Oilfield heating furnace coil pipe scaling multi-dimensional evaluation method based on soft measurement

Abstract

The invention discloses an oil field heating furnace coil pipe scaling multi-dimensional evaluation method based on soft measurement, which is characterized in that a scaling layer seriously affects the inner diameter of a pipe diameter and the heat transfer efficiency, a soft measurement method is utilized, and an axial temperature rise model and a pipeline pressure drop model of a coil pipe from an inlet to an outlet are established; according to the method, historical operation data and a clustering analysis means are combined, a pipeline heat transfer coefficient and a pipeline inner diameter representation value are solved, and the scaling condition of the inner wall of the water jacket furnace coil pipe is comprehensively evaluated from multiple dimensions by analyzing changes of the heat transfer coefficient and the pipeline inner diameter under different working conditions, so that a scale removal prompt is given in time. On the basis of a hydraulic thermal model, on the basis of historical operation data and a big data analysis technology, a soft measurement indirect judgment mode of coil pipe scaling is achieved, and certain representativeness is achieved. And by combining with the front end, the pipe diameter change and the heat transfer performance change caused by scaling are visually displayed, so that staff can conveniently judge the scale cleaning time.

Description

Technical Field

[0001] The invention relates to the technical field of oilfield heating furnaces, and in particular to a multi-dimensional evaluation method for scaling of oilfield heating furnace coils based on soft measurement. Background Art

[0002] The water jacket heater is a commonly used water bath heating equipment for crude oil and natural gas. The inside of the water jacket heater coil is usually a mixture of oil, gas and water. These mixtures contain high concentrations of calcium and magnesium ions, silicates, phosphates and corrosive gases such as H2S, CO2 and O2. Since the coil is in a high temperature and harsh environment for a long time, it is easy to form scales such as calcium carbonate and magnesium carbonate and deposit on the inner wall of the pipeline at high temperature, causing various effects. First, the scaling of the coil reduces the inner diameter of the coil and narrows the flow cross-sectional area, resulting in a decrease in displacement and an increase in transportation resistance. Severe scaling may cause pipeline blockage and even cause safety accidents such as explosions; secondly, since the thermal conductivity of the scaling layer is much lower than that of metal materials, the heat transfer efficiency is significantly reduced, which greatly increases the fuel consumption; finally, the unevenness of the scaling layer will cause the concentration of thermal and mechanical stresses, increase the risk of fatigue and cracks in the pipeline, aggravate equipment wear, and even cause perforation and accidents. Timely and effective scaling monitoring and descaling reminders can help restore the heating furnace's conveying performance and heat transfer efficiency, extend the heating furnace's service life, ensure the efficient and safe operation of the water jacket heating furnace, significantly reduce the operating costs of maintenance and equipment replacement, and improve production efficiency. Therefore, how to comprehensively evaluate the scaling of the inner wall of the water jacket furnace coil from multiple dimensions in order to give timely descaling reminders is an urgent problem to be solved in the industry. Summary of the invention

[0003] The purpose of the present invention is to address the shortcomings of the prior art and provide a multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement, so as to solve the problem of how to comprehensively evaluate the scaling of the inner wall of the water jacket furnace coil from multiple dimensions so as to provide scaling removal prompts in a timely manner.

[0004] The present invention provides a multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement, comprising:

[0005] The database / OPC communication module collects real-time data of each PLC point during the operation of the heating furnace from the heating furnace distributed control system, and stores the data in the database;

[0006] The sample data extraction module determines the stable operation window time according to the data, and extracts the data sample of the stable operation history of the heating furnace;

[0007] The cluster analysis module performs cluster analysis on the data samples and divides the data sample distribution into K clusters to simulate the distribution of data representing different operating conditions;

[0008] The mechanism analysis module establishes a pipeline axial temperature rise model, and determines the lower limit of the total heat transfer coefficient of K clusters based on the pipeline axial temperature rise model; establishes a pipeline pressure drop model, and determines the coil inner diameter representation value of K clusters based on the pipeline pressure drop model;

[0009] The online reasoning module determines whether the oilfield heating furnace coil is scaled according to the lower limit value of the total heat transfer coefficient and the value representing the inner diameter of the coil; the corresponding scaling ratio in the kth cluster in the past period of time in the statistical analysis library is calculated every day, and the weighted average is performed according to the weight of each cluster to obtain the scaling ratio value determined based on the weighted average of all clusters; it is judged whether the scaling ratio value exceeds the threshold value. If the scaling ratio value exceeds the threshold value, an alarm is issued for scaling, otherwise it is not scaling.

[0010] Furthermore, the sample data extraction module determines the stable operation window time according to the data, including:

[0011] Input the coil pipe length L and the pipe internal diameter D, and use the PLC real-time point medium flow G and pipe internal diameter D to calculate the time of the medium from inlet to outlet. The distribution graph of time T is analyzed to determine the fixed size value T of the window.

[0012] Furthermore, data samples of the heating furnace's historical stable operation are extracted, including:

[0013] With time T as the window size, the sliding data set is used to determine the stability of the point inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow rate, and furnace temperature through the pre-specified steady-state determination threshold. The steady-state determination expression is:

[0014] y ji are the i-th values ​​of characteristic variable j; j = 1, 2, 3, 4, 5, 6, representing the following characteristic variables: inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow, furnace temperature; n is the number of historical data to determine whether the process is stable; is the average value of the jth feature variable selected ε is a pre-specified steady-state judgment threshold, and its value range is (0, 0.05); when the threshold condition is met, the data in the time window segment is representative data, and the average value of each feature in the time window is taken as the representative value of the point, which constitutes an analysis sample and is entered into the database for subsequent analysis; the window is continuously sliding and the above process is repeated to obtain multiple analysis samples as data samples of the historical stable operation of the heating furnace.

[0015] Furthermore, the cluster analysis module performs cluster analysis on the data samples and divides the data sample distribution into K clusters to simulate the distribution of data representing different operating conditions, including:

[0016] For the data samples, after normalization, Kmeans algorithm cluster analysis is performed, and the number of clusters is determined to be K using the elbow method; wherein the normalization process limits each feature value to between (0, 1), and the specific normalization formula is:

[0017]

[0018] Among them, X min , X max Respectively represent the minimum and maximum values ​​of a feature;

[0019] For each determined cluster C k , find out the number of analysis samples that fall into this cluster, and establish the weight value according to the number of samples in this cluster Among them, N is the number of samples in the entire data set, N k Cluster C k The number of samples in .

[0020] Furthermore, the mechanism analysis module establishes a pipeline axial temperature rise model, and determines the lower limit of the total heat transfer coefficient of K clusters according to the pipeline axial temperature rise model, including:

[0021] Using the basic theory of thermodynamics, the pipeline axial temperature rise model is established. The specific formula is:

[0022]

[0023] Where P is the total heat transfer coefficient, D is the outer diameter of the pipe, L is the length of the pipe, ρ is the medium density, C is the specific heat capacity, G is the medium flow rate, T 炉 is the furnace temperature, T 入 is the inlet temperature, T 出 is the outlet temperature;

[0024] By using the above formula, cluster C is calculated k The heat transfer coefficient P value corresponding to each sample in the cluster is determined according to the box plot quartile method. The lower limit of P under the cluster is calculated as follows: P = Q1-IQR, where the interquartile range IQR = Q3-Q1, Q1 and Q3 represent the lower quartile and upper quartile, respectively. Cluster C is thus determined. k The corresponding lower limit of the total heat transfer coefficient P k ; Determine the lower limit of the total heat transfer coefficient P of K clusters in turn k .

[0025] Furthermore, a pipeline pressure drop model is established, and the coil inner diameter representation values ​​of K clusters are determined according to the pipeline pressure drop model, including:

[0026] According to the theory of fluid mechanics, the pipeline pressure drop model is established by Darcy's formula, and the friction loss formula along the pipeline is obtained as follows: And P 入 -P 出 =ρgH L , Q is the crude oil volume flow rate, P 入 , P 出 are the coil inlet pressure and outlet pressure respectively, and finally the formula for the inner diameter of the coil is obtained: Among them, the hydraulic friction coefficient λ = f(Re,ε) is the Reynolds number Relative roughness The relevant function; υ is the corresponding kinematic viscosity value, and the kinematic viscosity of crude oil measured at different temperatures in the laboratory is used to fit the relationship between temperature and kinematic viscosity υ=f(T i ), indicating that at temperature T i The corresponding kinematic viscosity value is υ; λ is the hydraulic friction coefficient;

[0027] By using the above formula, cluster C is calculated k The inner diameter D value of the coil corresponding to each sample in the cluster is determined according to the box plot quartile method. The calculation formula is: D = Q1-IQR, where the interquartile range IQR = Q3-Q1, Q1 and Q3 represent the lower quartile and upper quartile respectively, thus determining cluster C k The corresponding lower limit value D of the coil inner diameter k ; Determine the lower limit value D of the inner diameter of the coil of K clusters in turn k .

[0028] Furthermore, the online reasoning module determines whether the oilfield heating furnace coil is scaled according to the lower limit value of the total heat transfer coefficient and the value indicating the inner diameter of the coil, including:

[0029] During the prediction, the field data is input every time T, and the stability judgment condition in the sample data extraction module is used to judge whether the data meets the stable data. If not, the data of this period is discarded without any analysis, and the next cycle judgment is waited for. If the data is determined to be stable, the average value of the relevant characteristic variables is calculated, and then the cluster C to which the data belongs is obtained through the cluster analysis module. k According to the formula of the mechanism analysis module, the total heat transfer coefficient P and the representative value of the inner diameter D of the coil for this stable data are calculated respectively, and the lower limit value P of the total heat transfer coefficient corresponding to this cluster in the database is calculated. k , coil inner diameter indication value lower limit D kIf the value is lower than the lower limit, it is determined to be scaling, otherwise it is not scaling.

[0030] Furthermore, the scaling ratio corresponding to the kth cluster in the past period of time in the statistical analysis library is statistically analyzed every day, and the weighted average is performed according to the weight of each cluster to obtain the scaling ratio value determined based on the weighted average of all clusters, including:

[0031] In units of days, the corresponding scaling ratio in the kth cluster in the past period of time in the database is statistically analyzed every day N′ is the total number of records in the cluster. k is the number of records marked as scaling in the cluster; according to the weight β of each cluster previously k Perform weighted averaging to obtain the scaling ratio value r=β determined based on the weighted average of all clusters k r k .

[0032] Furthermore, it also includes: the P and D values ​​of different clusters at different times are visualized with the time axis as the X-axis and the corresponding P and D values ​​as the Y-axis to respectively display the change trend of the total heat transfer coefficient P and the inner diameter D over time.

[0033] The beneficial effects of the present invention are as follows: The multi-dimensional evaluation method for scaling of oilfield heating furnace coils based on soft measurement of the present invention cuts in from the two angles that the scaling layer seriously affects the inner diameter of the pipe diameter and the heat transfer efficiency. By using the soft measurement method, the axial temperature rise model and the pipeline pressure drop model from the inlet to the outlet of the coil pipeline are established. Combined with historical operation data and cluster analysis methods, the pipeline heat transfer coefficient and the pipeline inner diameter representation value are solved. By analyzing the changes in the heat transfer coefficient and the inner diameter of the pipeline under different working conditions, the scaling of the inner wall of the water jacket furnace coil is comprehensively evaluated from multiple dimensions, so as to give timely descaling prompts. Based on the hydraulic and thermal model, based on historical operation data and big data analysis technology, a soft measurement indirect judgment method for coil scaling is realized, which has a certain representativeness. Combined with the front end, the changes in pipe diameter and heat transfer performance caused by scaling are visualized, which is convenient for employees to judge the descaling time. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The present invention is a flow chart of the multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement. DETAILED DESCRIPTION

[0035] The following non-limiting embodiments can enable ordinary technicians in the field to understand the present invention more comprehensively, but do not limit the present invention in any way. It is necessary to point out that the following embodiments cannot be understood as limiting the scope of protection of the present invention. If technical personnel skilled in the art make some non-essential improvements and adjustments to the present invention based on the above-mentioned content of the present invention, they still fall within the scope of protection of the present invention.

[0036] See also Figure 1 ,The whole solution mainly includes several modules: database / OPC communication module, sample data extraction module, ,clustering analysis module, mechanism analysis module, and online reasoning module.

[0037] Step 1: The database / OPC communication module collects real-time data of each PLC point during the operation of the heating furnace from the heating furnace distributed control system, and stores the data in the database.

[0038] Step 2: The sample data extraction module determines the stable operation window time according to the data, and extracts the data sample of the stable operation history of the heating furnace.

[0039] Specifically, input the coil pipeline length L and the pipe internal diameter D, and use the PLC real-time point medium flow G and the pipe internal diameter D to calculate the time of the medium from the inlet to the outlet. The distribution graph of time T is analyzed to determine the fixed size value T of the window.

[0040] With time T as the window size, the sliding data set is used to determine the stability of the point inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow rate, and furnace temperature through the pre-specified steady-state determination threshold. The steady-state determination expression is:

[0041] y ji are the i-th values ​​of characteristic variable j; j = 1, 2, 3, 4, 5, 6, representing the following characteristic variables: inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow, furnace temperature; n is the number of historical data to determine whether the process is stable; is the average value of the jth feature variable selected ε is a pre-specified steady-state judgment threshold, and its value range is (0, 0.05); when the threshold condition is met, the data in the time window segment is representative data, and the average value of each feature in the time window is taken as the representative value of the point, which constitutes an analysis sample and is entered into the database for subsequent analysis; the window is continuously sliding and the above process is repeated to obtain multiple analysis samples as data samples of the historical stable operation of the heating furnace.

[0042] Step three: The cluster analysis module performs cluster analysis on the data samples and divides the data sample distribution into K clusters to simulate the distribution of data representing different operating conditions.

[0043] Specifically, for the data samples, after normalization processing, Kmeans algorithm clustering analysis is performed, and the number of clusters is determined to be K using the elbow method; wherein the normalization processing limits each feature value to between (0, 1), and the specific normalization formula is:

[0044]

[0045] Among them, X min , X max Respectively represent the minimum and maximum values ​​of a feature;

[0046] For each determined cluster C k , find out the number of analysis samples that fall into this cluster, and establish the weight value according to the number of samples in this cluster Among them, N is the number of samples in the entire data set, N k Cluster C k The number of samples in .

[0047] Step 4: The mechanism analysis module establishes a pipeline axial temperature rise model, and determines the lower limit of the total heat transfer coefficient of K clusters based on the pipeline axial temperature rise model; establishes a pipeline pressure drop model, and determines the representative value of the coil inner diameter of K clusters based on the pipeline pressure drop model.

[0048] Specifically, the basic theory of thermodynamics is used to establish the pipeline axial temperature rise model, and the specific formula is:

[0049]

[0050] Where P is the total heat transfer coefficient, D is the outer diameter of the pipe, L is the length of the pipe, ρ is the medium density, C is the specific heat capacity, G is the medium flow rate, T 炉 is the furnace temperature, T 入 is the inlet temperature, T 出 is the outlet temperature;

[0051] By using the above formula, cluster C is calculated k The heat transfer coefficient P value corresponding to each sample in the cluster is determined according to the box plot quartile method. The lower limit of P under the cluster is calculated as follows: P = Q1-IQR, where the interquartile range IQR = Q3-Q1, Q1 and Q3 represent the lower quartile and upper quartile, respectively. Cluster C is thus determined. k The corresponding lower limit of the total heat transfer coefficient P k ; Determine the lower limit of the total heat transfer coefficient P of K clusters in turn k , and save the relevant result information into the database for subsequent query.

[0052] According to the theory of fluid mechanics, the pipeline pressure drop model is established by Darcy's formula, and the friction loss formula along the pipeline is obtained as follows: And P入 -P 出 = ρgH L , where Q is the crude oil volume flow rate, and P 入 、P 出 are the inlet pressure and outlet pressure of the coil pipe respectively. Finally, the formula for the inner diameter of the coil pipe is obtained: where the hydraulic friction coefficient λ = f(Re, ε) is a function related to the Reynolds number and the relative roughness ; υ is the corresponding kinematic viscosity value, and the relationship between temperature and kinematic viscosity υ = f(T i ) is fitted according to the measured crude oil kinematic viscosity at different temperatures in the laboratory, indicating that the corresponding kinematic viscosity value at temperature T i is υ; λ is the hydraulic friction coefficient, and the specific calculation form is shown in Table 1.

[0053] Table 1 Division range of Reynolds number Re and calculation of hydraulic friction coefficient λ

[0054]

[0055] Note: 1. When 2000 < Re < 3000, it can be calculated according to the hydraulically smooth zone;

[0056] 2. Re1 - The critical Reynolds number for the transition from the hydraulically smooth zone to the mixed friction zone;

[0057] 3. Re2 - The critical Reynolds number for the transition from the mixed friction zone to the rough zone;

[0058] 4. e - The absolute (equivalent) roughness of the inner wall of the pipe. For straight-seam steel pipes, e takes 0.054 mm; for seamless steel pipes, e takes 0.06 mm; for spiral-seam steel pipes, DN250 - DN350, e takes 0.125 mm, and for DN400 and above, e takes 0.10 mm.

[0059] Through the above formula, the inner diameter D value of the coil pipe corresponding to each sample in cluster C k is calculated. According to the box plot quartile method, the lower limit value of D under this cluster is determined. The calculation formula is: D = Q1 - IQR, where the interquartile range IQR = Q3 - Q1, and Q1 and Q3 represent the lower quartile and upper quartile respectively. Thus, the lower limit value D k of the inner diameter representation value of the coil pipe corresponding to cluster C k is determined; the lower limit values D k of the inner diameter representation values of the coil pipes of K clusters are determined in sequence, and the relevant result information is saved in the database for subsequent query and use.

[0060] Step 5: The online reasoning module determines whether the oilfield heating furnace coil is scaled according to the lower limit value of the total heat transfer coefficient and the value representing the inner diameter of the coil; the corresponding scaling ratio in the kth cluster in the past period of time in the statistical analysis library is calculated every day, and the weighted average is performed according to the weight of each cluster to obtain the scaling ratio value determined based on the weighted average of all clusters; it is determined whether the scaling ratio value exceeds the threshold value. If the scaling ratio value exceeds the threshold value, an alarm is issued for scaling, otherwise it is not scaling.

[0061] Specifically, during the prediction, the field data is input every time T, and the stability judgment condition in the sample data extraction module is used to determine whether the data meets the stable data. If not, the data of this period is discarded without any analysis, and the next cycle judgment is waited for. If the data is determined to be stable, the average value of the relevant characteristic variables is calculated, and then the cluster C to which the data belongs is obtained through the cluster analysis module. k According to the formula of the mechanism analysis module, the total heat transfer coefficient P and the representative value of the inner diameter D of the coil for this stable data are calculated respectively, and the lower limit value P of the total heat transfer coefficient corresponding to this cluster in the database is calculated. k , coil inner diameter indication value lower limit D k If the value is lower than the lower limit, it is determined to be scaling, otherwise it is not scaling. The conclusions of the two dimensions: scaling or not scaling, and the corresponding heat transfer coefficient P and inner diameter D values ​​are entered into the relevant database for future use.

[0062] In units of days, the corresponding scaling ratio in the kth cluster in the past period of time in the analysis database is statistically analyzed every day, for example, 15 days N′ is the total number of records in the cluster. k is the number of records marked as scaling in the cluster; according to the weight β of each cluster previously k Perform weighted averaging to obtain the scaling ratio value r=β determined based on the weighted average of all clusters k r k For example, if the threshold is set to 0.5, if the ratio value r exceeds the threshold, the interface alarm will be scaled, otherwise it will not be scaled.

[0063] The P and D values ​​of different clusters at different times are visualized with the time axis as the X-axis and the corresponding P and D values ​​as the Y-axis to show the changing trends of the total heat transfer coefficient P and the inner diameter D over time, so that operators can further make reasonable fouling judgments.

[0064] The above-described embodiments of the present invention do not limit the protection scope of the present invention.

Claims

1. A multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement, characterized in that: include: The database / OPC communication module collects real-time data of each PLC point during the operation of the heating furnace from the heating furnace distributed control system, and stores the data in the database; The sample data extraction module determines the stable operation window time according to the data, and extracts the data sample of the stable operation history of the heating furnace; The cluster analysis module performs cluster analysis on the data samples and divides the data sample distribution into K clusters to simulate the distribution of data representing different operating conditions; The mechanism analysis module establishes a pipeline axial temperature rise model, and determines the lower limit of the total heat transfer coefficient of K clusters based on the pipeline axial temperature rise model; establishes a pipeline pressure drop model, and determines the coil inner diameter representation value of K clusters based on the pipeline pressure drop model; The online reasoning module determines whether the oilfield heating furnace coil is scaled according to the lower limit value of the total heat transfer coefficient and the value indicating the inner diameter of the coil; The scaling ratio corresponding to the kth cluster in the past period of time in the statistical analysis library is statistically analyzed every day, and the weighted average is performed according to the weight of each cluster to obtain the scaling ratio value determined based on the weighted average of all clusters; it is judged whether the scaling ratio value exceeds the threshold. If the scaling ratio value exceeds the threshold, an alarm is issued for scaling, otherwise it is not scaling.

2. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 1, characterized in that: The sample data extraction module determines the stable operation window time according to the data, including: Input the coil pipe length L and the pipe internal diameter D, and use the PLC real-time point medium flow G and pipe internal diameter D to calculate the time of the medium from inlet to outlet. The distribution graph of time T is analyzed to determine the fixed size value T of the window.

3. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 2, characterized in that: Extract the data samples of the heating furnace's historical stable operation, including: With time T as the window size, the sliding data set is used to determine the stability of the point inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow rate, and furnace temperature through the pre-specified steady-state determination threshold. The steady-state determination expression is: y ji are the i-th values ​​of characteristic variable j; j = 1, 2, 3, 4, 5, 6, representing the following characteristic variables: inlet temperature, outlet temperature, inlet pressure, outlet pressure, medium flow, furnace temperature; n is the number of historical data to determine whether the process is stable; is the average value of the jth feature variable selected ε is a pre-specified steady-state judgment threshold, and its value range is (0, 0.05); when the threshold condition is met, the data in the time window segment is representative data, and the average value of each feature in the time window is taken as the representative value of the point, which constitutes an analysis sample and is entered into the database for subsequent analysis; the window is continuously sliding and the above process is repeated to obtain multiple analysis samples as data samples of the historical stable operation of the heating furnace.

4. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 1, characterized in that: The cluster analysis module performs cluster analysis on the data samples and divides the data sample distribution into K clusters to simulate the distribution of data representing different operating conditions, including: For the data samples, after normalization, Kmeans algorithm cluster analysis is performed, and the number of clusters is determined to be K using the elbow method; wherein the normalization process limits each feature value to between (0, 1), and the specific normalization formula is: Among them, X min , X max Respectively represent the minimum and maximum values ​​of a feature; For each determined cluster C k , find out the number of analysis samples that fall into this cluster, and establish the weight value according to the number of samples in this cluster Among them, N is the number of samples in the entire data set, N k Cluster C k The number of samples in .

5. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 1, characterized in that: The mechanism analysis module establishes the pipeline axial temperature rise model and determines the lower limit of the total heat transfer coefficient of K clusters based on the pipeline axial temperature rise model, including: Using the basic theory of thermodynamics, the pipeline axial temperature rise model is established. The specific formula is: Where P is the total heat transfer coefficient, D is the outer diameter of the pipe, L is the length of the pipe, ρ is the medium density, C is the specific heat capacity, G is the medium flow rate, T 炉 is the furnace temperature, T 入 is the inlet temperature, T 出 is the outlet temperature; By using the above formula, cluster C is calculated k The heat transfer coefficient P value corresponding to each sample in the cluster is determined according to the box plot quartile method. The lower limit of P under the cluster is calculated as follows: P = Q1-IQR, where the interquartile range IQR = Q3-Q1, Q1 and Q3 represent the lower quartile and upper quartile, respectively. Cluster C is thus determined. k The corresponding lower limit of the total heat transfer coefficient P k ; Determine the lower limit of the total heat transfer coefficient P of K clusters in turn k .

6. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 1, characterized in that: A pipeline pressure drop model is established, and the coil inner diameter representation values ​​of K clusters are determined according to the pipeline pressure drop model, including: According to the theory of fluid mechanics, the pipeline pressure drop model is established by Darcy's formula, and the friction loss formula along the pipeline is obtained as follows: And P 入 -P 出 =ρgH L , Q is the crude oil volume flow rate, P 入 , P 出 are the coil inlet pressure and outlet pressure respectively, and finally the formula for the inner diameter of the coil is obtained: Among them, the hydraulic friction coefficient λ = f(Re,ε) is the Reynolds number Relative roughness The relevant function; υ is the corresponding kinematic viscosity value, and the kinematic viscosity of crude oil measured at different temperatures in the laboratory is used to fit the relationship between temperature and kinematic viscosity υ=f(T i ), indicating that at temperature T i The corresponding kinematic viscosity value is υ; λ is the hydraulic friction coefficient; By using the above formula, cluster C is calculated k The inner diameter D value of the coil corresponding to each sample in the cluster is determined according to the box plot quartile method. The calculation formula is: D = Q1-IQR, where the interquartile range IQR = Q3-Q1, Q1 and Q3 represent the lower quartile and upper quartile respectively, thus determining cluster C k The corresponding coil inner diameter lower limit D k ; Determine the lower limit value D of the inner diameter of the coil of K clusters in turn k .

7. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 1, characterized in that: The online reasoning module determines whether the oilfield heating furnace coil is scaled according to the lower limit value of the total heat transfer coefficient and the value indicating the inner diameter of the coil, including: During the prediction, the field data is input every time T, and the stability judgment condition in the sample data extraction module is used to judge whether the data meets the stable data. If not, the data of this period is discarded without any analysis, and the next cycle judgment is waited for. If the data is determined to be stable, the average value of the relevant characteristic variables is calculated, and then the cluster C to which the data belongs is obtained through the cluster analysis module. k According to the formula of the mechanism analysis module, the total heat transfer coefficient P and the representative value of the inner diameter D of the coil for this stable data are calculated respectively, and the lower limit value P of the total heat transfer coefficient corresponding to this cluster in the database is calculated. k , coil inner diameter indication value lower limit D k If the value is lower than the lower limit, it is determined to be scaling, otherwise it is not scaling.

8. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 7, characterized in that: In units of days, the scaling ratio corresponding to the kth cluster in the past period of time in the statistical analysis library is calculated every day, and the weighted average is performed according to the weight of each cluster to obtain the scaling ratio value determined based on the weighted average of all clusters, including: In units of days, the corresponding scaling ratio in the kth cluster in the past period of time in the database is statistically analyzed every day N′ is the total number of records in the cluster. k is the number of records marked as scaling in the cluster; according to the weight β of each cluster previously k Perform weighted averaging to obtain the scaling ratio value r=β determined based on the weighted average of all clusters k r k .

9. The multi-dimensional evaluation method for oilfield heating furnace coil scaling based on soft measurement according to claim 8, characterized in that: Also includes: The P and D values ​​of different clusters at different times are visualized with the time axis as the X-axis and the corresponding P and D values ​​as the Y-axis to show the changing trends of the total heat transfer coefficient P and the inner diameter D over time.

Images