Program cooling crystallization kettle and method capable of regulating and controlling sugar alcohol crystal granularity

Abstract

The invention relates to the technical field of crystallization control, and discloses a programmed cooling crystallization kettle and method capable of regulating and controlling sugar alcohol crystal granularity, and the method comprises the following steps: constructing a basic nonlinear cooling model in advance, and respectively collecting real-time torque and actual average chord length of the crystal by using a torque sensor and a focused light beam reflection measurement probe in the crystallization process; and the control system calculates the cooling rate correction according to the granularity deviation through a PID algorithm, calculates a rheology correction factor according to the torque change, and performs attenuation treatment on the sum of the basic rate and the correction by using the rheology correction factor to obtain a final cooling instruction to adjust the temperature in the kettle. Through dual feedback of coupling granularity and rheology, the contradiction between crystal growth and mass transfer resistance under a high-viscosity system is solved, the inclusion amount of the mother liquor is remarkably reduced while the target particle size is realized, and the purity and uniformity of the product are improved.

Description

A programmed cooling crystallizer and method for controlling the particle size of sugar alcohol crystals Technical Field

[0001] This invention relates to the field of crystallization control technology, specifically to a programmed cooling crystallization vessel and method for controlling the particle size of sugar alcohol crystals. Background Technology

[0002] Sugar alcohols, such as maltitol and xylitol, are important functional sweeteners. Their industrial production primarily involves cooling crystallization to precipitate solid crystals from high-concentration solutions. Unlike conventional inorganic salt systems, sugar alcohol solutions exhibit significant rheological properties; that is, the viscosity increases exponentially with decreasing temperature and increasing solid content. This high viscosity significantly increases the mass transfer resistance of solute molecules diffusing from the bulk solution to the crystal surface, making the thermodynamic and kinetic control of the crystallization process extremely complex.

[0003] Existing industrial sugar alcohol crystallization processes mostly rely on preset cooling curves for control, such as linear cooling or experience-based nonlinear cooling. This open-loop control mode only performs temperature regulation according to the time axis and cannot perceive the actual growth state of the crystals inside the reactor. Due to uncontrollable factors in actual production, such as fluctuations in raw material purity, changes in cooling water temperature, and differences in stirring efficiency, preset curves often fail to accurately match the actual crystallization kinetics requirements. When the cooling rate is too fast, causing the supersaturation to exceed the upper limit of the metastable region, it will trigger explosive nucleation, generating a large number of fine crystals that are difficult to separate; while when the cooling rate is too slow, it will significantly prolong the production cycle and reduce equipment utilization.

[0004] To address the aforementioned issues, some existing technologies have introduced online monitoring methods such as focused beam reflection measurement, attempting to adjust the cooling rate in real time based on changes in crystal particle size. However, single-parameter feedback control based solely on particle size deviation has significant limitations when processing high-viscosity sugar alcohol systems. In the later stages of crystallization, the system is in a high-viscosity solid-liquid mixture state, with severely impaired mass transfer. If the control system maintains a high cooling rate solely to catch up with the target particle size, it will lead to excessively high local supersaturation near the crystal growth interface, causing the crystal growth rate to outpace the rate of ordered arrangement of lattice molecules. This mismatch between kinetics and mass transfer rate easily results in the mother liquor being trapped inside the crystal, forming a cavitation phenomenon known as mother liquor entrapment. This not only reduces the final purity of the product but also leads to structural defects within the crystal, affecting subsequent drying and storage stability. Therefore, this invention provides a programmed cooling crystallizer and method with adjustable sugar alcohol crystal particle size to overcome the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a programmed cooling crystallizer and method for controlling the particle size of sugar alcohol crystals, thus solving the problems mentioned in the background art.

[0006] The first aspect of this invention provides a programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals. This method aims to solve the problem that traditional control methods struggle to balance crystal growth kinetics and mass transfer resistance during the crystallization of high-viscosity sugar alcohols. The method includes the following steps: First, a basic nonlinear cooling model suitable for the sugar alcohol crystallization system is pre-constructed; based on this model, the basic cooling rate at the current operating moment is obtained. During cooling crystallization, a torque sensor mounted on the stirring shaft collects real-time torque values, while a focused beam reflection measurement probe inserted into the crystal slurry collects the actual average chord length of the crystals. Based on the collected data, a dual feedback calculation is performed: a PID algorithm is used to calculate the cooling rate correction based on the deviation between the actual average chord length and the preset ideal particle size growth trajectory; a rheological correction factor is calculated based on the degree of change of the real-time torque value relative to the initial reference torque; the sum of the basic cooling rate and the cooling rate correction is attenuated using the rheological correction factor to calculate the final execution cooling rate command, and the temperature of the crystallization vessel is adjusted according to this command, thereby maintaining crystal growth while preventing localized supersaturation due to excessively high system viscosity.

[0007] In a preferred embodiment, the final cooling rate command is calculated as follows: the base cooling rate is added to the cooling rate correction to obtain a sum of rates, and then this sum of rates is divided by the rheological correction factor. If the calculation result is less than zero, the command is set to zero. This calculation logic ensures that when the system viscosity increases, causing the rheological correction factor to rise, the system can automatically reduce the cooling rate.

[0008] In a preferred embodiment, the rheological correction factor is calculated using a logarithmic continuous correction algorithm. This involves calculating the ratio of the filtered real-time torque value to the initial reference torque, taking the natural logarithm of this ratio, multiplying it by the rheological sensitivity coefficient to obtain the increment, and then adding this increment to the numerical value. This algorithm enables a smooth response to viscosity changes.

[0009] In a preferred embodiment, the cooling rate correction is calculated based on an incremental PID control algorithm, using the granularity deviations of the current time, the previous sampling time, and the time before that sampling time for weighted calculation.

[0010] In a preferred embodiment, the method further includes a nucleation anomaly intervention step. When the count rate of fine particles fed back by the focused beam reflection measurement probe shows an exponential increase within a preset time, it is determined that a secondary nucleation burst has occurred, and the system automatically executes a temperature recovery program to dissolve the excess fine crystal nuclei until the count rate returns to stability.

[0011] In a preferred embodiment, the crystallization process is considered complete when the following conditions are met simultaneously: the actual temperature inside the crystallization vessel drops to the endpoint temperature and is maintained for a preset time; the absolute value of the slope of the real-time torque value of the stirring shaft changing over time remains less than a preset threshold; the count rate of fine particles is lower than the endpoint threshold and the peak value of the main particle size reaches the target specification. After the determination is completed, the crystal growth step begins, maintaining the endpoint temperature constant and reducing the stirring speed.

[0012] A second aspect of the present invention provides a programmed cooling crystallizer with adjustable sugar alcohol crystal particle size. The device includes a vessel body, a composite stirring impeller, a torque sensor, a focused beam reflection measurement probe, and a central control unit. The vessel body is equipped with a heat exchange jacket and internal heat exchange coils; the composite stirring impeller is driven by a motor and installed inside the vessel; the torque sensor is connected to a drive shaft to collect torque data; the focused beam reflection measurement probe is immersed below the liquid surface to collect chord length data. The central control unit is configured to receive the sensor data and execute the calculation steps in the aforementioned method to output control commands to adjust the flow rate of the heat exchange medium.

[0013] In a preferred embodiment, the composite impeller employs a combination of an outer anchor-type blade and an inner spiral-belt blade. The outer anchor-type blade is close to the inner wall of the vessel to enhance wall scraping heat transfer, while the inner spiral-belt blade generates an axial circulating flow field. This structure is adapted to the high viscosity characteristics of sugar alcohols in the later stages of crystallization, and, in conjunction with the aforementioned control methods, ensures a uniform distribution of the temperature and concentration fields within the vessel.

[0014] This invention provides a programmed cooling crystallizer and method for controlling the particle size of sugar alcohol crystals. It offers the following advantages: 1. By constructing a feedback adjustment mechanism based on rheological properties, this invention effectively resolves the contradiction between crystal growth rate and mass transfer resistance in high-viscosity sugar alcohol crystallization systems. When the system viscosity increases significantly in the later stages of crystallization, the rheological correction factor can automatically decrease the cooling rate according to changes in real-time torque, preventing excessive local supersaturation caused by forced cooling. This avoids excessively rapid crystal growth that could trap the mother liquor within the crystal lattice, significantly reducing the amount of mother liquor trapped in the product and thus improving the purity of the final product.

[0015] 2. This invention utilizes a focused beam reflection measurement probe and a PID control algorithm to form a closed-loop control for crystal grain size. This allows for real-time capture of the deviation between the actual and ideal crystal growth trajectory, and dynamic adjustment of the supersaturation driving force, eliminating interference from equipment heat dissipation fluctuations or nucleation randomness. Compared to traditional open-loop control, this enables precise particle size targeting, significantly reducing the coefficient of variation in particle size distribution and resulting in crystal products with better uniformity.

[0016] 3. This invention improves batch stability and efficiency in the production process by employing multi-parameter coupled endpoint determination logic and corresponding hardware configuration. By comprehensively determining the temperature, torque change slope, and fine particle count rate, the system can accurately identify the true equilibrium point of crystallization thermodynamics and kinetics, avoiding premature termination leading to insufficient yield or premature termination leading to energy waste. Combined with the optimization of the high-viscosity flow field by the composite stirring paddle, it ensures the representativeness of sensor data and realizes standardized control of the crystallization process in industrial production. Attached Figure Description

[0017] Figure 1 is a perspective view of the crystallization vessel of the present invention; Figure 2 is a general process flow chart of the present invention; Figure 3 is a flow chart of the seed crystal addition step of the present invention; Figure 4 is a flow chart of the closed-loop control step of the present invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please refer to Figures 1-4. This invention provides a programmed cooling crystallizer system with adjustable sugar alcohol crystal particle size, which mainly includes a crystallizer body, a stirring system, an online sensing system, a programmed cooling unit, a seed crystal addition device, and a central control unit.

[0020] The crystallization vessel, used for the crystallization process of sugar alcohol solutions, is a vertical cylindrical structure made of stainless steel. A heat exchange jacket is installed on the outer wall of the crystallization vessel for introducing refrigerant or heat to regulate the vessel wall temperature. Internal heat exchange coils are installed within the cavity of the crystallization vessel, arranged spirally in the lower part of the vessel to increase the heat exchange area. The top of the crystallization vessel has a feed inlet and an exhaust outlet, while the bottom has a discharge outlet.

[0021] The stirring system is vertically installed at the top center of the crystallization vessel. The system includes a variable frequency drive motor, a stirring shaft, and a composite stirring paddle. The variable frequency drive motor is connected to the upper end of the stirring shaft, driving its rotation. A torque sensor is installed on the stirring shaft to collect the output torque value in real time. The composite stirring paddle is fixed to the lower end of the stirring shaft and extends into the interior of the crystallization vessel.

[0022] The composite agitator employs a combination of anchor blades and helical blades. The anchor blades are located on the outermost side, their rotation sweeping across the inner wall of the crystallizer. A 2-3 mm gap is maintained between the outer edge of the anchor blades and the inner wall of the crystallizer; this gap, combined with fluid shear force, prevents scaling on the wall surface. The helical blades are located inside the anchor blades, comprising an inner helical band and an outer helical band, with opposite rotation directions. This structure creates an axial circulating flow field within the crystallizer, allowing the high-viscosity sugar alcohol slurry to tumble vertically as a whole.

[0023] The online sensing system includes a focused beam reflection measurement probe mounted on the side wall of the crystallizer. The probe is inserted below the liquid surface inside the crystallizer at an angle of 30 to 45 degrees. A conical self-cleaning guide tube is coaxially fitted around the measurement window at the front end of the probe. This conical self-cleaning guide tube is a hollow frustum-shaped structure with openings at both ends; the larger opening faces the direction of fluid flow from the composite agitator, and the smaller opening faces the direction of fluid outflow. This structure utilizes Bernoulli's principle to create an accelerating flow field at the measurement window, which is used to clean the surface of the focused beam reflection measurement probe.

[0024] The programmed cooling unit includes a refrigerant circulation pump, a heat exchanger, and a flow control valve assembly. The outlet of the refrigerant circulation pump is connected via piping branches to the inlet of the heat exchange jacket and the inlet of the heat exchange inner coil. The flow control valve assembly is located on the piping and is used to independently regulate the refrigerant flow rate entering the heat exchange jacket and the heat exchange inner coil.

[0025] The seed crystal feeding device is located on the top end cap of the crystallization vessel. The seed crystal feeding device includes a storage hopper and a loss-in-weight weighing scale. The discharge port of the loss-in-weight weighing scale is connected to the interior of the crystallization vessel and is used to precisely feed micron-sized sugar alcohol seed crystals according to instructions.

[0026] The central control unit establishes electrical connections with the variable frequency drive motor, torque sensor, focused beam reflection measurement probe, refrigerant circulation pump, flow control valve group, and loss-in-weight weighing scale. The central control unit internally stores a nonlinear cooling curve model and rheological correction algorithm, used to receive feedback signals from the torque sensor and focused beam reflection measurement probe, and to send control commands to the programmable cooling unit.

[0027] This invention provides a programmed cooling crystallization method for sugar alcohol crystals based on rheological property correction, comprising the following steps: S100, conveying the pretreated sugar alcohol solution to the crystallization vessel, starting the stirring system and measuring the initial reference torque; S200, calling the preset nonlinear cooling curve and controlling the programmed cooling unit to start cooling the sugar alcohol solution; S300, monitoring the solution temperature in the crystallization vessel, and when the temperature reaches the upper limit of the metastable region, controlling the seed crystal addition device to add sugar alcohol seeds; S400, acquiring the chord length distribution data of the focused beam reflection measurement probe and the torque data of the stirring shaft in real time, and calculating the rheological correction factor and particle size deviation; S500, based on the rheological correction factor and particle size deviation, using a closed-loop control algorithm to calculate the corrected cooling rate command, and adjusting the programmed cooling unit to execute the rate; S600, when the solid content and crystal particle size are detected to reach the preset target values, switching to constant-rate crystal growth mode until discharge.

[0028] The initialization and baseline parameter determination step, as the starting point of the entire temperature-cooling crystallization process, directly affects the baseline accuracy of the subsequent rheological correction algorithm. This step mainly includes three sub-processes: material loading, flow field establishment, and rheological baseline calibration.

[0029] S110, Material Loading and Preheating. The sugar alcohol solution, processed in the previous step, is pumped into the crystallization vessel through the inlet. The initial concentration of the sugar alcohol solution is controlled between 72wt% and 82wt%, and the initial temperature is controlled between 65℃ and 85℃. During the feeding process, the temperature control loop of the heat exchange jacket is activated, and a heat medium matching the feed temperature is introduced to ensure that the temperature of the inner wall of the vessel is consistent with the solution temperature, with the temperature difference controlled within ±1℃, to prevent premature nucleation caused by the solution contacting the cold wall surface. The feed rate is determined based on the effective volume of the crystallization vessel, and the liquid level is typically controlled to reach 70% to 80% of the total height of the vessel, ensuring that the liquid level completely submerges the measurement window of the focused beam reflection measurement probe and the main body of the composite agitator. S120, Flow Field Establishment and Steady-State Confirmation. After the material loading is completed, the variable frequency drive motor of the stirring system is started. The variable frequency drive motor drives the composite agitator at a preset initial speed. Start rotating. Initial rotation speed. The setting follows the principle of ensuring that the stirring Reynolds number is... At 10 3 Up to 10 4The transition flow region ensures overall mixing without generating strong turbulent shear, typically set between 60 rpm and 120 rpm. The anchor blades of the composite agitator generate tangential flow near the vessel wall, scraping away the boundary layer, while the inner and outer helical blades generate axial circulation in the central region of the vessel, driving uniform mixing of the solution. For 5 to 10 minutes after stirring starts, a temperature sensor inside the vessel continuously monitors the solution temperature distribution. When the temperature deviation between the upper, middle, and lower points inside the vessel is less than 0.5℃, a stable homogeneous flow field is established, and the solution is in thermodynamic equilibrium; S130, rheological reference calibration. After confirming a stable flow field and a clear, unnucleated solution, the central control unit collects the real-time output torque signal from the agitator shaft via a torque sensor. To eliminate errors caused by signal fluctuations, the system continuously collects torque data for a period (e.g., 30 to 60 seconds), processes the raw data using a moving average filtering algorithm, and calculates its arithmetic mean, which is defined as the initial reference torque. Initial reference torque This characterizes the fluid resistance torque generated by a sugar alcohol solution without solid particles on the stirring impeller under specific initial concentration, initial temperature, and initial rotational speed conditions. This parameter will be stored in the control unit's memory as the denominator term in subsequent calculations of the rheological correction factor. The acquisition and processing of torque data can be achieved using conventional analog-to-digital converter circuits, and will not be elaborated upon here. Simultaneously, at this stage, the focused beam reflection measurement probe begins acquiring background chord length distribution data. If the detected fine particle count is lower than a preset background noise threshold, the system is confirmed to be in a clean state awaiting crystallization.

[0030] S210, Establishment of the basic nonlinear cooling model. Based on the solubility characteristics of sugar alcohols, the central control unit pre-constructs a basic nonlinear cooling curve. This model maintains the supersaturation of the system by controlling the cooling rate. Located in the middle of the metastable region, that is 1.15. The nonlinear cooling curve follows the following mathematical relationship: In the formula: Indicates cooling operation time The set temperature at any given time, in degrees Celsius; This indicates the starting temperature of the cooling process, which corresponds to the feed temperature in step S110. This indicates the running time of the cooling process, in minutes (min). This is the cooling rate coefficient, with units of °C·min. -β Its value is determined by fitting the slope of the solubility curve of a specific sugar alcohol at saturation temperature using the least squares method, and typically ranges from 0.6 to 1.5; This is a nonlinear exponential factor, dimensionless, used to adjust the curvature of the cooling curve, with a value ranging from 0.3 to 0.8. For sugar alcohols with a narrow metastable region (such as xylitol), Use a smaller value (0.3–0.5) to achieve slow initial cooling; for sugar alcohols with a wide metastable region (such as maltitol), Take the larger value (0.5~0.8); S220, execution of the cooling curve and temperature closed loop. The programmed cooling unit receives the data from the central control unit. The command adjusts the refrigerant flow rate using a cascaded PID control algorithm. The control unit collects real-time temperature feedback from the internal temperature sensor. Calculate its relationship with the set temperature The deviation is adjusted by regulating the valve group to change the refrigerant flow rate into the heat exchange jacket and inner coil. To improve the heat transfer response speed, when the absolute value of the deviation is greater than 2°C, the system simultaneously opens the dual-loop cooling of the jacket and inner coil; when the absolute value of the deviation is less than or equal to 2°C, only the jacket loop is used for fine-tuning. This graded control strategy can ensure that the actual cooling trajectory matches the preset nonlinear curve and prevent temperature overshoot; S230, confirmation of the metastable region. During the cooling process, the system continuously monitors the supersaturation state of the solution. The temperature is controlled by the preset curve, and combined with real-time FBRM background count monitoring, the system is indirectly confirmed to be in the metastable region. If the count rate of fine particles (<10μm) detected by FBRM exceeds the preset safety threshold (e.g., 500 particles / second) before the addition of seed crystals, it indicates that spontaneous nucleation has occurred in the system. At this time, the control unit will immediately execute the temperature recovery program, that is, control the heating medium to enter the jacket to raise the kettle temperature by 1-2°C until the count rate drops, ensuring that the crystallization process always takes place within the controlled metastable region.

[0031] S310, Determination and execution of seed crystal addition point. During the programmed cooling process, the central control unit compares the actual temperature inside the reactor in real time. With the preset nucleation temperature Nucleation temperature The sugar alcohol solution is set to be 3°C to 5°C below its saturation temperature at the current concentration, placing it in the upper limit of the metastable region. When monitored... Down to At this time, the central control unit sends a feeding command to the seed crystal feeding device. Based on the received command, the loss-in-weight weighing scale opens the feeding valve and evenly adds the seed crystals from the storage hopper into the crystallization vessel via a vibratory feeding method. The seed crystals are micron-sized dry powder of the same sugar alcohol, with an average particle size of... The addition amount should be controlled between 20μm and 50μm. The concentration is 0.196% to 0.5% of the total solution mass. During seed crystal addition, the stirring system is maintained at the initial rotation speed. During operation, the axial circulating flow field of the composite stirring impeller rapidly disperses the seed crystals throughout the solution system, preventing seed crystal agglomeration; S320, online identification of nucleation state. After the seed crystals are added, the system enters the nucleation monitoring stage, using a focused beam reflection measurement probe to capture changes in the number of particles in real time. The central control unit focuses on monitoring the count rate trend of fine particles with a chord length of less than 50 μm, and the system calculates the fine crystal count growth rate every 60 seconds. The calculation formula is: ;in, For the current fine grain count, This is the fine crystal count from the previous sampling time. If... The value shows a steady upward trend and remains between 5% and 15%, indicating effective nucleation without secondary nucleation bursts. The system maintains its current nonlinear cooling strategy. An exponential surge occurred, with the growth rate exceeding 50% in a single sampling period, indicating a secondary nucleation burst; S330, dynamic control of nucleation anomalies. In response to the secondary nucleation burst detected in S320, the central control unit automatically triggers temperature control intervention logic. The system immediately pauses the current cooling program and controls the program cooling unit to perform isothermal holding operation, maintaining the reactor temperature at the current point for 10 to 20 minutes. The isothermal operation utilizes the Oswald ripening effect to promote the dissolution of excess fine crystal nuclei, reducing the system's supersaturation. [Fine crystal count growth rate to be determined] After three consecutive sampling cycles, the temperature drops below 5%, the system automatically resumes the execution of the nonlinear cooling curve. This dynamic intervention based on FBRM feedback ensures that the particle number density is within a controllable range during the initial stage of crystal growth, resulting in a uniform particle size distribution in the final product.

[0032] S410, Real-time Acquisition and Preprocessing of Rheological Parameters. During the crystal growth stage, as the solid content increases and the temperature decreases, the fluid viscosity and non-Newtonian fluid characteristics of the sugar alcohol slurry gradually increase. The central control unit continuously reads the instantaneous torque value fed back by the torque sensor at a sampling period of 0.1 to 0.5 seconds. To filter out high-frequency noise caused by liquid surface fluctuations or mechanical vibrations, the system uses a moving average filtering algorithm to smooth the original torque signal to obtain the real-time torque value. At the same time, the system records the current running time. To prevent subsequent calculation errors caused by momentary signal drops, the system is set... The effective lower limit is That is, when the measured value is less than When, the forced value is S420, Construction of the calculation model for the rheological correction factor. To quantify the impact of increased system viscosity on mass transfer resistance, a dimensionless rheological correction factor is introduced. This factor reflects the degree of deviation of the current system fluidity from its initial state. The central control unit calculates it using the following logarithmic empirical formula.

[0033] In the formula: express The rheological correction factor at time t is a dimensionless value, and its minimum value is set to 1. express Real-time torque value after filtering and limiting, in Newton-meters (N·m); This represents the initial reference torque measured and stored in step S130, expressed in Newton-meters (N·m). This is the rheological sensitivity coefficient, used to adjust the algorithm's response to changes in viscosity. The value of is related to the chain length of the sugar alcohol molecule and the intrinsic viscosity of the solution. For disaccharide alcohols such as maltitol, The value ranges from 0.2 to 0.4; for monosaccharide alcohols such as xylitol, The value ranges from 0.1 to 0.2. The specific value of input is determined through preliminary pilot-scale experiments: at a constant stirring speed, the increase factor of torque at the end of crystallization relative to the initial torque is recorded. And based on the maximum cooling rate allowed by the target crystal grain size, the decay ratio Through relational formulas The physical meaning and application logic of S430, the rheological correction factor, are calculated. The numerical value directly characterizes the degree of mass transfer obstruction in the current crystallization environment.

[0034] when hour, This indicates that the system viscosity did not increase significantly, and the mass transfer resistance was small; when hour, This indicates that as the system viscosity increases, the diffusion of solute molecules to the crystal surface is hindered. This factor... It will be used as a correction factor in the calculation of subsequent cooling rate commands, that is, when As the temperature increases, the system will automatically decrease the cooling rate proportionally to allow sufficient time for crystal growth and prevent secondary nucleation caused by excessively rapid cooling leading to local accumulation of supersaturation.

[0035] S510 constructs the target trajectory for crystal grain size growth and calculates the deviation. To achieve precise control of crystal size, the central control unit pre-stores the ideal grain size growth trajectory calculated based on the crystal growth kinetics model. The trajectory is a monotonically increasing curve with time, representing the theoretical value of the average chord length of the crystal under ideal supersaturation and flow field conditions. During the cooling crystallization stage, the system reads the actual average chord length data fed back by the focused beam reflection measurement probe in real time with a control cycle of 30 to 60 seconds. The system compares the actual measured values ​​with the theoretical target values ​​and calculates the granularity deviation. The calculation formula is as follows: In the formula, for The particle size deviation at any given time, expressed in micrometers (μm); when When, it indicates that the current crystal growth is lagging behind the target trajectory; when This indicates that the current crystal growth is ahead of the target trajectory; S520, PID feedback adjustment calculation based on grain size deviation. The central control unit is equipped with an incremental PID control algorithm module, used to adjust the grain size deviation based on the PID feedback. Correction for calculating the cooling rate This correction factor compensates for the crystal growth rate by adjusting the cooling rate, which in turn alters the supersaturation. The PID calculation follows the discretization formula: In the formula: This is the cooling rate adjustment component calculated by PID control, in °C / min. , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively, and their values ​​are tuned using the Ziegler-Nichols method. These represent the granularity deviation values ​​at the current time, the previous sampling time, and the time before that sampling time, respectively. To prevent system overshoot due to integral saturation, an output limit is set in the algorithm module. The absolute value is limited to within 50% of the base cooling rate; S530, the rheological correction factor is calculated in conjunction with the cooling rate. This step aims to couple the influence of rheological properties on mass transfer to temperature control. The system first calculates the base nonlinear cooling rate at the current moment. According to the temperature model in step S210, Next, obtain the rheological correction factor calculated in step S420. The central control unit calculates the final cooling rate command using the following coupling formula.

[0036] In the formula: This is the final execution rate command sent to the programmable cooling unit, in °C / min; The theoretical cooling rate set for the basic model; This is a rheological correction factor, which acts as a decay term in the denominator. If the calculated... The system will force the setting. This means pausing cooling. The formula implements dual feedback control: the PID term... It is responsible for dynamically adjusting the driving force based on particle size deviation, while the rheological factor... It is responsible for globally constraining the overall rate based on the system viscosity. When the system viscosity is too high ( When the temperature is increased, even if the PID algorithm requires faster cooling, the final instruction will still be... It will also be forcibly attenuated, thereby preventing localized oversaturation and impurity trapping caused by poor mixing in high-viscosity environments; S540, executes command issuance and actuator action. The central control unit will calculate the... This is converted into a specific refrigerant flow adjustment signal. The system adjusts the flow rate based on the current actual temperature inside the reactor. and target rate Predict the target temperature for the next control cycle. The flow control valve assembly adjusts its opening degree according to the target temperature to control the refrigerant flow into the heat exchange jacket and inner coil. If the particle size deviation exceeds a certain threshold within three consecutive control cycles... Continuously less than -5μm (i.e., severely advanced growth), although Even though the value has been set to 0, the system will still automatically trigger the reheat procedure described in S230, introducing a heat transfer medium to raise the reactor temperature and dissolve the fine crystal nuclei. The specific electronic circuit implementation for PID parameter tuning and valve opening control can be implemented using existing industrial automation control technologies, and will not be elaborated upon here.

[0037] S610, a multi-parameter collaborative endpoint determination logic. At the end of the cooling crystallization process, the central control unit determines whether the process has ended based on a comprehensive index of temperature, torque, and particle size distribution. The system monitors the following three criteria in real time: first, the actual temperature inside the reactor... Reduced to the preset endpoint temperature (Typically 25℃ to 35℃), and maintained at this temperature for more than 30 minutes; secondly, the torque signal of the stirring shaft tends to be stable, that is, the derivative (or slope of change) of torque with respect to time is calculated over 10 consecutive minutes. When the absolute value of the slope remains less than 0.01 N·m / min, it is considered that the solid content and viscosity have reached equilibrium; thirdly, the count rate of fine particles (<10 μm) fed back by the focused beam reflection measurement probe is lower than the preset endpoint threshold, and the peak value of the main particle size is... The target specifications (e.g., 200μm to 300μm) are reached. The system only determines the crystallization process is complete and automatically exits the closed-loop control mode when all three conditions are met simultaneously; S620, crystal cultivation and solid-liquid separation. After the endpoint is determined, the system enters the crystal cultivation stage, aiming to eliminate lattice defects inside the crystal and improve crystal surface integrity. The control unit maintains the reactor temperature at... Keep the stirring speed constant and reduce it to the initial speed. The stirring speed is 50% to 60%, and the process continues for 60 to 120 minutes. This low-speed stirring state prevents crystal sedimentation and accumulation through fluid lifting force, and significantly reduces collision and wear between the crystals and the impeller and reactor wall. After crystal growth, the bottom discharge valve is opened, and the crystal slurry is transported to a centrifuge for solid-liquid separation by gravity or pumping. The separated mother liquor is returned to the upstream process for concentration and recovery through pipelines, while the wet crystals are transported to a vibrating fluidized bed dryer and dried using low-temperature dehumidifying hot air to obtain the final sugar alcohol crystal product; S630, in-situ cleaning and reset of the system. After discharge, in order to prevent residual crystal slurry from scaling on the reactor wall and probe surface and affecting the next batch of production, the system automatically starts the CIP cleaning program. The cleaning spray ball sprays hot water at 80°C to 90°C into the reactor, while the agitator rotates at high speed (e.g., 150 rpm) to generate strong turbulence to scour the reactor wall. Specifically, for the measurement window of the focused beam reflection measurement probe, a high-pressure water jet is sprayed using a directional cleaning nozzle installed inside the vessel, combined with the tangential shearing force generated by high-speed stirring, to thoroughly remove the adhering substances from the window surface. After cleaning for 15 to 30 minutes, the cleaning solution is drained, and the system resets the parameters of each sensor and the state of the control algorithm to prepare for the initialization operation of the next batch.

[0038] Example 1: The dual feedback closed-loop control of the present invention is strictly executed according to steps S100 to S600 in the above specific embodiments, wherein the cooling strategy is based on a nonlinear model + PID granular feedback + rheological factor correction, and the control logic is as follows: Seed strategy is to... It is added locally and has an abnormal temperature recovery mechanism in S330.

[0039] Comparative Example 1: Compared to Example 1, the difference lies in the use of a traditional linear cooling crystallization process, without performing the model building and closed-loop feedback control in steps S200 to S500. Specifically, after the seed crystals are added, the system is set to a constant cooling rate of 0.15℃ / min, from the nucleation temperature... linear cooling to the final temperature Furthermore, FBRM particle size data and torque rheology data are not collected, while the other hardware equipment and raw material parameters are the same.

[0040] Comparative Example 2: Compared to Example 1, the difference lies in that only the nonlinear cooling model in step S210 is used for open-loop control, without introducing the real-time feedback adjustment in steps S400 and S500. Specifically, the system strictly follows the preset curve. The cooling process is executed while ignoring particle size deviations and viscosity changes during actual operation; that is, the command is always executed. The rest are the same.

[0041] Comparative Example 3: Compared to Example 1, the difference lies in the introduction of only single-feedback control based on granularity deviation, without incorporating rheological property correction. Specifically, in the control algorithm of step S530, a rheological correction factor is forcibly set. That is, the formula for calculating the cooling rate command degenerates into The system adjusts the rate based solely on the particle size deviation feedback from the FBRM, without responding to changes in the system viscosity; all other aspects remain the same.

[0042] Comparative Example 4: Compared to Example 1, the difference lies in the introduction of only single feedback control based on rheological characteristics, without incorporating granularity deviation adjustment. Specifically, in the control algorithm of step S530, the PID adjustment term is forcibly set. That is, the formula for calculating the cooling rate command degenerates into The system only relies on the rheological factor decay rate fed back by the torque sensor and does not actively track the grain size trajectory of crystal growth; all other aspects are the same.

[0043] Comparative Example 5: The difference compared to Example 1 lies in the rheological correction factor. The calculation methods differ. Specifically, in step S420, instead of using a logarithmic attenuation formula, a linear threshold determination method is employed: when the real-time torque... Exceeding the initial reference torque When the cooling rate is 1.5 times that of the previous rate, the cooling rate is directly increased. Halve the value; if it does not exceed the limit, no correction is made. This comparative example is used to verify the superiority of the continuous rheological correction algorithm of this invention over simple threshold switching control; all other aspects are the same.

[0044] The experimental steps and conclusions based on Example 1 and Comparative Examples 1 to 5 are as follows: Experimental steps: Parallel experiments were conducted in a 500L pilot-scale crystallizer using the same batch of maltitol solution with an initial purity of 92%. The initial feed concentration, impeller type, and seed crystal size were controlled identically in each group of experiments. After the crystallization process, the resulting slurry was centrifuged and fluidized bed dried, and the final product was sampled and tested. The test indicators included: the average particle size was determined using a Malvern laser particle size analyzer (wet method). The product yield is calculated by measuring the ratio of the mass of dried crystals to the theoretical maximum amount of precipitation. After dissolving a fixed weight of crystals, the impurity content is determined by HPLC to determine the amount of mother liquor contained in the solution. This indicator reflects the degree of internal defects in the crystals. The total time from the start of cooling to the end of the determination of the endpoint is recorded as the time consumed per batch.

[0045]

[0046] Experimental Conclusions: Comparison of data from Example 1 with Comparative Examples 1 and 2 shows that the closed-loop control strategy is crucial for the stability of the crystallization process. Comparative Example 1 used linear cooling; due to the mismatch between the cooling rate and the solubility curve, the proportion of fine crystals reached 13.2%, with a CV value of 46.5%, indicating explosive nucleation and a wide particle size distribution. Although Comparative Example 2 used a nonlinear model, the lack of real-time feedback prevented it from handling thermodynamic fluctuations in actual operation, resulting in a product particle size of 225 μm, failing to meet the predetermined target. Example 1 achieved the target particle size and uniformity of distribution through real-time adjustments. Comparison of data from Example 1 with Comparative Examples 3 and 4 verifies the synergistic effect of rheological-particle size dual feedback. Comparative Example 4 only had rheological constraints; the system passively decelerated as viscosity increased, leading to low production efficiency (14.2 h) and small particle size. Comparative Example 3 only had particle size feedback; in the high-viscosity environment of the later stage of crystallization, the control algorithm maintained a high cooling rate to catch up with the particle size target, resulting in local mass transfer obstruction and an increase in mother liquor inclusion to 580 ppm. Example 1 combines the advantages of both, utilizing rheological factors By moderately reducing the cooling rate in the later stages, the inclusion content was controlled at 115 ppm while maintaining the particle size (268 μm), thus balancing macroscopic size and microscopic purity. Comparing the data from Example 1 and Comparative Example 5 demonstrates the advantages of the continuous correction algorithm. Comparative Example 5 uses a threshold-switching correction method; rate jumps near the threshold lead to supersaturation fluctuations, resulting in a lower CV value (23.6%) for its final product compared to Example 1 (18.2%). Example 1 employs a logarithmic continuous correction formula, achieving smooth adjustment of the cooling rate, which is more conducive to maintaining a stable crystal growth environment.

Claims

1. A programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals, characterized in that, Includes the following steps: S1: Pre-construct a basic nonlinear cooling model for sugar alcohol crystallization and obtain the basic cooling rate at the current moment based on it; S2: During the cooling crystallization process, the real-time torque value of the stirring shaft is collected in real time using a torque sensor, and the actual average chord length of the crystals in the crystal slurry is collected in real time using a focused beam reflection measurement probe; S3: Based on the deviation between the actual average chord length and the preset ideal particle size growth trajectory, the cooling rate correction amount is calculated using a PID algorithm; S4: Calculate the rheological correction factor based on the degree of change of the real-time torque value relative to the initial reference torque; S5: The sum of the basic cooling rate and the cooling rate correction amount is attenuated using the rheological correction factor to calculate the final cooling rate command, and the temperature of the crystallizer is adjusted according to the command.

2. The programmed cooling crystallization vessel and method for adjustable sugar alcohol crystal particle size according to claim 1, characterized in that, The calculation method for the final execution cooling rate command in step S5 is as follows: add the basic cooling rate and the cooling rate correction amount to obtain the rate sum value, and then divide the rate sum value by the rheological correction factor. The quotient obtained is the final execution cooling rate command. When the calculated value of the final execution cooling rate command is less than zero, the system forcibly sets the final execution cooling rate command to zero.

3. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 2, characterized in that, The rheological correction factor in step S4 is calculated as follows: First, the ratio of the real-time torque value after filtering to the initial reference torque is calculated. Then, the natural logarithm of the ratio is calculated. Subsequently, the natural logarithm is multiplied by a preset rheological sensitivity coefficient to obtain the rheological increment. Finally, the rheological increment is added to the value one to obtain the rheological correction factor. The system sets the minimum value of the rheological correction factor to one. When the real-time torque value is less than the initial reference torque, the real-time torque value is forced to be equal to the initial reference torque in the calculation.

4. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 2, characterized in that, The calculation steps for the cooling rate correction in step S3 include: firstly, calculating the particle size deviation at the current moment, where the particle size deviation is equal to the theoretical value of the ideal particle size growth trajectory at the current moment minus the actual average chord length at the current moment; then, based on the incremental PID control algorithm, using preset proportional coefficients, integral coefficients, and derivative coefficients, and combining the particle size deviations at the current moment, the previous sampling moment, and the previous sampling moment, performing a weighted calculation to obtain the cooling rate correction.

5. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 1, characterized in that, In step S1, the target temperature trajectory set by the basic nonlinear cooling model is a power function curve with respect to the running time. Specifically, the current target temperature is equal to the cooling start temperature minus the temperature decrease term, which is the cooling amplitude coefficient multiplied by the nonlinear exponent of the current running time, and the nonlinear exponent ranges from 1.5 to 2.

5. The basic cooling rate corresponds to the absolute value of the first derivative of the target temperature trajectory with respect to the running time.

6. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 1, characterized in that, It also includes seed crystal addition and nucleation anomaly intervention steps: when the temperature inside the crystallization vessel reaches the preset nucleation temperature, seed crystals are added; after adding seed crystals, the microparticle count rate fed back by the focused beam reflection measurement probe is continuously monitored; if the microparticle count rate shows an exponential increase within a preset time, it is determined that a secondary nucleation burst has occurred, and the system automatically executes a temperature recovery program, introducing a heat medium to raise the temperature inside the crystallization vessel to dissolve excess microparticles until the microparticle count rate returns to stability.

7. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 1, characterized in that, It also includes a process endpoint determination step, which determines the end of the crystallization process when the following three conditions are met simultaneously: Condition 1: The actual temperature inside the crystallization vessel drops to the preset endpoint temperature and is maintained for a preset time; Condition 2: The absolute value of the slope of the real-time torque value of the stirring shaft changing with time is continuously less than 0.01 N·m per minute; Condition 3: The microparticle count rate fed back by the focused beam reflection measurement probe is lower than the preset endpoint threshold, and the peak value of the main particle size reaches the target specification.

8. The programmed cooling crystallization method for controlling the particle size of sugar alcohol crystals according to claim 1, characterized in that, After the crystallization process is completed, a crystal growth step is also included: maintaining the temperature inside the crystallization vessel at the endpoint temperature and reducing the stirring speed to 50% to 60% of the initial speed, and continuing to run for 60 to 120 minutes.

9. A temperature-controlled crystallizing reactor for adjusting the particle size of sugar alcohol crystals, applied to the temperature-controlled crystallization method for adjusting the particle size of sugar alcohol crystals as described in any one of claims 1-8, characterized in that, The system includes: a vessel body with an external heat exchange jacket and an internal heat exchange coil; a composite stirring paddle installed at the center of the vessel body and driven by a stirring motor; a torque sensor connected between the stirring motor and the drive shaft of the composite stirring paddle for acquiring the real-time torque value; a focused beam reflection measurement probe inserted into the vessel body and submerged below the liquid surface for acquiring the actual average chord length; and a central control unit electrically connected to the torque sensor, the focused beam reflection measurement probe, and a regulating valve group controlling the refrigerant flow of the heat exchange jacket and the internal heat exchange coil; the central control unit is configured to receive the real-time torque value and the actual average chord length, and execute calculation steps to output the final cooling rate command.

10. A programmed cooling crystallizer with adjustable sugar alcohol crystal particle size according to claim 9, characterized in that, The composite stirring impeller includes: an outer anchor-type impeller whose rotation trajectory is close to the inner wall of the vessel for scraping off wall crystals; and an inner spiral-type impeller coaxially disposed inside the anchor-type impeller for generating an axial circulating flow field.

Images