Production process of sebacic acid polycarbonate polyether polyol and circulating flow rate optimization control method

Abstract

The invention relates to the technical field of chemistry and chemical engineering, and particularly discloses a production process of sebacic acid polycarbonate polyether polyol and a circulation flow rate optimization control method, raw materials are premixed and then fed into a reaction pipeline at one time, CO2 is introduced into the reaction pipeline for reaction, and in the middle and rear sections of the reaction, the reaction temperature is controlled to be 20-30 DEG C, and the reaction time is controlled to be 3-5 hours; the conversion of sebacic acid is promoted by replenishing a catalyst, and meanwhile, after the catalyst is replenished, real-time monitoring and data analysis are carried out on the material density by using an intelligent industrial control technology so as to extract a time sequence correlation characteristic of a material density monitoring value and a time sequence correlation characteristic of a short-time variation of the material density; and the reaction process is comprehensively understood in combination with multi-dimensional material density characteristic information, so that intelligent recommendation of the circulation flow rate is performed. Through the mode, uniform mixing and efficient conversion of the materials can be ensured, residues of sebacic acid are reduced, and the acid value of the product is reduced, so that the stability and the high efficiency of the reaction process are improved.

Description

Technical Field

[0001] The present application relates to the field of chemical engineering technology, and more specifically, to a production process of sebacic acid polycarbonate polyether polyol and a circulation flow rate optimization control method. Background Art

[0002] Polycarbonate polyether polyols are a class of polyols with carbonate groups in the molecule and hydroxyl groups at the ends of the molecular chain. In the preparation process of this material, carbon dioxide is one of the raw materials. It is widely used as a chemical fixation of carbon dioxide due to its low cost, easy access, non-toxicity and non-flammability. Since the reaction process does not require high temperature and has low energy consumption, the regulated copolymerization of carbon dioxide and epoxide has broad application prospects and significant industrial value.

[0003] Sebacic acid polycarbonate polyether polyol is a polyol synthesized by polymerization of carbon dioxide and propylene oxide with DMC catalyst using sebacic acid as a small molecule initiator. However, due to the large acid value of the sebacic acid product, the acid value of the product is basically around 1.5 mgKOH / g. A higher acid value often limits the use of polyols in the synthesis of polyurethanes, causing synthesis difficulties and affecting its downstream applications. In the preparation process, the acid value of the sebacic acid product needs to be controlled below 0.3 mgKOH / g to make its downstream applications easy and widespread. Therefore, improving the conversion rate of sebacic acid and synthesizing a sebacic acid product with a low acid value have become current technical difficulties.

[0004] In the middle and later stages of the reaction, by adding a catalyst, more active centers can be provided for the reaction system, so as to react with more sebacic acid molecules and promote their conversion into polycarbonate polyether polyols, thereby reducing their residual amount in the final product and reducing the acid value of the product. After adding the catalyst, the number of active centers in the reaction system will increase significantly. At this time, in order to make better use of these active centers, accelerating the circulation flow rate of the material will help to more evenly distribute the newly added catalyst, enhance the fluidity and mixing degree of the reaction materials in the reaction pipeline, and reduce the uneven reaction caused by the addition of the catalyst. However, the traditional circulation flow rate adjustment method often relies on manual experience to set a fixed flow rate value, which is difficult to adapt to the dynamic changes in the reaction system in real time, and is easily affected by the personal experience and skills of the operator. It is not only inefficient, but also difficult to ensure the stability and repeatability of the reaction.

[0005] Therefore, an optimized production process of sebacic acid polycarbonate polyether polyol and a circulation flow rate optimization control method are expected. Summary of the invention

[0006] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides a novel production process and circulation flow rate optimization control method of sebacic acid polycarbonate polyether polyol, which feeds the raw materials into the tube-and-tube reaction device at one time after premixing, and passes CO2 into the reaction pipeline for reaction. In the middle and later stages of the reaction, the conversion of sebacic acid is promoted by adding a catalyst. At the same time, after adding the catalyst, the material density is monitored and analyzed in real time using intelligent industrial control technology to extract the time series correlation characteristics of the material density monitoring value and the time series correlation characteristics of the short-term change of the material density, and the reaction process is fully understood in combination with the multi-dimensional material density characteristic information, so as to carry out intelligent recommendation of the circulation flow rate. In this way, the uniform mixing and efficient conversion of the materials can be ensured, thereby effectively increasing the conversion rate of sebacic acid, reducing the residue of sebacic acid, and reducing the acid value of the product, so as to improve the stability and efficiency of the reaction process.

[0007] Accordingly, according to one aspect of the present application, a production process of sebacic acid polycarbonate polyether polyol is provided, wherein the production process is carried out in a tube-in-tube reaction device, wherein the tube-in-tube reaction device comprises a reaction pipe and a circulation pump, wherein a jacket is provided on the outside of the reaction pipe, and the circulation pump is connected to the reaction pipe through a connecting pipe, so as to allow the material to circulate in the reaction pipe; The molecular formula of the sebacic acid polycarbonate polyether polyol is: Formula (1); Wherein, in formula (1), R1 and R2 are selected from formula (2) or OH; Formula (2); Wherein, the molecular formula satisfies: m / (m+n)+n / (m+n)=1, 0.1<n / (m+n)<0.9, and * indicates the connection position; The production process comprises: a premixing step: mixing the epoxy compound, CO2, sebacic acid and the catalyst in a premixing kettle to obtain a premix; a feeding step: feeding the premix into the tube-in-tube reaction device at one time, and after the feeding is completed, introducing CO2 into the tube-in-tube reaction device until the pressure in the tube-in-tube reaction device is 2-6MPa, and maintaining the pressure; a reaction step: starting the circulation pump, adjusting the temperature of the medium in the jacket, heating the reaction pipeline to an initial temperature of 80-95°C for reaction, wherein the circulation pump is maintained at the first circulation flow rate, and when the material density is monitored to reach 1g / cm 3 When the material density reaches 1 g / cm 3, adding catalyst to the reaction pipeline, and adjusting the rate of the circulation pump to a second circulation flow rate, including: acquiring a time queue of the material density collected by a densitometer; respectively extracting the absolute quantity time series change characteristics and the change quantity local neighborhood characteristics of the time queue of the material density to obtain a sequence of material density time series correlation feature vectors and material density change quantity local neighborhood feature vectors; performing attention query joint encoding on the sequence of the material density time series correlation feature vectors and the material density change quantity local neighborhood feature vectors to obtain a material density-material density change joint query code representation vector; and generating a recommended value for the second circulation flow rate based on the material density-material density change joint query code representation vector.

[0008] In the above-mentioned production process of sebacic acid polycarbonate polyether polyol, the absolute quantity time series change characteristics and the local neighborhood characteristics of the change quantity of the time queue of the material density are extracted respectively to obtain the material density time series correlation feature vector and the sequence of the local neighborhood feature vector of the material density change quantity, including: inputting the time queue of the material density into the material density time series feature extractor based on the Bi-LSTM model to obtain the material density time series correlation feature vector; calculating the difference between the material density of each two adjacent predetermined time points in the time queue of the material density to obtain the time queue of the material density time series neighborhood change quantity; inputting the time queue of the material density time series neighborhood change quantity into the material density change quantity local neighborhood feature extractor based on the 1D-CNN model to obtain the sequence of the material density change quantity local neighborhood feature vector.

[0009] In the above-mentioned production process of sebacic acid polycarbonate polyether polyol, the material density time series association feature vector and the sequence of the local neighborhood feature vector of the material density change are subjected to attention query joint encoding to obtain a material density-material density change joint query code representation vector, including: performing local neighborhood implicit optimization on the sequence of the local neighborhood feature vector of the material density change to obtain a sequence of local neighborhood semantic association feature vectors of the material density change; performing attention matching query and multi-granularity query feature fusion on the material density time series association feature vector, the sequence of the local neighborhood feature vector of the material density change, and the sequence of the local neighborhood semantic association feature vector of the material density change to obtain the material density-material density change joint query code representation vector.

[0010] In the above-mentioned production process of sebacic acid polycarbonate polyether polyol, the sequence of local neighborhood feature vectors of material density variation is subjected to local neighborhood implicit optimization to obtain a sequence of local neighborhood semantic association feature vectors of material density variation, including: performing one-dimensional convolution encoding on the sequence of local neighborhood feature vectors of material density variation to obtain a sequence of local neighborhood semantic association feature vectors of material density variation, wherein the scale of the one-dimensional convolution encoding is equal to an integer multiple of the length of the local neighborhood feature vector of the material density variation.

[0011] In the production process of the above-mentioned sebacic acid polycarbonate polyether polyol, the material density time series association feature vector, the sequence of the local neighborhood feature vector of the material density change, and the sequence of the local neighborhood semantic association feature vector of the material density change are subjected to attention matching query and multi-granularity query feature fusion to obtain the material density-material density change joint query coding representation vector, including: using the value embedding coding matrix to process the material density time series association feature vector to obtain a value feature vector; using the material density time series association feature vector as the query vector, the value feature vector as the value vector and the sequence of the local neighborhood feature vector of the material density change as the sequence of key vectors, inputting the query vector, the value vector and the sequence of the key vector into the first multi-head attention query module based on the converter structure to obtain a sequence of material density-material density change association query implicit vectors; using the material density time series association feature vector as the query vector, the value feature vector as the value vector and the sequence of the local neighborhood feature vector of the material density change as the sequence of key vectors. A query vector, the value feature vector as a value vector and a sequence of local neighborhood semantic association feature vectors of material density change as a sequence of key vectors, the query vector, the value vector and the sequence of key vectors are input into a second multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change local neighborhood association query implicit vectors; the positional mean vectors of the sequence of material density-material density change association query implicit vectors and the sequence of material density-material density change local neighborhood association query implicit vectors are respectively calculated to obtain a material density-material density change association query matching feature vector and a material density-material density change local neighborhood association query matching feature vector; the material density-material density change association query matching feature vector and the material density-material density change local neighborhood association query matching feature vector are fused to obtain the material density-material density change joint query encoding representation vector.

[0012] In the above-mentioned production process of sebacic acid polycarbonate polyether polyol, the material density time series association feature vector is used as the query vector, the value feature vector is used as the value vector, and the sequence of the local neighborhood feature vectors of the material density change is used as the sequence of key vectors. The query vector, the value vector and the sequence of the key vector are input into the first multi-head attention query module based on the converter structure to obtain a sequence of material density-material density change association query implicit vectors, including: multiplying the query vector by the transposed vector of the key vector and dividing it by the square root of the scale of the key vector to obtain an attention score matrix; passing the attention score matrix through a softmax function and then multiplying it with the value vector to obtain the material density-material density change association query implicit vector.

[0013] In the production process of the above-mentioned sebacic acid polycarbonate polyether polyol, the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector are fused to obtain the material density-material density change joint query code representation vector, including: calculating the position point addition, position point subtraction and position point multiplication between the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector to obtain a first material density-material density change query fusion feature vector, a second material density-material density change query fusion feature vector and a third material density-material density change query fusion feature vector; after cascading the first material density-material density change query fusion feature vector, the second material density-material density change query fusion feature vector and the third material density-material density change query fusion feature vector into a material density-material density change multi-scale fusion feature vector, one-dimensional convolution processing and maximum pooling processing are performed on it to obtain the material density-material density change joint query code representation vector.

[0014] In the above-mentioned production process of sebacic acid polycarbonate polyether polyol, based on the material density-material density change joint query code representation vector, a recommended value of the second circulation flow rate is generated, including: inputting the material density-material density change joint query code representation vector into a decoder-based circulation flow rate optimization module to obtain an optimization instruction, and the optimization instruction includes a decoded value of the recommended value of the second circulation flow rate.

[0015] In the production process of the above-mentioned sebacic acid polycarbonate polyether polyol, the reaction step includes: during the reaction process, monitoring the reaction temperature and material density in the reaction pipe, and performing the following steps: when it is monitored that the CO2 consumption value in the reaction pipe is greater than 0.1MPa, and the reaction temperature rise value exceeds 2°C, reducing the medium temperature in the jacket to 70-80°C, until the reaction temperature is monitored to be reduced to 80-95°C, then increasing the medium temperature in the jacket to 80-95°C; when it is monitored that the material density reaches the reaction end point, terminating the reaction, wherein the reaction end point is: the material density is 1.1-1.3g / cm 3 , and / or, the fluctuation value of the material density within 5 minutes is not greater than 0.01g / cm 3 .

[0016] According to another aspect of the present application, a method for optimizing and controlling a circulation flow rate is provided, which includes: obtaining a time queue of material density collected by a densitometer; respectively extracting the absolute quantity time series change characteristics and the change quantity local neighborhood characteristics of the time queue of the material density to obtain a sequence of material density time series association feature vectors and material density change quantity local neighborhood feature vectors; performing attention query joint encoding on the sequence of the material density time series association feature vectors and the material density change quantity local neighborhood feature vectors to obtain a material density-material density change joint query code representation vector; and generating a recommended value for a second circulation flow rate based on the material density-material density change joint query code representation vector.

[0017] Compared with the prior art, the production process and circulation flow rate optimization control method of the sebacic acid polycarbonate polyether polyol described in the present application, which feeds the raw materials into the tube reactor at one time after premixing, passes CO2 into the reaction pipeline for reaction, and in the middle and later stages of the reaction, promotes the conversion of sebacic acid by adding a catalyst, and at the same time, after adding the catalyst, uses intelligent industrial control technology to monitor and analyze the material density in real time, so as to extract the time series correlation characteristics of the material density monitoring value and the time series correlation characteristics of the short-term change of the material density, and combines the multi-dimensional material density characteristic information to fully understand the reaction process, so as to make intelligent recommendations for the circulation flow rate. In this way, it is possible to ensure the uniform mixing and efficient conversion of materials, thereby effectively increasing the conversion rate of sebacic acid, reducing the residue of sebacic acid, and reducing the acid value of the product, so as to improve the stability and efficiency of the reaction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 It is an overall schematic diagram of a tube-in-tube reaction device for producing sebacic acid polycarbonate polyether polyol according to an embodiment of the present application.

[0020] Figure 2 The flowchart is a production process of sebacic acid polycarbonate polyether polyol according to an embodiment of the present application.

[0021] Figure 3 The present invention is a flow chart of adding a catalyst to the reaction pipe and adjusting the rate of the circulation pump to a second circulation flow rate in the production process of sebacic acid polycarbonate polyether polyol according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0023] The present application proposes a production process of sebacic acid polycarbonate polyether polyol. Specifically, the sebacic acid polycarbonate polyether polyol uses sebacic acid as a small molecule initiator, propylene oxide and CO2 under the action of DMC catalyst to synthesize the sebacic acid polycarbonate polyether polyol. The molecular formula of the sebacic acid polycarbonate polyether polyol is: Formula (1); Wherein, in formula (1), R1 and R2 are selected from formula (2) or OH; Formula (2); The molecular formula satisfies the following conditions: m / (m+n)+n / (m+n)=1, 0.1<n / (m+n)<0.9, and * indicates the connection position.

[0024] Specifically, the production process is carried out in a tube-in-tube reaction device. Figure 1 Schematic diagram of the overall structure of the tube-in-tube reaction device for producing sebacic acid polycarbonate polyether polyol according to the embodiment of the present application. Figure 1As shown, the tube-in-tube reaction device comprises a premixing kettle 101 and a reaction unit, and the raw materials are premixed in the premixing kettle. The premixing kettle 101 is provided with a stirrer M. The premixed raw materials are introduced into the reaction unit and reacted in the reaction pipe 109 in the reaction unit.

[0025] Specifically, in a preferred embodiment, the reaction unit includes a reaction pipe 109, a circulation pump 110 and a carbon dioxide gas replenishment port 102. There are multiple reaction pipes 109, and multiple reaction pipes 109 are arranged in parallel. In each reaction pipe 109, the total volume of the material is 50-70% of the volume of the reaction pipe, and a jacket 106 is provided on the outside of the reaction pipe 109, and the reaction pipe 109 is temperature-controlled by passing steam, water, heat transfer oil and other media into the jacket. The space between the multiple reaction pipes 109 is filled with a heat exchange medium, and an insulation layer is provided on the outside. The carbon dioxide gas replenishment port 102 is connected to the reaction pipe 109 for replenishing carbon dioxide into the reaction pipe 109. The feed end and the discharge end of the reaction pipe 109 are respectively connected to the liquid distributor and the circulation pump 110 through a connecting pipe, so that the material circulates in the reaction pipe 109.

[0026] Further preferably, the space in the reaction pipe 109 is divided into a gas-liquid mixing heat conduction area and a solid-liquid mixing area. The gas-liquid mixing heat conduction area is the CO2 space at the top of the reaction pipe. The loop between the gas-liquid mixing heat conduction area, the reaction pipe 109 and the circulating pump 110 together constitutes a gas-liquid mixing heat conduction path, so that the mixed material with a low CO2 content in the mixed material re-dissolves CO2, and heat exchange is performed in the gas-liquid mixing heat conduction unit. A static mixer 107 is provided in the solid-liquid mixing area. The loop between the static mixer 107, the reaction pipe 109 and the circulating pump 110 together constitutes a solid-liquid mixing path, so that the solid catalyst is evenly dispersed in the epoxy compound, and the generated reaction heat is dispersed at the same time, reducing the temperature difference between the center temperature in the reaction pipe and the pipe wall, and avoiding the generation of an explosion point.

[0027] Further preferably, the upper dispersion sheet 105 is arranged at a position of 10 to 50 mm near the top of the reaction pipe 109, and it is provided with a plurality of dispersion holes; the lower dispersion sheet 108 is arranged at a position of 10 to 50 mm near the bottom of the reaction pipe 109, and it is provided with a plurality of dispersion holes. By setting the upper dispersion sheet 105, the reaction materials can be evenly dispersed, and flow down in a water droplet shape at the end, increasing the mass transfer and heat transfer area, and setting the lower dispersion sheet 108 for fixing the static mixer 107, and further dispersing and mixing the products flowing down. The dispersion sheet cooperates with the built-in static mixer 107 and the circulating pump 110 of the reaction pipe 109, and the entire circulating dispersion system can replace the traditional stirring process to continuously mix the reaction materials during the circulation process. At the same time, the dispersion sheet can effectively enable the reaction materials to perform better gas-liquid mass transfer and release excess heat in the gas area.

[0028] Preferably, a temperature sensor is provided inside the reaction pipe 109, and a density meter is provided at the outlet, and the reaction temperature is measured by the temperature sensor, and the material density is measured by the density meter. The jacket temperature is adjusted according to the material density and the internal temperature of the tube-in-tube reaction device to control the reaction rate.

[0029] Furthermore, Figure 2 Flow chart of the production process of sebacic acid polycarbonate polyether polyol according to the embodiment of the present application. Figure 2 As shown, the production process of sebacic acid polycarbonate polyether polyol described in the present application includes the following steps: premixing step S1: mixing epoxy compound, CO2, sebacic acid and catalyst in a premixing kettle to obtain a premix; feeding step S2: feeding the premix into the tube-in-tube reaction device at one time, and after the feeding is completed, introducing CO2 into the tube-in-tube reaction device until the pressure of the tube-in-tube reaction device is 2~6MPa, and maintaining the pressure; reaction step S3: starting the circulation pump, adjusting the temperature of the medium in the jacket, and heating the reaction pipe 109 to an initial temperature of 89~91°C for reaction, wherein the circulation pump is maintained at the first circulation flow rate, and when the material density is monitored to reach 1g / cm 3 When the catalyst is added to the reaction pipe 109, the rate of the circulation pump is adjusted to a second circulation flow rate, and the rate of the second circulation flow rate is greater than the first circulation flow rate.

[0030] Specifically, in the premixing step S1, in the premixing zone, carbon dioxide is provided by a carbon dioxide storage tank or carbon dioxide recovered and purified by a carbon dioxide purification device, and is passed into a premixing kettle, and the propylene oxide raw material enters the premixing kettle in the carbon dioxide atmosphere through a metering pump. A fixed amount of sebacic acid and a catalyst are added to the premixing kettle using a weighing bucket. After the raw materials are mixed in the premixing kettle, a premix is ​​obtained.

[0031] In one embodiment of the present application, in the premixing step S1, the catalyst is a DMC catalyst; the dosage ratio of the epoxy compound to the sebacic acid is 15:1 to 70:1, and the dosage of the catalyst is 400-900ppm of the dosage of the epoxy compound; the pressure of CO2 in the premixing kettle is 0.1~2MPa; the premixing temperature is 0~40°C, and the premixing time is 1~2h; preferably, the pressure in the premixing kettle is 0.1~0.5MPa, the premixing temperature is 40°C, and the premixing time is 1h.

[0032] Specifically, in the feeding step S2, the premix is ​​fed into the reaction pipe 109 at one time through the feed port. When the raw material enters the reaction pipe 109, the reaction unit is controlled to meet the following conditions: the temperature of the reaction pipe 109 is 35-45°C and the CO2 pressure is 0.1-2MPa. More preferably, during the feeding process, a lower temperature is maintained to avoid the initiator from decomposing into small molecules. When the raw material completely enters the reaction unit, the feeding is completed. Carbon dioxide is supplemented through the carbon dioxide replenishment port 102 to a pressure of 2-6 Mpa in the reaction pipe 109. More preferably, the pressure in the reaction pipe 109 is 3MPa. When the pressure is too low, a solenoid valve is used to automatically introduce carbon dioxide from the carbon dioxide replenishment port 102 to supplement the pressure.

[0033] Specifically, in the reaction step S3, the circulation pump 110 is turned on to adjust the temperature of the medium in the jacket 106 to raise the temperature of the reaction pipe 109 to an initial temperature of 89-91° C. (for example, to an initial temperature of 90° C.) to carry out the reaction. During the entire reaction process, the reaction temperature and material density in the reaction pipe 109 are monitored by a temperature sensor and a density meter.

[0034] When the CO2 consumption value in the reaction pipe 109 is monitored to be greater than 0.1MPa, and the material reaction temperature rises by more than 2°C, the medium temperature in the jacket is reduced to 79-81°C (for example, to 80°C), until the material temperature is monitored to drop to 89-91°C (for example, to 90°C), and then the medium temperature in the jacket is increased to keep the material temperature at the reaction temperature. Since the sebacic acid product itself requires a high reaction temperature, the CO2 consumption value and the increase in reaction temperature can be used to determine the possibility of a too fast reaction rate, and temperature control can be performed in time to avoid the risk of implosion.

[0035] When the material density reaches 1g / cm 3 When the catalyst is added from the catalyst adding tank 103 to the reaction pipe 109 through the feed mixer 104, the rate of the circulation pump 110 is increased from the first circulation flow rate to the second circulation flow rate, and the rate of the second circulation flow rate is greater than the first circulation flow rate. The material density reaches 1g / cm 3When the reaction is carried out to the middle and later stages, some of the initially added catalysts may have reduced activity due to various reasons (such as deactivation, dilution, etc.). By adding catalysts, more active centers can be provided for the reaction system to react with more sebacic acid molecules, promote their conversion into polycarbonate polyether polyols, thereby reducing their residual amount in the final product and reducing the acid value of the product. After adding catalysts, the number of active centers in the reaction system will increase significantly. At this time, in order to better utilize these active centers, accelerating the circulation flow rate of the material helps to more evenly distribute the newly added catalyst, enhance the fluidity and mixing degree of the reaction materials in the reaction pipe 109, which is not only conducive to mass transfer between gas and liquid, but also helps to effectively transfer the reaction heat, prevent local overheating, and maintain the stability of the reaction temperature.

[0036] In one embodiment of the present application, considering that the traditional circulation flow rate adjustment method often relies on manual experience judgment, it is easily affected by the personal experience and skills of the operator, which is not only inefficient, but also difficult to ensure the stability and repeatability of the reaction. In order to overcome this problem, the present application proposes a circulation flow rate intelligent adjustment method based on deep learning technology, which extracts the time series correlation characteristics of the material density monitoring value and the time series correlation characteristics of the short-term change of the material density by real-time monitoring and data analysis of the material density, and comprehensively understands the reaction process by combining multi-dimensional material density feature information, thereby realizing intelligent recommendation of the circulation flow rate to improve the stability and efficiency of the reaction process.

[0037] Figure 3 The present invention is a flow chart of adding catalyst to the reaction pipe 109 and adjusting the speed of the circulation pump to the second circulation flow rate in the production process of sebacic acid polycarbonate polyether polyol according to an embodiment of the present application. Figure 3 As shown, the reaction step S3 also includes: S310, obtaining the time queue of the material density collected by the densitometer; S320, respectively extracting the absolute quantity time series change characteristics and the change quantity local neighborhood characteristics of the time queue of the material density to obtain a sequence of material density time series correlation feature vectors and material density change quantity local neighborhood feature vectors; S330, performing attention query joint encoding on the sequence of the material density time series correlation feature vectors and the material density change quantity local neighborhood feature vectors to obtain a material density-material density change joint query code representation vector; S340, generating a recommended value of the second circulation flow rate based on the material density-material density change joint query code representation vector.

[0038] Specifically, first, through the step S310, the time queue of the material density collected by the density meter is obtained. It should be understood that in the production process of sebacic acid polycarbonate polyether polyol, the material density and the rate of change of the material density can be used as important indicators of the reaction process. For example, in the early stage of the reaction, the material density is low and the rate of increase is fast, indicating that the reaction has just started and the reactants are rapidly converted into products; after the catalyst is added, the material density is high and the rate of increase is improved compared to before the catalyst is added, indicating that the number of active centers in the reaction system increases, the reaction rate is accelerated, and the circulation flow rate also needs to be accelerated accordingly to ensure the uniformity of the reaction; and when the material density is high and the rate of change slows down, or even tends to be stable, it means that the reaction is close to completion, and the circulation flow rate can be appropriately reduced at this time to avoid unnecessary energy consumption and possible side reactions. Based on this, the present application monitors the material density in real time to grasp the reaction process, and dynamically adjusts the circulation flow rate accordingly to ensure that the circulation flow rate matches the reaction process.

[0039] In order to achieve this goal, the present application further adopts time series analysis technology to extract the absolute quantity time series change characteristics and the local neighborhood characteristics of the change quantity of the time series of the material density. In the step S320, first, in order to mine the time series feature information of the material density and ensure the comprehensiveness and accuracy of the feature extraction, the present application inputs the time series of the material density into the material density time series feature extractor based on the Bi-LSTM model for time series encoding, so as to utilize the bidirectional learning ability of the Bi-LSTM model, and consider the forward and backward time series context information in the time series data, so as to better capture the long-term dependency of the material density, mine the time series change law of the material density, and thus generate the material density time series correlation feature vector.

[0040] Next, in order to further obtain the rate of change information of the material density. In the step S320, the difference between the material density at each two adjacent predetermined time points in the time queue of the material density is calculated to quantify the change in material density, and obtain the time queue of the time series neighborhood change in material density. Next, in order to understand the subtle changes in the material density change in more detail, so as to capture transient events and rapid changes in the reaction process (such as the sudden increase in reaction rate after adding a catalyst), the present application further inputs the time queue of the material density time series neighborhood change into the material density change local neighborhood feature extractor based on the 1D-CNN model for feature extraction, so as to utilize the local receptive field characteristics of the 1D-CNN model, focus on capturing the local features of the material density change, thereby refining the granularity of data analysis and improving the accuracy of feature representation, so as to generate a sequence of local neighborhood feature vectors of the material density change.

[0041] Correspondingly, the step S320 includes: inputting the time queue of the material density into the material density time series feature extractor based on the Bi-LSTM model to obtain the material density time series correlation feature vector; calculating the difference between the material density at each two adjacent predetermined time points in the time queue of the material density to obtain the time queue of the material density time series neighborhood change amount; inputting the time queue of the material density time series neighborhood change amount into the material density change amount local neighborhood feature extractor based on the 1D-CNN model to obtain a sequence of the material density change amount local neighborhood feature vectors.

[0042] Then, in step S330, in order to comprehensively utilize the complementary information of the sequence of the material density time series association feature vector and the local neighborhood feature vector of the material density change, the present application proposes a feature fusion method based on the attention mechanism, by performing local neighborhood implicit optimization on the sequence of the local neighborhood feature vector of the material density change to enhance the sensitivity to the material density change rate, thereby improving the accuracy of feature fusion. Specifically, firstly, a one-dimensional convolution kernel whose scale is an integer multiple of the length of the local neighborhood feature vector of the material density change is used to perform one-dimensional convolution encoding on the sequence of the local neighborhood feature vector of the material density change to capture the contextual information within a specific local neighborhood range in the sequence, and generate a sequence of semantic association feature vectors of the local neighborhood of the material density change, thereby better reflecting the feature correlation of the local neighborhood. Next, a multi-head attention mechanism based on a converter structure is used for feature matching fusion, which constructs a query vector and a value vector based on the material density temporal association feature vector, and uses the sequence of the local neighborhood feature vector of the material density change and the sequence of the local neighborhood semantic association feature vector of the material density change as the sequence of the key vector, respectively, to learn the semantic association between multi-dimensional material density information from different perspectives, and at the same time focus on the local neighborhood features of the material density change and the association characteristics between local neighborhoods, so as to comprehensively consider local and global information and improve the accuracy of feature matching fusion. Next, the feature matching fusion information of different granularity levels is summarized by calculating the position mean, and the material density-material density change association query matching feature vector and the material density-material density change local neighborhood association query matching feature vector of different granularity levels are generated, and the two are interactively fused in multi-dimensional semantics to integrate the information of different granularity levels and generate the material density-material density change joint query encoding representation vector, so as to fully reflect the complex relationship between material density and its change rate, so as to more accurately understand the reaction process.

[0043] In a specific example of the present application, the step S330 includes: performing local neighborhood implicit optimization on the sequence of local neighborhood feature vectors of the material density change to obtain a sequence of local neighborhood semantic association feature vectors of the material density change; performing attention matching query and multi-granularity query feature fusion on the material density time series association feature vector, the sequence of local neighborhood feature vectors of the material density change, and the sequence of local neighborhood semantic association feature vectors of the material density change to obtain the material density-material density change joint query encoding representation vector.

[0044] Specifically, the sequence of local neighborhood feature vectors of material density variation is subjected to local neighborhood implicit optimization to obtain a sequence of local neighborhood semantic association feature vectors of material density variation, including: performing one-dimensional convolution coding on the sequence of local neighborhood feature vectors of material density variation to obtain a sequence of local neighborhood semantic association feature vectors of material density variation, wherein the scale of the one-dimensional convolution coding is equal to an integer multiple of the length of the local neighborhood feature vector of the material density variation.

[0045] Specifically, the material density time series association feature vector is subjected to attention matching query and multi-granularity query feature fusion with the sequence of local neighborhood feature vectors of the material density change and the sequence of local neighborhood semantic association feature vectors of the material density change to obtain the material density-material density change joint query coding representation vector, including: using the value embedding coding matrix to process the material density time series association feature vector to obtain a value feature vector; using the material density time series association feature vector as the query vector, the value feature vector as the value vector and the sequence of local neighborhood feature vectors of the material density change as the sequence of key vectors, inputting the query vector, the value vector and the sequence of key vectors into a first multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change association query implicit vectors; using the material density time series association feature vector as the query vector, the value feature vector as the value vector and the sequence of local neighborhood feature vectors of the material density change as the sequence of key vectors. The feature vector is used as the value vector and the sequence of the local neighborhood semantic association feature vectors of the material density change is used as the sequence of key vectors, and the sequence of the query vector, the value vector and the key vector is input into the second multi-head attention query module based on the converter structure to obtain a sequence of material density-material density change local neighborhood association query implicit vectors; the positional mean vectors of the sequence of the material density-material density change association query implicit vectors and the sequence of the material density-material density change local neighborhood association query implicit vectors are calculated respectively to obtain the material density-material density change association query matching feature vector and the material density-material density change local neighborhood association query matching feature vector; the material density-material density change association query matching feature vector and the material density-material density change local neighborhood association query matching feature vector are fused to obtain the material density-material density change joint query encoding representation vector.

[0046] More specifically, the material density time series association feature vector is used as the query vector, the value feature vector is used as the value vector, and the sequence of local neighborhood feature vectors of the material density change is used as the sequence of key vectors. The query vector, the value vector and the sequence of key vectors are input into a first multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change association query implicit vectors, including: multiplying the query vector by the transposed vector of the key vector and dividing the result by the square root of the scale of the key vector to obtain an attention score matrix; passing the attention score matrix through a softmax function and then multiplying it with the value vector to obtain the material density-material density change association query implicit vector.

[0047] More specifically, the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector are fused to obtain the material density-material density change joint query code representation vector, including: calculating the point-by-point addition, point-by-point subtraction and point-by-point multiplication between the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector to obtain a first material density-material density change query fusion feature vector, a second material density-material density change query fusion feature vector and a third material density-material density change query fusion feature vector; after cascading the first material density-material density change query fusion feature vector, the second material density-material density change query fusion feature vector and the third material density-material density change query fusion feature vector into a material density-material density change multi-scale fusion feature vector, one-dimensional convolution processing and maximum value pooling processing are performed on it to obtain the material density-material density change joint query code representation vector.

[0048] Accordingly, the step S330 includes: processing the sequence of the material density time series association feature vector and the material density change local neighborhood feature vector using the following joint query coding formula to obtain the material density-material density change joint query coding representation vector, wherein the joint query coding formula is:

[0049]

[0050]

[0051]

[0052]

[0053]

[0054] ,

[0055]

[0056] in, represents the sequence of local neighborhood feature vectors of the material density variation, , , and are respectively the first, second, and third in the sequence of local neighborhood feature vectors of the material density variation. and The local neighborhood feature vector of material density variation, The value of is the number of local neighborhood feature vectors of the material density variation. represents one-dimensional convolution processing, represents the length of the one-dimensional convolution kernel, A sequence of semantically associated feature vectors representing the local neighborhood of the material density variation, , , and are the first, second, and third in the sequence of the local neighborhood semantic association feature vectors of the material density variation. and The local neighborhood semantic association feature vector of material density variation, The value of is the number of semantic association feature vectors in the local neighborhood of the material density variation. represents the time series correlation feature vector of the material density, is the value embedding encoding matrix, is the bias term, is a value vector, is the normalized exponential function, The value of is the scale of the key vector, represents the transpose of a vector, Indicates Material density-material density change association query implicit vector, Indicates Material density-material density change local neighborhood association query implicit vector, Represents the matching feature vector of material density-material density change association query, Represents the material density-material density change local neighborhood association query matching feature vector, represents dot product, Indicates point addition, Indicates point reduction, Indicates cascade operation, represents the maximum pooling operation, represents a one-dimensional convolution operation, Represents the material density-material density change joint query coding representation vector.

[0057] Furthermore, the intelligent adjustment of the circulation flow rate is guided based on the material density-material density change joint query code representation vector. In a specific example of the present application, the step S340 includes: inputting the material density-material density change joint query code representation vector into a decoder-based circulation flow rate optimization module to obtain an optimization instruction, and the optimization instruction includes a decoded value of the recommended value of the second circulation flow rate. That is, the material density-material density change joint query code representation vector is processed using a decoder model to utilize the powerful decoding regression capability of the decoder model, and in a step-by-step progressive manner, the key information in the material density-material density change joint query code representation vector is gradually extracted, and converted into a specific circulation flow rate recommendation value, thereby obtaining the decoded value of the recommended value of the second circulation flow rate, thereby achieving optimized control of the circulation flow rate.

[0058] In the technical solution of the present application, the sequence of the material density time series correlation feature vector and the material density change local neighborhood feature vector respectively represent the time series correlation feature of the absolute amount of material density and the local time series correlation feature of the material density change in the local time domain window. When the sequence of the material density time series correlation feature vector and the material density change local neighborhood feature vector are input into the attention gated query network based on 1D convolution local neighborhood implicit optimization, the absolute amount of the material density and the relative amount of the material density have a data underlying correlation and the time window alignment offset of each material density change local neighborhood feature vector in the sequence of the material density time series correlation feature vector and the material density change local neighborhood feature vector will make the feature manifold of the material density-material density change joint query encoding representation vector obtained by the attention gated query encoding have expression sparsity and uneven distribution in the high-dimensional feature space, thereby affecting the accuracy of the optimization instructions obtained by the decoder-based cyclic flow rate optimization module.

[0059] Based on this, in a preferred embodiment, the material density-material density change joint query code representation vector is input into a decoder-based circulation flow rate optimization module to obtain an optimization instruction, including: Determine the number of zero eigenvalues ​​in the material density-material density change joint query code representation vector, and respectively calculate the reciprocal of the logarithm of the number of zero eigenvalues ​​with base two and the exponential value of the reciprocal of the number of zero eigenvalues ​​with base natural constants to obtain a first material density-material density change joint query representation value and a second material density-material density change joint query representation value; Calculate each eigenvalue of the material density-material density change joint query code representation vector as a power function with the inverse of the number of zero eigenvalues ​​as an exponent to obtain a material density-material density change joint query code structure state vector; Performing a dot multiplication of the material density-material density change joint query code structure state vector and the first material density-material density change joint query code representation value to obtain a material density-material density change joint query code information representation vector; The autocorrelation matrix of the material density-material density change joint query coding structure state vector is respectively multiplied by the second material density-material density change joint query coding representation value and weight hyperparameter to obtain a material density-material density change joint query coding regression understanding matrix; Performing matrix multiplication of the material density-material density change joint query coded information representation vector and the material density-material density change joint query coded regression understanding matrix to obtain an optimized material density-material density change joint query coded representation vector; The optimized material density-material density change joint query encoding representation vector is passed through a decoder-based circulation flow rate optimization module to obtain the optimization instruction.

[0060] That is, the first material density-material density change joint query code representation value and the second material density-material density change joint query code representation value are expressed as:

[0061] in, is the number of zero eigenvalues ​​in the material density-material density change joint query code representation vector, is the first material density-material density change joint query code representation value, is the second material density-material density change joint query code representation value, It is an exponential function with a natural constant as its base.

[0062] And the optimized material density-material density change joint query code representation vector is expressed as:

[0063]

[0064]

[0065] in, is the material density-material density change joint query code representation vector, is the weight hyperparameter, is the material density-material density change joint query encoding regression comprehension matrix, is the material density-material density change joint query coded information representation vector, is the number of zero eigenvalues ​​in the material density-material density change joint query code representation vector, is the first material density-material density change joint query code representation value, is the second material density-material density change joint query code representation value, represents a power function in which each eigenvalue of the material density-material density change joint query code representation vector is calculated with the inverse of the number of zero eigenvalues ​​as the exponent, represents the transpose of a vector, It is the point multiplication by position. is matrix multiplication, It is the optimized material density-material density change joint query coding representation vector.

[0066] Based on this, in the above preferred example, by modeling the zero dimension of the vector field of the material density-material density change joint query code representation vector after the material density-material density change joint query code representation vector is expanded as the global field redundant dependency, the effective long-range modeling of the material density-material density change joint query code representation vector under the field linear complexity is realized, so that in the feature regression process of the material density-material density change joint query code representation vector, the structured correlation self-distillation between the information representation and the decoding regression understanding is performed in the selective feature enhancement state space based on the feature distribution, so as to maintain the overall information complexity of the feature distribution of the material density-material density change joint query code representation vector while completing the feature structure modulation to make the feature more in line with the requirements of the objective function, reduce redundant information, and maximize the retention of useful information, so as to improve the learning efficiency and prediction performance, that is, to improve the accuracy of the harmful instructions obtained by the material density-material density change joint query code representation vector through the decoder-based cyclic flow rate optimization module.

[0067] When the material density reaches the reaction end point, the reaction is terminated. The reaction end point is: the material density is 1.1~1.3 g / cm 3 , and / or, the fluctuation value of material density within 5 minutes is not greater than 0.01g / cm 3 .

[0068] Further preferably, in the above reaction process, the reaction mixture product at the discharge end of the reaction pipeline 109 converges to the circulation pump 110 via a connecting pipeline at one end, and converges to the liquid distributor at the top of the reaction pipeline 109 through the connecting pipeline of the circulation pump 110 and the other end, and then passes through the solid-liquid mixing unit and the gas-liquid mixing heat conduction unit of the reaction pipeline 109 to complete a cycle. The circulation pump 110 is controlled to complete a cycle within 1 to 20 minutes. Further preferably, the circulation pump is controlled to complete a cycle within 10 to 15 minutes. More preferably, the circulation pump is controlled to complete a cycle within 10 minutes. By regulating the circulation flow rate of the circulation pump 110, it is ensured that the reaction process is fully carried out by heat conduction, gas-liquid mass transfer, and solid-liquid mixing, so as to ensure that the reaction process is stable and improve the quality and production efficiency of the product.

[0069] After the reaction is completed, the reaction product (the reaction product is a mixture containing polycarbonate polyether polyol, cyclic carbonate and unreacted raw material propylene oxide) flows out from the discharge port. The single discharge of the discharge port is not more than 80% of the volume of the reaction pipe 109. The reaction product is then passed into the separation unit 111 in the post-processing area, and the separation unit 111 separates the propylene oxide and dissolved CO2 in the mixture to obtain a crude product, and adjusts the viscosity of the crude product at the same time. The crude product is passed into the refining unit 112, and a refining agent is added for adsorption treatment. After filtering, a crude product with a catalyst metal content of less than 5ppm is obtained. The refined crude product is passed into the distillation unit 113 to separate the main product polycarbonate polyether polyol and the by-product cyclic carbonate, that is, a polycarbonate polyether polyol product is obtained.

[0070] It should be noted that the above-mentioned separation unit, refining unit, distillation unit, etc. can all refer to the existing structure and will not be described in detail here.

[0071] Correspondingly, the present application also provides a circulation flow rate optimization control method, which includes: obtaining a time queue of material density collected by a densitometer; respectively extracting the absolute quantity time series change characteristics and the change quantity local neighborhood characteristics of the time queue of the material density to obtain a sequence of material density time series correlation feature vectors and material density change quantity local neighborhood feature vectors; performing attention query joint encoding on the sequence of the material density time series correlation feature vectors and the material density change quantity local neighborhood feature vectors to obtain a material density-material density change joint query code representation vector; based on the material density-material density change joint query code representation vector, generating a recommended value for the second circulation flow rate.

[0072] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned circulation flow rate optimization control method have been described in detail in the description of the production process of the above-mentioned sebacic acid polycarbonate polyether polyol, and therefore, its repeated description will be omitted.

[0073] In summary, the production process and the circulation flow rate optimization control method of the sebacic acid polycarbonate polyether polyol according to the embodiment of the present application are explained, which feeds the raw materials into the reaction pipe at one time after premixing, and introduces CO2 into the reaction pipe for reaction. In the middle and later stages of the reaction, the conversion of sebacic acid is promoted by adding a catalyst. At the same time, after adding the catalyst, the material density is monitored and analyzed in real time using intelligent industrial control technology to extract the time series correlation characteristics of the material density monitoring value and the time series correlation characteristics of the short-term change of the material density, and the multi-dimensional material density characteristic information is combined to fully understand the reaction process, so as to make intelligent recommendations for the circulation flow rate. In this way, the uniform mixing and efficient conversion of the materials can be ensured, thereby effectively increasing the conversion rate of sebacic acid, reducing the residue of sebacic acid, and reducing the acid value of the product, so as to improve the stability and efficiency of the reaction process.

[0074] The present application will be further described below in conjunction with specific embodiments.

[0075] Example 1 This embodiment provides a sebacic acid polycarbonate polyether polyol, using the above-mentioned tube-in-tube reaction device, the reaction process is as follows: 8L propylene oxide, 4g DMC catalyst, and 660g sebacic acid were added to the premixing kettle respectively, and 0.4Mpa of carbon dioxide was introduced and mixed and stirred at a temperature of 40°C. After stirring for 1h, the mixture was pressed into an 8L×2 tube reactor, and the diameter of a single reaction pipe was 100mm and the length of the reaction pipe was 1000mm. After all the reaction pipes were pressed in, carbon dioxide was introduced into the reaction pipe to 3Mpa and the pressure was maintained. Turn on the circulation pump, control the flow rate at 600g / min (the first circulation flow rate), control the medium temperature in the jacket of the reaction pipe, and raise the temperature of the reaction pipe to the initial temperature of 90°C. When the reaction temperature is monitored to be higher than 92°C (increased by more than 2°C) and the CO2 consumption is greater than 0.1MPa, reduce the temperature in the jacket to 80°C to reduce the reaction temperature. When the reaction temperature drops to 90°C, adjust the temperature in the jacket to 85°C. It was monitored that the material density reached 1.0 g / cm 3 When the reaction entered the middle and late stages, 0.2 g of DMC catalyst was added. The flow rate of the circulation pump was adjusted to 900 g / min (the second circulation flow rate, i.e., the second circulation flow rate adjustment strategy in this embodiment was fixed flow rate control). After reacting at 90°C for 3 hours, the density was uniformly increased to 1.12 g / cm 3 The reaction was maintained for 5 minutes, at which time the reaction was completed, and the temperature was lowered to take out the product. After the catalyst was removed by the catalyst filter, the main and by-products were separated using a scraper evaporator.

[0076] The DMC catalyst is a zinc-cobalt double metal cyanide complex catalyst obtained by reacting water-soluble metal salts of zinc and cobalt in a water-soluble solvent. The specific preparation method is as follows: potassium hexacyanocobaltate and zinc bromide in a molar ratio of 1:4 are weighed, dissolved in an aqueous solvent containing water and tert-butyl alcohol, and stirred continuously, and the total mass of the metal salt (i.e., cobalt salt and zinc salt) and the mass ratio of the aqueous solvent are 1:5. Inorganic acid and organic acid are added, the inorganic acid is dilute hydrochloric acid, the pH value is 2, the organic acid is glutaric acid, the molar ratio of the inorganic acid to the organic acid is 5:1, and the total molar number of the metal salt to the molar number of the acid is 4:1. Stir for several hours at a temperature of 10-100°C, and precipitate is continuously generated. The above precipitate is filtered and dried to obtain a filter cake. The filter cake is re-slurried and washed with an aqueous solvent at a temperature of 10-100°C, specifically, the washing temperature is 100°C, washing for 3 minutes, stirring for several hours, and then filtered and dried to obtain a filter cake, and the above slurrying, washing, and drying steps are repeated for multiple times at a temperature of 10-100°C until the system liquid pH is 6-7, specifically, the temperature is 60°C, and each time for 6 minutes. The solid product is further dried under vacuum conditions at 80-100°C to obtain the final catalyst, and the catalyst is processed into powder particles by mechanical grinding under anhydrous drying conditions before use.

[0077] Example 2 In Example 2, the second circulation flow rate adjustment strategy is a circulation flow rate optimization control strategy, which is: using intelligent industrial control technology to perform real-time monitoring and data analysis on material density to extract the time series correlation characteristics of material density monitoring values ​​and the time series correlation characteristics of short-term changes in material density, and combining multi-dimensional material density feature information to fully understand the reaction process, so as to make intelligent recommendations for circulation flow rates. Other conditions are the same as those in Example 1.

[0078] The purified product is sampled and analyzed, specifically, the polymerization product (polycarbonate-polyether polyol and cyclic propylene carbonate and unreacted propylene oxide) is sampled and collected in a container, and the polymerization product logistics sample is characterized by nuclear magnetic hydrogen spectrum to calculate the ratio of polymer to cyclic small molecules in the crude product. After the polymer is purified, the nuclear magnetic hydrogen spectrum test is performed to calculate the ratio of polycarbonate chain segments to polyether chain segments on the polymer main chain. There are only two structures of polycarbonate chain segments and polyether chain segments on the polymer main chain, and the sum of the percentages of the two is 100%. The final product with a carbonate connection ratio of 50-65% is a qualified product, and 55-60% is a superior product.

[0079] With the help of 1The amount of carbon dioxide introduced in the obtained polycarbonate polyether polyol (carbonate segment content) and the ratio of propylene carbonate (cyclic carbonate) to polycarbonate polyether polyol were determined by H-NMR (Bruker, DPX400, 400 MHz; pulse program zg30, waiting time d1:10 s, 64 scans). In each case, the sample was dissolved in deuterated chloroform. 1 The relevant resonant proton peaks in H-NMR (based on TMS=0ppm) are as follows: Among them, 5.0ppm and 4.2ppm belong to the proton peaks of the methine and methylene on the polycarbonate chain, 4.9ppm, 4.5ppm and 4.1ppm belong to the proton peaks of the methine and methylene in the five-membered ring carbonate, and 3.5-3.8ppm belong to the proton peak of the ether chain. The capital letter A plus a numerical subscript represents the integrated area of ​​the peak at a certain ppm on the nuclear magnetic resonance hydrogen spectrum, for example, A 5.0 represents the integrated area of ​​the peak at 5.0 ppm. 1 The integral area of ​​the H-NMR spectrum and its related proton peaks is used to calculate the carbonate chain ratio (molar ratio) F in the copolymerization reaction. CO2 and cyclic carbonate content mass fraction W PC , the amount of carbon dioxide embedded M CO2 , the specific calculation method is as follows: F CO2 =(A 5.0 +A 4.2 -2×A 4.6 ) / [(A 5.0 +A 4.2 -2×A 4.6 )+A 3.5 ]×100%; W PC =102×A 1.5 / [102×(A 5.0 +A 4.2 -2×A 4.6 +A1.5)+58×A 3.5 ]×100%; M CO2 =44×F CO2 / [102×F CO2 +58×(1-F CO2 )]×100%.

[0080] The coefficient 44 is the molar mass of carbon dioxide, the coefficient 58 is the molar mass of PO (propylene oxide), and the coefficient 102 is the sum of the molar masses of carbon dioxide (molar mass 44 g / mol) and PO (molar mass 58 g / mol).

[0081] The formula for calculating the product conversion rate n is n=m×(1-M CO2 ) / (Mm×M CO2 ), wherein m is the mass of the product after removing propylene oxide, and M is the total sample mass containing propylene oxide.

[0082] The polymer number average molecular weight (Mn) and polymer molecular weight polydispersity index (PDI) are determined by gel permeation chromatography (GPC). The smaller the PDI of the final product, the better the quality. A PDI below 2 is a qualified product, and a PDI below 1.5 is a superior product.

[0083] The molecular weight difference index ΔM is used to characterize the difference between the theoretical molecular weight of the product calculated according to the feed ratio and the actual molecular weight of the product measured after the reaction. The smaller the molecular weight difference index, the smaller the difference between the actual product molecular weight and the theoretical molecular weight, indicating that the controllability of the production equipment and process is better, the yield of the target product is higher, and the product quality is higher. The molecular weight difference index ΔM can be calculated using the following formula: ΔM=|g 实际 -g 理论 | / g 理论 ×100%, where g 实际 Represents the measured number average molecular weight of the product, g 理论 represents the theoretical molecular weight.

[0084] The acid value is determined according to the method in GB / T 12008.5-2010 "Plastic polyether polyols Part 5: Determination of acid value". In order to meet the subsequent production requirements, the acid value of the product should be controlled below 0.3 mgKOH / g, and the smaller the better.

[0085] With reference to the preparation method and test method of Example 1, the test results are shown in Table 1 below: Table 1

[0086] As can be seen from Table 1, the method of adding an additional catalyst in the second half of the reaction of the production process, combined with temperature control and reaction cycle speed control, can stabilize the synthetic product, and the product acid value is stable below 0.3 mgKOH / g, and the synthesis requirements of polyurethane can be met without subsequent treatment. Compared with Example 1, Example 2 has no changes in other parameters and reaction materials. The only change is that the circulation flow rate optimization control method described above is adopted in Example 2. It can be seen from the data of Example 2 that the circulation flow rate optimization control method described above can make the raw materials evenly mixed and efficiently converted, reduce the residue of sebacic acid, and reduce the product acid value. In Example 1, the reaction end time is 3 hours, while in Example 2, the reaction end time is 2.2 hours, and the reaction efficiency is improved by 26.7%; after testing, the reaction products have a similar carbonate chain ratio and PDI value between the two, and the molecular weight difference index ΔM is reduced by 27.3%, and the acid value is further reduced. Therefore, the circulation flow rate optimization control method provided by the present application can improve the reaction efficiency and ensure the quality of the reaction product.

[0087] Finally, it should be noted that the embodiments described above are only some embodiments of the present invention, rather than all embodiments. The detailed description of the embodiments of the present invention is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Claims

1. A production process of sebacic acid polycarbonate polyether polyol, the production process of sebacic acid polycarbonate polyether polyol is carried out in a tube-in-tube reaction device, the tube-in-tube reaction device comprises a reaction pipeline and a circulation pump, a jacket is provided on the outside of the reaction pipeline, the circulation pump is connected to the reaction pipeline through a connecting pipeline, and is used to make the material circulate in the reaction pipeline, characterized in that: The production process comprises: Premixing step: mixing the epoxy compound, CO2, sebacic acid and the catalyst in a premixing kettle to obtain a premix; Feeding step: feeding the premixture into the tubular reaction device at one time. After the feeding is completed, CO2 is introduced into the reaction pipeline until the pressure of the tubular reaction device reaches 2-6 MPa, and the pressure is maintained; Reaction step: start the circulation pump, adjust the temperature of the medium in the jacket, and heat the reaction pipe to an initial temperature of 80-95°C for reaction, wherein the circulation pump is maintained at the first circulation flow rate, and when the material density reaches 1g / cm 3 When the catalyst is added to the reaction pipeline, the speed of the circulation pump is adjusted to the second circulation flow rate; Among them, when the material density is monitored to reach 1g / cm 3 , adding catalyst to the reaction pipeline, and adjusting the rate of the circulation pump to a second circulation flow rate, including: acquiring a time queue of the material density collected by a densitometer; respectively extracting the absolute quantity time series change characteristics and the change quantity local neighborhood characteristics of the time queue of the material density to obtain a sequence of material density time series correlation feature vectors and material density change quantity local neighborhood feature vectors; performing attention query joint encoding on the sequence of the material density time series correlation feature vectors and the material density change quantity local neighborhood feature vectors to obtain a material density-material density change joint query code representation vector; and, based on the material density-material density change joint query code representation vector, generating a recommended value for the second circulation flow rate.

2. The production process of sebacic acid polycarbonate polyether polyol according to claim 1, characterized in that: The absolute quantity time series change characteristics and the local neighborhood characteristics of the change quantity of the time queue of the material density are extracted respectively to obtain a sequence of the material density time series correlation feature vector and the material density change quantity local neighborhood feature vector, including: Inputting the time queue of the material density into a material density time series feature extractor based on a Bi-LSTM model to obtain a time series correlation feature vector of the material density; Calculate the difference between the material density at each two adjacent predetermined time points in the material density time queue to obtain the time queue of the material density time series neighborhood change amount; The time queue of the material density temporal neighborhood variation is input into the material density variation local neighborhood feature extractor based on the 1D-CNN model to obtain a sequence of the material density variation local neighborhood feature vectors.

3. The production process of sebacic acid polycarbonate polyether polyol according to claim 2, characterized in that: Performing attention query joint encoding on the sequence of the material density time series correlation feature vector and the local neighborhood feature vector of the material density change to obtain a material density-material density change joint query encoding representation vector, including: Performing local neighborhood implicit optimization on the sequence of local neighborhood feature vectors of material density variation to obtain a sequence of local neighborhood semantic association feature vectors of material density variation; The material density time series association feature vector, the sequence of the material density change local neighborhood feature vectors, and the sequence of the material density change local neighborhood semantic association feature vectors are subjected to attention matching query and multi-granularity query feature fusion to obtain the material density-material density change joint query encoding representation vector.

4. The production process of sebacic acid polycarbonate polyether polyol according to claim 3, characterized in that: Performing local neighborhood implicit optimization on the sequence of local neighborhood feature vectors of the material density variation to obtain a sequence of local neighborhood semantic association feature vectors of the material density variation, including: One-dimensional convolution coding is performed on the sequence of local neighborhood feature vectors of material density variation to obtain a sequence of local neighborhood semantic association feature vectors of material density variation, wherein the scale of the one-dimensional convolution coding is equal to an integer multiple of the length of the local neighborhood feature vector of material density variation.

5. The production process of sebacic acid polycarbonate polyether polyol according to claim 4, characterized in that: The material density time series association feature vector, the sequence of the material density change local neighborhood feature vectors, and the sequence of the material density change local neighborhood semantic association feature vectors are subjected to attention matching query and multi-granularity query feature fusion to obtain the material density-material density change joint query encoding representation vector, including: Using a value embedding coding matrix, the material density time series correlation feature vector is processed to obtain a value feature vector; Using the material density time series associated feature vector as a query vector, the value feature vector as a value vector, and the sequence of the local neighborhood feature vectors of the material density change as a sequence of key vectors, the query vector, the value vector, and the sequence of the key vectors are input into a first multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change associated query implicit vectors; Using the material density time series association feature vector as a query vector, the value feature vector as a value vector, and the sequence of the local neighborhood semantic association feature vectors of the material density change as a sequence of key vectors, the query vector, the value vector, and the sequence of the key vectors are input into a second multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change local neighborhood association query implicit vectors; Respectively calculating the position mean vectors of the sequence of the material density-material density change associated query implicit vectors and the sequence of the material density-material density change local neighborhood associated query implicit vectors to obtain the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector; The material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector are fused to obtain the material density-material density change joint query encoding representation vector.

6. The production process of sebacic acid polycarbonate polyether polyol according to claim 5, characterized in that: Taking the material density time series associated feature vector as the query vector, the value feature vector as the value vector and the sequence of the local neighborhood feature vectors of the material density change as the sequence of key vectors, the query vector, the value vector and the sequence of the key vectors are input into a first multi-head attention query module based on a converter structure to obtain a sequence of material density-material density change associated query implicit vectors, including: Multiplying the query vector by the transposed vector of the key vector and dividing by the square root of the scale of the key vector to obtain an attention score matrix; The attention score matrix is ​​passed through a softmax function and then multiplied by the value vector to obtain the material density-material density change association query implicit vector.

7. The production process of sebacic acid polycarbonate polyether polyol according to claim 6, characterized in that: The material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector are integrated to obtain the material density-material density change joint query encoding representation vector, including: Calculate the point-by-point addition, point-by-point subtraction, and point-by-point multiplication between the material density-material density change associated query matching feature vector and the material density-material density change local neighborhood associated query matching feature vector to obtain a first material density-material density change query fusion feature vector, a second material density-material density change query fusion feature vector, and a third material density-material density change query fusion feature vector; After the first material density-material density change query fusion feature vector, the second material density-material density change query fusion feature vector and the third material density-material density change query fusion feature vector are cascaded into a material density-material density change multi-scale fusion feature vector, one-dimensional convolution processing and maximum value pooling processing are performed on it to obtain the material density-material density change joint query encoding representation vector.

8. The production process of sebacic acid polycarbonate polyether polyol according to claim 7, characterized in that: Based on the material density-material density change joint query code representation vector, a recommended value of the second circulation flow rate is generated, including: The material density-material density change joint query code representation vector is input into a decoder-based circulation flow rate optimization module to obtain an optimization instruction, wherein the optimization instruction includes a decoded value of a recommended value of the second circulation flow rate.

9. The production process of sebacic acid polycarbonate polyether polyol according to claim 1, characterized in that: The reaction step comprises: during the reaction process, monitoring the reaction temperature and material density in the reaction pipe, and performing the following steps: When it is monitored that the CO2 consumption value in the reaction pipeline is greater than 0.1MPa and the reaction temperature rises by more than 2°C, the medium temperature in the jacket is reduced to 70-80°C, until the reaction temperature is monitored to drop to 80-95°C, then the medium temperature in the jacket is increased to 80-95°C; When the density of the material reaches the reaction end point, the reaction is terminated, wherein the reaction end point is: the material density is 1.1-1.3 g / cm 3 , and / or, the fluctuation value of the material density within 5 minutes is not greater than 0.01g / cm 3 .

Images