following technical scheme: A system for detecting process parameters of an oil production process, comprising:

[0005] Data acquisition module: uses sensors to collect various process parameters in the oil production process, including but not limited to temperature, pressure, flow, concentration, and vibration parameters;

[0006] Data processing unit: used to receive multi-dimensional data from the data acquisition module, and process the collected multi-dimensional data in real time through data fusion algorithm to obtain optimized process parameter information;

[0007] User terminal module: By connecting with the data processing unit, it can display process parameter monitoring data, trend curves, early warning information, etc. in real time, and support functions such as data query, historical data playback, and report generation;

[0008] Optimization control module: Based on real-time monitoring data, it provides process adjustment suggestions or automated control decisions to optimize the oil production process.

[0009] As a preferred embodiment, the output end of the data acquisition module is connected to the input end of the data processing unit, the output end of the data processing unit is connected to the input end of the user terminal module, and the output end of the data processing unit is also connected to the input end of the optimization control module.

[0010] As a preferred implementation, the data processing unit includes:

[0011] Data preprocessing module: denoises, removes bias, and normalizes the data collected from the sensor to ensure the accuracy and reliability of the data;

[0012] Data fusion module: adopts multi-sensor data fusion algorithm to comprehensively process the data collected by different sensors to improve detection accuracy and reliability;

[0013] Data analysis module: analyzes the fused data, builds models based on historical data, performs trend prediction and anomaly detection, supports real-time monitoring and feedback, and can mark abnormal data according to set thresholds and output relevant diagnostic information;

[0014] Early warning module: When the data analysis module detects that a process parameter deviates from the set range, the early warning mechanism is triggered and a warning signal is sent to the operator in real time, prompting him to make necessary operational adjustments.

[0015] As a preferred embodiment, the data preprocessing module cleans, denoises, standardizes and normalizes the raw data collected from different sensors or monitoring devices to ensure that the quality of the input data is suitable for further analysis and modeling. The steps of data preprocessing include:

[0016] Data denoising and outlier detection: Identify and remove outliers in data through statistical methods;

[0017] Data filling: fill in missing data;

[0018] Data standardization and normalization: Different sensors have different measurement units. The standardization and normalization steps ensure that the data has a uniform scale, making subsequent analysis more reliable. Standardization uses Z-score standardization, and normalization converts the data to a specified range.

[0019] Data conversion: Convert certain specific data to ensure that they meet the requirements of model input, for example, converting timestamps to a unified date format, or converting data from a periodically changing state to a rate of change, such as pressure or temperature.

[0020] As a preferred implementation, the data fusion module fuses data from different sensors or monitoring devices to form a unified and more representative information output. Data fusion technology can eliminate possible errors in a single sensor and improve the reliability and accuracy of the data.

[0021] The steps of data fusion include:

[0022] Multi-sensor data alignment: Since the timestamps of different sensors may not be completely aligned, the data needs to be time-aligned to ensure that the data is consistent in time;

[0023] Weighted average: weighted average of data output by multiple sensors;

[0024] Data fusion algorithm: Use Kalman filter data fusion algorithm to perform real-time estimation in dynamic systems;

[0025] Data fusion output: The fused data will serve as input for subsequent analysis, modeling and prediction.

[0026] As a preferred implementation, the data analysis module is responsible for analyzing the fused multi-dimensional data to achieve trend prediction and anomaly detection. The specific composition and workflow of this module are as follows:

[0027] Data cleaning and standardization: After data fusion, the data is first cleaned and standardized. The accuracy and consistency of the input data are ensured by denoising, removing outliers and standardizing various parameters;

[0028] Establish a forecasting model based on historical data: Use the ARIMA model to forecast the trend of oil production process parameters;

[0029] Anomaly detection and diagnosis: Use the Gaussian distribution model to detect anomalies in the current process parameters. By assuming that the data conforms to the normal distribution, the probability density of each data point is calculated, and abnormal values ​​that deviate from the normal range are detected based on the set threshold.

[0030] As a preferred implementation, it also includes a cloud computing and edge computing fusion architecture, which is bidirectionally data connected to the data acquisition module set and the data processing unit.

[0031] The connection and working mode between the data acquisition module and the edge computing node are as follows:

[0032] The data acquisition module is responsible for collecting various process parameters in the oil production process through sensors. These data are transmitted to the edge computing node located on site. The edge computing node includes a local processing unit, which is responsible for receiving, preprocessing and preliminarily analyzing the data from the data acquisition module;

[0033] The connection and working mode between the data processing unit and the edge computing node are as follows:

[0034] After the edge computing node completes the data preprocessing, the processed data is transmitted to the data processing unit, and the data processing unit processes the multidimensional data from the edge computing node.

[0035] After adopting the above technical solution, the beneficial effects of the present invention are as follows: the present invention adopts multi-sensor data fusion technology to make up for the problem of insufficient monitoring accuracy of a single sensor and provide more accurate real-time monitoring data. Through the analysis and optimization functions of the intelligent data processing unit, it supports automated process control and can automatically adjust production parameters when an abnormality occurs, thereby improving production efficiency and safety. The system integrates multi-dimensional data preprocessing, analysis, fusion, and optimization functions to form a comprehensive closed loop of process parameter monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0037] Figure 1 The present invention is a system block diagram of a petroleum production process process parameter detection system.

[0038] In the figure, 100 is a data acquisition module, 200 is a data processing unit, 300 is a user terminal module, and 400 is an optimization control module. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] See also Figure 1 The present invention provides a technical solution: a system for detecting process parameters of an oil production process, comprising:

[0041] Data acquisition module 100: uses sensors to collect various process parameters in the oil production process, including but not limited to temperature, pressure, flow, concentration, and vibration parameters;

[0042] The data processing unit 200 is used to receive the multi-dimensional data from the data acquisition module 100, and to process the collected multi-dimensional data in real time through a data fusion algorithm to obtain optimized process parameter information;

[0043] User terminal module 300: by connecting with the data processing unit 200, it can display the process parameter monitoring data, trend curve, warning information, etc. in real time, and support functions such as data query, historical data playback, and report generation;

[0044] Optimization control module 400: Based on real-time monitoring data, it provides process adjustment suggestions or automated control decisions to optimize the oil production process.

[0045] The output end of the data acquisition module 100 is connected to the input end of the data processing unit 200 , the output end of the data processing unit 200 is connected to the input end of the user terminal module 300 , and the output end of the data processing unit 200 is also connected to the input end of the optimization control module 400 .

[0046] The data processing unit 200 includes:

[0047] Data preprocessing module: denoises, removes bias, and normalizes the data collected from the sensor to ensure the accuracy and reliability of the data;

[0048] Data fusion module: adopts multi-sensor data fusion algorithm to comprehensively process the data collected by different sensors to improve detection accuracy and reliability;

[0049] Data analysis module: analyzes the fused data, builds models based on historical data, performs trend prediction and anomaly detection, supports real-time monitoring and feedback, and can mark abnormal data according to set thresholds and output relevant diagnostic information;

[0050] Early warning module: When the data analysis module detects that a process parameter deviates from the set range, the early warning mechanism is triggered and a warning signal is sent to the operator in real time, prompting him to make necessary operational adjustments.

[0051] As an embodiment of the present invention, it is embodiment 1:

[0052] The data preprocessing module cleans, denoises, standardizes and normalizes the raw data collected from different sensors or monitoring devices to ensure that the input data quality is suitable for further analysis and modeling. The steps of data preprocessing include:

[0053] Data denoising and outlier detection: Identify and remove outliers in data through statistical methods;

[0054] Data filling: fill in missing data;

[0055] Data standardization and normalization: Different sensors have different measurement units. The standardization and normalization steps ensure that the data has a uniform scale, making subsequent analysis more reliable. Standardization uses Z-score standardization, and normalization converts the data to a specified range.

[0056] Data conversion: Convert certain specific data to ensure that they meet the requirements of model input, for example, converting timestamps to a unified date format, or converting data from a periodically changing state to a rate of change, such as pressure or temperature.

[0057] The data fusion module fuses data from different sensors or monitoring devices to form a unified and more representative information output. Data fusion technology can eliminate possible errors in a single sensor and improve the reliability and accuracy of the data.

[0058] The steps of data fusion include:

[0059] Multi-sensor data alignment: Since the timestamps of different sensors may not be completely aligned, the data needs to be time-aligned to ensure that the data is consistent in time;

[0060] Weighted average: weighted average of data output by multiple sensors;

[0061] Data fusion algorithm: Use Kalman filter data fusion algorithm to perform real-time estimation in dynamic systems;

[0062] Data fusion output: The fused data will serve as input for subsequent analysis, modeling and prediction.

[0063] Suppose we have a monitoring system for an oil production site, which includes multiple sensors for measuring different process parameters such as temperature, pressure, and flow. These sensor data are noisy, have missing values, and have inconsistent units. Our goal is to process this data through the data preprocessing module and fuse the data from multiple sensors through the data fusion module to obtain accurate process parameters.

[0064] Raw data:

[0065] Table 1 shows some of the data collected from two temperature sensors, one pressure sensor and one flow sensor (unit: temperature °C, pressure MPa, flow m 3 / h). Assume that the sensors have different timestamps and there are missing data and outliers.

[0066]

[0067] Table 1

[0068] The data of temperature sensor 1 on 2024-11-04 has an abnormal value (20.0°C).

[0069] The data of temperature sensor 2 on 2024-11-03 is missing (NaN).

[0070] The flow sensor data on 2024-11-04 is missing (NaN).

[0071] Step 1: Data Preprocessing

[0072] 1.1 Data denoising and outlier detection:

[0073] Abnormal value of temperature sensor 1: The temperature data recorded on 2024-11-04 is 20.0°C, which is significantly lower than the surrounding data (around 35°C). We regard it as an abnormal value and remove it or fill it with the mean of adjacent data.

[0074] Eliminate abnormal data: The data of 2024-11-04 of temperature sensor 1 is abnormal and will be deleted.

[0075] 1.2 Data filling:

[0076] Interpolate and fill the missing data (2024-11-03) of temperature sensor 2 using linear interpolation:

[0077]

[0078] Perform mean filling on the missing data of the flow sensor (2024-11-04) and calculate the mean of the flow sensor:

[0079]

[0080] The flow data filled on 2024-11-04 is 51.5m3 / h.

[0081] 1.3 Data standardization and normalization:

[0082] Standardize each parameter using the Z-score standardization formula:

[0083]

[0084] Here, μ is the mean and σ is the standard deviation.

[0085] ,Assume that the mean of temperature data is 35.4 and the standard deviation is 0.2, the mean of pressure data is 10.7 and the standard deviation is 0.2, and the mean of flow data is 51.5 and the standard deviation is 1.0.

[0086] Normalize the data of temperature sensor 1:

[0087]

[0088] The same method was used to standardize other data.

[0089] Step 2: Data Fusion

[0090] 2.1 Data Alignment:

[0091] Align all data to the same timestamp. Time-align temperature, pressure, and flow data using interpolation methods.

[0092] 2.2 Weighted average:

[0093] Assume that the weights of temperature sensors 1 and 2 are 0.6 and 0.4 respectively, and the weights of pressure sensor and flow sensor are 0.5 respectively.

[0094] Take a weighted average of the temperature data:

[0095] Similar weighted fusion is performed for other parameters (such as pressure and flow).

[0096] T 融合 =0.6*T 温度传感器1 +0.4*T 温度传感器2

[0097] 2.3 Output after fusion:

[0098] The weighted average temperature fusion value is 36.2℃, the pressure fusion value is 10.7MPa, and the flow fusion value is 51.8m 3 / h. The fused data is used for subsequent trend prediction and anomaly detection.

[0099] As an embodiment of the present invention, this is Embodiment 2:

[0100] The data analysis module is responsible for analyzing the fused multi-dimensional data to achieve trend prediction and anomaly detection. The specific composition and workflow of this module are as follows:

[0101] Data cleaning and standardization: After data fusion, the data is first cleaned and standardized. The accuracy and consistency of the input data are ensured by denoising, removing outliers and standardizing various parameters;

[0102] Establish a forecasting model based on historical data: Use the ARIMA model to forecast the trend of oil production process parameters;

[0103] Anomaly detection and diagnosis: Use the Gaussian distribution model to detect anomalies in the current process parameters. By assuming that the data conforms to the normal distribution, the probability density of each data point is calculated, and abnormal values ​​that deviate from the normal range are detected based on the set threshold.

[0104] Trend prediction algorithm: ARIMA model

[0105] Assume that the time series of a process parameter in the oil production process is X = {x1, x2, ..., x t}, where x t Represents the process parameters at time t (such as temperature, pressure, flow, etc.). The basic form of the ARIMA model can be expressed as:

[0106] X t =φ1X t-1 +φ2X t-2 +…+φ p X t-p +θ1∈ t-1 +θ2∈ t-2 +…+θ q ∈ t-q +∈ t in,

[0107] φ1,φ2,…,φ p is the parameter of the autoregressive (AR) term

[0108] θ1,θ2,…,θ q are the parameters of the moving average (MA) term,

[0109] ∈ t is the white noise error term.

[0110] The core idea of ​​the ARIMA model is to predict future data by modeling the autoregression and moving average relationship of historical data.

[0111] The three main parameters of the ARIMA model are:

[0112] p: the order of the autoregressive term,

[0113] d: difference order, used to make the non-stationary series become a stationary series,

[0114] q: The order of the moving average term.

[0115] Trend prediction formula

[0116] We use historical data X history ={x1,x2,…,x n An ARIMA model was trained, and its prediction formula is:

[0117]

[0118] here, It represents the predicted value of the process parameters at the next time point t=n+1. is the prediction error term.

[0119] Anomaly Detection and Diagnosis Methods: Gaussian Distribution Model

[0120] Anomaly detection is to find abnormal values ​​that deviate from the normal range by comparing the distribution of current data with historical data. In this invention, we assume that the process parameter data obeys the normal distribution (Gaussian distribution) and determine whether it is abnormal by calculating the probability density of the current data point.

[0121] Basic assumptions of the Gaussian distribution model

[0122] Assume that a process parameter data x t It obeys Gaussian distribution, and its probability density function is:

[0123]

[0124] Process parameter value.

[0125] Outlier Detection Formula

[0126] In order to detect outliers, we calculate the standardized value (Z-score) of each data point to determine whether it deviates from the normal distribution range. The Z-score formula is as follows:

[0127]

[0128] When | Z t |Exceeding a certain threshold Z threshold When the data point x t are considered as outliers.

[0129] Anomaly detection and diagnosis

[0130] Assume that the temperature data in a certain oil production process has been trend predicted by the ARIMA model, and after data fusion, we get the following process temperature data (unit: ℃):

[0131] X history ={35.1,35.3,35.2,35.5,35.4,35.7,36.0,35.6,50.2,35.8,35.9}

[0132] 1. Calculate the mean and variance:

[0133] First, calculate the mean and variance of the data: Mean μ = 35.57 Variance σ 2 =0.21·Standard deviationσ=√0.21≈0.458

[0134] 2. Calculate Z-score:

[0135] Now, calculate the Z-score for a data point in the temperature data (for example, 50.2°C):

[0136]

[0137] 3. Detect anomalies:

[0138] Assume that the threshold value is set to Z threshold =3, since Z=31.99 far exceeds the threshold value 3, this data point (50.2°C) is determined to be an outlier.

[0139] As an embodiment of the present invention, this is Embodiment 3:

[0140] The working principle of the oil production process parameter detection system is completed through four main modules: data preprocessing module, data fusion module, data analysis module and early warning module. These four modules work together to ensure that the system can monitor, predict and diagnose key parameters in the oil production process in real time, and ensure the safety and stability of production.

[0141] Specific:

[0142] Data preprocessing module: cleans, denoises, fills in missing values ​​and standardizes the raw data collected from various sensors.

[0143] Data fusion module: aligns and weightedly fuses data from multiple sensors to provide more accurate process parameters.

[0144] Data analysis module: predict future data through trend prediction algorithms (such as ARIMA) and identify abnormal data points through anomaly detection algorithms (such as Gaussian distribution model).

[0145] Early warning module: Based on the prediction results and anomaly detection, the system will trigger an early warning to prompt operators to take corresponding control measures.

[0146] In the oil production process, temperature, pressure and flow are key process parameters. Real-time monitoring of these parameters is essential to ensure the safety of the production process. Assume that we have three sensors to monitor these three parameters and use the oil production process process parameter detection system for data analysis and early warning, as shown in Table 2.

[0147]

[0148] Table 2

[0149] Step 1: Data Preprocessing

[0150] 1. Data denoising and outlier detection:

[0151] The 2024-11-04 data of temperature sensor 1 (20.0℃) is much lower than the surrounding temperature data (around 35℃), so we regard it as an outlier and remove it.

[0152] The data of 2024-11-03 of temperature sensor 2 is missing (NaN), and we use linear interpolation to fill it:

[0153]

[0154] The data of 2024-11-04 of the flow sensor is missing (NaN), so we use the mean filling method to fill it. The mean of the flow data is:

[0155] Average flow rate = (50 + 52 + 51 + 53) / 4 = 51.5 m 3 / h

[0156] Data Standardization:

[0157] The means and standard deviations of temperature, pressure, and flow are as follows:

[0158] Temperature mean: 35.4℃, standard deviation: 0.2

[0159] Pressure mean: 10.7MPa, standard deviation: 0.2

[0160] Average flow rate: 51.5m 3 / h, standard deviation: 1.0

[0161] Z-score normalization was performed for each parameter.

[0162] Step 2: Data Fusion

[0163] Data alignment:

[0164] Ensure that the timestamps of different sensors are aligned. For example, if the data of temperature sensor 1 is missing on 2024-11-03, fill it in through interpolation.

[0165] Weighted Average:

[0166] Assume that the weights of temperature sensors 1 and 2 are 0.6 and 0.4 respectively, and the weights of pressure sensor and flow sensor are 0.5 respectively.

[0167] Take a weighted average of the temperature data:

[0168] T 融合 =0.6*T 温度传感器1 +0.4*T 温度传感器2

[0169] Step 3: Data Analysis

[0170] Trend Forecasting (using ARIMA):

[0171] Assume that the temperature data is modeled by the ARIMA model, and the future temperature trend is predicted based on the past temperature data. For example, through historical data and the ARIMA model, we can predict that the temperature on 2024-11-06 will be 35.5℃.

[0172] Anomaly detection (using Gaussian distribution model):

[0173] Use the Gaussian distribution model to detect anomalies in temperature, pressure, and flow data. If the predicted temperature value is 35.5°C, and the value measured by a sensor is 20°C, the Z-score value will be very large (close to or greater than 3), so the value is judged to be abnormal.

[0174] Step 4: Early Warning

[0175] If the predicted temperature trend deviates significantly from the real-time data, or if the data point of a sensor deviates from the normal value (Z-score is greater than the set threshold), the system will trigger an early warning. For example, when the real-time data of the flow sensor exceeds the set upper limit (such as 55m 3 / h), the system will automatically sound an alarm to remind the operator to check the equipment.

[0176] In the above, the data preprocessing module ensures the accuracy of the data by cleaning, interpolating and filling missing data. The data fusion module provides more accurate process parameters by combining the data of multiple sensors through weighted average. The data analysis module uses the ARIMA model for trend prediction and the Gaussian distribution model for anomaly detection, and the early warning module automatically generates early warnings based on the prediction and anomaly detection results.

[0177] As an embodiment of the present invention, this is Embodiment 4:

[0178] During the oil production process, changes in pressure and temperature will affect the working state of the equipment. We have multiple sensors monitoring these two key parameters, and the system processes data and provides early warnings, as shown in Table 3.

[0179]

[0180] Step 1: Data Preprocessing

[0181] 1. Data denoising and outlier detection:

[0182] The 2024-11-04 data (20.0°C) of temperature sensor 1 is significantly lower than other temperature data, so we treat it as an outlier and delete it. For the missing (NaN) data of pressure sensor 1 on 2024-11-03, we use interpolation to fill it:

[0183]

[0184] 2. Data Standardization:

[0185] Assume that the mean and standard deviation of the temperature and pressure data are as follows:

[0186] Pressure mean: 10.7MPa, standard deviation: 0.2

[0187] Temperature mean: 35.4℃, standard deviation: 0.2

[0188] The data were standardized to ensure that the data in different units were on the same scale.

[0189] Step 2: Data Fusion

[0190] 1. Weighted average:

[0191] Assume that the weights of pressure sensor 1 and sensor 2 are 0.6 and 0.4 respectively, and the weights of temperature sensor 1 and sensor 2 are 0.5 respectively.

[0192] Perform a weighted average on the pressure data:

[0193] P 融合 =0.6×P 压力传感器1 +0.4×P 庄

[0194] Take a weighted average of the temperature data:

[0195] T 融合 =0.5×T 温度传感器1 +0.5×T 温1

[0196] Step 3: Data Analysis

[0197] Trend Forecasting (using ARIMA):

[0198] Use the ARIMA model to model the historical data of temperature and pressure and predict future data trends. Assume that the predicted future temperature is 35.6℃ and the predicted pressure is 10.95MPa.

[0199] Anomaly detection (using Gaussian distribution model):

[0200] Anomaly detection is performed based on the difference between the predicted value and the actual measured value. If the difference between the actual value and the predicted value exceeds the set threshold, the value is considered anomaly.

[0201] Step 4: Early Warning

[0202] When the real-time data of pressure or temperature exceeds the set safety threshold, the system will trigger an early warning. For example, when the predicted pressure exceeds 11.0MPa, the system will automatically alarm and prompt the operator to make adjustments.

[0203] As an embodiment of the present invention, it is embodiment 5:

[0204] It also includes a cloud computing and edge computing fusion architecture, which has a bidirectional data connection with a data acquisition module set and a data processing unit.

[0205] The connection and working mode between the data acquisition module and the edge computing node are as follows:

[0206] The data acquisition module is responsible for collecting various process parameters in the oil production process through sensors. These data are transmitted to the edge computing node located on site. The edge computing node includes a local processing unit, which is responsible for receiving, preprocessing and preliminarily analyzing the data from the data acquisition module;

[0207] The connection and working mode between the data processing unit and the edge computing node are as follows:

[0208] After the edge computing node completes the data preprocessing, the processed data is transmitted to the data processing unit, and the data processing unit processes the multidimensional data from the edge computing node.

[0209] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A process parameter detection system for oil production process, characterized in that: include: Data acquisition module: uses sensors to collect various process parameters in the oil production process, including but not limited to temperature, pressure, flow, concentration, and vibration parameters; Data processing unit: used to receive multi-dimensional data from the data acquisition module, and process the collected multi-dimensional data in real time through data fusion algorithm to obtain optimized process parameter information; User terminal module: By connecting with the data processing unit, it can display process parameter monitoring data, trend curves, and early warning information in real time, and support data query, historical data playback, and report generation functions; Optimization control module: Based on real-time monitoring data, it provides process adjustment suggestions or automated control decisions to optimize the oil production process.

2. A petroleum production process process parameter detection system as claimed in claim 1, characterized in that: The output end of the data acquisition module is connected to the input end of the data processing unit, the output end of the data processing unit is connected to the input end of the user terminal module, and the output end of the data processing unit is also connected to the input end of the optimization control module.

3. A petroleum production process process parameter detection system as claimed in claim 2, characterized in that: The data processing unit comprises: Data preprocessing module: denoises, removes bias, and normalizes the data collected from the sensor to ensure the accuracy and reliability of the data; Data fusion module: adopts multi-sensor data fusion algorithm to comprehensively process the data collected by different sensors to improve detection accuracy and reliability; Data analysis module: analyzes the fused data, builds models based on historical data, performs trend prediction and anomaly detection, supports real-time monitoring and feedback, and can mark abnormal data according to set thresholds and output relevant diagnostic information; Early warning module: When the data analysis module detects that a process parameter deviates from the set range, the early warning mechanism is triggered and a warning signal is sent to the operator in real time, prompting him to make necessary operational adjustments.

4. A petroleum production process process parameter detection system as claimed in claim 3, characterized in that: The data preprocessing module cleans, denoises, standardizes and normalizes the raw data collected from different sensors or monitoring devices to ensure that the input data quality is suitable for further analysis and modeling. The steps of data preprocessing include: Data denoising and outlier detection: Identify and remove outliers in data through statistical methods; Data filling: fill in missing data; Data standardization and normalization: Different sensors have different measurement units. The standardization and normalization steps ensure that the data has a uniform scale, making subsequent analysis more reliable. Standardization uses Z-score standardization, and normalization converts the data to a specified range. Data conversion: Convert certain specific data to ensure that they meet the requirements of model input, for example, converting timestamps to a unified date format, or converting data from a periodically changing state to a rate of change, such as pressure or temperature.

5. A petroleum production process process parameter detection system as claimed in claim 3, characterized in that: The data fusion module fuses data from different sensors or monitoring devices to form a unified and more representative information output. Data fusion technology can eliminate the possible errors of a single sensor and improve the reliability and accuracy of the data. The steps of data fusion include: Multi-sensor data alignment: Since the timestamps of different sensors may not be completely aligned, the data needs to be time-aligned to ensure that the data is consistent in time; Weighted average: weighted average of data output by multiple sensors; Data fusion algorithm: Use Kalman filter data fusion algorithm to perform real-time estimation in dynamic systems; Data fusion output: The fused data will serve as input for subsequent analysis, modeling and prediction.

6. A petroleum production process process parameter detection system as claimed in claim 3, characterized in that: The data analysis module is responsible for analyzing the fused multi-dimensional data to achieve trend prediction and anomaly detection. The specific composition and workflow of this module are as follows: Data cleaning and standardization: After data fusion, the data is first cleaned and standardized; by denoising, removing outliers and standardizing various parameters, the accuracy and consistency of the input data are ensured; Establish a forecasting model based on historical data: Use the ARIMA model to forecast the trend of oil production process parameters; Anomaly detection and diagnosis: Use the Gaussian distribution model to detect anomalies in the current process parameters. By assuming that the data conforms to the normal distribution, the probability density of each data point is calculated, and abnormal values ​​that deviate from the normal range are detected based on the set threshold.

7. A petroleum production process process parameter detection system as claimed in claim 6, characterized in that: It also includes a cloud computing and edge computing fusion architecture, which has a bidirectional data connection with a data acquisition module set and a data processing unit.

8. A petroleum production process process parameter detection system as claimed in claim 7, characterized in that: The connection and working mode between the data acquisition module and the edge computing node are as follows: The data acquisition module is responsible for collecting various process parameters in the oil production process through sensors. These data are transmitted to the edge computing node located on site. The edge computing node includes a local processing unit, which is responsible for receiving, preprocessing and preliminarily analyzing the data from the data acquisition module; The connection and working mode between the data processing unit and the edge computing node are as follows: After the edge computing node completes the data preprocessing, the processed data is transmitted to the data processing unit, and the data processing unit processes the multidimensional data from the edge computing node.

Images