Method and system for predicting the purity of alpha-olefins refined in a partition wall column

Abstract

The present application provides a kind of prediction method of partition wall column purifying alpha-olefin purity, based on the experimental data of separating alpha-olefin raw material using partition wall column, using nonlinear regression algorithm, establishes the prediction model of the purity of the first extraction channel and the second extraction channel product of separating alpha-olefin component using partition wall column.By selecting the key variables that have greater impact on product purity, the change rule of alpha-olefin purity refined by partition wall column can be accurately grasped, and the calculation result has higher accuracy.The present application also provides a kind of partition wall column alpha-olefin purity prediction system and a kind of data processing device.

Description

Technical Field

[0001] This invention belongs to the field of α-olefin mixture refining technology, specifically relating to a method for predicting the purity of α-olefins refined by a partition wall tower. Background Technology

[0002] Alpha-olefins are important comonomers of polyethylene. Alpha-olefin copolymer polyethylene has excellent tensile strength, tear strength, transparency and environmental stress cracking resistance due to its long branched chain structure.

[0003] In actual production, the separation of crude α-olefins adopts traditional refining methods, which require multiple distillation columns in series for refining. This results in a large number of distillation columns, a large equipment scale, and low energy utilization rate in refining.

[0004] Currently, in the actual production of α-olefin units, the energy consumption of the refining process accounts for about 50% of the entire production process. Many efficient energy-saving measures have not yet been widely adopted, and the energy-saving potential of the refining unit is huge.

[0005] Indirect-wall distillation columns can be used to separate three-component mixtures. By embedding partitions within the column, they achieve the effect of separating multi-component mixtures in a single column. Indirect-wall distillation technology can effectively solve the problem of backmixing of intermediate components within the column and enhance the heat transfer process. Compared with conventional two-column refining sequences, indirect-wall distillation columns have significant advantages in terms of energy efficiency, equipment investment, and plant footprint. Regarding the refining feedstock, the complexity of the composition of α-olefin feedstocks and the frequent changes in their sources, coupled with the small batch sizes of each feedstock and their frequent mixing, lead to changes in the physical properties of the feedstocks, especially the mass fraction of the target product component. This can easily result in problems such as low product purity, high energy consumption, and poor economic efficiency, posing considerable challenges to the operation of indirect-wall distillation columns for α-olefin refining.

[0006] The purification of α-olefin feedstocks with different compositions using indirect-wall column purification exhibits significant performance differences. Under the premise of ensuring product yield, in terms of α-olefin product purity, the relative volatility between α-olefins and other components in the feedstock has the greatest impact, followed by the mass fraction of α-olefins in the feedstock, and then the process conditions of the indirect-wall column. By using predictive technology for the purity of α-olefins purified by indirect-wall column purification, the purity of the purified product under different operating conditions can be predicted for variations in α-olefin feedstocks in chemical plants. This allows for the identification of optimal operating conditions for producing the target product, thereby increasing its yield and providing a basis for maximizing the value of α-olefin feedstocks and the operation of the indirect-wall column. Furthermore, it provides technical support for the quality screening, modification direction, and optimized utilization of α-olefin feedstocks.

[0007] A Chinese national invention, "A Soft Measurement Method for Concentration in an Isolated Wall Distillation Column" (Publication No.: CN111474856A), proposes a soft measurement method for concentration in an isolated wall distillation column, belonging to the category of soft measurement methods in chemical processes. This concentration soft measurement method, integrating a data-driven model, adds soft measurement to the cascade control of temperature and concentration in an isolated wall distillation column, thereby predicting product concentration by detecting easily detectable variables. In the cascade control, the inner layer uses proportional-integral control to control the temperature of the sensitive plate in the isolated wall distillation column; the outer layer uses model predictive control to control product concentration. This invention fully utilizes easily detectable auxiliary variables to predict product concentration, avoiding the high cost and long latency of real-time online concentration detection; it also fully utilizes the combination of proportional-integral control and model predictive control to achieve rapid and error-free control of the isolated wall distillation column. However, this technology mainly focuses on building models for Kaibel isolated wall distillation columns to predict the chain reaction of product purity and operating conditions, using a tetraol system as the separated raw material, and does not consider the influence of the raw material's physical properties.

[0008] Chinese National Invention No. CN104965967A, entitled "A Real-Time Yield Prediction Method for Atmospheric and Vacuum Distillation Units," discloses a method for real-time yield prediction of atmospheric and vacuum distillation units. This method employs data adjustment techniques to process real-time field data and combines an improved differential evolution algorithm to perform real-time correction of model parameters for the distillation process. This ensures the established model can adapt to different production conditions and achieves process simulation of actual operating conditions. Based on the corrected model, the method analyzes the impact of key operating / process conditions on the yield of distillation column side-stream products. Using neural network modeling technology, it correlates operating conditions with the yield of key products in the atmospheric and vacuum distillation unit, generating corresponding key product yields that conform to production operations, thus achieving real-time yield prediction. This invention can quickly and in real-time update crude oil distillation unit yield data in other optimization models, improving the optimization accuracy of planning models such as PIMS models. Summary of the Invention

[0009] To address the problem of purity prediction for α-olefin products separated by indirect-wall column, this invention proposes a method for predicting the purity of purified α-olefins using indirect-wall column based on machine learning principles. The method includes: collecting sample data from the indirect-wall column α-olefin separation operation to construct an initial dataset, which is then divided into a training set and a test set; the sample data includes the purified α-olefin purity data from the first and second extraction channels of the indirect-wall column during the α-olefin separation operation; training a machine learning model using the training set to obtain an initial prediction model; and testing the initial prediction model using the test set. If the error of the test result relative to the data in the test set is less than or equal to an accuracy threshold, then the initial prediction model is adopted as the purity prediction model for purified α-olefins using indirect-wall column.

[0010] Furthermore, this machine learning model is a nonlinear regression model. Specifically, the least squares principle is used to calculate the nonlinear regression equation and perform curve fitting to obtain the initial prediction model:

[0011] y ni =f(Pc i α ij ESI i M i m i , R1, R2, N, T, P, L, V)

[0012] y ni Pc represents the purity of component i in the nth sampling channel. i α represents the saturated vapor pressure of component i. ij ESI represents the relative volatility of component i to component j. i M represents the separation difficulty coefficient of component i. i m represents the molecular weight of component i. i R1 represents the mass fraction of component i in the feed, R2 represents the top reflux ratio, N represents the number of trays, T represents the temperature, P represents the pressure, L represents the liquid-liquid ratio, and V represents the gas-liquid ratio.

[0013] Preferably, if the error of the test result relative to the data of the test set is greater than the accuracy threshold, the initial prediction model is discarded, the machine learning model is retrained to obtain a new initial prediction model, and the model is tested until the error of the test result of the new initial prediction model relative to the data of the test set is less than or equal to the accuracy threshold.

[0014] Furthermore, the accuracy threshold includes: for components purified by the indirect wall column, the accuracy threshold is 5% for components with a purity greater than or equal to 1%; the accuracy threshold is 20% for components with a purity less than 1% but greater than 0.01%; and the relative error of components with a purity less than or equal to 0.01% is ignored.

[0015] Preferably, the sample data is obtained through a small-scale or pilot-scale test of an indirect-wall tower, or through an experimental evaluation device, or through an indirect-wall tower simulation evaluation experimental device.

[0016] Preferably, both the training set and the test set contain the physical properties of all selected α-olefin feedstocks, and the sample data size of the test set is greater than or equal to 20% of the sample data size of the initial dataset.

[0017] This invention also proposes a system for predicting the purity of purified α-olefins using a transmural tower, comprising: a data acquisition module for collecting sample data from the operation of separating α-olefins using a transmural tower, constructing an initial dataset, and dividing the initial dataset into a training set and a test set; the sample data includes purified α-olefin purity data from the first and second extraction channels of the transmural tower during the α-olefin separation operation; a model training module for training a machine learning model using the training set to obtain an initial prediction model; and a model testing module for testing the initial prediction model using the test set. If the error of the test result relative to the data in the test set is less than or equal to an accuracy threshold, then the initial prediction model is adopted as the purification purity prediction model for transmural tower α-olefins.

[0018] The present invention also proposes a computer-readable storage medium storing computer-executable instructions, characterized in that, when the computer-executable instructions are executed, the method for predicting the purity of α-olefins purified by the indirect wall tower as described above is implemented.

[0019] The present invention also proposes a data processing device, including a computer-readable storage medium as described above, wherein when the processor of the data processing device retrieves and executes computer-executable instructions in the computer-readable storage medium, it realizes the prediction of the purity of α-olefins purified by the inter-wall tower.

[0020] This invention utilizes the mapping relationship between the dependent variable of α-olefin purity obtained by indirect wall tower purification and independent variables such as key process parameters of the indirect wall tower and key physical property parameters of α-olefin feedstock. It provides a method for predicting the purity of α-olefin obtained by indirect wall tower purification, which solves the problem of lacking timely and effective evaluation of feedstock purification performance due to changes in α-olefin feedstock. It can serve as a guiding tool for optimizing the operation of α-olefin plants and improve the economic efficiency of α-olefin plants. Attached Figure Description

[0021] Figure 1 A schematic diagram illustrating the process of establishing a purity prediction model for α-olefin purification using a partition wall tower according to the present invention.

[0022] Figure 2 This is a schematic diagram of the data processing device of the present invention. Detailed Implementation

[0023] The following provides a detailed description of the embodiments of the present invention. These embodiments are implemented based on the technical solution of the present invention and provide detailed implementation methods and processes. However, the scope of protection of the present invention is not limited to the following embodiments. Experimental methods in the following embodiments that do not specify specific conditions are generally performed under conventional conditions.

[0024] This invention relates to a method for predicting the purity of α-olefins purified by a partition wall column. Based on experimental data from the separation of α-olefin feedstocks using a partition wall column, a nonlinear regression algorithm is employed to establish a predictive model for the purity of the products from the first and second extraction channels of the α-olefin components separated by the partition wall column. By selecting key variables that significantly influence product purity, the method can accurately grasp the variation pattern of the purity of α-olefins purified by the partition wall column, and the calculation results have high accuracy.

[0025] Figure 1 This is a schematic diagram illustrating the process for establishing a purity prediction model for α-olefin purification using a partition wall tower, as described in this invention. Figure 1 As shown, the method for predicting the purity of α-olefins purified by the indirect wall column of the present invention mainly includes:

[0026] (1) Data Collection: This includes the physical properties of α-olefin feedstock and experimental data on the separation of α-olefin components using a partition wall column. The experimental data on the separation of α-olefin components using a partition wall column includes the composition of the distillation feed, the product composition of each outflow channel, and the distillation process conditions. The physical properties of the feedstock and the experimental data on the separation of α-olefin components using a partition wall column are matched one-to-one and divided into two groups: modeling data and validation data. There is no overlap between the two groups, and each group includes the physical properties of all selected feedstocks.

[0027] In the embodiments of the present invention, the raw materials and equipment are from the 5000-ton / year hexene-1 unit of Daqing Chemical Research Center, and the evaluation and analysis method is gas chromatography.

[0028] Sample data were obtained by statistically analyzing the composition, mass fraction, saturated vapor pressure, molecular weight, and indirect wall tower process conditions of different α-olefin feedstocks. The physical properties of the selected α-olefin feedstock are as follows: the mass fraction of the target product ranges from 15-70 ω%, there is no azeotropic or miscible relationship between the components, the saturated vapor pressure ranges from [value missing], and the molecular weight ranges from 28-160. The intermediate wall column has 10-50 theoretical plates in the pre-fractionation section, 10-50 theoretical plates in the side stream section, 10-70 theoretical plates in the rectification section, and 10-40 theoretical plates in the stripping section. The column operating pressure is 0.1-1.5 MPaG. The top condenser temperature is 35-80℃. The bottom reboiler temperature is 150-250℃. The top reflux ratio is 1-50, the side stream reflux ratio is 1-10, the top reflux liquid separation ratio is 0.1-0.4, and the bottom stripping gas separation ratio is 0.1-0.5.

[0029] Multiple α-olefin feedstocks were subjected to simulated indirect-wall tower purification evaluation, and the product purity under corresponding conditions was calculated. This embodiment utilizes the indirect-wall tower simulation evaluation test device at the Daqing Chemical Research Center of China National Petroleum Corporation to conduct α-olefin purification experiments and obtain experimental data on α-olefin product purity. Relevant purification experimental analysis instruments and equipment include: HP-6890 gas chromatography for analyzing the composition of α-olefins (ethylene, 1-butene, 1-hexene, 1-octene, 1-decene, cyclohexane, and hexene), and HP-7890 gas chromatography for analyzing the composition and mass content of α-olefin products.

[0030] The sample data was divided into two groups: modeling data and validation data. There was no overlap between the two groups, and each group included the physical properties of all selected α-olefin feedstocks. The validation data comprised no less than 20% of the total data.

[0031] In embodiments of the present invention, α-olefins mainly include, but are not limited to, one or a mixture of several of ethylene, propylene, 1-butene, 1-hexene, 1-octene, and 1-decene. The physical properties of the α-olefin raw materials mainly include saturated vapor pressure, relative volatility, separation difficulty coefficient, molecular weight, and mass fraction.

[0032] Furthermore, experimental data for the purification of α-olefins using indirect-wall column refining can be obtained through indirect-wall column small-scale or pilot-scale test evaluation devices and indirect-wall column simulation evaluation test devices. The process conditions described mainly include the overhead reflux ratio, side stream reflux ratio, number of trays, temperature, pressure, overhead reflux liquid-to-liquid ratio, and bottom stripping gas-to-gas ratio. Alternatively, an indirect-wall column refining α-olefin component simulation evaluation test device can be used to simulate and evaluate the purity of products from various compositions of α-olefin feedstocks purified by indirect-wall column refining.

[0033] The purity of products from the first extraction channel mainly includes the purity of 1-butene, 1-hexene, cyclohexane, hexene, 1-octene, and 1-decene; the purity of products from the second extraction channel mainly includes the purity of 1-butene, 1-hexene, cyclohexane, hexene, 1-octene, and 1-decene.

[0034] (2) Establish a prediction model: Based on the modeling data in step (1), a nonlinear regression algorithm is used to model the model, and the least squares principle is preferred to calculate the regression equation and fit the curve. A prediction model for the purity of the first and second extraction channels of the α-olefin components separated by the wall tower is established.

[0035] Based on the physical properties of α-olefin feedstock and key process parameters of the indirect-wall tower as independent variables, and the product purity of the outflow channel as the dependent variable, a logical mapping relationship is established. The input variables of the model are the physical properties of the α-olefin feedstock and the process conditions of the indirect-wall tower, and the output variables are the product purity of the first outflow channel and the product purity of the second outflow channel. Using 1STOPT software, and based on the least squares principle, the model association for predicting α-olefin product purity is performed, namely:

[0036] y ni =f(Pc i α ij ESI i M i m i , R1, R2, N, T, P, L, V)

[0037] Where y ni Pc represents the purity of component i in the nth sampling channel. i α represents the saturated vapor pressure of component i. ij ESI represents the relative volatility of component i to component j. i M represents the separation difficulty coefficient of component i. i m represents the molecular weight of component i. i R1 represents the mass fraction of component i in the feed, R2 represents the top reflux ratio, N represents the number of trays, T represents the temperature, P represents the pressure, L represents the liquid-liquid ratio, and V represents the gas-liquid ratio.

[0038] The prediction model calculation results are the results after data normalization.

[0039] Predictive models can also be built using software such as ASPEN PLUS, FORTRAN, and 1STOPT.

[0040] (3) Model verification: Verify the prediction model obtained in step (2) using the verification data in step (1). Calculate the relative error between the prediction model calculation results and the verification data to ensure that the prediction model calculation results are basically consistent with the verification data.

[0041] The relative error between the prediction model's calculation results and the validation data is calculated by subtracting the validation data from the prediction model's calculation results for the corresponding product purity under the same α-olefin feedstock properties and indirect wall tower process conditions, and dividing the result by the corresponding validation data for product purity.

[0042] ε yi =|Prediction model calculation result - validation data| ÷ validation data × 100%, ε yi This indicates the relative error in the yield of different product components;

[0043] The relative error was calculated using the prediction model calculation results and verification data of the same α-olefin raw material under the same process conditions. Since the α-olefin product refined by the indirect wall tower is composed of multiple components, the following conditions must be met to determine whether the prediction model is applicable: (1) The relative error of the purity of the main component in the product is within 5%, which means that the purity of the main component should be greater than or equal to 1%; (2) The relative error of the component with a purity of 0.01%-1% in the product is within 20%; (3) The relative error of the component with a purity of less than 0.01% in the product is negligible because its content is too small (in the experiment, the purity is usually less than a few ppm due to entrainment) and does not affect the product quality; if the relative error meets the above three conditions, the model is considered applicable.

[0044]

[0045] Table 1. Process Conditions and Raw Material Properties

[0046]

[0047]

[0048] Table 2 Comparison of Product Purity Prediction Results and Verification Data

[0049] The physical properties of the selected α-olefin feedstock and the corresponding verification data and predicted purity (partial data) under the operating conditions of the indirect wall tower are shown in Tables 1 and 2.

[0050] Figure 2 This is a schematic diagram of the data processing apparatus of the present invention. Figure 2 As shown, embodiments of the present invention also provide a computer-readable storage medium and a data processing apparatus. The computer-readable storage medium of the present invention stores computer-executable instructions, which, when executed by the processor of the data processing apparatus, realize the above-mentioned prediction of the purity of α-olefins purified by the intermembranous column. Those skilled in the art will understand that all or part of the steps in the above method can be implemented by a program instructing related hardware (e.g., processor, FPGA, ASIC, etc.), and the program can be stored in a readable storage medium, such as a read-only memory, a disk, or an optical disk. All or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Accordingly, each module in the above embodiments can be implemented in hardware, for example, by implementing its corresponding function through an integrated circuit, or it can be implemented in the form of a software functional module, for example, by a processor executing a program / instruction stored in memory to implement its corresponding function. Embodiments of the present invention are not limited to any particular combination of hardware and software.

[0051] This technical solution establishes a purity model for α-olefins purified by a two-wall column based on the physical properties of the α-olefin feedstock and experimental data from the two-wall column purification process. By selecting key variables that significantly affect product purity, it can accurately grasp the variation pattern of α-olefin product purity, and its calculation results have high accuracy. Furthermore, this technical solution has the advantages of requiring less experimental data from the two-wall column, easy convergence of calculation results, and fast computation speed. It can quickly model the process and provide guidance for the optimized operation of the two-wall column, enabling efficient utilization of the current α-olefin feedstock to maximize the purity of the target product.

[0052] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the claims of the present invention.

Claims

1. A method for predicting the purity of α-olefins purified by a partition wall column, characterized in that, include: The composition, mass fraction, saturated vapor pressure, and molecular weight of each component in different α-olefin feedstocks for the separation of α-olefins using a partition wall tower, as well as the process conditions of the partition wall tower and the purity of the α-olefin product, were collected to obtain sample data. An initial dataset was constructed and divided into a training set and a test set. The purity of the α-olefin product includes the purity data of the purified α-olefins in the first and second outflow channels of the indirect wall column during the α-olefin separation operation. The machine learning model is trained using this training set. The least squares principle is used to calculate the equation and fit the curve for nonlinear regression, resulting in an initial prediction model. y ni =f(Pc i ,α ij ,ESI i ,M i ,m i ,R1,R2,N,T,P,L,V) Among them, y ni Pc represents the purity of component i in the nth sampling channel. i α represents the saturated vapor pressure of component i. ij ESI represents the relative volatility of component i to component j. i M represents the separation difficulty coefficient of component i. i Indicates the molecular weight of component i, ranging from 28 to 160; m i R1 represents the mass fraction of component i in the feed, ranging from 15-70 ω%. R2 represents the reflux ratio at the top of the column, ranging from 1-50. R3 represents the side reflux ratio, ranging from 1-10. N represents the number of trays, where the pre-fractionation section of the intermediate wall column has 10-50 theoretical trays, the side reflux section has 10-50 theoretical trays, the rectification section has 10-70 theoretical trays, and the stripping section has 10-40 theoretical trays. T represents the temperature, with the top condenser temperature at 35-80℃ and the bottom reboiler temperature at 150-250℃. P represents the pressure, ranging from 0.1-1.5 MPaG. L represents the reflux liquid separation ratio at the top of the column, ranging from 0.1-0.4, and V represents the stripping gas separation ratio at the bottom of the column, ranging from 0.1-0.

5. The initial prediction model is tested using the test set. If the error of the test result relative to the data in the test set is less than or equal to the accuracy threshold, then the initial prediction model is adopted as the purity prediction model for α-olefins purified by the indirect wall tower. The accuracy threshold includes: for components of α-olefins purified by the indirect wall tower, the accuracy threshold is 5% for components with a purity greater than or equal to 1%; the accuracy threshold is 20% for components with a purity less than 1% but greater than 0.01%; and the relative error of components with a purity less than or equal to 0.01% is ignored.

2. The method for predicting the purity of α-olefins purified by a partition wall column as described in claim 1, characterized in that, This machine learning model is a non-linear regression model.

3. The method for predicting the purity of α-olefins purified by a partition wall column as described in claim 1, characterized in that, If the error of the test result relative to the data in the test set is greater than the accuracy threshold, the initial prediction model is discarded, the machine learning model is retrained to obtain a new initial prediction model, and the model is tested until the error of the test result of the new initial prediction model relative to the data in the test set is less than or equal to the accuracy threshold.

4. The method for predicting the purity of α-olefins purified by a partition wall column as described in claim 1, characterized in that, The sample data is obtained through small-scale or pilot-scale testing of the partition wall tower, or through an experimental evaluation device, or through a partition wall tower simulation evaluation experimental device.

5. The method for predicting the purity of α-olefins purified by a partition wall column as described in claim 1, characterized in that, Both the training set and the test set contain the physical properties of all selected α-olefin feedstocks, and the sample data size of the test set is greater than or equal to 20% of the sample data size of the initial dataset.

6. A system for predicting the purity of α-olefins purified by a partition wall column, characterized in that, include: The data acquisition module is used to collect the composition, mass fraction, saturated vapor pressure, and molecular weight of each component in the separation of α-olefins by the indirect wall tower, as well as the process conditions of the indirect wall tower and the purity of the α-olefin product. It obtains sample data, constructs an initial dataset, and divides the initial dataset into a training set and a test set. The purity of the α-olefin product includes the purified α-olefin purity data of the first and second outflow channels of the indirect wall column during the α-olefin separation operation. The model training module is used to train the machine learning model using the training set to obtain an initial prediction model; wherein, the least squares principle is used to calculate the equation and fit the curve for nonlinear regression to obtain the initial prediction model: y ni =f(Pc i ,α ij ,ESI i ,M i ,m i ,R1,R2,N,T,P,L,V) Among them, y ni Pc represents the purity of component i in the nth sampling channel. i α represents the saturated vapor pressure of component i. ij ESI represents the relative volatility of component i to component j. i M represents the separation difficulty coefficient of component i. i The molecular weight of component i is denoted as ω%, ranging from 28 to 160; mi represents the mass fraction of component i in the feed, ranging from 15 to 70 ω%; R1 represents the reflux ratio at the top of the column, ranging from 1 to 50; R2 represents the side reflux ratio, ranging from 1 to 10; N represents the number of trays, where the pre-fractionation section of the intermediate wall column has 10 to 50 theoretical trays, the side reflux section has 10 to 50 theoretical trays, the rectification section has 10 to 70 theoretical trays, and the stripping section has 10 to 40 theoretical trays; T represents the temperature, with the top condenser temperature at 35 to 80°C and the bottom reboiler temperature at 150 to 250°C; P represents the pressure, ranging from 0.1 to 1.5 MPaG; L represents the reflux liquid separation ratio at the top of the column, ranging from 0.1 to 0.4; and V represents the stripping gas separation ratio at the bottom of the column, ranging from 0.1 to 0.

5. The model testing module is used to test the initial prediction model with the test set. If the error of the test result relative to the data in the test set is less than or equal to the accuracy threshold, then the initial prediction model is adopted as the purity prediction model for α-olefins purified by the indirect wall tower. The accuracy threshold includes: for components of α-olefins purified by the indirect wall tower, the accuracy threshold is 5% for components with a purity greater than or equal to 1%; the accuracy threshold is 20% for components with a purity less than 1% but greater than 0.01%; and the relative error is ignored for components with a purity less than or equal to 0.01%.

7. A computer-readable storage medium storing computer-executable instructions, characterized in that, When the computer-executable instructions are executed, the method for predicting the purity of α-olefins purified by a partition wall tower as described in any one of claims 1 to 5 is implemented.

8. A data processing apparatus comprising the computer-readable storage medium as claimed in claim 7, wherein when a processor of the data processing apparatus retrieves and executes computer-executable instructions in the computer-readable storage medium, it realizes the prediction of the purity of α-olefins purified by a partition wall column.

Images