Intelligent preparation method of high-solid-content low-viscosity waterborne polyurethane based on industrial control system

Abstract

The invention discloses an intelligent preparation method of high-solid-content low-viscosity waterborne polyurethane based on an industrial control system, and relates to the technical field of preparation of waterborne polyurethane materials, and the intelligent preparation method comprises the following steps: in the preparation process of the high-solid-content low-viscosity waterborne polyurethane, deploying various types of online sensors and monitoring devices through the industrial control system; collecting mixing process data generated in the mixing and emulsifying stages of the organic phase and the water phase in real time; and preprocessing the obtained original mixed data, systematically summarizing and sorting the preprocessed data, and constructing a dynamic data set for analyzing mixed behaviors in different process states. According to the method, the microbubble generation state is identified through multi-source sensing and machine learning, the shearing strategy is dynamically adjusted, the microbubble retention risk is actively intervened in the emulsification process, the stability of a dispersion system is improved, the problems of later agglomeration and phase separation are avoided, and the particle size distribution, viscosity control and product consistency of waterborne polyurethane are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of waterborne polyurethane material preparation, and particularly relates to an intelligent preparation method of high-solid and low-viscosity waterborne polyurethane based on an industrial control system. BACKGROUND

[0002] The intelligent preparation of high-solid and low-viscosity waterborne polyurethane based on an industrial control system refers to the use of industrial automation control technologies (such as PLC, DCS system, sensor network, etc.) to monitor and intelligently adjust the key parameters (including temperature, pH value, reaction time, shear rate, dispersion stability, etc.) in the whole process of waterborne polyurethane synthesis, to accurately control the monomer feeding rate, emulsification rhythm, water phase addition ratio and post-processing process in the reaction stage, so as to effectively improve the solid content (such as greater than 45%) of the polyurethane dispersion and simultaneously reduce the system viscosity (such as controlled within 500-2000 mPa·s) under the premise of ensuring the stability of the system, to achieve the optimal balance of product performance and production efficiency. This method breaks the technical bottleneck of "high solid content easy to increase viscosity, low viscosity difficult to stabilize" in the traditional preparation process, and is suitable for the industrialized preparation of high-performance environmentally friendly materials such as functional coatings, adhesives, textile finishing agents, etc.

[0003] The prior art has the following disadvantages: In the preparation process of high-solid and low-viscosity waterborne polyurethane, it is usually necessary to realize the full mixing and stable emulsification of the organic phase and the water phase through high-shear emulsification or water phase dispersion. However, the existing technology often focuses on the control of macroscopic parameters (such as shear rate, temperature, feeding ratio, etc.), and does not fully consider the micro gas-liquid interface behavior induced by high shear and its potential impact. Especially in high-solid and low-surface tension systems, high-speed shearing can easily cause instability of the gas-liquid shear interface at the microscale, and then form a large number of invisible microbubbles. Due to the low viscosity of the system and the incomplete formation of the glue core, these microbubbles often cannot escape or break in time, and finally remain in the dispersion system in the form of hidden cavity structure, and the industrial control system is difficult to effectively identify or respond to the adjustment under the current technical architecture. In this context, when the dispersion system enters the later stage of the reaction (such as chain extension and curing stage) or suddenly encounters temperature fluctuations and other disturbances, the above-mentioned residual bubbles will quickly break due to the sudden change of interfacial tension or pressure imbalance, causing local severe pressure disturbance. This disturbance will directly destroy the stable balance between the glue cores, causing sudden agglomeration between the polyurethane glue core particles, out-of-control particle size and mutual crosslinking, and finally leading to phase separation, emulsion flocculation or abnormal increase in viscosity of the dispersion system, and even causing the whole system to be scrapped, pipeline blockage or customer product failure, etc.

[0004] The above information disclosed in the BACKGROUND section merely to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present application is to provide an intelligent preparation method of high solid content and low viscosity waterborne polyurethane based on an industrial control system, to identify the micro-bubble generation state through multi-source perception and machine learning, to dynamically adjust the shearing strategy, to actively intervene in the micro-bubble retention risk in the emulsification process, to improve the stability of the dispersion system, to avoid the problems of late agglomeration and phase separation, and to significantly improve the particle size distribution, viscosity control and product consistency of the waterborne polyurethane, so as to solve the problems in the above background.

[0006] In order to achieve the above purpose, the present application provides the following technical scheme: an intelligent preparation method of high solid content and low viscosity waterborne polyurethane based on an industrial control system, comprising the following steps:

[0007] In the preparation process of high solid content and low viscosity waterborne polyurethane, a plurality of types of online sensors and monitoring devices are deployed through an industrial control system to collect mixed process data generated in the mixing and emulsification stages of organic phase and water phase in real time;

[0008] The obtained original mixed data is pre-processed, and the pre-processed data is systematically summarized and arranged to construct a dynamic data set for analyzing the mixing behavior under different process conditions;

[0009] On the basis of the data set, the core characteristic index of micro-bubble generation caused by unstable gas-liquid shearing interface is automatically mined from multi-source data through signal analysis technology, and the core characteristic index extracted is comprehensively analyzed to quantify the authenticity of micro-bubble generation;

[0010] The core characteristic index after analysis and quantization processing is input into a machine learning model trained in advance using historical process data, and the current mixing state is intelligently evaluated and distinguished through the model;

[0011] When it is identified that micro-bubbles are generated in the mixing process, the running mode of the shearing equipment is dynamically switched from constant high shearing state to periodic fluctuation shearing mode, that is, after running at constant high shearing for several seconds, it enters the "intervention window" stage, and is switched to low shearing for a short time to provide conditions for micro-bubble rupture or escape.

[0012] Preferably, in the preparation process of high solid content and low viscosity waterborne polyurethane, the specific steps of collecting mixed process data in real time based on the industrial control system are as follows:

[0013] Determine the key monitoring parameters of the mixing and emulsification stages;

[0014] Selecting and arranging corresponding online sensors and monitoring devices at key parts of the reaction equipment;

[0015] All sensor signals are uniformly accessed, collected and dispatched through an industrial control system to build a standardized data channel and realize continuous data stream acquisition;

[0016] A visual interface is configured in the industrial control interface to display the dynamic change trend of various mixing parameters and key alarm states in real time, thereby realizing online perception and tracking control of the whole process of organic phase and water phase mixing behavior.

[0017] Preferably, on the basis of the data set, the core characteristic index of micro-bubble generation caused by instability of the gas-liquid shear interface is automatically mined from the multi-source data through signal analysis technology. The extracted characteristic index includes the bifurcation of the internal liquid flow velocity distribution, the occurrence frequency of the re-turbulent event, and the instantaneous fluctuation amplitude of the energy consumption of the shear device per unit time. The bifurcation of the internal liquid flow velocity distribution and the occurrence frequency of the re-turbulent event are comprehensively analyzed under the detection window to generate the local flow field transition reference value and the shear energy consumption fluctuation reference value, respectively. The local flow field transition reference value and the shear energy consumption fluctuation reference value are used to quantify the authenticity of micro-bubble generation.

[0018] Preferably, the specific steps of comprehensively analyzing the bifurcation of the internal liquid flow velocity distribution and the occurrence frequency of the re-turbulent event under the detection window to generate the local flow field transition reference value are as follows:

[0019] In the detection window, the instantaneous liquid flow velocity vectors at multiple positions in space are obtained by a plurality of distributed liquid flow velocity sensors. Based on the obtained instantaneous liquid flow velocity vectors, a bifurcation disturbance intensity parameter for quantifying the mutation intensity of the liquid flow path is constructed, and the calculation formula is as follows:

[0020]

[0021] In the formula, is the instantaneous liquid flow velocity vector of the i th monitoring point, is the instantaneous liquid flow velocity vector of the i+1 th monitoring point, i.e. the instantaneous liquid flow velocity vector of the next monitoring point, is the normal vector of the instantaneous liquid flow velocity vector of the i th monitoring point, N is the total number of monitoring points, and δ i is the velocity modulus mutation, which represents the absolute value of the difference in liquid flow velocity between adjacent two monitoring points, tanh(·) is the hyperbolic tangent function, and Λ is the bifurcation disturbance intensity parameter;

[0022] On the basis of obtaining the bifurcation disturbance intensity parameter Λ, a disturbance amplification coefficient related to liquid disturbance is introduced, and a local flow field transition reference value is constructed in combination with a nonlinear jump enhancement factor. The calculation expression is as follows:

[0023] LFTRV = ln(l + Δ · tanh(Λ γ )

[0024] wherein, LFTRV is the local flow field transition reference value, Δ is the disturbance amplification coefficient, γ is the nonlinear compression factor, and tanh(·) is the hyperbolic tangent function.

[0025] Preferably, the specific steps of comprehensively analyzing the instantaneous fluctuation amplitude of the energy consumption of the shearing equipment per unit time in the detection window to generate a shearing energy consumption fluctuation reference value are as follows:

[0026] In the detection window, the instantaneous power sequence data of the shearing equipment is collected based on the industrial control system, denoted as P i wherein i represents a sampling point, the shearing power jump amplitude factor is constructed by analyzing the power fluctuation amplitude between adjacent multiple time series sampling points, and the calculation expression is as follows:

[0027]

[0028] wherein, P j is the instantaneous energy consumption value of the shearing equipment at the jth sampling point, P j-1 and P j+1 are the instantaneous energy consumption values of the shearing equipment at the previous sampling point and the next sampling point respectively, ε is a very small positive real number for avoiding zero in the denominator, and G j is the shearing power jump amplitude factor.

[0029] On the basis of the shearing power jump amplitude factor, the change trend of adjacent sampling points in the sequence is quantitatively analyzed, and a shearing energy consumption fluctuation reference value for representing the disturbance trend intensity of the shearing is constructed, and the calculation expression is as follows:

[0030]

[0031] wherein, SEFRV is the shearing energy consumption fluctuation reference value, β is the trend enhancement coefficient, G j-1 is the shearing power jump amplitude factor of the j-1th sampling point, i.e., the shearing power jump amplitude factor of the previous sampling point, and n is the total number of sampling points.

[0032] Preferably, the local flow field transition reference value and the shearing energy consumption fluctuation reference value after analysis and quantification are input into a machine learning model trained in advance using historical process data, a microbubble generation risk coefficient is generated through the model, and the current mixing state is intelligently evaluated and discriminated based on the microbubble generation risk coefficient.

[0033] Preferably, the microbubble generation risk coefficient generated by intelligently evaluating the current mixing state through a machine learning model trained in advance with historical process data is compared with a pre-set microbubble generation risk coefficient reference threshold to determine whether microbubble generation exists in the mixing process, and the determination logic is as follows:

[0034] If the microbubble generation risk coefficient is greater than the pre-set microbubble generation risk coefficient reference threshold, it is determined that microbubble generation exists in the mixing process; if the microbubble generation risk coefficient is less than or equal to the pre-set microbubble generation risk coefficient reference threshold, it is determined that microbubble generation does not exist in the mixing process.

[0035] Preferably, when it is identified that microbubble generation exists in the mixing process, the running mode of the shearing equipment is dynamically switched from constant high shearing state to periodic fluctuation shearing mode, that is, after continuous high shearing operation for several seconds, it enters the "intervention window" stage and is switched to low shearing operation for a short time, and the specific steps for providing conditions for microbubble rupture or escape are as follows:

[0036] The microbubble generation risk coefficient is compared with the microbubble generation risk coefficient reference threshold to generate an intervention trigger identification variable, and the logic is as follows:

[0037]

[0038] Wherein, MGRF is the microbubble generation risk coefficient, MGRF ref is the microbubble generation risk coefficient reference threshold, and a is the shearing mode control logic identification.

[0039] When a = 1, that is, microbubble generation exists in the mixing process, it enters the intervention state, and according to the relationship between the current MGRF value and the threshold value and the historical maximum risk value, the shearing risk normalization factor λ and the shearing power adjustment coefficient η are calculated, and the calculation expression is as follows:

[0040]

[0041] In the formula, λ is the shearing risk normalization factor, MGRF max is the allowed maximum risk reference value, η base is the basic shearing power coefficient, and η is the shearing power adjustment coefficient.

[0042] After obtaining the shearing control parameters, a periodic fluctuation shearing operation instruction is generated to replace the original constant high shearing operation mode, and in each cycle, the shearing device target running power is generated according to the high shearing duration ratio, and the calculation expression is as follows:

[0043] P ctrl = P high ·(1-α)+[P high• θ + η · P high • (1 - β) · α

[0044] In the formula, P ctrl is the target power controlled by the shearing device in the current period, P high is the rated high-shearing operation power of the device, and θ is the proportion of the high-shearing period in the period.

[0045] In the above technical solution, the present application has the following technical effects and advantages:

[0046] The present application can realize early identification and dynamic intervention control of micro-bubble generation behavior in the emulsification process of high-solid and low-viscosity water-based polyurethane, significantly improving the structural stability of the dispersion system and the controllability of the preparation process. Specifically, the method fuses multi-source online sensing, signal feature extraction and machine learning intelligent discrimination technology, first converts the gas-liquid shearing interface instability behavior at the microscopic scale into a quantifiable and predictable control object, and after identifying the micro-bubble generation risk, actively breaks the bubble accumulation trend through a periodic fluctuating shear strategy, providing conditions for its rupture or escape, from the root cause, avoiding the phenomena of glue core agglomeration, abnormal particle size expansion and phase separation caused by bubble burst in the later stage of the emulsion, and ultimately realizing the overall optimization of product viscosity, particle size distribution and storage stability, effectively improving the consistency and quality stability of industrial production of the product. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments or prior art of the present application, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0048] Figure 1 The method flowchart of the present application is based on the intelligent preparation method of high-solid and low-viscosity water-based polyurethane of the industrial control system. DETAILED DESCRIPTION

[0049] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the scope of example implementations to those skilled in the art.

[0050] The present application provides an intelligent preparation method of high-solid and low-viscosity water-based polyurethane based on an industrial control system, as shown in Figure 1 The present application provides an intelligent preparation method of high-solid and low-viscosity water-based polyurethane based on an industrial control system, as shown in

[0051] In the preparation process of high solid content and low viscosity waterborne polyurethane, various types of online sensors and monitoring devices are deployed through industrial control system to collect the mixing process data generated during the mixing and emulsification stage of organic phase and aqueous phase in real time.

[0052] In the preparation process of high solid content and low viscosity waterborne polyurethane, the mixing and emulsification stage of organic phase and aqueous phase is the most critical initial stage in the entire reaction process. This stage is characterized by intense mixing, high shear rate, and complex interface changes, which can easily lead to problems such as micro-bubble accumulation, uneven emulsification, or abnormal gel nucleus formation. Therefore, it is necessary to deploy various types of online sensors and monitoring devices through industrial control systems (such as PLC, DCS, SCADA, etc.) to collect multi-dimensional process data in real time, including shear power, stirring speed, temperature, pressure, acoustic signal (acoustic emission), optical turbidity, local vibration, pH, conductivity, and rheological parameters. The collection of these data can convert the "invisible" mixing dynamic behavior and interface state into "quantifiable" physical indicators, providing data support for subsequent identification of micro-bubble generation, judgment of mixing uniformity, and dynamic adjustment of process parameters.

[0053] Specifically, the above deployment can be achieved through the following systems: ① PLC system (Programmable Logic Controller) is suitable for local control of the reaction kettle, which can directly control the stirring motor, shear head, and other equipment, while connecting temperature sensors, torque sensors, pressure transmitters, etc.; ② DCS system (Distributed Control System) is suitable for centralized scheduling of the entire polyurethane production line, integrating data streams from various subsystems and realizing logical linkage; ③ SCADA system (Supervisory Control And Data Acquisition) can communicate with the field bus system, remotely collect data from optical scattering sensors, acoustic acquisition units, online turbidity meters, rheometer modules, etc., and realize dynamic visualization and process modeling of the mixing state. These systems cooperate to realize the closed-loop control logic from "data collection - real-time judgment - active control", which is the infrastructure for intelligent polyurethane preparation.

[0054] The purpose of this step is to capture various "invisible behaviors" involving gas-liquid shear interface changes during the preparation process in the form of multi-dimensional physical signals in real time, laying the foundation for subsequent data analysis and abnormal behavior identification, and realizing the "macroscopic visualization" of micro-bubble generation behavior.

[0055] In the preparation process of high solid content and low viscosity waterborne polyurethane, the specific steps of real-time acquisition of mixing process data based on industrial control system are as follows: first, determine the key monitoring parameters in the mixing and emulsifying stage, such as shear power, rotation speed, reaction temperature, stirring torque, pressure in the kettle, turbidity, acoustic emission signal, etc.; second, select and install corresponding online sensors and monitoring devices at key positions of the reaction equipment, such as installing torque sensor at stirring shaft, setting optical turbidimeter and acoustic wave sensor in emulsification area, and installing temperature and pressure integrated transmitter on kettle wall; third, all sensor signals are uniformly accessed and collected by industrial control system (such as PLC or DCS) for scheduling, building standardized data channel and realizing continuous data stream acquisition; fourth, configure visual interface in industrial control interface (such as host computer or SCADA system) to display dynamic change trend and key alarm state of various mixing parameters in real time, so as to realize online perception and tracking control of the whole process of organic phase and water phase mixing behavior.

[0056] The obtained original mixing data is pretreated, and the pretreated data is systematically summarized and arranged to build a dynamic data set for analyzing the mixing behavior under different process conditions.

[0057] The obtained original mixing data often contains various noises, irrelevant fluctuations and sensor errors, so it needs to be systematically pretreated. It includes signal denoising (such as wavelet filtering), outlier rejection, data synchronization alignment, standardization and normalization, and feature enhancement operations. The core function of this step is to improve the accuracy and stability of the data, eliminate the data distortion caused by instrument drift, environmental change and other non-target factors, provide high-quality, structured data set basis for the extraction of subsequent key indicators, make the data have consistency, comparability and the attribute of being used for subsequent algorithm modeling.

[0058] This data set not only contains multi-period, multi-variable process parameters, but also has labels such as "microbubble generation / non-generation" according to experimental records or artificial experience. Through the accumulation of such large-scale, structured data set, it can provide rich foundation for subsequent data mining and feature modeling, support high confidence abnormal behavior identification and risk prediction.

[0059] On the basis of the data set, through signal analysis technology, the core feature index of microbubble generation caused by unstable gas-liquid shear interface is automatically mined from multi-source data, and the extracted core feature index is comprehensively analyzed to quantify the authenticity of microbubble generation.

[0060] On the basis of the data set, the core characteristic index of micro-bubble generation caused by the instability of gas-liquid shear interface is automatically mined from multi-source data by signal analysis technology. The extracted characteristic indexes include the bifurcation of the liquid flow velocity distribution in the system, the occurrence frequency of re-turbulent events, and the instantaneous fluctuation amplitude of the energy consumption of the shear device per unit time. The bifurcation of the liquid flow velocity distribution in the system, the occurrence frequency of re-turbulent events, and the instantaneous fluctuation amplitude of the energy consumption of the shear device per unit time are comprehensively analyzed under the detection window to generate the local flow field transition reference value and the shear energy consumption fluctuation reference value, respectively. The local flow field transition reference value and the shear energy consumption fluctuation reference value quantify the authenticity of micro-bubble generation.

[0061] The authenticity of micro-bubble generation is quantified by the local flow field transition reference value and the shear energy consumption fluctuation reference value. The core role is to convert the originally implicit and non-directly observable micro-bubble generation behavior into quantifiable and assessable numerical indicators, thereby realizing early diagnosis and dynamic perception of the instability state of the gas-liquid shear interface. The local flow field transition reference value reflects the frequency and intensity of the transition of the liquid flow in the system from laminar flow to turbulent flow, re-turbulent or vortex structure, and reveals the degree of disorder of the micro-scale flow field caused by the disturbance of the gas-liquid interface. The shear energy consumption fluctuation reference value reveals the instantaneous nonlinear disturbance characteristics of the device power load variation caused by bubble generation and rupture during the shearing process. Joint modeling of these two indicators not only accurately reflects the activity level of micro-bubble generation in the current system, but also serves as a key decision basis for the process control system to identify the risk window, trigger shear adjustment or flow field optimization, and provides a high-precision judgment basis for subsequent intelligent intervention.

[0062] The surge in the bifurcation of the liquid flow velocity distribution and the occurrence frequency of re-turbulent events in the system indeed indicates the instability of the gas-liquid shear interface and induces micro-bubble generation behavior. In the mixing and emulsification process of high-solid and low-viscosity waterborne polyurethane, the shear field should maintain a certain degree of flow order to ensure the balance of interfacial tension and the stable diffusion of droplets. However, when the liquid flow velocity frequently appears bifurcation (i.e., flow direction mutation) and re-turbulent (i.e., small-scale disturbance enhancement) phenomena in local areas, it indicates that the gas-liquid interface is being strongly disturbed, and the interface layer cannot maintain a stable structure. This disturbance easily causes interface rupture or compression, thereby inducing cavitation, desorption, or interface fracture behavior of the liquid in local areas, and further rapidly generating a large number of micro-bubbles. In addition, the generation of micro-bubbles itself will in turn enhance the disturbance of the local flow field, forming a positive feedback mechanism of "shear-disturbance-bubble". Therefore, the abnormal increase in the bifurcation and re-turbulent frequency of the liquid flow velocity distribution is an important physical signal that the interface stability is destroyed and the system enters a high micro-bubble generation state, with high representativeness and engineering monitoring value.

[0063] The specific steps of generating the local flow field transition reference value by comprehensively analyzing the bifurcation and the frequency of re-turbulent events of the internal liquid flow velocity distribution in the detection window are as follows:

[0064] In the detection window, the instantaneous liquid flow velocity vectors at multiple positions in space are obtained through multiple groups of distributed liquid flow velocity sensors, and based on the obtained instantaneous liquid flow velocity vectors, a bifurcation disturbance intensity parameter for quantifying the mutation intensity of the liquid flow path is constructed, and the calculation formula is as follows:

[0065]

[0066] In the formula, is the instantaneous liquid flow velocity vector of the i-th monitoring point, is the instantaneous liquid flow velocity vector of the i+1-th monitoring point, i.e. the instantaneous liquid flow velocity vector of the next monitoring point, is the normal vector of the instantaneous liquid flow velocity vector of the i-th monitoring point, indicating a unit vector perpendicular to , N is the total number of monitoring points, and δ i is the velocity modulus mutation variable, indicating the absolute value of the difference between the liquid flow velocity (modulus) of the adjacent two monitoring points, and the calculation formula is as follows: If δ i is small, the flow velocity changes smoothly, and if δ i is large, it indicates that the flow velocity is suddenly changed in the local spatial range, which may be caused by bubble-induced local flow field disorder, tanh(·) is the hyperbolic tangent function, and Λ is the bifurcation disturbance intensity parameter, indicating the total intensity of the combined occurrence of direction deviation and velocity mutation in the liquid flow velocity field within the entire detection window. The larger the value is, the stronger the non-continuity of the liquid flow path is, and the more likely it is in the shear interface instability state;

[0067] The core purpose of introducing the hyperbolic tangent function tanh(·) is to achieve the dual processing effect of nonlinear compression and mutation trend enhancement of local flow disturbance intensity. In the preparation process of high solid content and low viscosity waterborne polyurethane, the directional deviation (i.e. inner product term) between the velocity vectors of the liquid flow may appear extreme large value or high frequency fluctuation. If directly used to calculate the disturbance intensity, it is easy to cause the abnormal amplification of single point to the overall disturbance level, thereby interfering with the true discrimination of micro-bubble generation trend. The hyperbolic tangent function has the mathematical characteristics of linear in the central region and saturation in the edge region: it can sensitively capture the initial flow field instability trend with high resolution for small to moderate disturbance changes; and it automatically "clamps" the extreme or minimum directional deviation to the range of [-1, 1], avoiding the dominance of abnormal values in the calculation results. Therefore, the tanh(·) function plays a filtering role similar to "mutation enhancement + extreme value suppression" in this regard, making the flow field disturbance intensity discrimination both stable and sensitive to real interface mutation behavior, improving the robustness and engineering adaptability of the entire transition recognition mechanism.

[0068] Through the above formula, the non-continuous mutation behavior of the liquid flow velocity distribution in the mixed system under the micro spatial scale can be effectively captured, thereby identifying the local flow structure instability phenomenon induced by shear, and providing a disturbance basis quantity for the construction of the flow field transition index.

[0069] On the basis of obtaining the bifurcation disturbance intensity parameter Λ, a disturbance amplification coefficient related to the liquid disturbance is introduced, and a local flow field transition reference value is constructed in combination with a nonlinear jump enhancement factor, and the calculation expression is as follows:

[0070] LFTRV = ln(1 + Δ · tanh(Λ γ )

[0071] In the formula, LFTRV is the local flow field transition reference value, Δ is the disturbance amplification coefficient, which is used to amplify the normalized disturbance response, and is adjusted according to other micro abnormal signals in the system, and can be normalized by the liquid density fluctuation amplitude, conductivity rapid change amplitude or high frequency fluctuation rate in the acoustic emission spectrum, γ is the nonlinear compression factor, which is used for nonlinear compression processing of the exponential mapping of the disturbance intensity parameter Λ, controls the system to be more sensitive to the initial small scale transition behavior and respond in advance, rather than triggering intervention when the disturbance is severe, the value range is 0-1, the closer to 0 indicates stronger compression, the closer to 1 indicates more original disturbance amplitude, which can be adjusted according to the complexity of the system, tanh(·) is the hyperbolic tangent function.

[0072] In the construction of the local flow field transition index, the core role of introducing the hyperbolic tangent function tanh(·) is to achieve nonlinear normalization and dynamic enhancement processing of the disturbance intensity. Specifically, tanh(Λ γ)The compressed disturbance intensity value can be mapped to the interval [0, 1), which has the following two key advantages: one is sensitive to small disturbance response, that is, in the initial stage of disturbance, the function is approximately linear, which is beneficial to timely identify the early signs of gas-liquid shear interface instability; the second is to suppress saturation of large disturbance, when the disturbance value is too large, the output of tanh gradually tends to 1, avoiding the exponential amplification or system false triggering caused by extreme value. Therefore, the tanh function in this model not only plays the role of "normalizer", but also as a "risk trend enhancer", ensuring that the system has good stability and identification accuracy under complex flow conditions.

[0073] In the calculation formula of the local flow field transition index, ln represents the natural logarithm function, that is, the logarithm function with the natural constant e (about equal to 2.718) as the base, which plays a balancing mechanism in the index construction to adjust the non-linear compression and amplification. Specifically, ln(1+x) has the mathematical property of "rapid growth at the beginning and gradual flattening at the end": when x is small, ln(1+x) is approximately linear, which can amplify subtle disturbances; when x is large, the growth rate slows down, effectively suppressing the expansion of extreme values, preventing abnormal model output. Therefore, the introduction of the ln function in the index model helps to convert the shear disturbance signal into a more stable and reasonable scale index value, ensuring that the system can be sensitive to risk conditions and not be disturbed by severe disturbances, improving the reliability and numerical stability of the overall judgment.

[0074] The reference value can realize the quantitative characterization of the spatial evolution trend of the liquid flow velocity, and the larger the value is, the stronger the instability of the liquid flow structure in the mixed system is, and the easier the shear interface disorder is to cause and promote the generation of a large number of micro-bubbles. As a response index of micro-bubble generation behavior, the reference value can be used as an input variable of the real-time bubble identification model and as a core control basis for subsequent shear energy consumption regulation or mixing structure adjustment.

[0075] The larger the local flow field transition reference value generated by the comprehensive analysis of the bifurcation and re-turbulence event frequency of the liquid flow velocity distribution in the detection window, the higher the bifurcation and re-turbulence event frequency of the liquid flow velocity distribution in the system, which means that the flow state of the local area in the mixing process is turning from stable laminar flow or controllable turbulent flow to highly unstable nonlinear disturbance mode. Such disturbance is often caused by shear imbalance of the gas-liquid interface, which can lead to rupture of the interface film structure and increase of cavitation micro-disturbances, thereby greatly increasing the probability of micro-bubble generation. Therefore, within the monitoring window, the increase of the transition reference value directly reflects the entry of the gas-liquid shear interface into an unstable state, which is a typical precursor signal of micro-bubble formation. On the contrary, when the reference value is low, it means that the liquid flow state is stable and the interface is not disturbed, and the system remains in the uniform emulsification interval, which is not easy to induce micro-bubbles.

[0076] The abnormal increase in the instantaneous fluctuation amplitude of the energy consumption of the shearing device per unit time can usually be used as an effective indirect feature of the instability of the gas-liquid shearing interface leading to the generation of micro-bubbles. In the high-shear emulsification process for preparing water-based polyurethane, when the gas-liquid interface is stable, the flow state of the system is relatively uniform, and the power consumed by the shearing device is relatively stable. When the interface is unstable, for example, due to excessive shearing, interfacial tension fluctuation or flow field disturbance causing micro-cavitation, bubbles will be rapidly generated, expanded and broken in the local area, causing sudden changes in the local medium density and viscosity distribution, and then causing nonlinear changes in the load of the stirrer in a short period of time. These sudden changes in mechanical resistance directly reflect the abnormal increase in the power fluctuation of the shearing device. Although the micro-bubble generation process itself is extremely small in size, its group effect is enough to disturb the overall energy consumption curve. Therefore, from the process signal level, the abnormal increase in the instantaneous fluctuation amplitude of the shearing energy consumption can be regarded as a signal output of the severe disturbance of the gas-liquid interface and the batch generation of micro-bubbles, which is helpful for early identification and accurate characterization of the instability risk of the interface.

[0077] The specific steps of comprehensively analyzing the instantaneous fluctuation amplitude of the energy consumption of the shearing device per unit time in the detection window to generate the shearing energy fluctuation reference value are as follows:

[0078] In the detection window, the instantaneous power sequence data of the shearing device is collected based on the industrial control system, denoted as P i , the power fluctuation amplitude between adjacent multiple time sequence sampling points is analyzed, the shearing power jump amplitude factor is constructed, and the calculation expression is as follows:

[0079]

[0080] In the formula, P j is the instantaneous energy consumption value of the shearing device at the jth sampling point, P j-1 and P j+1 are the instantaneous energy consumption values of the shearing device at the previous sampling point and the next sampling point respectively, ε is a very small positive real number for avoiding zero in the denominator, and the value range is 10 -6 -10 -10 , G j is the shearing power jump amplitude factor, which quantifies the "mutation amplitude" of the shearing energy at the current time, i.e., whether there is a significant non-stationary behavior, and the larger the value, the more likely it is that the shearing energy fluctuates sharply and asymmetrically before and after the time point, which is caused by the disturbance of the gas-liquid interface, the aggregation or rupture of micro-bubbles, and the abnormal shearing response;

[0081] By constructing the shearing power jump amplitude factor G j , the asymmetric fluctuation behavior caused by the disturbance of the gas-liquid interface can be captured. This factor can sensitively reflect the abnormal shearing response caused by the generation of micro-bubbles, providing key basic data for subsequent disturbance trend analysis and fluctuation reference value construction.

[0082] On the basis of the shear power jump amplitude factor, the change trend of adjacent sampling points in the sequence is quantitatively analyzed, and a shear energy fluctuation reference value for representing the trend intensity of shear disturbance is constructed. The calculation expression is as follows:

[0083]

[0084] In the formula, SEFRV is the shear energy fluctuation reference value, β is the trend enhancement coefficient, which is used to enhance or scale the variation amplitude of the local disturbance trend, and the value range is 1-5. By adjusting the coefficient, the sensitivity of the overall formula to the perturbation change can be controlled. When β takes a larger value, the response to the small fluctuation will be amplified, and the identification sensitivity will be improved. When β takes a smaller value, the system pays more attention to the macroscopic violent fluctuation, and the noise interference is filtered. G j-1 is the shear power jump amplitude factor of the j-1th sampling point, that is, the shear power jump amplitude factor of the previous sampling point, and n is the total number of sampling points.

[0085] In the construction of the shear energy fluctuation reference value, the core role of the hyperbolic tangent function tanh is to realize the nonlinear amplification and boundary compression control of the disturbance trend change. Specifically, when the shear power jump amplitude factor (G j -G j-1 ) changes slightly, the tanh function shows a nearly linear response, which can finely retain the small fluctuation characteristics of the system and is used to identify the early microbubble generation signal. When the disturbance amplitude changes sharply, the tanh function quickly approaches ±1, effectively compressing the abnormal sharp value within a limited range, preventing the single-point extreme value from causing irrational amplification of the overall reference value. Through this "S-shaped" response mechanism, the tanh function has the dual functions of sensitive identification of small disturbances and suppression of extreme value influence, making the shear energy fluctuation reference value more stable, continuous and having physical trend expression ability, and more suitable for describing the nonlinear disturbance characteristics caused by microbubble generation under the unstable state of the gas-liquid interface.

[0086] Through the above steps, the nonlinear disturbance trend of the shear device energy consumption can be comprehensively reflected. If the reference value significantly increases in the monitoring window, it indicates that there is a batch generation behavior of microbubbles induced by the instability of the gas-liquid interface in the current shear process. Conversely, if the reference value changes smoothly, it means that the system is running smoothly, the interface disturbance is weak, and the risk of microbubble generation is low.

[0087] The greater the shear energy consumption fluctuation reference value generated by comprehensive analysis of the instantaneous fluctuation amplitude of the energy consumption of the shearing device per unit time in the detection window, the more it can be used to represent the degree of instability of the gas-liquid shear interface and indirectly reflect the possibility of microbubble generation. In the process of high-shear mixing, if the gas-liquid interface is in a stable state, the liquid flow resistance and viscosity distribution are relatively uniform, the power output of the shearing device is stable during operation, and the energy consumption fluctuation amplitude is small; but when the interface is unstable, such as the interface tension fluctuates sharply, the mutual disturbance is enhanced, and micro-cavitation occurs frequently, a large number of microbubbles will be formed and broken in the local area, further causing flow field mutation and instantaneous resistance increase. This local disturbance will directly affect the mechanical load of the shearing device, causing a large nonlinear fluctuation in the energy consumption per unit time. By mathematical modeling and statistical quantification of the change characteristics of the energy consumption curve, the shear energy consumption fluctuation reference value obtained can be used as a sensitive representation of the interface stability. Therefore, the greater the shear energy consumption fluctuation reference value, the more unstable the shearing process, and the more active the microbubble generation; on the contrary, the smaller the reference value, the more stable the system operation, the gas-liquid interface is stable, and the risk of microbubble generation is low.

[0088] The core feature index after analysis and quantification is input into the machine learning model trained in advance using historical process data, and the current mixing state is intelligently evaluated and distinguished by the model;

[0089] The local flow field transition reference value and the shear energy consumption fluctuation reference value after analysis and quantification are input into the machine learning model trained in advance using historical process data, and the microbubble generation risk coefficient is generated by the model, and the current mixing state is intelligently evaluated and distinguished based on the microbubble generation risk coefficient.

[0090] The "machine learning model trained in advance using historical process data" refers to a prediction model with autonomous learning ability constructed and trained based on a large amount of mixed working condition data and microbubble state label data accumulated in the past waterborne polyurethane production process before actually putting into the intelligent identification system, so that it can master the complex nonlinear mapping relationship between the gas-liquid shear interface behavior and the microbubble generation. In this training stage, the technical personnel collect multi-dimensional process parameters including shear power, local temperature fluctuation, flow field disturbance intensity, turbidity response, acoustic emission intensity, etc. through an industrial data platform or an experimental system, and combine image recognition, optical detection or manual labeling to label and classify whether the mixed state appears microbubble and the bubble generation intensity, such as "no microbubble generation", "microbubble generation", etc. Then, the training set is constructed using these labeled sample data, and machine learning algorithms such as random forest, support vector machine, XGBoost, convolutional neural network (CNN) are used for training, so that the model can automatically judge the gas-liquid interface state and potential microbubble risk of the mixed system by inputting multi-dimensional process parameters. This process can be regarded as a process of "teaching the model to identify risk signs from historical experience", which ensures that the model has the ability to accurately predict and identify in the future when new data is input.

[0091] After the model training is completed, it can be embedded into the real-time control system as an intelligent identification module for online reasoning. When the system collects key characteristic values such as local flow field transition index and shear energy consumption fluctuation index in actual production in real time, these parameters are input into the trained machine learning model as the input of the model, and the model quickly outputs a microbubble generation risk coefficient based on the "feature-label" mapping rule mastered during the training. The coefficient represents the probability or risk level of microbubble generation or aggregation under the current process state. Through continuous evaluation of the risk coefficient, the system can determine whether the current mixing is unstable, whether the shear intervention needs to be started, whether there is shear energy consumption anomaly, etc., thereby realizing dynamic intelligent control of the mixing process. In summary, the pre-trained machine learning model not only identifies "whether there is bubble", but also learns the rules from a large amount of historical process knowledge, thereby providing reliable and real-time decision support for predictive management and process adaptive adjustment of the gas-liquid interface behavior.

[0092] The machine learning model is not limited here, and machine learning models that can comprehensively analyze the local flow field transition reference value LFTRV and the shear energy consumption fluctuation reference value SEFRV to generate the microbubble generation risk coefficient MGRF are all applicable. To realize the technical solution of the present application, the present application provides a specific implementation manner;

[0093] The micro-bubble generation risk factor MGRF is generated according to the following formula: MGRF = K1·LFTRV + K2·SEFRV, wherein K1 and K2 are preset proportional coefficients of the local flow field transition reference value LFTRV and the shear energy fluctuation reference value SEFRV, and K1 and K2 are greater than 0.

[0094] The "preset proportional coefficient" refers to two constant factors K1 and K2 in the calculation formula of the micro-bubble generation risk factor MGRF, which are multiplied by the local flow field transition reference value LFTRV and the shear energy fluctuation reference value SEFRV respectively. The role of these coefficients is to weight and normalize the contribution of two different physical quantities, so that they have comparable numerical scales and can reasonably participate in the linear combination of the risk factor.

[0095] In actual modeling, because the physical phenomena represented by the local flow field transition reference value LFTRV and the shear energy fluctuation reference value SEFRV are different, the units are different, and the fluctuation ranges are different, direct addition will lead to imbalance. Therefore, K1 and K2 are artificially set to "stretch" or "compress" the numerical weight of each, ensuring that the actual influence degree of each in the micro-bubble generation risk is reflected when synthesizing MGRF. This "preset" is usually based on historical process data analysis, experience model fitting results or engineers' understanding of system sensitivity parameters, and is a key parameter adjustment mechanism in the process of mapping physical characteristics to risk indicators.

[0096] According to the micro-bubble generation risk factor, the greater the local flow field transition reference value generated by comprehensively analyzing the bifurcation of the liquid flow velocity distribution and the occurrence frequency of re-turbulent events in the detection window, and the greater the shear energy fluctuation reference value generated by comprehensively analyzing the instantaneous fluctuation amplitude of the energy consumption of the shearing device in the detection window, the greater the micro-bubble generation risk factor generated by intelligently evaluating the current mixing state by using the machine learning model trained in advance using historical process data, indicating that the probability of generating micro-bubbles in the mixing process is greater, and vice versa.

[0097] The micro-bubble generation risk factor generated by intelligently evaluating the current mixing state by using the machine learning model trained in advance using historical process data is compared and analyzed with the preset micro-bubble generation risk factor reference threshold to determine whether micro-bubbles are generated in the mixing process, and the judgment logic is as follows:

[0098] If the micro-bubble generation risk factor is greater than the preset micro-bubble generation risk factor reference threshold, it is determined that micro-bubbles are generated in the mixing process; if the micro-bubble generation risk factor is less than or equal to the preset micro-bubble generation risk factor reference threshold, it is determined that micro-bubbles are not generated in the mixing process.

[0099] When it is identified that there is microbubble generation in the mixing process, the operation mode of the shearing device is dynamically switched from a constant high-shear state to a periodic fluctuation shear mode, that is, after several seconds of continuous high-shear operation, a "intervention window" stage is entered, and a short time is switched to low-shear operation (for example, for 2-5 seconds), which provides conditions for microbubble rupture or escape, breaks the bubble retention dilemma caused by continuous high-shear operation, and realizes active intervention on the risk of microbubble accumulation.

[0100] By dynamically switching the operation mode of the shearing device, the traditional "constant high-shear" mode is changed to a more flexible and intelligent response "periodic fluctuation shear" mechanism, thereby effectively breaking the dilemma of long-term retention of microbubbles in a high-solid low-viscosity system, and realizing active regulation and intervention on the risk of microbubble generation. In the constant high-shear state, due to the continuous action of shear force, microbubbles are continuously formed, but it is difficult for them to escape, but to be trapped in the liquid phase, embedded in the colloid core structure, forming a hidden cavity, and then causing quality problems such as colloid core aggregation or separation instability in the later reaction stage. By setting an "intervention window", that is, after several seconds of high-shear operation in each cycle, a short time (such as 2-5 seconds) is switched to a low-shear state, which can significantly reduce the fluid disturbance intensity in the system, allowing the trapped microbubbles to float, diffuse or rupture and escape under the action of buoyancy. At the same time, the low-shear environment also allows the liquid system to restore part of the flow field stability, which helps the colloid core to self-organize and arrange and release the bubble interfacial tension, thereby actively reducing the microbubble accumulation density at the micro level. This strategy not only improves the stability and uniformity of the dispersion system, but also lays the foundation for the intelligentization and high-reliability control of the entire preparation process, and is a key operation mechanism for preventing and controlling the process risk induced by microbubbles.

[0101] When it is identified that there is microbubble generation in the mixing process, the operation mode of the shearing device is dynamically switched from a constant high-shear state to a periodic fluctuation shear mode, that is, after several seconds of continuous high-shear operation, a "intervention window" stage is entered, and a short time is switched to low-shear operation (for example, for 2-5 seconds), which provides conditions for microbubble rupture or escape, breaks the bubble retention dilemma caused by continuous high-shear operation, and realizes active intervention on the risk of microbubble accumulation.

[0102] The microbubble generation risk coefficient is compared with the microbubble generation risk coefficient reference threshold value, and an intervention trigger identification variable is generated, and the logic is as follows:

[0103]

[0104] Wherein, MGRF is the microbubble generation risk coefficient, MGRF refis a micro-bubble generation risk coefficient reference threshold, a is a shear mode control logic identifier, a = 1 indicates that there is micro-bubble generation in the mixing process, and the dynamic shear intervention mode needs to be started, and a = 0 indicates that the mixing process is in a stable state, and a constant shear operation can be maintained;

[0105] When a = 1, that is, there is micro-bubble generation in the mixing process, the intervention state is entered, the shear risk normalization factor λ and the shear power adjustment coefficient η are calculated according to the relationship between the current MGRF value and the threshold value and the historical maximum risk value, and the calculation expression is as follows:

[0106]

[0107] In the formula, λ is a shear risk normalization factor, indicating the relative position ratio of the current risk level in MGRF ref and MGRF max , the value range is 0-1, MGRF max is the maximum risk reference value allowed, which can be understood as the limit tolerable micro-bubble risk level, η base is a basic shear power coefficient, the initial power setting reference value in the low shear stage, and η is a shear power adjustment coefficient, which is a parameter for controlling the shear power intensity in the “intervention window” stage.

[0108] λ is dynamically modified to realize flexible matching of shear intensity and risk level.

[0109] After obtaining the shear control parameters, a periodic fluctuating shear operation instruction is generated to replace the original constant high shear operation mode, and in each cycle, the target operation power of the shear device is generated according to the high shear duration ratio, and the calculation expression is as follows:

[0110] P ctrl =P high ·(1-α)+[P high ·θ+η·P high ·(1-β)]·α

[0111] In the formula, P ctrl is the target power of the shear device in the current cycle, indicating the actual operation power of the shear device that should be applied in the current control cycle, P high is the rated high shear operation power of the device, indicating the constant high shear power setting value of the device under normal operation, which is a fixed maximum operation parameter and is used as a reference power for the control model, participating in adjusting the target power level of the high shear section and the low shear section, θ is the proportion of the high shear period in the cycle, indicating the proportion of the time maintaining the high shear operation in a control cycle, and the value range is 0-1.

[0112] The control formula can realize the periodic fluctuation operation of the shear device between high shear and low shear, provide conditions for microbubble rupture and escape in the intervention window stage (such as 2-5 seconds of low shear time), effectively eliminate the microbubble retention phenomenon caused by continuous high shear, and actively reduce the risk of microbubble accumulation.

[0113] The present application can realize early identification and dynamic intervention control of the microbubble generation behavior in the emulsification process of high solid content and low viscosity waterborne polyurethane, and significantly improve the structural stability of the dispersion system and the controllability of the preparation process. Specifically, the method first converts the gas-liquid shear interface instability behavior at the microscale into a quantifiable and predictable control object by fusing multi-source online sensing, signal feature extraction and machine learning intelligent discrimination technology, and after identifying the microbubble generation risk, actively breaks the bubble accumulation trend through the periodic fluctuation shear strategy, provides conditions for bubble rupture or escape, avoids the phenomena of glue core agglomeration, abnormal particle size expansion and phase separation caused by bubble burst in the later stage of emulsion from the root, and finally realizes the overall optimization of product viscosity, particle size distribution and storage stability, effectively improves the consistency of the product and the quality stability of industrial production.

[0114] The above formulas are dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0115] The above only describes certain exemplary embodiments of the present application by way of illustration, and it is self-evident that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present application.

[0116] It should be noted that in this paper, if there are relationship terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitation, the element defined by the sentence "including a…" does not exclude the presence of other identical elements in the process, method, article or equipment including the element.

[0117] It should be understood that the size of the sequence number of the above processes does not mean the order of execution in various embodiments of the present application, and the execution order of the processes should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0118] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0120] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment of the present application according to actual needs.

[0121] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically, or two or more units can be integrated into one unit.

[0122] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0123] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without deviating from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of the claims of the present application.

Claims

1. An intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system, characterized in that: The following steps are involved: During the preparation of high-solids, low-viscosity waterborne polyurethane, various types of online sensors and monitoring devices are deployed through the industrial control system to collect real-time mixing process data generated during the mixing and emulsification stages of the organic and aqueous phases. Preprocess the acquired raw mixing data, systematically summarize and organize the preprocessed data, and construct a dynamic data set for analyzing mixing behavior under different process states; Based on the data set, signal analysis technology is used to automatically mine the core characteristic indicators of microbubble generation caused by gas-liquid shear interface instability from multi-source data. The extracted core characteristic indicators are comprehensively analyzed to quantify the authenticity of microbubble generation. The core characteristic indicators that have been analyzed and quantified are input into a machine learning model that has been pre-trained using historical process data, and the model is used to intelligently evaluate and identify the current mixing state. When microbubble generation is detected during the mixing process, the shear device's operating mode is dynamically switched from a constant high-shear state to a periodic fluctuating shear mode. That is, after a few seconds of continuous high-shear operation, it enters the "intervention window" stage and briefly switches to low-shear operation to provide conditions for the microbubbles to break or escape.

2. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 1, characterized in that: In the preparation process of high-solids, low-viscosity waterborne polyurethane, the specific steps for real-time collection of mixing process data based on the industrial control system are as follows: Identify key monitoring parameters during the mixing and emulsification stages; Select and deploy corresponding online sensors and monitoring devices at key locations of the reaction equipment; All sensor signals are uniformly accessed and collected through the industrial control system, building a standardized data channel to achieve continuous data flow acquisition; A visualization interface is configured in the industrial control interface to display the dynamic change trends and key alarm status of various mixing parameters in real time, thereby realizing online perception and tracking control of the entire process of mixing behavior of organic and aqueous phases.

3. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 1, characterized in that: Based on the data set, signal analysis technology is used to automatically mine the core characteristic indicators of microbubble generation caused by gas-liquid shear interface instability from multi-source data. The extracted characteristic indicators include the frequency of bifurcation and re-turbulence events in the liquid flow velocity distribution within the system and the instantaneous fluctuation amplitude of the energy consumption of the shear equipment per unit time. After a comprehensive analysis of the frequency of bifurcation and re-turbulence events in the liquid flow velocity distribution within the system and the instantaneous fluctuation amplitude of the energy consumption of the shear equipment per unit time within the detection window, a local flow field transition reference value and a shear energy consumption fluctuation reference value are generated respectively. The local flow field transition reference value and the shear energy consumption fluctuation reference value are used to quantify the authenticity of microbubble generation.

4. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 3, characterized in that: The specific steps for comprehensively analyzing the frequency of bifurcation and re-turbulence events in the system's internal flow velocity distribution within the detection window to generate a local flow field transition reference value are as follows: Within the detection window, multiple sets of distributed liquid flow velocity sensors are used to obtain the instantaneous liquid flow velocity vectors at multiple locations in space. Based on the obtained instantaneous liquid flow velocity vectors, a bifurcation disturbance intensity parameter is constructed to quantify the intensity of the sudden change in the liquid flow path. The calculation formula is as follows: Where, is the instantaneous liquid flow velocity vector at the i-th monitoring point, is the instantaneous liquid flow velocity vector of the i+1th monitoring point, that is, the instantaneous liquid flow velocity vector of the next monitoring point, is the normal vector of the instantaneous liquid flow velocity vector at the i-th monitoring point, N is the total number of monitoring points, δ i is the velocity modulus mutation, which represents the absolute value of the difference in liquid velocity between two adjacent measuring points, tanh(·) is the hyperbolic tangent function, and Λ is the bifurcation disturbance intensity parameter; On the basis of obtaining the bifurcation disturbance intensity parameter Λ, the disturbance amplification coefficient related to the liquid disturbance is introduced, and the local flow field transition reference value is constructed in combination with the nonlinear jump enhancement factor. The calculation expression is as follows: LFTRV=ln(1+Δ·tanh(Λ γ )) where LFTRV is the local flow field transition reference value, Δ is the disturbance amplification coefficient, γ is the nonlinear compression factor, and tanh(·) is the hyperbolic tangent function.

5. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 3, characterized in that: The specific steps for comprehensively analyzing the instantaneous fluctuation amplitude of the shearing equipment's energy consumption per unit time within the detection window to generate a shearing energy consumption fluctuation reference value are as follows: Within the detection window, the instantaneous power sequence data of the shearing equipment is collected based on the industrial control system and is recorded as P i , where i represents the sampling point. By analyzing the power fluctuation amplitude between multiple adjacent time series sampling points, the shear power jump amplitude factor is constructed. The calculation expression is as follows: Where, P j is the instantaneous energy consumption of the shearing equipment at the jth sampling point, P j-1 and P j+1 are the instantaneous energy consumption values ​​of the shearing equipment at the previous sampling point and the next sampling point, respectively. ε is a very small positive real number used to avoid the denominator being zero. G j is the shear power jump amplitude factor; Based on the shear power jump amplitude factor, the change trend of adjacent sampling points in the sequence is quantitatively analyzed to construct a shear energy consumption fluctuation reference value used to characterize the intensity of the shear disturbance trend. The calculation expression is as follows: Where, SEFRV is the shear energy fluctuation reference value, β is the trend enhancement coefficient, G j-1 is the shear power jump amplitude factor of the j-1th sampling point, that is, the shear power jump amplitude factor of the previous sampling point, and n is the total number of sampling points.

6. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 3, characterized in that: The local flow field transition reference value and shear energy consumption fluctuation reference value that have been analyzed and quantified are input into a machine learning model that has been pre-trained using historical process data. The microbubble generation risk coefficient is generated through the model, and intelligent evaluation and judgment are performed based on the current mixing state of the microbubble generation risk coefficient.

7. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 6, characterized in that: The microbubble generation risk coefficient generated by intelligently evaluating the current mixing state using a machine learning model pre-trained using historical process data is compared with a pre-set reference threshold for the microbubble generation risk coefficient to determine whether microbubbles are generated during the mixing process. The judgment logic is as follows: If the microbubble generation risk coefficient is greater than a preset microbubble generation risk coefficient reference threshold, it is determined that microbubbles are generated during the mixing process; if the microbubble generation risk coefficient is less than or equal to the preset microbubble generation risk coefficient reference threshold, it is determined that no microbubble is generated during the mixing process.

8. The intelligent preparation method of high-solid-content, low-viscosity waterborne polyurethane based on an industrial control system according to claim 7, characterized in that: When microbubble generation is detected during the mixing process, the shear device's operating mode is dynamically switched from a constant high shear state to a periodic fluctuating shear mode. That is, after a few seconds of continuous high shear operation, the "intervention window" phase is entered, and low shear operation is briefly switched to provide conditions for the microbubbles to break or escape. The specific steps are as follows: The microbubble generation risk factor is compared with the microbubble generation risk factor reference threshold to generate an intervention trigger identification variable. The logic is as follows: Among them, MGRF is the microbubble generation risk factor, MGRF ref is the reference threshold of the microbubble generation risk factor, α is the shear mode control logic identifier; When α = 1, that is, microbubbles are generated during the mixing process, the intervention state is entered. According to the relationship between the current MGRF value, the threshold value and the historical maximum risk value, the shear risk normalization factor λ and the shear power adjustment coefficient η are calculated. The calculation expression is as follows: Where λ is the shear risk normalization factor, MGRF max is the maximum risk reference value allowed, η base is the basic shear power coefficient, η is the shear power adjustment coefficient; After obtaining the shear control parameters, a periodic fluctuating shear operation instruction is generated to replace the original constant high shear operation mode. In each cycle, the shearer target operating power is generated according to the high shear duration ratio. The calculation expression is as follows: P ctrl =P high ·(1-a)+[P high ·θ+η·P high ·(1-b)]·a Where, P ctrl is the target power controlled by the shear in the current cycle, P high is the rated high shear operating power of the equipment, and θ is the proportion of the high shear period in the cycle.

Images