Production process and bubbling control method of succinic acid type polycarbonate polyether polyol

Abstract

The invention relates to the technical field of chemistry and chemical engineering, and particularly discloses a production process and a bubbling control method of succinic acid type polycarbonate polyether polyol, an oxygen compound, succinic acid and a DMC catalyst are premixed by adopting a carbon dioxide bubbling mode, and meanwhile, a camera is used for monitoring the mixing distribution state of the premixture in real time; and performing grid granularity local feature extraction on a real-time mixed distribution state image acquired by a camera by using an intelligent industrial control technology, and further performing mixed distribution uniformity evaluation based on semantic difference of each grid granularity local mixed distribution state feature relative to the whole image so as to realize intelligent control on bubbling time. Compared with a traditional production process, the process steps are effectively simplified, the controllability and stability of the premixing effect in the production process are improved, dependence and errors on manual operation are reduced, and the method is suitable for industrial large-scale production.

Description

Technical Field

[0001] The present application relates to the field of chemical engineering technology, and more specifically, to a production process and a bubbling control method for succinic acid type polycarbonate polyether polyol. Background Art

[0002] Succinic acid type polycarbonate polyether polyol is a kind of polyol with carbonate groups in the molecule and hydroxyl groups at the end of the molecular chain. One of the raw materials, carbon dioxide, is cheap and easy to obtain, non-toxic and non-flammable, has a clear carbon dioxide chemical fixation effect, and the reaction requires low temperature and low energy consumption, which makes the carbon dioxide and epoxy compound regulation copolymerization method have broad application prospects and high industrial value.

[0003] In the synthesis process of succinic acid-based polycarbonate polyether polyols, the double metal cyanide (abbreviated as DMC) catalyst is easy to prepare and has low cost. The residual heavy metals in the product are below 30ppm, and polycarbonate polyether polyols can be prepared with high activity. However, the reaction of polycarbonate polyether polyols is an exothermic reaction. It is necessary to ensure that the reaction does not overheat in the reaction system. If the temperature is too high, it will lead to low carbonate chain content, wide molecular weight distribution, and high proportion of by-product cyclic carbonate, which will reduce the quality of the product. Compared with the reaction system of polyether polyols, the reaction system of polycarbonate polyether polyols has a large amount of CO2 gas raw materials in addition to the large amount of heat generated by the epoxy compound activation catalyst, which leads to high temperature and high pressure in the system. This leads to higher pressure in the reaction system, and the temperature and pressure resistance requirements of the reactor are extremely stringent.

[0004] Based on the reaction mechanism, in the prior art, in order to obtain polycarbonate polyether polyol products with target molecular weight on the one hand, and to ensure production safety and avoid explosion and equipment safety hazards caused by large amounts of heat release on the other hand, a two-step production process of batch feeding must be adopted. The process steps are cumbersome, time-consuming, and have low economic benefits.

[0005] In addition, during the premixing process of epoxy compounds, catalysts and initiators, the three are mixed by bubbling with carbon dioxide, which can effectively improve the uniformity between the components. However, in this process, the control of the bubbling time is crucial to the smooth progress of the reaction. If the bubbling time is too long, the reactants will over-react in the premixing unit, affecting the molecular weight distribution of the final product; if the bubbling time is too short, the uniform mixing of the reactants cannot be guaranteed, which will also have an adverse effect on the polymerization reaction. Traditional bubbling time control methods are usually based on the preset premixing time and the operator's experience adjustment, lacking accurate monitoring methods, which often lead to instability and poor repeatability of the reaction.

[0006] Therefore, an optimized production process and bubbling control method for succinic acid-type polycarbonate polyether polyols are desired. Summary of the invention

[0007] 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 bubbling control method of succinic acid type polycarbonate polyether polyol, which uses carbon dioxide bubbling to premix oxygen compounds, succinic acid, and DMC catalysts, and uses a camera to monitor the mixed distribution state of the premix in real time, and uses intelligent industrial control technology to extract local features of the grid granularity of the real-time mixed distribution state image collected by the camera, and then evaluates the uniformity of the mixed distribution based on the semantic difference of the local mixed distribution state features of each grid granularity relative to the overall image, so as to realize intelligent control of the bubbling time. Compared with the traditional production process, the process steps are effectively simplified, the controllability and stability of the premixing effect are improved, the dependence and error on manual operation are reduced, and it is suitable for industrial large-scale production.

[0008] Accordingly, according to one aspect of the present application, a production process of succinic acid type polycarbonate polyether polyol is provided, and the production process of succinic acid type polycarbonate polyether polyol is carried out in a tube-in-tube reaction device, and the reaction device includes a premixing unit and a reaction zone; wherein the production process includes: mixing an epoxy compound, succinic acid, and a DMC catalyst in the premixing unit to obtain a premix; feeding the premix into the reaction zone at one time; introducing carbon dioxide into the reaction zone to supplement the reaction pressure, wherein the introduction direction of the carbon dioxide is opposite to the flow direction of the material in the reaction pipe; and obtaining succinic acid type polycarbonate polyether polyol by polymerization reaction; wherein the premixing time is 1h-2h; the maximum pressure in the premixing unit does not exceed 1MPa, and the premixing unit uses carbon dioxide bubbling to mix the epoxy compound, succinic acid, and DMC catalyst, and Carbon dioxide is slowly bubbled in the premixing unit at a rate of 40-60 bubbles / min; wherein the epoxy compound, succinic acid and DMC catalyst are mixed in the premixing unit to obtain a premix, comprising: obtaining a real-time mixed distribution state image of the premix captured by a camera; performing mixed distribution state feature extraction based on network granularity on the real-time mixed distribution state image of the premix to obtain a set of local mixed distribution state feature vectors at a grid granularity; performing feature significant aggregation processing based on clustering contribution on the set of local mixed distribution state feature vectors at a grid granularity to obtain a global mixed state aggregation representation vector; and determining whether to stop bubbling based on the semantic difference distribution of each grid granularity local mixed distribution state feature vector in the set of local mixed distribution state feature vectors at a grid granularity relative to the global mixed state aggregation representation vector.

[0009] In the production process of the succinic acid type polycarbonate polyether polyol, the premixing temperature of the premixing unit is 30°C-40°C; in the premixing unit, the temperature of the carbon dioxide is 40°C; In the production process of the above-mentioned succinic acid-type polycarbonate polyether polyol, the real-time mixed distribution state image of the premixture is subjected to mixed distribution state feature extraction based on network granularity to obtain a set of grid granularity local mixed distribution state feature vectors, including: grid segmentation of the real-time mixed distribution state image to obtain a set of mixed distribution state grid granularity images; inputting each mixed distribution state grid granularity image in the set of mixed distribution state grid granularity images into a mixed distribution state feature extractor based on a void convolutional layer to obtain a set of grid granularity local mixed distribution state feature vectors.

[0010] In the above-mentioned production process of succinic acid type polycarbonate polyether polyol, the set of the grid-granularity local mixed distribution state feature vectors is subjected to feature significant aggregation processing based on clustering contribution to obtain a global mixed state aggregation representation vector, including: clustering analysis is performed on the set of the grid-granularity local mixed distribution state feature vectors to obtain a mixed distribution state self-supervised clustering representation vector; the implicit clustering contribution factor of each grid-granularity local mixed distribution state feature vector in the sequence of the grid-granularity local mixed distribution state feature vectors relative to the mixed distribution state self-supervised clustering representation vector is calculated to obtain a sequence of grid-granularity local mixed distribution state feature implicit clustering contribution factors; based on the sequence of grid-granularity local mixed distribution state feature implicit clustering contribution factors, the sequence of the grid-granularity local mixed distribution state feature vectors is subjected to feature aggregation to obtain the global mixed state aggregation representation vector.

[0011] In the production process of the above-mentioned succinic acid type polycarbonate polyether polyol, the implicit clustering contribution factor of each grid granularity local mixed distribution state feature vector in the sequence of the grid granularity local mixed distribution state feature vector is calculated relative to the mixed distribution state self-supervised clustering representation vector to obtain a sequence of grid granularity local mixed distribution state feature implicit clustering contribution factors, including: calculating the dot division vector between the grid granularity local mixed distribution state feature vector and the mixed distribution state self-supervised clustering representation vector, and calculating the base-two logarithmic value of the absolute value of each eigenvalue in the dot division vector to obtain an implicit clustering contribution weight vector; calculating the dot product vector between the implicit clustering contribution weight vector and the grid granularity local mixed distribution state feature vector, and calculating the exponential function value with base e and the sum of each eigenvalue of the dot product vector as the exponent to obtain the grid granularity local mixed distribution state feature implicit clustering contribution factor.

[0012] In the production process of the above-mentioned succinic acid type polycarbonate polyether polyol, based on the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics of the grid granularity, the sequence of local mixed distribution state feature vectors of the grid granularity is feature aggregated to obtain the global mixed state aggregate representation vector, including: arranging the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics of the grid granularity into a mixed distribution state local feature clustering contribution field distribution vector; performing explicit modeling based on the self-attention mechanism on the mixed distribution state local feature clustering contribution field distribution vector to obtain a mixed distribution state local feature clustering contribution field modulation weight vector; using each eigenvalue in the mixed distribution state local feature clustering contribution field modulation weight vector as a weight, calculating the position-weighted sum of the sequence of local mixed distribution state feature vectors of the grid granularity to obtain the global mixed state aggregate representation vector.

[0013] In the production process of the above-mentioned succinic acid polycarbonate polyether polyol, the local feature clustering contribution field distribution vector of the mixed distribution state is explicitly modeled based on the self-attention mechanism to obtain the local feature clustering contribution field modulation weight vector of the mixed distribution state, including: using the query embedding matrix, key embedding matrix and value embedding matrix to embed the local feature clustering contribution field distribution vector of the mixed distribution state to obtain a query vector, a key vector and a value vector; 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 multiplying it with the value vector to obtain the local feature clustering contribution field modulation weight vector of the mixed distribution state.

[0014] In the above-mentioned production process of succinic acid-type polycarbonate polyether polyol, whether to stop bubbling is determined based on the semantic difference distribution of each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector, including: calculating the mixing state semantic difference coefficient between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors to obtain a set of mixing state semantic difference coefficients; calculating the variance of the set of mixing state semantic difference coefficients; and determining whether to stop bubbling based on a comparison between the variance and a preset threshold.

[0015] In the production process of the above-mentioned succinic acid polycarbonate polyether polyol, the mixing state semantic difference coefficient between the global mixing state aggregation representation vector and each grid-size local mixing distribution state feature vector in the set of the grid-size local mixing distribution state feature vector is calculated to obtain a set of mixing state semantic difference coefficients, including: calculating the Euclidean distance between the global mixing state aggregation representation vector and each grid-size local mixing distribution state feature vector in the set of the grid-size local mixing distribution state feature vector as the mixing state semantic difference coefficient to obtain the set of mixing state semantic difference coefficients.

[0016] In the production process of the above-mentioned succinic acid-type polycarbonate polyether polyol, the reaction pressure of the reaction zone is 3MPa-4MPa. When the pressure in the reaction zone is too low, a solenoid valve is used to automatically introduce carbon dioxide to supplement the pressure. The direction of introduction of the carbon dioxide is opposite to the flow direction of the material.

[0017] According to another aspect of the present application, a bubbling control method is provided, which includes: acquiring a real-time mixed distribution state image of a premixture captured by a camera; performing network granularity-based mixed distribution state feature extraction on the real-time mixed distribution state image of the premixture to obtain a set of grid granularity local mixed distribution state feature vectors; performing clustering contribution-based feature significance aggregation processing on the set of grid granularity local mixed distribution state feature vectors to obtain a global mixed state aggregate representation vector; determining whether to stop bubbling based on the semantic difference distribution of each grid granularity local mixed distribution state feature vector in the set of grid granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector.

[0018] Compared with the prior art, the production process and bubbling control method of succinic acid-type polycarbonate polyether polyol described in the present application uses carbon dioxide bubbling to premix oxygen compounds, succinic acid, and DMC catalysts, and uses a camera to monitor the mixing distribution state of the premix in real time, and uses intelligent industrial control technology to extract local features of the grid granularity of the real-time mixing distribution state image collected by the camera, and then evaluates the mixing distribution uniformity based on the semantic difference of the local mixing distribution state features of each grid granularity relative to the overall image, so as to achieve intelligent control of the bubbling time. Compared with traditional production processes, it effectively simplifies the process steps, improves the controllability and stability of the premixing effect, reduces the dependence on and errors of manual operations, and is suitable for industrial large-scale production. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] 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.

[0020] Figure 1 This is an overall schematic diagram of a succinic acid-type polycarbonate polyether polyol tubular reaction device according to an embodiment of the present application, wherein the green arrow represents the flow direction of carbon dioxide gas, the black arrow represents the flow direction of the raw material propylene oxide, the blue arrow represents the flow direction of the reaction materials in the circulation reaction in the reaction unit, and the purple arrow represents the flow direction of the product after the reaction is completed.

[0021] Figure 2 The flowchart is a production process of succinic acid-type polycarbonate polyether polyol according to an embodiment of the present application.

[0022] Figure 3 The present invention is a flow chart of mixing epoxy compound, succinic acid and DMC catalyst in the premixing unit to obtain a premix in the production process of succinic acid type polycarbonate polyether polyol according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] 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.

[0024] Figure 1 The overall schematic diagram of the succinic acid type polycarbonate polyether polyol tubular reaction device according to the embodiment of the present application is shown in the figure: 001, carbon dioxide; 002, carbon dioxide purification unit; 003, premixing unit; 004, propylene oxide; 005, propylene oxide purification unit; 006, metering pump; 007, succinic acid; 008, DMC catalyst; 009, reaction zone; 010, separation unit; 011, refining unit; 012, refining agent; 013, distillation unit; 014, filling.

[0025] The present application proposes a production process of a succinic acid type polycarbonate polyether polyol. Specifically, the production process is carried out in a tube-in-tube reaction device, such as Figure 1As shown, the reaction device comprises a premixing unit (003) and a reaction zone (009). The reaction zone (009) comprises a gas-liquid mixing heat conduction unit and a solid-liquid mixing unit. The epoxy compound, succinic acid, and DMC catalyst are fully mixed in the premixing unit (003), and then fed into the reaction zone (009) at one time. Carbon dioxide is introduced into the reaction zone (009) to supplement the pressure. The carbon dioxide reacts with propylene oxide through the gas replenishment port. The reaction area is a reaction pipeline. The direction of introducing carbon dioxide is opposite to the flow direction of the material in the reaction pipeline. The succinic acid type polycarbonate polyether polyol is obtained through a polymerization reaction. The general formula of the polymerization reaction is: .

[0026] The gas-liquid mixed heat transfer unit includes a dispersion sheet, a circulation pump and a gas-liquid mass transfer area, which are used to control the flow rate and flow direction of the material in the reaction system, disperse and control the material in the gas-liquid mass transfer area so as not to contact the pipe wall; The solid-liquid mixing unit comprises a reaction pipeline, a static mixer and a circulation pump. The ratio of the length L to the diameter d of the reaction pipeline is 10≤L:d≤40.

[0027] The tube-in-tube reaction device described in the present application uniformly mixes the raw materials at one time through the premixing unit (003), and realizes the one-time feeding production process of succinic acid type polycarbonate polyether polyol by combining the parameter settings of the heat transfer unit and the solid-liquid mixing unit in the reactor, so that the molecular weight distribution of the prepared succinic acid type polycarbonate polyether polyol is narrow, the actual product molecular weight is less different from the target theoretical value, the production process is safe, no implosion occurs, and there is no amplification effect. Specifically, firstly, the epoxy compound, DMC catalyst, and initiator are fully mixed at 30-40°C by continuous bubbling of carbon dioxide in the premixing unit (003), ensuring that succinic acid has a high solubility while also inhibiting the occurrence of premature reaction, providing a certain amount of initial carbon dioxide for the formal reaction, and premixing in the low-temperature carbon dioxide continuous bubbling atmosphere so that the initiator and DMC are in a low-temperature dynamic motion process, reducing the initiator's wrapping of the DMC catalyst and not causing the catalyst active center to be passivated. In the heat transfer unit, a specific dispersion sheet structure is used to match the material flow rate of the process. While maintaining the dispersed material, the material is made to fall in the form of large liquid drops without splashing onto the tube wall. While ensuring gas-liquid mass transfer, the evaporation of propylene oxide on the tube wall in the gas-liquid mass transfer zone is reduced, and the ability of subsequent propylene oxide to wash down the solid material that may be precipitated is increased. The gas-liquid mass transfer zone described in the present application enables the material to reabsorb carbon dioxide and release heat of generation, and does not cause excessive atomization of the material to cause evaporation and precipitation of solids on the tube wall, causing material imbalance. While consuming carbon dioxide, low-temperature carbon dioxide is supplemented to 3MPa, so that the area forms a lower temperature area, so that the heat of generation is better released, and the decomposition of the succinic acid initiator caused by the high temperature of the reaction heat is reduced; the solid-liquid mixing unit enables the catalyst to be fully mixed and evenly mixed when the material flows in the reactor, so that the temperature of the pipeline wall is better conducted to the inside, and the heat of generation during the reaction can be dispersed in time, and it is not easy to produce an explosion point.

[0028] In one embodiment, the DMC catalyst is the three-ligand bimetallic complex catalyst, and its structure is M' a [M``(CN)6] b ·(t-BuOH) m (Poly) n ·[M` c (PO4) d ] y ; wherein the first ligand t-BuOH is tert-butyl alcohol, the second ligand Poly is selected from a polyether polyol or a derivative thereof having a number average molecular weight Mn of 300 to 2000, the third ligand is a phosphate, and the general structural formula of the polyether polyol is expressed as Formula 1: Formula 1, In Formula 1, the value of e ranges from 0 to 40, the value of f ranges from 0 to 40, the value of e+f ranges from 2 to 45, R1 to R2 are independently hydrogen atoms or monovalent hydrocarbon groups having 1 to 5 carbon atoms; the atomic molar ratio of M` to M`` is 1.5 to 3:1, and M` is selected from Zn 2+ , Fe 3+ 、Ni 2 + Sn 4+ , Pb 2+ ,M`` is selected from Co 3+ , Fe 3+ , Ir 3+ , a, b, c, d are ionic coefficients that satisfy the balance of positive and negative charges, m, n, y are any positive numbers, and the weight proportions of the ligands are as follows: the weight of the first ligand t-BuOH is 0.05-10% of the total weight of the three-ligand bimetallic complex catalyst; the weight of the second ligand Poly is 5-20% of the total weight of the three-ligand bimetallic complex catalyst; the weight of the third ligand phosphate is 4-30% of the total weight of the three-ligand bimetallic complex catalyst.

[0029] In one embodiment, the dispersion sheet is located at a position 10 to 50 mm near the top of the reaction pipe.

[0030] In one embodiment, the reaction pressure of the reaction zone (009) is 3-4 MPa; more preferably, the reaction pressure is 3 MPa; when the pressure is too low, a solenoid valve is used to automatically introduce carbon dioxide to supplement the pressure; that is, carbon dioxide is introduced from the gas inlet at the bottom of the reaction pipe to form convection with the material to continue heat exchange and increase the carbon dioxide content in the material.

[0031] In one embodiment, the ratio of the total volume of the material to each tube is 0<V 物料 ≤80%, more preferably 50-80%; preferably, the volume of the gas-liquid mass transfer zone accounts for 20%≤V 气液传质区 <100%, more preferably, 20%-50%.

[0032] In one embodiment, the feed ratio of the epoxy compound and succinic acid is propylene oxide:succinic acid=15:1-70:1, and the DMC catalyst is 500 ppm.

[0033] In one embodiment, the premixing temperature of the premixing unit (003) is 30-40°C; more preferably, the premixing temperature is 40°C; the premixing time is 1-2h; more preferably, the premixing time is 1h; In one embodiment, CO2 is used for slow bubbling in the premixing unit (003), the temperature of the CO2 is 40°C, and the maximum pressure in the premixing unit (003) does not exceed 1 MPa; more preferably, the premixing pressure is 0.1-0.5 MPa; more preferably, the premixing pressure is 0.4 MPa; the bubbling speed is 40-60 bubbles / min.

[0034] In one embodiment, the reaction temperature of the reaction zone is 70-90°C; more preferably, the reaction temperature is 80°C; In one embodiment, the reaction pressure of the reaction zone is 3-4 MPa; more preferably, the reaction pressure is 3 MPa; In one embodiment, the reaction time of the reaction zone is based on the density as the reaction end point, and the density reaches 1.12 g / cm 3 The reaction is terminated.

[0035] In one embodiment, the reaction zone of the tube-in-tube reaction device is composed of more than one reaction pipes connected in parallel, wherein the liquid circulation path formed between each reaction pipe and the connecting pipe and the circulation pump is the same.

[0036] In one embodiment, a temperature sensor is provided inside the reaction pipe, and a density detector is provided at the outlet. The jacket temperature is adjusted according to the material density and the internal temperature of the reactor to control the reaction rate.

[0037] In one embodiment, there are two or more reaction pipes connected in parallel, and the center distance between the nearest neighbor pipes is less than or equal to twice the pipe diameter.

[0038] In one embodiment, the reaction pipes are independent of each other and can be closed and isolated by valves.

[0039] Figure 2 Flow chart of the production process of succinic acid type polycarbonate polyether polyol according to the embodiment of the present application. Figure 2 As shown, the production process of the succinic acid type polycarbonate polyether polyol comprises the following steps: S1, mixing the epoxy compound, succinic acid and DMC catalyst in the premixing unit to obtain a premix; S2, feeding the premix into the reaction zone at one time; S3, introducing carbon dioxide into the reaction zone to supplement the reaction pressure, wherein the introduction direction of the carbon dioxide is opposite to the flow direction of the material in the reaction pipe; S4, obtaining the succinic acid type polycarbonate polyether polyol through polymerization reaction.

[0040] Specifically, in step S1, carbon dioxide is provided by a carbon dioxide storage tank (001) or carbon dioxide recovered and purified by a carbon dioxide purification unit (002), and is passed into a premixing unit (003). A propylene oxide raw material (004) is purified by a propylene oxide purification unit (005), and enters a premixing unit (003) in a carbon dioxide atmosphere through a metering pump (006). A fixed amount of succinic acid (007) and a catalyst (008) are added to the premixing unit (003) using a weighing bucket. The premixing temperature range is 30-40° C., the pressure range is ≤1 MPa, and the premixing time is 1-2 hours. The premixed raw materials are passed into a feed port of a reaction zone (009).

[0041] Preferably, considering that in the premixing step S1, the control of the bubbling time is crucial for the smooth progress of the reaction. If the bubbling time is too long, the reactants will be over-reacted in the premixing unit, affecting the molecular weight distribution of the final product; and if the bubbling time is too short, the uniform mixing of the reactants cannot be guaranteed, which will also have an adverse effect on the polymerization reaction. Traditional bubbling time control methods are generally based on preset premixing time and operator experience adjustment, lack of accurate monitoring means, which often leads to instability and poor repeatability of the reaction. In this regard, the present application further uses a camera to monitor the mixed distribution state of the premix in real time, and uses computer vision technology based on deep learning to extract local features of the real-time mixed distribution state image collected by the camera for grid granularity, and then based on the semantic difference of each grid granularity local mixed distribution state feature relative to the overall image, the mixed distribution uniformity is evaluated, so as to realize the intelligent control of the bubbling time, so as to ensure that the reactants are evenly mixed in the premixing unit, avoiding the problem of overreaction or uneven mixing.

[0042] Figure 3 The present invention is a flow chart of mixing epoxy compound, succinic acid and DMC catalyst in the premixing unit to obtain a premix in the production process of succinic acid type polycarbonate polyether polyol according to an embodiment of the present application. Figure 3 As shown, the premixing step S1 also includes: S110, acquiring a real-time mixed distribution state image of the premixture captured by a camera; S120, performing a mixed distribution state feature extraction based on network granularity on the real-time mixed distribution state image of the premixture to obtain a set of grid granularity local mixed distribution state feature vectors; S130, performing a feature significant aggregation processing based on clustering contribution on the set of grid granularity local mixed distribution state feature vectors to obtain a global mixed state aggregation representation vector; S140, determining whether to stop bubbling based on the semantic difference distribution of each grid granularity local mixed distribution state feature vector in the set of grid granularity local mixed distribution state feature vectors relative to the global mixed state aggregation representation vector.

[0043] Specifically, in the technical solution of the present application, first, through the step S110, a real-time mixed distribution state image of the premixture captured by a camera is obtained to monitor the mixing state of the premixture. Then, in order to analyze the uniformity of the mixture in a finer granularity, local features are further extracted from the real-time mixed distribution state image of the premixture.

[0044] Accordingly, in a specific example of the present application, the step S120 includes: gridding the real-time mixed distribution state image to obtain a set of mixed distribution state grid granularity images; inputting each mixed distribution state grid granularity image in the set of mixed distribution state grid granularity images into a mixed distribution state feature extractor based on a hole convolution layer to obtain a set of grid granularity local mixed distribution state feature vectors.

[0045] That is, by gridding the real-time mixed distribution state image, the whole image is decomposed into local images of multiple grid granularities. In this way, the overall uniformity of the mixture can be judged by measuring the similarity of each grid granularity image. Next, in order to extract the mixed distribution state features of each mixed distribution state grid granularity image as the basic data for uniformity evaluation, in the step S120, the present application adopts a hole convolution neural network (CNN) model with excellent performance in the field of image processing to extract image features. By performing hole convolution processing on each mixed distribution state grid granularity image in the set of mixed distribution state grid granularity images, the unique hole convolution structure and large receptive field feature extraction capability of the hole convolution neural network can be used to capture a wider range of image context information without losing spatial detail information, thereby effectively capturing the mixed distribution state features such as texture changes and color distribution in each mixed distribution state grid granularity image, so as to obtain a set of grid granularity local mixed distribution state feature vectors.

[0046] Next, in order to achieve the similarity evaluation of the local mixed distribution state features of each grid granularity, the present application further uses the overall mixed state information of the real-time mixed distribution state image as a metric benchmark, and evaluates the uniformity of the mixture based on the feature offset of each local feature relative to the whole. Specifically, firstly, the set of the local mixed distribution state feature vectors of the grid granularity is subjected to feature aggregation processing through the step S130 to obtain the mixed distribution state representation of the image as a whole. In the step S130, in order to improve the effect of feature aggregation, a feature significant aggregation method based on clustering contribution is proposed, which first uses an unsupervised learning method to perform cluster analysis on the set of the local mixed distribution state feature vectors of the grid granularity to reveal the intrinsic correlation distribution between local features, and generate a self-supervised cluster representation vector of the mixed distribution state, thereby representing the average mixed distribution state of the mixture. Then, by calculating the implicit clustering contribution factor of the local mixed distribution state feature vector of each grid granularity relative to the self-supervised cluster representation vector of the mixed distribution state, to reflect its contribution to the clustering feature, quantify its role in the clustering process, and thus provide a basis for subsequent feature selection and weight allocation. Next, the calculated implicit clustering contribution factors are integrated into vector form, and the self-attention mechanism is used to explicitly model them. A more targeted weight vector is generated through global autocorrelation analysis, thereby achieving weighted aggregation of local mixed distribution state features of each grid granularity and obtaining a global mixed state aggregation representation vector. In this way, it can be ensured that the contribution of each local feature to the overall mixed state is fully considered during the aggregation process, thereby obtaining a more representative representation of the overall mixed distribution state of the image.

[0047] In a specific example of the present application, the step S130 includes: performing cluster analysis on the set of grid-granularity local mixed distribution state feature vectors to obtain a mixed distribution state self-supervised cluster representation vector; calculating the implicit cluster contribution factor of each grid-granularity local mixed distribution state feature vector in the sequence of grid-granularity local mixed distribution state feature vectors relative to the mixed distribution state self-supervised cluster representation vector to obtain a sequence of grid-granularity local mixed distribution state feature implicit cluster contribution factors; based on the sequence of grid-granularity local mixed distribution state feature implicit cluster contribution factors, performing feature aggregation on the sequence of grid-granularity local mixed distribution state feature vectors to obtain the global mixed state aggregate representation vector.

[0048] Specifically, the implicit clustering contribution factor of each grid-granularity local mixed distribution state feature vector in the sequence of the grid-granularity local mixed distribution state feature vectors relative to the mixed distribution state self-supervised clustering representation vector is calculated to obtain a sequence of grid-granularity local mixed distribution state feature implicit clustering contribution factors, including: calculating the dot division vector between the grid-granularity local mixed distribution state feature vector and the mixed distribution state self-supervised clustering representation vector, and calculating the base-two logarithmic value of the absolute value of each eigenvalue in the dot division vector to obtain an implicit clustering contribution weight vector; calculating the dot product vector between the implicit clustering contribution weight vector and the grid-granularity local mixed distribution state feature vector, and calculating an exponential function value with base e and the sum of each eigenvalue of the dot product vector as an exponent to obtain the grid-granularity local mixed distribution state feature implicit clustering contribution factor.

[0049] Specifically, based on the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics at the grid granularity, feature aggregation is performed on the sequence of local mixed distribution state feature vectors at the grid granularity to obtain the global mixed state aggregate representation vector, including: arranging the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics at the grid granularity into a mixed distribution state local feature clustering contribution field distribution vector; performing explicit modeling on the mixed distribution state local feature clustering contribution field distribution vector based on the self-attention mechanism to obtain a mixed distribution state local feature clustering contribution field modulation weight vector; using each eigenvalue in the mixed distribution state local feature clustering contribution field modulation weight vector as a weight, calculating the position-weighted sum of the sequence of local mixed distribution state feature vectors at the grid granularity to obtain the global mixed state aggregate representation vector.

[0050] More specifically, the distribution vector of the local feature clustering contribution field in the mixed distribution state is input into the clustering contribution field explicit modeling module based on the self-attention mechanism to obtain the domain modulation weight vector of the local feature clustering contribution field in the mixed distribution state, including: using the query embedding matrix, the key embedding matrix and the value embedding matrix to embed the distribution vector of the local feature clustering contribution field in the mixed distribution state to obtain the query vector, the key vector and the value vector; 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 the attention score matrix; passing the attention score matrix through the softmax function and multiplying it with the value vector to obtain the domain modulation weight vector of the local feature clustering contribution field in the mixed distribution state.

[0051] Accordingly, the step S130 includes: processing the set of the grid granularity local mixed distribution state feature vectors using the following feature significant aggregation formula to obtain the global mixed state aggregation representation vector, wherein the feature significant aggregation formula is:

[0052]

[0053]

[0054]

[0055]

[0056] , ,

[0057]

[0058] in, represents the set of characteristic vectors of the local mixed distribution state of the grid granularity, , , , and are the first, second, and third in the set of local mixed distribution state feature vectors of the grid granularity. , and The local mixed distribution state eigenvector of the grid granularity, The value of is the number of characteristic vectors of the local mixed distribution state of the grid granularity, represents the self-supervised clustering representation vector of the mixed distribution state, Indicates the The first characteristic vector of the local mixed distribution state at the grid granularity eigenvalues, The first vector representing the self-supervised clustering representation of the mixed distribution state eigenvalues, represents the absolute value, represents the logarithmic function with base 2, represents the length of the characteristic vector of the local mixed distribution state of the grid granularity, represents the exponential function with base e, , , and Respectively represent the first, second, and and The implicit clustering contribution factor of the local mixed distribution state characteristics of the grid granularity, Represents the distribution vector of the local feature clustering contribution field of the mixed distribution state, , and denote the query embedding matrix, key embedding matrix and value embedding matrix respectively, , and denote the query vector, key vector and value vector respectively, represents the transpose of a vector, represents the normalized exponential function, represents matrix multiplication, Represents the local feature clustering contribution field modulation weight vector of the mixed distribution state, represents the first in the modulation weight vector of the local feature clustering contribution field of the mixed distribution state eigenvalues, Represents the global mixed state aggregate representation vector.

[0059] Then, in the step S140, the overall mixing uniformity of the material is evaluated based on the semantic difference distribution of each grid-granularity local mixing distribution state feature vector in the set of grid-granularity local mixing distribution state feature vectors relative to the global mixing state aggregate representation vector.

[0060] Accordingly, in a specific example of the present application, the step S140 includes: calculating the mixed state semantic difference coefficient between the global mixed state aggregation representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors to obtain a set of mixed state semantic difference coefficients; calculating the variance of the set of mixed state semantic difference coefficients; and determining whether to stop bubbling based on a comparison between the variance and a preset threshold.

[0061] That is, by calculating the mixed state semantic difference coefficient between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors, the consistency of the mixed state representing each local area relative to the overall mixed state is quantified. In an embodiment of the present application, the Euclidean distance can be used as a metric to calculate the mixed state semantic difference coefficient between each grid-granularity local mixed distribution state feature vector and the global mixed state aggregate representation vector to obtain a set of mixed state semantic difference coefficients.

[0062] Next, in order to measure the mixing uniformity between each local area, the variance of the set of semantic difference coefficients of the mixing state is further calculated to reveal the overall uniformity of the mixture. It should be understood that variance is a statistical indicator used to measure the discreteness of a set of data. In the technical solution of the present application, the smaller the variance, the higher the consistency of the mixing state of each local area relative to the overall mixing state, and the better the uniformity of the mixture as a whole. Conversely, a larger variance indicates a higher degree of uneven mixing. In this way, based on the comparison between the variance and the preset threshold, it can be determined whether the uniformity of the mixture has reached the expected goal, thereby determining whether to stop bubbling.

[0063] In particular, in the technical solution of the present application, each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors respectively represents the semantic coding features of the mixed state image of the local area of ​​the image determined by grid segmentation of the real-time mixed distribution state image. When clustering analysis is performed on the set of grid-granularity local mixed distribution state feature vectors, the clustering network performs significant clustering based on the clustering contribution of each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors to obtain the global mixed state aggregate representation vector. However, since the global mixed state aggregate representation vector is a cluster representation, it has characteristic manifold distribution density differences and edge blurring relative to each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors, which affects the calculation accuracy of the mixed state semantic difference coefficient between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors.

[0064] Based on this, in a preferred example, before calculating the mixed state semantic difference coefficient between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of the grid-granularity local mixed distribution state feature vectors, the global mixed state aggregate representation vector is subjected to feature manifold modulation to obtain an optimized global mixed state aggregate representation vector, and the feature manifold modulation process includes: Performing L random sampling on the global mixed state aggregate representation vector to obtain L global mixed state random perturbation eigenvalues; Arrange the L global mixed state random perturbation eigenvalues ​​in order from large to small to obtain a global mixed state random perturbation ordered vector; Calculating the inner product of the global mixed state aggregate representation vector to obtain a global mixed state energy modulation eigenvalue; Based on the global mixed state energy modulation characteristic value, the global mixed state randomly perturbed ordered vector is semantically scaled to obtain a semantically enhanced global mixed state randomly perturbed ordered vector; Calculating the product of the semantically enhanced global mixed state random perturbation ordered vector and the transposed vector of the global mixed state aggregate representation vector to obtain a global mixed state best sample prompt orthogonal modulation matrix; The matrix product between the global mixed state best sample hint orthogonal modulation matrix and the global mixed state aggregate representation vector is calculated to obtain an optimized global mixed state aggregate representation vector.

[0065] The above optimization process is expressed as follows:

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073] in, represents a normal distribution, represents the Gaussian density probability function, and denote the mean and variance of the global mixed state aggregate representation vector, respectively, represents a random variable that follows a normal distribution, represents the number of random samplings, Indicates The global mixed state random perturbation eigenvalues, represents the global mixed state aggregate representation vector, represents the first eigenvalues, It means to calculate the average value of the vector. represents the length of the global mixed state aggregate representation vector, It means that the random perturbation eigenvalues ​​of the global mixed state are arranged in order from large to small. represents the ordered vector of random perturbations in the global mixed state, represents the energy modulation eigenvalue of the global mixed state, represents the square of the vector's two-norm, It means point multiplication by position. represents matrix multiplication, represents the transpose of a vector, represents the best sample hint orthogonal modulation matrix of the global mixed state, Represents the optimized global mixed state aggregate representation vector.

[0074] Accordingly, by performing L random sampling on the global mixed state aggregate representation vector to obtain a set of new feature values ​​based on the original features but containing slight changes, randomness is introduced to explore different aspects of the feature space, and by simulating the noise or variation in the actual data, the model is better generalized to unseen data. Then, these global mixed state random perturbation feature values ​​are ordered from large to small to form a global mixed state random perturbation ordered vector to identify which perturbation directions may be the most important for the current problem. Such an ordering not only helps the algorithm focus on the feature differences that can best distinguish different categories, but also helps to reveal hidden data structures or patterns. Subsequently, the self-inner product of the global mixed state aggregate representation vector is calculated to obtain the global mixed state energy modulation eigenvalue, which reflects the overall strength or "energy" of the original feature vector and provides a benchmark for measuring the importance of the feature. Using this global mixed state energy modulation eigenvalue as a reference point, the previously formed global mixed state random perturbation ordered vector is semantically scaled to construct a semantically enhanced global mixed state random perturbation ordered vector. In this way, those disturbances that appear to be more important in the original feature space are amplified, thereby enhancing the model's sensitivity to changes in key features, while also retaining relatively small but potentially useful disturbance information.

[0075] Then, the product of the semantically enhanced global mixed state random perturbation ordered vector and the transpose of the global mixed state aggregate representation vector is calculated to obtain the global mixed state best sample prompt orthogonal modulation matrix, the purpose of which is to find the way in which the features interact with each other from two different perspectives. The formed matrix can be regarded as a conversion tool for capturing and representing the complex linear relationship between features. Finally, the final feature structure modulation is completed by applying the global mixed state best sample prompt orthogonal modulation matrix constructed above to the original global mixed state aggregate representation vector, so that the features are more in line with the requirements of the objective function, reducing redundant information, while maximizing the retention of useful information, so as to improve learning efficiency and prediction performance.

[0076] In summary, the feature manifold optimization process of the global mixed state aggregate representation vector is explained, which extracts the best hint sample based on the fine-grained target retrieval of random perturbations, and then guides adversarial robustness learning based on the best hint sample to complete the structure of the high-dimensional manifold of the feature vector in the feature space, so as to improve the detail and structure of the feature expression. In this way, the feature manifold consistency and edge alignment between the global mixed state aggregate representation vector and each grid-size local mixed distribution state feature vector in the set of the grid-size local mixed distribution state feature vector are improved, thereby improving the calculation accuracy of the mixed state semantic difference coefficient between the global mixed state aggregate representation vector and each grid-size local mixed distribution state feature vector in the set of the grid-size local mixed distribution state feature vector.

[0077] Specifically, in step S2, the raw material enters the reaction zone (009), wherein the reaction zone has a reaction unit, including a reaction pipe, a heating medium, a heating jacket, and an insulation layer. The outside of the reaction pipe is provided with a heating jacket, and the space between the reaction pipes is filled with a heating medium. When the raw material enters the reaction pipe, the reaction unit is controlled to meet the conditions that the temperature of the reaction pipe is 40° C. and the carbon dioxide pressure is 0.1-2Mpa.

[0078] Specifically, in step S3, after the raw material is completely introduced, carbon dioxide is supplemented by bubbling through the carbon dioxide gas replenishing port to make the pressure in the reaction pipeline 3-4Mpa. At the same time, the circulation pump is turned on to control the temperature of the reaction unit to 70-90°C, and the raw material passes through the solid-liquid mixing unit and the gas-liquid mixing heat conduction unit; the solid-liquid mixing unit includes a reaction pipeline, a static mixer and a circulation pump, and the solid-liquid mixing unit makes the solid catalyst evenly dispersed in the epoxy compound, and at the same time disperses the generated reaction heat to reduce the temperature difference between the center temperature in the pipeline and the wall of the pipeline; the gas-liquid mixing heat conduction unit includes a dispersion sheet and a circulation pump, a gas-liquid heat transfer area, and the high-efficiency heat conduction gas-liquid mixing unit makes the mixed material with low CO2 content in the reaction unit dissolve CO2 again, so that there is enough CO2 to participate in the reaction continuously, and at the same time, heat exchange is carried out in the gas-liquid mixing heat conduction unit, so that the raw material with low carbon dioxide content at the bottom of the reaction pipeline returns to the gas area at the top of the reaction pipeline to fully dissolve carbon dioxide. The specific process is as follows: the synthetic product in the reaction unit generates reaction heat, which is transferred to the CO2 gas in the gas-liquid heat transfer zone. After the reactant CO2 is consumed, the gas inlet replenishes CO2 to keep the pressure in the reaction unit constant. The newly replenished CO2 is directly replenished from the bottom to form convection with the material to continue heat exchange, which plays a role in cooling down and at the same time increases the carbon dioxide content in the reacting material in the lower part of the pipeline.

[0079] During the reaction process, the reaction mixture containing polycarbonate polyether polyol and propylene carbonate flows through the connecting pipe located at the other side of the feed end of the reaction pipe, passes through the circulation pump arranged at the confluence of the connecting pipe, and converges at the liquid distributor at the top of the reaction pipe, passes through the solid-liquid mixing unit and the gas-liquid mixing heat conduction unit to complete a cycle. The circulation flow rate is controlled within 1-20min to complete a cycle. By observing the density sensor data, when the density reaches 1.12 g / cm 3 , the reaction is deemed to be completed and flows out through the discharge port. The single feed of the feed port is not more than 80% of the volume of the reaction pipeline.

[0080] Specifically, in step S4, after the reaction is completed, the mixture containing polycarbonate polyether polyol, cyclic carbonate and unreacted raw material propylene oxide is passed into the separation unit (010) of the post-processing area, and the separation unit (010) 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 (011), and a refining agent (012) is added for adsorption treatment, and a crude product with a catalyst metal content of less than 5 ppm is obtained after filtration; the refined crude product is passed into the distillation unit (013), and the main product polycarbonate polyether polyol and the by-product cyclic carbonate are separated, that is, the polycarbonate polyether polyol product is obtained and filled (014).

[0081] Correspondingly, the present application also provides a bubbling control method, which includes: acquiring a real-time mixed distribution state image of a premixture captured by a camera; performing network granularity-based mixed distribution state feature extraction on the real-time mixed distribution state image of the premixture to obtain a set of grid granularity local mixed distribution state feature vectors; performing feature significance aggregation processing based on clustering contribution on the set of grid granularity local mixed distribution state feature vectors to obtain a global mixed state aggregate representation vector; determining whether to stop bubbling based on the semantic difference distribution of each grid granularity local mixed distribution state feature vector in the set of grid granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector.

[0082] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned bubbling control method have been described in detail in the description of the production process of the above-mentioned succinic acid-type polycarbonate polyether polyol, and therefore, its repeated description will be omitted.

[0083] Furthermore, the present application also provides a succinic acid type polycarbonate polyether polyol prepared by the production process as described above, and the structural formula of the succinic acid type polycarbonate polyether polyol is shown in Formula 2: Formula 2: In Formula 2, R1 and R2 are hydroxyl groups or structures as shown in Formula 3, but R1 and R2 are not all hydroxyl groups. Formula three, In formula 3, 0.1<m / m+n<0.9, 0.1<n / m+n<0.9, m and n are both positive integers; the molecular weight Mn of the succinic acid type polycarbonate polyether polyol is 800-4200; the molecular weight distribution coefficient PDI of the succinic acid type polycarbonate polyether polyol is ≤2; the molecular weight difference index ΔM of the succinic acid type polycarbonate polyether polyol is <2%; preferably, the molecular weight difference index ΔM≤1%; the molecular weight difference index 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 measured by the product obtained 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 device 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:

[0084] Among them, g 实际 Represents the measured number average molecular weight of the product, g 理论 represents the theoretical molecular weight, which is calculated as:

[0085] Where D is the molecular weight of the initiator.

[0086] Where A is the average molecular weight of the linker, calculated as:

[0087] Where X is the proportion of ester linkages in the product, and Y is the proportion of ether linkages in the product.

[0088] Where B is the number of chain links, calculated as:

[0089] Wherein K is the molar ratio of the initiator feed, C is the conversion rate, and Z is the mass ratio of polypropylene carbonate.

[0090] The conversion rate C is calculated as follows: C=M 反应po / M 投料po Among them, M 反应po The actual mass of converted PO is calculated as follows:

[0091] where n po is the mass proportion of po in the product, which is calculated as follows:

[0092] Where X1 is the carbonate linkage NMR integral, Y1 is the ether linkage integral, and Z1 is the polypropylene carbonate integral.

[0093] In summary, the production process and bubbling control method of the succinic acid type polycarbonate polyether polyol according to the embodiment of the present application are explained, which uses carbon dioxide bubbling to premix oxygen compounds, succinic acid, and DMC catalysts, and uses a camera to monitor the mixed distribution state of the premix in real time, and uses intelligent industrial control technology to extract local features of the grid granularity of the real-time mixed distribution state image collected by the camera, and then evaluates the uniformity of the mixed distribution based on the semantic difference of the local mixed distribution state features of each grid granularity relative to the overall image, so as to achieve intelligent control of the bubbling time. Compared with the traditional production process, it effectively simplifies the process steps, improves the controllability and stability of the premixing effect, reduces the dependence and error on manual operation, and is suitable for industrial large-scale production.

[0094] Example 1 8L propylene oxide, 4g DMC catalyst and 409g succinic acid were added to the premixing kettle respectively, and 0.4Mpa carbon dioxide was introduced and mixed at 40°C. After stirring for 1h, the mixture was pressed into an 8L×2 reactor array, and the diameter of a single reaction pipe was 100mm and the length of the reaction pipe was 1000mm. After all the mixture was pressed into the reactor, carbon dioxide was introduced into the reaction pipe to 3Mpa and the pressure was maintained. The circulation pump was turned on and the flow rate was controlled at 600g / min. After reacting at 80°C for 3h, the density was uniformly increased to 1.12g / cm 3 The reaction is complete and the product is taken out after the temperature is lowered. After the catalyst is removed by the catalyst filter, the main and by-products are separated by a scraper evaporator.

[0095] Example 2 Compared with Example 1, other parameters and reaction materials in Example 2 are unchanged. The only change is that the bubbling control method described above is adopted in Example 2.

[0096] The crude product after the catalyst is removed is sampled and analyzed. Specifically, the crude product is sampled and collected in a container, and the polymerization reaction product logistics sample is characterized by nuclear magnetic resonance spectroscopy to calculate the ratio of polymer to cyclic small molecules in the crude product. The polymer is purified and then subjected to nuclear magnetic resonance spectroscopy testing to calculate the ratio of polycarbonate chain segments to polyether chain segments on the polymer main chain. There are only two structures, polycarbonate chain segments and polyether chain segments, on the polymer main chain, and the sum of the two percentages is 100%.

[0097] 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 resonances in H-NMR (based on TMS = 0 ppm) 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 number subscript represents the integrated area of ​​the peak at a certain ppm on the nuclear magnetic resonance hydrogen spectrum. A is the abbreviation of Area in English, for example, A 5.0 represents the integrated area of ​​the peak at 5.0 ppm. 1 The integrated area of ​​the H-NMR spectrum and its related proton peaks, the ratio of carbonate chains in the copolymerization reaction (molar ratio) (F CO2 ) and the mass fraction of cyclic carbonate content (W PC ), the amount of carbon dioxide embedded (M CO2 ) is calculated as follows: Among them, 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%; The coefficient 44 is the molar mass of carbon dioxide, the coefficient 58 is the molar mass of PO, and the coefficient 102 is the sum of the molar masses of carbon dioxide (molar mass 44 g / mol) and the molar masses of PO (molar mass 58 g / mol).

[0098] The conversion rate of propylene oxide is to cool down the product, separate the unreacted propylene oxide through distillation in a primary separation device and calculate the actual conversion rate. The calculation formula 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 containing propylene oxide.

[0099] The polymer number average molecular weight (Mn) and polymer molecular weight polydispersity index (PDI) were determined by gel permeation chromatography (GPC). The test results are shown in Table 1 below: Table 1

[0100] In Example 2, compared with Example 1, other parameters and reaction materials have not been changed. The only change is that the bubbling time control method described above is adopted in Example 2. As can be seen from the data in Table 1, the adaptive bubbling time control strategy provided in the embodiment of the present application can monitor and analyze the mixed distribution state of the premix in real time, thereby flexibly controlling the bubbling time. Example 2 adopts the adaptive bubbling time control strategy to reduce the premixing time by 50%, and can make the mixing more uniform, thereby making the reaction more uniform, reducing the molecular weight difference index, and being able to synthesize the target product more accurately.

[0101] It can be seen from the data of Example 2 that the bubbling time control as described above can make the raw materials fully react. It should be noted that in Example 1, the premixing end time is 2 hours, while in Example 2, the premixing end time is 1 hour, and the reaction products are tested and the quality between the two is equivalent. Therefore, the use of 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.

[0102] 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 for succinic acid type polycarbonate polyether polyol, characterized in that: The production process of the succinic acid type polycarbonate polyether polyol is carried out in a tube-in-tube reaction device, which includes a premixing unit and a reaction zone; Wherein, the production process comprises: mixing the epoxy compound, succinic acid, and DMC catalyst in the premixing unit to obtain a premix; feeding the premixture into the reaction zone at one time; Introducing carbon dioxide into the reaction zone to supplement the reaction pressure, wherein the direction of the carbon dioxide introduction is opposite to the flow direction of the material in the reaction pipe; A succinic acid type polycarbonate polyether polyol is obtained through polymerization reaction; The maximum pressure of the premixing unit does not exceed 1MPa, and the premixing time is 1h-2h; the premixing unit uses carbon dioxide bubbling to mix the epoxy compound, succinic acid, and DMC catalyst, and carbon dioxide is slowly bubbled in the premixing unit, and the bubbling speed is 40-60 bubbles / min; The epoxy compound, succinic acid and DMC catalyst are mixed in the premixing unit to obtain a premix, comprising: Acquire a real-time mixed distribution state image of the premixture collected by a camera; Performing a network-granularity-based mixed distribution state feature extraction on the real-time mixed distribution state image of the premixture to obtain a set of grid-granularity local mixed distribution state feature vectors; Performing feature significant aggregation processing based on clustering contribution on the set of local mixed distribution state feature vectors of the grid granularity to obtain a global mixed state aggregation representation vector; Whether to stop bubbling is determined based on the semantic difference distribution of each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector.

2. The production process of succinic acid type polycarbonate polyether polyol according to claim 1, characterized in that: The real-time mixed distribution state image of the premixture is subjected to mixed distribution state feature extraction based on network granularity to obtain a set of local mixed distribution state feature vectors of grid granularity, including: Grid-slicing the real-time mixed distribution state image to obtain a set of mixed distribution state grid granularity images; Each mixed distribution state grid granularity image in the set of mixed distribution state grid granularity images is input into a mixed distribution state feature extractor based on a dilated convolutional layer to obtain a set of grid granularity local mixed distribution state feature vectors.

3. The production process of succinic acid type polycarbonate polyether polyol according to claim 2, characterized in that: The set of the grid-granularity local mixed distribution state feature vectors is subjected to feature significant aggregation processing based on clustering contribution to obtain a global mixed state aggregation representation vector, including: Performing cluster analysis on the set of local mixed distribution state feature vectors of the grid granularity to obtain a mixed distribution state self-supervised cluster representation vector; Calculating the implicit clustering contribution factor of each grid granularity local mixed distribution state feature vector in the sequence of grid granularity local mixed distribution state feature vectors relative to the mixed distribution state self-supervised clustering representation vector to obtain a sequence of grid granularity local mixed distribution state feature implicit clustering contribution factors; Based on the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics at the grid granularity, feature aggregation is performed on the sequence of local mixed distribution state feature vectors at the grid granularity to obtain the global mixed state aggregation representation vector.

4. The production process of succinic acid type polycarbonate polyether polyol according to claim 3, characterized in that: Calculating the implicit clustering contribution factor of each grid granularity local mixed distribution state feature vector in the sequence of grid granularity local mixed distribution state feature vectors relative to the mixed distribution state self-supervised clustering representation vector to obtain a sequence of grid granularity local mixed distribution state feature implicit clustering contribution factors, including: Calculate the dot-wise division vector between the grid-granularity local mixed distribution state feature vector and the mixed distribution state self-supervised clustering representation vector, and calculate the base-two logarithm of the absolute value of each feature value in the dot-wise division vector to obtain an implicit clustering contribution weight vector; Calculate the dot product vector between the implicit clustering contribution weight vector and the grid granularity local mixed distribution state characteristic vector, and calculate the exponential function value with e as the base and the sum of the eigenvalues ​​of the dot product vector as the exponent to obtain the implicit clustering contribution factor of the grid granularity local mixed distribution state characteristic.

5. The production process of succinic acid type polycarbonate polyether polyol according to claim 4, characterized in that: Based on the sequence of implicit clustering contribution factors of the local mixed distribution state characteristics of the grid granularity, feature aggregation is performed on the sequence of local mixed distribution state feature vectors of the grid granularity to obtain the global mixed state aggregation representation vector, including: Arranging the sequence of the implicit clustering contribution factors of the local mixed distribution state characteristics of the grid granularity into a mixed distribution state local characteristic clustering contribution field distribution vector; Performing explicit modeling on the local feature clustering contribution field distribution vector of the mixed distribution state based on the self-attention mechanism to obtain a local feature clustering contribution field modulation weight vector of the mixed distribution state; The position-weighted sum of the sequence of the local mixed distribution state feature vectors of the grid granularity is calculated using each eigenvalue in the mixed distribution state local feature cluster contribution field modulation weight vector as a weight to obtain the global mixed state aggregation representation vector.

6. The production process of succinic acid type polycarbonate polyether polyol according to claim 5, characterized in that: The local feature clustering contribution field distribution vector of the mixed distribution state is explicitly modeled based on the self-attention mechanism to obtain the local feature clustering contribution field modulation weight vector of the mixed distribution state, including: Using a query embedding matrix, a key embedding matrix and a value embedding matrix respectively, the local feature clustering contribution field distribution vector of the mixed distribution state is embedded and transformed to obtain a query vector, a key vector and a value vector; 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 local feature clustering contribution field modulation weight vector of the mixed distribution state.

7. The production process of succinic acid type polycarbonate polyether polyol according to claim 6, characterized in that: Determining whether to stop bubbling based on the semantic difference distribution of each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector includes: Calculating the mixed state semantic difference coefficients between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors to obtain a set of mixed state semantic difference coefficients; Calculating the variance of the set of mixed-state semantic difference coefficients; Based on the comparison between the variance and a preset threshold, it is determined whether to stop bubbling.

8. The production process of succinic acid type polycarbonate polyether polyol according to claim 7, characterized in that: Calculating the mixed state semantic difference coefficients between the global mixed state aggregate representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors to obtain a set of mixed state semantic difference coefficients, including: The Euclidean distance between the global mixed state aggregation representation vector and each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors is calculated as the mixed state semantic difference coefficient to obtain the set of the mixed state semantic difference coefficients.

9. The production process of succinic acid type polycarbonate polyether polyol according to claim 1, characterized in that: The reaction pressure of the reaction zone is 3MPa-4MPa. When the pressure in the reaction zone is too low, a solenoid valve is used to automatically introduce carbon dioxide to supplement the pressure. The direction of introduction of the carbon dioxide is opposite to the flow direction of the material.

10. A bubbling control method, characterized in that: include: Acquire a real-time mixed distribution state image of the premixed mixture captured by a camera; Performing a network-granularity-based mixed distribution state feature extraction on the real-time mixed distribution state image of the premixture to obtain a set of grid-granularity local mixed distribution state feature vectors; Performing feature significant aggregation processing based on clustering contribution on the set of local mixed distribution state feature vectors of the grid granularity to obtain a global mixed state aggregation representation vector; Whether to stop bubbling is determined based on the semantic difference distribution of each grid-granularity local mixed distribution state feature vector in the set of grid-granularity local mixed distribution state feature vectors relative to the global mixed state aggregate representation vector.

Images