A method and system for realizing low-energy consumption continuous production control of polydextrose

Abstract

This invention discloses a method and system for continuous polydextrose production with low energy consumption, relating to the field of production control technology. It involves detecting and removing foreign matter from coarse powder materials, dynamically adjusting polymerization reaction control parameters based on the type and quantity of foreign matter; finely pulverizing and sieving clean materials, adjusting fine pulverization control parameters based on the proportion of material remaining on the sieve; returning the material remaining on the sieve to fine pulverization, and magnetically removing the material passing through the sieve to obtain the finished product. This invention avoids excessive heating during polymerization and coking byproducts through foreign matter feedback, and matches pulverization power to demand through sieving feedback, achieving closed-loop collaborative optimization throughout the entire process. This reduces overall energy consumption and improves resource utilization efficiency while ensuring stable product quality.

Description

Technical Field

[0001] This invention relates to the field of production control technology, specifically to a method and system for achieving low-energy continuous production control of polydextrose. Background Technology

[0002] Polydextrose, as a functional dietary fiber, is widely used in the food, health product, and pharmaceutical fields due to its low calorie content, high water solubility, and ability to regulate intestinal function. The production of polydextrose involves multiple processes, including raw material polymerization, cooling, and pulverization, with energy consumption primarily concentrated in the polymerization and pulverization stages.

[0003] In current production processes, the polymerization stage typically employs an open-loop control mode with fixed temperature and time. Operating parameters rely on pre-set batch formulations, lacking online monitoring and feedback adjustment for the generation of reaction byproducts. To ensure monomer conversion, temperatures are generally set at high levels and maintained for extended periods. This fixed-parameter control method cannot dynamically adjust the heat supply according to the actual reaction progress, resulting in significant energy wastage. Excessively high temperature settings also exacerbate localized overheating, generating coked, hard particles. In the pulverization stage, fixed speed or power operation is commonly used. To ensure the material is pulverized to the required particle size, fine pulverizing equipment is often set to high speed and high power for continuous operation. While this ensures pulverization effectiveness, it easily leads to prolonged high energy consumption, causing a mismatch between the actual energy required for pulverization and the equipment's output energy, resulting in wasted electricity.

[0004] Overall, the unnecessary high temperatures in the polymerization process and the unnecessary high rotation speeds in the pulverization process are the main reasons for the high energy consumption in polydextrose production, which also increases production costs and reduces resource utilization efficiency. Summary of the Invention

[0005] The purpose of this invention is to solve the problem of high energy consumption in production caused by excessive heating during the polymerization stage and excessive power output during the crushing stage, as mentioned in the background art, and to propose a method and system for continuous production control of polydextrose with low energy consumption.

[0006] A first aspect of this invention provides a method for controlling the continuous production of polydextrose with low energy consumption, the method comprising:

[0007] The polymerization reaction of raw and auxiliary materials is controlled according to preset control parameters;

[0008] The molten material obtained from polymerization is cooled to obtain a solidified material.

[0009] The cooled and solidified material is coarsely crushed to obtain coarse powder.

[0010] The coarse powder material is inspected for and removed to obtain clean material; and the control parameters of the polymerization reaction are adjusted according to the results of the foreign matter inspection.

[0011] The clean material is subjected to fine pulverization to obtain fine powder material;

[0012] The fine powder material is sieved and classified to obtain oversize and undersize material; and the control parameters for fine grinding are adjusted according to the sieving results.

[0013] The material on the sieve is returned and finely pulverized together with the clean material.

[0014] The screened material is magnetically attracted to remove metal impurities, thus obtaining qualified material.

[0015] Optionally, a steel strip cooler is used for cooling, and the method further includes:

[0016] The initial temperature of the molten material as it enters the steel strip cooler is acquired in real time.

[0017] The cooling medium supply corresponding to the initial temperature is determined according to a preset temperature supply mapping function or lookup table.

[0018] Optionally, the foreign matter inspection of the coarse powder material includes:

[0019] Acquire visible light and near-infrared images of coarse powder materials;

[0020] The visible light image and the near-infrared band image are registered and stitched together in the channel dimension to form a multi-channel input image;

[0021] The multi-channel input image is input into a pre-trained foreign object recognition model to obtain recognition results; the recognition results include the type of foreign object and its location.

[0022] Optionally, the foreign object recognition model is an improved recognition model based on YOLOv8; specific improvements include:

[0023] The number of input layer channels of the model was increased from 3 to 4 to receive the stitched multi-channel image;

[0024] The C2f module is lightweighted by using partial convolution to obtain the PC2f module; the C2f modules of the backbone network and neck network are replaced with PC2f modules to reduce computational redundancy.

[0025] Add a small target detection head to improve the detection accuracy of small targets;

[0026] A parameterless attention module is inserted at the connection between the neck network and the detection head to enhance the feature response of the foreign object region and suppress background noise.

[0027] Optionally, the C2f module includes at least one bottleneck unit (Bottleneck); the lightweight improvement of the C2f module using partial convolution to obtain the PC2f module includes:

[0028] By replacing the standard convolutions in the bottleneck unit Bottleneck with partial convolutions, a lightweight bottleneck unit PBottleneck is obtained.

[0029] Replace all bottleneck units (Bottleneck) in the C2f module with lightweight bottleneck units (PBottleneck) to obtain the PC2f module.

[0030] Optionally, adjusting the control parameters of the polymerization reactor based on the foreign matter inspection results includes:

[0031] The results of foreign object identification generated during the foreign object inspection process are statistically analyzed to obtain statistical data; the statistical data includes at least the type and quantity of foreign objects.

[0032] When the number of foreign objects exceeds a preset threshold, a first warning signal is generated;

[0033] In response to the first warning signal, the control parameters of the polymerization reactor are adjusted according to the type of foreign matter to suppress overpolymerization or promote the polymerization reaction.

[0034] The control parameters include at least one of the following: the setpoint for the mid-to-late stage of the polymerization reaction temperature curve, the reaction time, and the stirring speed.

[0035] Optionally, the adjustment rules for the control parameters of the polymerization reactor include:

[0036] When the foreign matter is a hard, coking foreign matter produced by over-reaction, implement inhibition regulation: reduce the temperature setpoint in the later stages of the polymerization reaction or shorten the reaction time;

[0037] When the foreign matter is a soft agglomerate resulting from insufficient reaction, implement acceleration adjustment: increase the temperature setpoint in the later stages of the polymerization reaction or extend the reaction time;

[0038] When the hard coking foreign matter and the soft lumps are present at the same time, the polymerization uniformity is determined to be abnormal, and uniformity adjustment is performed: the stirring speed is increased and the current temperature setting is maintained; if the abnormality is not resolved within the preset monitoring period, a second warning signal is sent to the monitoring terminal.

[0039] Optionally, adjusting the control parameters for fine grinding based on the sieving results includes:

[0040] Within a statistical period, the mass of the material on the screen and the total mass of the material entering the fine grinding process within that period are obtained.

[0041] The proportion of the material over the screen to the total mass of the material is calculated as a load index;

[0042] When the load index exceeds the preset range, the parameters of the fine grinding process are adjusted to bring the load index back to the preset range.

[0043] The control parameters for fine grinding include at least one of the following: grinding mill speed, feed rate, and grinding disc gap.

[0044] Optionally, the adjustment rules for the control parameters of the fine grinding include:

[0045] When the load index is higher than the upper limit threshold of the preset range, at least one of the following shall be performed: increase the speed of the crusher, decrease the feed speed, or decrease the grinding disc gap to enhance the crushing effect;

[0046] When the load index is lower than the lower limit threshold of the preset range, at least one of the following shall be executed: reduce the speed of the crusher, increase the feeding speed, or increase the grinding disc gap, so as to reduce the crushing energy consumption.

[0047] A second aspect of this invention provides a continuous polydextrose production control system with low energy consumption, the system comprising:

[0048] The polymerization module is used to control the polymerization reaction of raw and auxiliary materials according to preset control parameters;

[0049] The cooling module is used to cool the molten material obtained from polymerization to obtain a solidified material;

[0050] The coarse crushing module is used to coarsely crush the cooled solidified material to obtain coarse powder.

[0051] The foreign object inspection module is used to inspect and remove foreign objects from the coarse powder material to obtain clean material.

[0052] The fine grinding module is used to finely grind the clean material to obtain fine powder material.

[0053] The sieving module is used to sieve and classify the fine powder material to obtain oversize material and undersize material.

[0054] The rework module is used to return the material on the screen and perform fine crushing treatment together with the clean material.

[0055] The magnetic attraction module is used to magnetically attract the undersize material, remove metal impurities, and obtain qualified material.

[0056] The control and adjustment module is used to adjust the control parameters of the polymerization reaction based on the foreign matter inspection results; and to adjust the control parameters of fine grinding based on the sieving results.

[0057] The beneficial effects of this invention are:

[0058] 1. By adjusting the polymerization reaction parameters after foreign matter inspection, the polymerization process is transformed from fixed parameter open-loop control to dynamic closed-loop control. The polymerization parameters can be adaptively adjusted according to overpolymerization or underreaction, thereby reducing excessive heat supply and local overheating, reducing heat consumption and improving polymerization stability.

[0059] 2. By monitoring the cyclic load index of the fine grinding process online and adaptively adjusting the grinding parameters accordingly, the grinding energy output is precisely matched with the actual needs of the material. This overcomes the excessive energy consumption under the traditional constant high speed and high power operation mode, significantly reduces energy waste in the grinding process, and effectively reduces equipment wear and extends equipment service life.

[0060] 3. By constructing a multi-process collaborative control mechanism, an intelligent closed-loop collaborative optimization system covering the entire polydextrose production process is formed. Under the premise of ensuring stable product quality, the ineffective energy consumption of each link is systematically reduced, and the resource utilization efficiency and production economy are comprehensively improved. Attached Figure Description

[0061] Figure 1 A flowchart illustrating a method for controlling the continuous production of polydextrose with low energy consumption, provided as an embodiment of the present invention;

[0062] Figure 2 A network architecture diagram of a foreign object recognition model provided in an embodiment of the present invention;

[0063] Figure 3 This is a schematic diagram of a PC2f structure provided in an embodiment of the present invention. Detailed Implementation

[0064] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention is provided in conjunction with the accompanying drawings and preferred embodiments.

[0065] This invention provides a method for controlling the continuous production of polydextrose with low energy consumption. See also... Figure 1 , Figure 1 A flowchart illustrating a method for controlling the continuous production of polydextrose with low energy consumption, provided as an embodiment of the present invention. The method includes the following steps:

[0066] S1 controls the polymerization reaction of raw and auxiliary materials according to preset control parameters.

[0067] S2, the molten material obtained from polymerization is cooled to obtain a solidified material.

[0068] S3, the cooled solidified material is coarsely crushed to obtain coarse powder.

[0069] S4, foreign matter is inspected and removed from the coarse powder material to obtain clean material.

[0070] S5, adjust the control parameters of the polymerization reaction based on the results of foreign matter inspection.

[0071] S6, finely pulverizes the clean material to obtain fine powder.

[0072] S7, the fine powder material is screened and classified to obtain the oversize material and the undersize material.

[0073] S8, adjust the control parameters for fine grinding based on the sieving results.

[0074] S9 returns the material on the screen to be finely crushed together with the clean material.

[0075] S10 uses magnetic attraction to remove metal impurities from the screened material, resulting in qualified material.

[0076] Foreign matter includes hard, charred foreign matter produced by excessive reaction and soft lumps produced by insufficient reaction.

[0077] This invention provides a method for controlling the continuous production of polydextrose with low energy consumption. During the polymerization stage, reaction parameters are dynamically adjusted based on foreign matter inspection results to avoid overheating, coking, or incomplete reaction. During the pulverization stage, the fine pulverization speed and power are adaptively controlled based on sieving results, ensuring that equipment output matches actual needs and reducing ineffective energy consumption. This invention significantly reduces overall energy consumption and improves resource utilization efficiency while ensuring stable product quality through multi-stage process optimization.

[0078] In one embodiment, a steel strip cooler is used for cooling. The method further includes:

[0079] The initial temperature of the molten material entering the steel strip cooler is obtained in real time, and the cooling medium supply of the steel strip cooler is controlled based on the feedforward of the temperature.

[0080] In one implementation, a lookup table can be established based on historical cooling data, and the corresponding cooling medium supply can be determined according to the temperature range of the initial temperature. Further, a temperature-supply curve can be fitted based on historical cooling data to obtain a temperature-supply mapping function, and the corresponding cooling medium supply can be determined based on this function.

[0081] This embodiment introduces a data-driven dynamic adjustment method, which can pre-match the cooling intensity according to the actual thermal state of the material entering the cooling section. This avoids the problems of overcooling or undercooling caused by traditional fixed cooling amounts, improves cooling efficiency and temperature uniformity, and reduces the ineffective consumption of cooling medium, thus lowering energy consumption. Furthermore, this method helps stabilize the physical properties of the solidified material, reduces subsequent coarse crushing load fluctuations, and thereby improves the overall stability and energy efficiency of continuous production.

[0082] In one embodiment, machine vision is used for foreign object inspection. Specifically, foreign object inspection of coarse powder materials includes:

[0083] Step 1: Acquire visible light and near-infrared images of the coarse powder material.

[0084] Step two involves registering the visible light image and the near-infrared band image and stitching them together along the channel dimension to form a multi-channel input image.

[0085] Step 3: Input the multi-channel input image into the pre-trained foreign object recognition model to obtain the recognition result; the recognition result includes the type of foreign object and its location.

[0086] In one implementation, the foreign object recognition model can be a deep learning-based target detection model, such as the YOLO series models, Faster R-CNN models, etc.

[0087] This embodiment utilizes visible light to acquire surface morphology information and near-infrared light to enhance the identification capability of material composition differences, thereby improving the differentiation accuracy between coking hard particles and unreacted soft clumps and reducing the probability of false detection and missed detection in single-spectrum detection. By directly outputting the type and location of foreign objects through the model, online, rapid, and non-contact detection can be achieved, providing a highly reliable data foundation for parameter feedback adjustment, thus improving the accuracy of the system's closed-loop control and the overall energy efficiency optimization effect.

[0088] In one embodiment, the foreign object detection model is an improved version of YOLOv8. See also... Figure 2 , Figure 2 This diagram illustrates the network architecture of a foreign object recognition model provided in an embodiment of the present invention. The CBS, SPPF, Upsample, Concat, and Detect modules in the diagram are inherent modules of the YOLOv8 model and will not be elaborated upon here. Specific improvements include: increasing the number of input layer channels from 3 to 4 to receive the stitched multi-channel images; replacing the C2f modules in the backbone and neck networks with lightweight PC2f modules to reduce computational redundancy; and adding a small target detection head (see...). Figure 2The part within the dashed box in the middle is used to improve the detection accuracy of small targets; SimAM (parametric attention module) is inserted at the connection between the neck network and the detection head to enhance the feature response of the foreign object region and suppress background noise.

[0089] In one implementation, PConv (partial convolution) is used to lightweightly improve the C2f module, resulting in the PC2f module. See [link to implementation details]. Figure 3 , Figure 3 This is a schematic diagram of a PC2f structure provided in an embodiment of the present invention. CBS, Split, Concat, and Add in the diagram are inherent modules of the original C2f and will not be described in detail here. Figure 3 (a) shows the original structure of C2f, where the C2f module includes at least one Bottleneck unit. Figure 3 (b) shows the specific structure of Bottleneck. Figure 3 As shown in (d), replacing the standard convolutions in the bottleneck unit Bottleneck with partial convolutions yields the lightweight bottleneck unit PBottleneck. Figure 3 As shown in (c), replacing all bottleneck units in the C2f module with lightweight bottleneck units PBottleneck yields the PC2f module.

[0090] Partial convolution PConv selects only a subset of input feature channels for standard convolution computation, while maintaining identity mapping or lightweight processing for the remaining channels. The convolution result is then fused with the channels not involved in convolution along the channel dimension, achieving a balance between feature computation and information preservation. This embodiment, by embedding partial convolution PConv into the model, significantly reduces the amount of convolution computation and parameter size, lowers model computational complexity and memory usage, while preserving the integrity of original feature information as much as possible. This allows the PC2f module to maintain strong feature representation capabilities while remaining lightweight, thereby improving overall inference efficiency.

[0091] In one implementation, the small target detection head adds a detection branch of a high-resolution feature layer (P2) to the existing multi-scale detection structure. This allows shallow high-resolution features to directly participate in target prediction and performs independent classification and regression for small-scale features. The small target detection head utilizes the fine-grained spatial information retained in the shallow feature map to independently detect small, low-contrast foreign objects, thus compensating for the loss of small target information during downsampling in the original detection head. This significantly improves the detection rate of small coking particles and other foreign objects, reduces the probability of missed detections, and enhances the model's detection stability.

[0092] In one implementation, SimAM is a parameterless attention mechanism that does not introduce additional learnable parameters. Instead, it directly calculates attention weights based on feature statistics to enhance key foreign object regions (such as the edges of coking particles or regions with material differences) while suppressing background noise and irrelevant texture responses. This approach improves feature selection capabilities with almost no increase in model computational overhead, enhances the distinction between foreign object regions and the background, and strengthens the model's robustness and anti-interference ability against complex powder backgrounds.

[0093] The foreign object recognition model proposed in this embodiment achieves multimodal information fusion through input channel expansion, lightweighting of the model through PC2f, enhances the detection capability of small foreign objects through a small target detection head, and enhances the feature representation capability of key regions through SimAM, thereby significantly reducing computational costs while ensuring detection accuracy.

[0094] In one embodiment, step S5, adjusting the control parameters of the polymerization reactor based on the foreign matter inspection results, includes:

[0095] Step 1: Compile statistics on the foreign object identification results generated during the foreign object inspection process; the statistics should include at least the type and quantity of foreign objects.

[0096] Step two: When the number of foreign objects exceeds a preset threshold, a first warning signal is generated;

[0097] Step 3: In response to the first warning signal, adjust the control parameters of the polymerization reactor according to the type of foreign matter to suppress overpolymerization or promote the polymerization reaction;

[0098] The control parameters include at least one of the following: the setpoint for the mid-to-late stage of the polymerization reaction temperature curve, the reaction time, and the stirring speed.

[0099] In one implementation, a preset threshold can be determined based on historical statistics. For example, statistical data on the identification results of each type of foreign object are collected for N consecutive cycles under normal operating conditions, the mean u and standard deviation σ of the number of foreign objects detected are calculated, and the preset threshold is set to u+3σ.

[0100] In one implementation, the adjustment rules for the control parameters of the polymerization reactor include:

[0101] Scenario 1: When the foreign object is a hard, coking foreign object produced by over-reaction, implement inhibition regulation: reduce the temperature setpoint in the later stage of the polymerization reaction or shorten the reaction time.

[0102] Scenario 2: When the foreign object is a soft mass resulting from insufficient reaction, implement acceleration adjustment: increase the temperature setpoint in the later stages of the polymerization reaction or extend the reaction time.

[0103] Scenario 3: When hard coking foreign matter and soft lumps are present at the same time, the polymerization uniformity is determined to be abnormal, and uniformity adjustment is performed: increase the stirring speed and maintain the current temperature set value; if the abnormality is not resolved within the preset monitoring period, a second warning signal is sent to the monitoring terminal to prompt manual intervention.

[0104] In one implementation, specific adjustment rules and corresponding adjustment increments are pre-set in the control system's configuration table, allowing operators to select and fine-tune them according to actual process requirements. For example, in scenario one, the temperature setpoint for the later stages of the polymerization reaction is gradually decreased in 2°C increments, up to a maximum decrease of 5°C, or the polymerization reaction time is shortened by 10 minutes until the amount of foreign matter falls below the threshold. In scenario two, the temperature setpoint for the later stages of the polymerization reaction is gradually increased in 2°C increments, up to a maximum increase of 5°C, or the polymerization reaction time is extended by 10 minutes until the amount of foreign matter falls below the threshold. In scenario three, within the allowable range of stirring power, the stirring speed is first increased by 15%, maintaining the current temperature setpoint, and monitoring is continued for a period of time. If the abnormality persists, a second warning signal is generated and pushed to the monitoring terminal, prompting the operator to check the raw material ratio or equipment status.

[0105] This embodiment achieves dynamic adjustment of polymerization parameters by providing feedback on the type and quantity of foreign matter, thereby reducing the probability of coking byproduct formation, improving polymerization stability, and reducing heating redundancy caused by fixed value settings. This results in production control with lower energy consumption and higher product consistency.

[0106] In one embodiment, step S8, adjusting the control parameters for fine grinding based on the sieving results, includes:

[0107] Step 1: Within a statistical period, obtain the mass of the material on the sieve and the total mass of the material processed by fine grinding within that period.

[0108] Step 2: Calculate the proportion of material on the sieve to the total material mass, which will be used as the circulating load index.

[0109] Step 3: When the circulating load index exceeds the preset range, adjust the parameters of the fine grinding process to bring the circulating load index back to the preset range.

[0110] The control parameters for fine grinding include at least one of the following: grinding mill speed, feed rate, and grinding disc gap.

[0111] In one implementation, during the equipment commissioning phase, multiple sets of experiments are conducted by changing the pulverizer speed, feed rate, and grinding disc clearance to determine the corresponding cyclic load index value and unit product pulverization power consumption under each operating condition. Under the condition that the product pulverization is qualified, the optimal value of the cyclic load index is determined with the lowest unit pulverization power consumption as the optimization objective. Combined with multiple optimal values, the fluctuation range is determined to obtain the optimized cyclic load range for the pulverizer's operation, which serves as the preset range.

[0112] In one implementation, the adjustment rules for the control parameters of fine grinding include:

[0113] Scenario 1: When the cyclic load index is higher than the upper limit threshold of the preset range, perform at least one of the following: increase the pulverizer speed, decrease the feed speed, or decrease the grinding disc gap to enhance the pulverizing effect.

[0114] Scenario 2: When the cyclic load index is lower than the lower limit threshold of the preset range, perform at least one of the following: reduce the speed of the crusher, increase the feeding speed, or increase the grinding disc gap to reduce crushing energy consumption.

[0115] This embodiment dynamically adjusts the fine grinding process parameters based on the feedback of the screening results, thereby achieving adaptive control of the grinding process. This ensures that the grinding intensity matches the actual needs, guaranteeing that the particle size meets the standards while reducing energy consumption and equipment wear, and avoiding the energy waste caused by traditional fixed high-speed or high-power operation.

[0116] In one implementation, to more clearly demonstrate the adaptive adjustment process of the fine grinding control parameters, an application scenario is used as an example. The preset optimization range for the cyclic load index is determined to be [12%, 20%] based on historical data. When the cyclic load index continuously exceeds this range for two consecutive statistical periods, adjustment is triggered:

[0117] Scenario 1: If the circulating load index is greater than 20%, it indicates insufficient crushing intensity and excessive material on the screen. Preferred actions: Increase the crusher speed by 5% (e.g., from 3000 rpm to 3150 rpm), or decrease the feed speed by 5% (e.g., from 100 kg / h to 95 kg / h), or decrease the grinding disc clearance by 0.05 mm. In practical applications, one or more combinations can be selected based on equipment conditions.

[0118] Scenario 2: If the circulating load index is less than 12%, it indicates over-grinding and energy waste. Preferred actions: Reduce the grinder speed by 5%, increase the feed speed by 5%, or increase the grinding disc clearance by 0.05mm to reduce unnecessary energy consumption.

[0119] After adjustment, continuously monitor the next cycle. If the cyclic load index returns to the preset range, maintain the current parameters; if it still exceeds the limit, continue the next round of adjustment according to the above step size until the index stabilizes within the preset range. The limit adjustment range can be set to ±20% of the upper and lower speed limits to prevent the equipment from deviating from safe operating conditions.

[0120] This invention provides a low-energy-consumption continuous production control system for polydextrose. The system includes:

[0121] The polymerization module is used to control the polymerization reaction of raw and auxiliary materials according to preset control parameters.

[0122] The cooling module is used to cool the molten material obtained from polymerization to obtain solidified material.

[0123] The coarse crushing module is used to coarsely crush the cooled and solidified material to obtain coarse powder.

[0124] The foreign object inspection module is used to inspect and remove foreign objects from coarse powder materials to obtain clean materials.

[0125] The fine grinding module is used to finely grind clean materials to obtain fine powder.

[0126] The sieving module is used to sieve and classify fine powder materials to obtain oversize and undersize materials.

[0127] The rework module is used to return the material on the screen to be finely crushed together with the clean material.

[0128] The magnetic attraction module is used to magnetically attract the undersize material, remove metal impurities, and obtain qualified material.

[0129] The control and adjustment module is used to adjust the control parameters of the polymerization reaction based on the foreign matter inspection results; and to adjust the control parameters of fine grinding based on the sieving results.

[0130] This invention provides a low-energy continuous production control system for polydextrose. During the polymerization stage, reaction parameters are dynamically adjusted based on foreign matter inspection results to avoid overheating, coking, or incomplete reaction. During the pulverization stage, the fine pulverization speed and power are adaptively controlled based on sieving results, ensuring that equipment output matches actual needs and reducing ineffective energy consumption. This invention significantly reduces overall energy consumption and improves resource utilization efficiency while maintaining stable product quality through multi-stage processing and synergistic optimization.

[0131] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention are within the scope of the claims of the present invention.

Claims

1. A method for controlling the continuous production of polydextrose with low energy consumption, characterized in that, The method includes: The polymerization reaction of raw and auxiliary materials is controlled according to preset control parameters; The molten material obtained from polymerization is cooled to obtain a solidified material. The cooled and solidified material is coarsely crushed to obtain coarse powder. The coarse powder material is inspected for and removed to obtain clean material; and the control parameters of the polymerization reaction are adjusted according to the results of the foreign matter inspection. The clean material is subjected to fine pulverization to obtain fine powder material; The fine powder material is sieved and classified to obtain oversize and undersize material; and the control parameters for fine grinding are adjusted according to the sieving results. The material on the sieve is returned and finely pulverized together with the clean material. The screened material is magnetically attracted to remove metal impurities, thus obtaining qualified material.

2. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 1, characterized in that, The method further includes using a steel strip cooler for cooling treatment. The initial temperature of the molten material as it enters the steel strip cooler is acquired in real time. The cooling medium supply corresponding to the initial temperature is determined according to a preset temperature supply mapping function or lookup table.

3. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 1, characterized in that, The foreign matter inspection of the coarse powder material includes: Acquire visible light and near-infrared images of coarse powder materials; The visible light image and the near-infrared band image are registered and stitched together in the channel dimension to form a multi-channel input image; The multi-channel input image is input into a pre-trained foreign object recognition model to obtain recognition results; the recognition results include the type of foreign object and its location.

4. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 3, characterized in that, The foreign object recognition model is an improved version of YOLOv8; the specific improvements include: The number of input layer channels of the model was increased from 3 to 4 to receive the stitched multi-channel image; The C2f module is lightweighted by using partial convolution to obtain the PC2f module; the C2f modules of the backbone network and neck network are replaced with PC2f modules to reduce computational redundancy. Add a small target detection head to improve the detection accuracy of small targets; A parameterless attention module is inserted at the connection between the neck network and the detection head to enhance the feature response of the foreign object region and suppress background noise.

5. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 4, characterized in that, The C2f module includes at least one bottleneck unit (Bottleneck); the lightweight improvement of the C2f module using partial convolution to obtain the PC2f module includes: By replacing the standard convolutions in the bottleneck unit Bottleneck with partial convolutions, a lightweight bottleneck unit PBottleneck is obtained. Replace all bottleneck units (Bottleneck) in the C2f module with lightweight bottleneck units (PBottleneck) to obtain the PC2f module.

6. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 1, characterized in that, The adjustment of the control parameters of the polymerization reactor based on the foreign matter inspection results includes: The results of foreign object identification generated during the foreign object inspection process are statistically analyzed to obtain statistical data; the statistical data includes at least the type and quantity of foreign objects. When the number of foreign objects exceeds a preset threshold, a first warning signal is generated; In response to the first warning signal, the control parameters of the polymerization reactor are adjusted according to the type of foreign matter to suppress overpolymerization or promote the polymerization reaction. The control parameters include at least one of the following: the setpoint for the mid-to-late stage of the polymerization reaction temperature curve, the reaction time, and the stirring speed.

7. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 6, characterized in that, The adjustment rules for the control parameters of the polymerization reactor include: When the foreign matter is a hard, coking foreign matter produced by over-reaction, implement inhibition regulation: reduce the temperature setpoint in the later stages of the polymerization reaction or shorten the reaction time; When the foreign matter is a soft agglomerate resulting from insufficient reaction, implement acceleration adjustment: increase the temperature setpoint in the later stages of the polymerization reaction or extend the reaction time; When the hard coking foreign matter and the soft lumps are present at the same time, the polymerization uniformity is determined to be abnormal, and uniformity adjustment is performed: the stirring speed is increased and the current temperature setting is maintained; if the abnormality is not resolved within the preset monitoring period, a second warning signal is sent to the monitoring terminal.

8. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 1, characterized in that, The adjustment of the control parameters for fine grinding based on the sieving results includes: Within a statistical period, the mass of the material on the screen and the total mass of the finely ground material within that period are obtained. The proportion of the material over the screen to the total mass of the material is calculated and used as a circulating load index. When the circulating load index exceeds the preset range, the parameters of the fine grinding process are adjusted to bring the circulating load index back to the preset range. The control parameters for fine grinding include at least one of the following: grinding mill speed, feed rate, and grinding disc gap.

9. The method for controlling the continuous production of polydextrose with low energy consumption according to claim 8, characterized in that, The adjustment rules for the control parameters of the fine grinding include: When the cyclic load index is higher than the upper limit threshold of the preset range, at least one of the following shall be performed: increase the speed of the crusher, decrease the feed speed, or decrease the grinding disc gap to enhance the crushing effect; When the cyclic load index is lower than the lower limit threshold of the preset range, at least one of the following shall be executed: reduce the speed of the crusher, increase the feeding speed, or increase the grinding disc gap to reduce the crushing energy consumption.

10. A continuous production control system for polydextrose with low energy consumption, characterized in that, The system includes: The polymerization module is used to control the polymerization reaction of raw and auxiliary materials according to preset control parameters; The cooling module is used to cool the molten material obtained from polymerization to obtain a solidified material; The coarse crushing module is used to coarsely crush the cooled solidified material to obtain coarse powder. The foreign object inspection module is used to inspect and remove foreign objects from the coarse powder material to obtain clean material. The fine grinding module is used to finely grind the clean material to obtain fine powder material. The sieving module is used to sieve and classify the fine powder material to obtain oversize material and undersize material. The rework module is used to return the material on the screen and perform fine crushing treatment together with the clean material. The magnetic attraction module is used to magnetically attract the undersize material, remove metal impurities, and obtain qualified material. The control and adjustment module is used to adjust the control parameters of the polymerization reaction based on the foreign matter inspection results; and to adjust the control parameters of fine grinding based on the sieving results.

Images