KR20230144879A - Apparatus and method for predicting quality of polymers - Google Patents

Abstract

A polymer quality prediction device according to an embodiment of the present invention is a device for predicting the quality of a polymer in a process in which a polymer is produced from raw materials flowing into a reactor, and is used to measure measurable process variables for the raw materials and the reactor. configured measuring unit; And configured to calculate polymerization variables for the raw material and the polymer from the process variables using a preset first model, and predict the quality of the polymer from the process variables and the polymerization variables using a preset second model. Includes a control unit.

Description

Apparatus and method for predicting polymer quality {APPARATUS AND METHOD FOR PREDICTING QUALITY OF POLYMERS}

The present invention relates to an apparatus and method for predicting polymer quality, and more specifically, to an apparatus and method that can predict polymer quality in advance based on process variables and polymerization variables.

Styrene-acrylonitrile (SAN) resin, a copolymer resin made by polymerizing styrene (SM), an aromatic hydrocarbon, and acrylonitrile (AN), an unsaturated nitrile, has excellent transparency, chemical resistance, and rigidity, and is used for electrical and electronic purposes and household use. , It is widely used in office and automobile parts. In addition, the SAN resin is applied to ABS resin with low heat resistance and is used to reinforce heat resistance.

However, because the heat distortion temperature of SAN resin is around 100°C to 105°C, there are limits to its application to products requiring high heat resistance. Accordingly, a method of introducing α-methylstyrene (AMS) monomer is used to provide excellent heat resistance to ABS.

Figure 1 is a diagram schematically showing the SAN resin processing process.

For example, in the embodiment of FIG. 1, the raw material 10 may include α-methylstyrene (AMS), acrylonitrile (AN), styrene (SM), and an initiator. The raw material 10 passes through the reactor 20 and the volatilization tank 30, and finally, styrene-acrylonitrile (SAN) resin can be produced. Since it is realistically difficult to judge the quality of the SAN resin produced during the SAN resin process in real time, the general method is to secure the SAN resin manufactured for about 4 hours and judge the quality of the secured SAN resin. It is being adopted.

However, according to the conventional quality judgment method, there is a problem that the quality of the polymer resin cannot be judged during the period when the polymer resin is not secured. Therefore, there is a need to develop technology that predicts the quality of polymer resins in real time to monitor defects and process conditions of polymer products being produced, and to control and optimize the process according to the monitoring results.

The present invention was devised to solve the above problems, and its purpose is to provide a polymer quality prediction device and method that can predict polymer quality in real time.

Other objects and advantages of the present invention can be understood from the following description and will be more clearly understood by practicing the present invention. In addition, it will be readily apparent that the objects and advantages of the present invention can be realized by means and combinations thereof indicated in the claims.

An apparatus for predicting the quality of a polymer in a process in which a polymer is produced from raw materials introduced into a reactor according to one aspect of the present invention, comprising: a measuring unit configured to measure measurable process variables for the raw materials and the reactor; And configured to calculate polymerization variables for the raw material and the polymer from the process variables using a preset first model, and predict the quality of the polymer from the process variables and the polymerization variables using a preset second model. It may include a control unit.

The measuring unit may be configured to measure the flow rate of the raw material and the temperature and pressure of the reactor.

The process variables may be configured to include the flow rate of the raw material and the temperature and pressure of the reactor.

The first model may be a model preset to simulate the polymer process based on a thermodynamic model.

The control unit may be configured to input the process variables into the first model and calculate the polymerization variables representing chemical properties of the raw material and the polymer based on the output of the first model.

The polymerization variable may be configured to include at least one of the molar flow rate and mole fraction of the raw materials and the flow rate, moment, degree of polymerization, molecular weight, dispersion of molecular weight distribution, number of LCBs per time, and number of CBs per predetermined raw material. there is.

The second model may be a preset model to predict quality variables of the polymer based on machine learning.

The control unit may be configured to input the process variables and the polymerization variables into the second model and predict the quality of the polymer based on the output of the second model.

The control unit may be configured to predict at least one of fluidity, color, and refractive index of the polymer as quality of the polymer.

The control unit may be configured to predict the quality of the polymer by determining a plurality of main components for the process variable and the polymerization variable and inputting the determined plurality of main components into the second model.

The control unit calculates performance indicators for each of a plurality of models capable of determining the quality of the polymer using learning data corresponding to the process variables and the polymerization variables, and predicts the quality of the polymer based on the calculated performance indicators. It may be configured to set the model with the best performance as the second model.

A polymer processing system according to another aspect of the present invention may include a polymer quality prediction device according to an aspect of the present invention.

A polymer quality prediction method according to another aspect of the present invention is a method for predicting the quality of the polymer in a process in which a polymer is produced from raw materials flowing into a reactor, and includes measuring measurable process variables for the raw materials and the reactor. measurement step; A polymerization variable calculation step of calculating polymerization variables for the raw material and the polymer from the process variables using a preset first model; And it may include a quality prediction step of predicting the quality of the polymer from the process variables and the polymerization variables using a preset second model.

According to one aspect of the present invention, there is an advantage that the quality of the polymer can be predicted and evaluated in real time.

Additionally, according to one aspect of the present invention, there is an advantage of being able to predict quality indicators such as fluidity and/or color for which no theory-based formula exists using a machine learning-based model.

Additionally, according to one aspect of the present invention, since both process variables and polymerization variables are considered in the polymer quality prediction process, the accuracy of polymer quality prediction results can be improved.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

The following drawings attached to this specification serve to further understand the technical idea of the present invention along with the detailed description of the invention described later, and therefore the present invention should not be construed as limited to the matters described in such drawings.
Figure 1 is a diagram schematically showing the SAN resin processing process.
Figure 2 is a diagram schematically showing a polymer quality prediction device according to an embodiment of the present invention.
Figure 3 is a diagram schematically showing the process of predicting the quality of a polymer according to an embodiment of the present invention.
Figure 4 is a diagram schematically showing an example of polymerization variables according to an embodiment of the present invention.
Figure 5 is a diagram schematically showing the results of calculating performance indicators of a plurality of models for polymer fluidity by a polymer quality prediction device according to an embodiment of the present invention.
FIG. 6 is a diagram showing the experimental results for GPR among the experimental results of FIG. 5.
Figure 7 is a diagram schematically showing the results of calculating performance indices of a plurality of models for the color of a polymer by a polymer quality prediction device according to an embodiment of the present invention.
FIG. 8 is a diagram showing the experimental results for GPR among the experimental results of FIG. 7.
Figure 9 is a diagram schematically showing a polymer quality prediction method according to another embodiment of the present invention.

Terms or words used in this specification and claims should not be construed as limited to their common or dictionary meanings, and the inventor may appropriately define the concept of terms in order to explain his or her invention in the best way. It should be interpreted with meaning and concept consistent with the technical idea of the present invention based on the principle that it is.

Accordingly, the embodiments described in this specification and the configurations shown in the drawings are only one of the most preferred embodiments of the present invention and do not represent the entire technical idea of the present invention, so at the time of filing this application, various alternatives are available to replace them. It should be understood that equivalents and variations may exist.

Additionally, when describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description will be omitted.

Terms containing ordinal numbers, such as first, second, etc., are used for the purpose of distinguishing one of the various components from the rest, and are not used to limit the components by such terms.

Throughout the specification, when a part is said to “include” a certain element, this means that it does not exclude other elements, but may further include other elements, unless specifically stated to the contrary.

Additionally, throughout the specification, when a part is said to be "connected" to another part, this refers not only to the case where it is "directly connected" but also to the case where it is "indirectly connected" with another element in between. Includes.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings.

Figure 2 is a diagram schematically showing a polymer quality prediction device 100 according to an embodiment of the present invention.

The polymer quality prediction device 100 is a device for predicting the quality (v3) of a polymer in a process in which a polymer is produced from the raw material 10 flowing into the reactor 20.

For example, in the process for producing the SAN resin according to the embodiment of FIG. 1, the polymer quality prediction device 100 can predict the quality of the SAN resin in advance.

Referring to FIG. 2, the polymer quality prediction device 100 may include a measurement unit 110 and a control unit 120.

The measuring unit 110 may be configured to measure a measurable process variable (v1) for the raw material 10 and the reactor 20.

Specifically, the measuring unit 110 can measure various process variables (v1) that can be measured through a sensor during the polymer processing process.

For example, the measuring unit 110 may be configured to measure the flow rate of the raw material 10 and the temperature and pressure of the reactor 20. That is, the process variable v1 may be configured to include the flow rate of the raw material 10 and the temperature and pressure of the reactor 20.

In the embodiment of FIG. 1, the measuring unit 110 can measure the flow rate of the raw material 10 including AMS, AN, SM, and an initiator. Additionally, the measuring unit 110 can further measure the flow rates of MMX and MMR. Additionally, the measuring unit 110 can measure the temperature and pressure of the first and second reactors 22. Additionally, the measuring unit 110 may further measure the current of the first reactor 21 and the level and current of the second reactor 22.

In addition to the process variable (v1) described above, factors that can be sensed during the polymer processing process are measured by the measuring unit 110 and may be included in the process variable (v1).

Figure 3 is a diagram schematically showing the process of predicting polymer quality (v3) according to an embodiment of the present invention. Specifically, Figure 3 is a diagram schematically showing the operation process of the control unit 120 for predicting the quality (v3) of the polymer.

The control unit 120 may be configured to calculate the polymerization variable (v2) for the raw material 10 and the polymer from the process variable (v1) using a preset first model (m1).

Here, the first model (m1) may be a model preset to simulate the polymer process based on a thermodynamic model. For example, the first model (m1) may be a polymer non-random two-liquid (POLYNRTL) thermodynamic model. As a specific example, the ASPEN PLUS model may be applied to the first model (m1).

And, the polymerization variable (v2) may be a variable representing the chemical properties of the raw material 10 and the polymer. For example, the polymerization variable (v2) is the molar flow rate and mole fraction of the raw material (10), the flow rate of the polymer, moment, degree of polymerization, molecular weight, dispersion of molecular weight distribution, the number of long chain branches (LCB) per hour, and the predetermined raw material (10) It may be configured to include at least one of the number of CBs (chain branches).

Figure 4 is a diagram schematically showing an example of a polymerization variable (v2) according to an embodiment of the present invention. For example, referring to the non-limiting example of FIG. 4, the polymerization variable (v2) is the flow rate of polymer produced per hour (Rp), zero to second order moments (ZMOM, FMOM, SMOM), and monomer i per hour (e.g., i is the molar flow rate (SFLOW) of Sm, AN and AMS), mole fraction of monomer i (SFRAC), number average degree of polymerization (DP), mass average degree of polymerization (DPW), number average molecular weight (MWN), mass average molecular weight ( MWW), dispersion of polymer molecular weight distribution (PDI), number of LCBs per hour (LCB), and number of CBs per predetermined raw material (e.g., per 1,000 monomers) (FLCB).

Specifically, the control unit 120 inputs the process variable (v1) into the first model (m1), and the polymerization variable (v2) representing the chemical properties of the raw material 10 and the polymer based on the output of the first model (m1). ) can be configured to calculate. For example, in the embodiment of FIG. 3, the first model (m1) may be preset to receive a process variable (v1) and output a polymerization variable (v2).

The control unit 120 may be configured to predict the quality (v3) of the polymer from the process variable (v1) and the polymerization variable (v2) using a preset second model (m2).

Here, the second model (m2) may be a preset model to predict the quality (v3) variable of the polymer based on machine learning. For example, partial least squares (PLS), Tree, support vector machine (SVM), artificial neural network (ANN), and Gaussian process regression (GPR) may be applied to the second model (m2).

For example, in the embodiment of FIG. 3, the second model (m2) may be preset to receive process variables (v1) and polymerization variables (v2) and output polymer quality (v3). That is, not only the polymerization variable (v2) output from the first model (m1) but also the process variable (v1) sensed by the measuring unit 110 may be input to the second model (m2).

Specifically, the control unit 120 inputs the process variable (v1) and the polymerization variable (v2) into the second model (m2) and predicts the quality of the polymer (v3) based on the output of the second model (m2). It can be configured. For example, in the embodiment of FIG. 3, the second model (m2) may be preset to receive process variables (v1) and polymerization variables (v2) and output polymer quality (v3).

For example, the control unit 120 may be configured to predict at least one of polymer fluidity, color, and refractive index as polymer quality (v3).

The polymer quality prediction device 100 according to an embodiment of the present invention uses a machine learning-based model (second model (m2)) to measure the process variable (v1) and the thermodynamic model ( The quality (v3) of the polymer can be predicted from the polymerization parameter (v2) obtained using the first model (m1)). In other words, since the quality of the polymer (v3) can be predicted immediately when the process variable (v1) is measured, the quality can be predicted in real time during the polymer manufacturing process. In other words, because the quality (v3) of the polymer can be monitored in real time, the operation of various equipment in the process can be controlled and process optimization can proceed in real time. Accordingly, by using the polymer quality prediction device 100, the yield of the polymer can be significantly improved.

Meanwhile, the control unit 120 provided in the polymer quality prediction device 100 includes a processor, an application-specific integrated circuit (ASIC), another chipset, a logic circuit, and a processor known in the art to execute various control logics performed in the present invention. It may optionally include registers, communication modems, data processing devices, etc. Additionally, when the control logic is implemented as software, the control unit 120 may be implemented as a set of program modules. At this time, the program module is stored in memory and can be executed by the control unit 120. The memory may be internal or external to the control unit 120 and may be connected to the control unit 120 by various well-known means.

Additionally, the polymer quality prediction device 100 may further include a storage unit 130. The storage unit 130 may store data or programs necessary for each component of the polymer quality prediction device 100 to perform operations and functions, or data generated in the process of performing operations and functions. There is no particular limitation on the type of the storage unit 130 as long as it is a known information storage means that is known to be capable of recording, erasing, updating, and reading data. As an example, information storage means may include RAM, flash memory, ROM, EEPROM, registers, etc. Additionally, the storage unit 130 may store program codes in which processes executable by the control unit 120 are defined.

For example, the process variable v1 measured by the measurement unit 110 may be directly transmitted to the control unit 120 or stored in the storage unit 130. Additionally, mathematical expressions and relational expressions related to the first model (m1) and the second model (m2) may be stored in advance in the storage unit 130. The control unit 120 may access the storage unit 130 to obtain information related to the process variable (v1), the first model (m1), and the second model (m2).

The control unit 120 may be configured to determine a plurality of principal components for the process variable (v1) and the polymerization variable (v2).

For example, the control unit 120 may determine a plurality of principal components from a combination of the process variable (v1) and the polymerization variable (v2) using principal component analysis (PCA). For this purpose, the process variable (v1) and the polymerization variable (v2) may be data having orthogonality.

Here, each principal component refers to a combination of raw data (process variable (v1) and polymerization variable (v2)), with the first principal component (PC1) preserving the distribution of the original data the most, and the second principal component (PC2) can preserve the distribution of the original data next to the first principal component (PC1).

In addition, the control unit 120 may be configured to predict the quality (v3) of the polymer by inputting the determined plurality of main components into the second model (m2).

That is, the polymer quality prediction device 100 can more accurately predict the quality of the polymer (v3) based on a plurality of main components for the process variable (v1) and the polymerization variable (v2).

In order to more quickly predict the quality (v3) of the polymer and prevent unnecessary waste of system resources, the control unit 120 may input only some of the plurality of determined main components into the second model (m2).

Specifically, the control unit 120 may select a plurality of determined main components one by one until the cumulative distribution of the selected main components exceeds the threshold. For example, the control unit 120 may select the main components one by one in the order of the first main component (PC1), the second main component (PC2), and the third main component (PC3). And, if the cumulative distribution of the selected main component exceeds the threshold, the main component selection can be stopped.

For example, assume that the threshold is 60% and the distributions of the first to third principal components (PC1, PC2, PC3) are 36%, 19%, and 12%, respectively. Since the cumulative distribution is 0%, the control unit 120 can select the first principal component (PC1). Since the cumulative distribution is 36%, the control unit 120 can select the second principal component (PC2). Since the cumulative distribution is 55%, the control unit 120 can select the third principal component (PC3). Since the cumulative distribution is 67%, the control unit 120 can end the main component selection. And, the control unit 120 can predict the quality (v3) of the polymer by inputting the selected first to third main components (PC1, PC2, PC3) into the second model (m2).

That is, the polymer quality prediction device 100 can predict the quality (v3) of the polymer in real time using some of the plurality of main components. In addition, the polymer quality prediction device 100 reduces the number of main components to a level where reliability of the predicted polymer quality (v3) can be secured, thereby unnecessarily wasting system resources in the polymer quality (v3) prediction process. You can prevent it from happening.

The control unit 120 may be configured to calculate performance indicators for each of a plurality of models that can determine the quality (v3) of the polymer using learning data corresponding to the process variable (v1) and the polymerization variable (v2). In addition, the control unit 120 may be configured to set the model with the best polymer quality (v3) prediction performance as the second model (m2) based on the calculated performance index.

Figure 5 is a diagram schematically showing the results of the polymer quality prediction device 100 calculating performance indicators of a plurality of models for polymer fluidity according to an embodiment of the present invention. Specifically, the performance index for the process variable in FIG. 5 is the performance index when only the process variable is input to each model. In addition, the performance indicators for the process variables and comprehensive variables are performance indicators when the process variables and polymerization variables are respectively input into the model according to the present invention. The increase/decrease rate represents the increase/decrease rate of the result in which only the process variable is input and the result in which both the process variable and polymerization variable are input.

FIG. 6 is a diagram showing the experimental results for GPR among the experimental results of FIG. 5. In Figure 6, the X-axis may represent time. The Y-axis is fluidity and can represent the melting index (MI).

R ² is a coefficient of determination used in statistics and may be an index for evaluating the suitability of an estimated model. The value of R ² can be between 0 and 1, and the closer it is to 1, the better the performance of the model.

RMSE (root mean square error) is the root mean square error and may be an indicator used when dealing with the difference between the value predicted by the estimation model and the actual measured value. RMSE is known to be an appropriate indicator for indicating the precision of an estimation model. The closer the RMSE is to 0, the better the model may be.

MAPE (mean absolute percentage error) is the mean absolute percentage error, also called mean absolute percentage deviation (MAPD). MAPE means the average ratio of the absolute value of the error compared to the actual measured value. The closer MAPE is to 0, the better the model may be.

For example, in the embodiment of Figure 5, the control unit 120 can predict the fluidity of the polymer using PLS, Tree, SVM, ANN, and GPR. Here, relational expressions for PLS, Tree, SVM, ANN, and GPR may be stored in advance in the storage unit 130. As a result of the experiment, since R ² of GPR is the largest and RMSE and MAPE are the smallest, the control unit 120 can set the GPR model as the second model (m2) for predicting the fluidity of the polymer.

Figure 7 is a diagram schematically showing the results of the polymer quality prediction device 100 calculating performance indices of a plurality of models for the color of a polymer according to an embodiment of the present invention. The performance indices and increase/decrease rates for the process variables, process variables, and polymerization variables in FIG. 7 are the same as previously described with reference to FIG. 5 .

FIG. 8 is a diagram showing the experimental results for GPR among the experimental results of FIG. 7. In Figure 8, the X-axis may represent time. The Y axis is a color and can represent the b value of CIELAB color space.

For example, in the embodiment of Figure 7, the control unit 120 can predict the color of the polymer using PLS, Tree, SVM, ANN, and GPR. As a result of the experiment, since R ² of GPR is the largest and RMSE and MAPE are the smallest, the control unit 120 can set the GPR model as the second model (m2) for predicting the color of the polymer.

Referring to Figures 5 and 7, the control unit 120 may set the second model (m2) for predicting the fluidity and color of the polymer to the GPR model.

Unlike this embodiment, when the optimal model for each polymer quality (v3) is different, the control unit 120 may set the second model (m2) for each polymer quality (v3) differently. That is, the control unit 120 can improve the prediction accuracy for the quality by setting the optimal second model (m2) for each quality of the polymer.

Furthermore, the control unit 120 can not only independently set the second model (m2) for each quality of the polymer, but also set the optimal principal component corresponding to each second model (m2) to the second model (m2). ) can also be set independently. Therefore, the control unit 120 can reduce unnecessary waste of system resources in predicting various qualities of the polymer and increase the accuracy of the prediction results.

A polymer processing system according to another embodiment of the present invention may include a polymer quality prediction device 100 according to an embodiment of the present invention.

For example, a polymer processing system may include a raw material 10 inlet line through which the raw material 10 passes; A reactor (20) connected to a raw material (10) inlet line; A volatilization tank (30) connected to the reactor (20); A circulation line through which the non-volatile raw material 10 is connected to the raw material 10 inlet line; and a polymer quality prediction device 100.

Specifically, the measuring unit 110 of the polymer quality prediction device 100 is provided in each of the raw material 10 inlet line, reactor 20, volatilization tank 30, and circulation line to measure the process variable (v1). there is. The control unit 120 may be connected to the measurement unit 110 by wire and/or wirelessly and receive the process variable v1 from the measurement unit 110. And, as in the embodiment of FIG. 3, the control unit 120 may calculate the polymerization variable (v2) by inputting the process variable (v1) into the first model (m1). Then, the control unit 120 inputs the process variable (v1) and the polymerization variable (v2) into the second model (m2), and the polymer produced from the raw material 10 through the reactor 20 and the volatilization tank 30 is Quality (v3) can be predicted in real time.

According to the polymer processing system, the predicted polymer quality (v3) can be monitored in real time, so the operation of various equipment (e.g. valves, pumps, etc.) in the process can be controlled and the optimization of the process can be done in real time. It can proceed. Therefore, the quality (v3) of the produced polymer is uniform and can be improved.

Figure 9 is a diagram schematically showing a polymer quality prediction method according to another embodiment of the present invention.

Preferably, the polymer quality prediction method is a method for predicting the quality (v3) of the polymer in a process in which a polymer is produced from the raw material 10 flowing into the reactor 20, and each step of this method is performed by a polymer quality prediction device ( 100). Hereinafter, for convenience of explanation, content that overlaps with the content described above will be omitted or briefly described.

Referring to FIG. 9, the polymer quality prediction method may include a measurement step (S100), a polymerization variable (v2) calculation step (S200), and a quality prediction step (S300).

The measurement step (S100) is a step of measuring a measurable process variable (v1) for the raw material 10 and the reactor 20, and may be performed by the measurement unit 110.

Specifically, the measuring unit 110 can measure various process variables (v1) that can be measured through a sensor during the polymer processing process.

For example, the measuring unit 110 may be configured to measure the flow rate of the raw material 10 and the temperature and pressure of the reactor 20. That is, the process variable v1 may be configured to include the flow rate of the raw material 10 and the temperature and pressure of the reactor 20.

The polymerization variable (v2) calculation step (S200) is a step of calculating the polymerization variable (v2) for the raw material 10 and the polymer from the process variable (v1) using a preset first model (m1), and the control unit 120 It can be performed by .

Specifically, the control unit 120 inputs the process variable (v1) into the first model (m1), and the polymerization variable (v2) representing the chemical properties of the raw material 10 and the polymer based on the output of the first model (m1). ) can be configured to calculate. For example, in the embodiment of FIG. 3, the first model (m1) may be preset to receive a process variable (v1) and output a polymerization variable (v2).

The quality prediction step (S300) is a step of predicting the quality of the polymer (v3) from the process variable (v1) and the polymerization variable (v2) using a preset second model (m2) and can be performed by the control unit 120. there is.

Specifically, the control unit 120 inputs the process variable (v1) and the polymerization variable (v2) into the second model (m2) and predicts the quality of the polymer (v3) based on the output of the second model (m2). It can be configured. For example, in the embodiment of FIG. 3, the second model (m2) may be preset to receive process variables (v1) and polymerization variables (v2) and output polymer quality (v3).

More specifically, the control unit 120 may be configured to determine a plurality of principal components for the process variable (v1) and the polymerization variable (v2). In addition, the control unit 120 may be configured to predict the quality (v3) of the polymer by inputting the determined plurality of main components into the second model (m2).

The embodiments of the present invention described above are not only implemented through devices and methods, but may also be implemented through a program that realizes the function corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded. The implementation can be easily implemented by an expert in the technical field to which the present invention belongs based on the description of the embodiments described above.

Although the present invention has been described above with limited examples and drawings, the present invention is not limited thereto, and the technical idea of the present invention and the description below will be understood by those skilled in the art to which the present invention pertains. Of course, various modifications and variations are possible within the scope of equivalence of the patent claims.

In addition, the present invention described above is capable of various substitutions, modifications and changes without departing from the technical spirit of the present invention to those of ordinary skill in the technical field to which the present invention pertains. It is not limited by the drawings, and all or part of each embodiment can be selectively combined so that various modifications can be made.

10: Raw materials
20: reactor
21: first reactor
22: second reactor
30: Volatilization tank
100: Polymer quality prediction device
110: measuring unit
120: control unit
130: storage unit

Claims (10)

In a device for predicting the quality of the polymer in the process in which the polymer is produced from raw materials flowing into the reactor,
a measuring unit configured to measure measurable process variables for the raw material and the reactor; and
A control unit configured to calculate polymerization variables for the raw material and the polymer from the process variables using a preset first model, and predict the quality of the polymer from the process variables and the polymerization variables using a preset second model. A polymer quality prediction device comprising a.
According to paragraph 1,
The measuring unit,
configured to measure the flow rate of the raw material and the temperature and pressure of the reactor,
The process variables are,
A polymer quality prediction device, characterized in that it is configured to include the flow rate of the raw material and the temperature and pressure of the reactor.
According to paragraph 1,
The first model is,
It is a model preset to simulate the polymer process based on a thermodynamic model,
The control unit,
A polymer quality prediction device, characterized in that it is configured to input the process variables into the first model and calculate the polymerization variables representing chemical properties of the raw material and the polymer based on the output of the first model.
According to paragraph 3,
The polymerization variable is,
Polymer quality prediction, characterized in that it is configured to include at least one of the molar flow rate and mole fraction of the raw materials and the flow rate, moment, degree of polymerization, molecular weight, dispersion of molecular weight distribution, the number of LCBs per hour, and the number of CBs per predetermined raw material. Device.
According to paragraph 1,
The second model is,
It is a preset model to predict quality variables of the polymer based on machine learning,
The control unit,
A polymer quality prediction device, characterized in that it is configured to input the process variables and the polymerization variables into the second model and predict the quality of the polymer based on the output of the second model.
According to clause 5,
The control unit,
A polymer quality prediction device, characterized in that it is configured to predict at least one of fluidity, color, and refractive index of the polymer as the quality of the polymer.
According to clause 5,
The control unit,
A polymer quality prediction device, characterized in that it is configured to predict the quality of the polymer by determining a plurality of main components for the process variable and the polymerization variable and inputting the determined plurality of main components into the second model.
In clause 7,
The control unit,
Using the learning data corresponding to the process variables and the polymerization variables, performance indicators of each of a plurality of models capable of determining the quality of the polymer are calculated, and based on the calculated performance indicators, the quality prediction performance of the polymer is the best. A polymer quality prediction device, characterized in that it is configured to set a model to the second model.
A polymer processing system comprising a polymer quality prediction device according to any one of claims 1 to 8.
In a method for predicting the quality of the polymer in a process in which the polymer is produced from raw materials flowing into the reactor,
A measurement step of measuring measurable process variables for the raw material and the reactor;
A polymerization variable calculation step of calculating polymerization variables for the raw material and the polymer from the process variables using a preset first model; and
A polymer quality prediction method comprising a quality prediction step of predicting the quality of the polymer from the process variables and the polymerization variables using a preset second model.

Images