Automatic cocoa grinding control system and method

Abstract

The invention relates to the technical field of cocoa grinding, in particular to an automatic cocoa grinding control system and method. The method comprises the steps that a cocoa initial state set and a machine default state set are obtained, cocoa particle state changes are tracked based on the cocoa initial state set and the machine default state set, and a grinding transition starting time node is determined; according to the grinding transition starting time node, a temperature change performance parameter set in the grinding transition stage is tracked; acquiring a cocoa flavor target data set, analyzing cocoa target flavor temperature interval requirements in a grinding transition stage based on the cocoa flavor target data set and the temperature change expression parameter set, predicting cocoa temperature change characteristics, and determining a machine dynamic control strategy; and according to the machine dynamic control strategy, the grinding machine is controlled to conduct corresponding dynamic adjustment, and an adjustment report is output. In the cocoa grinding process, the flavor stability is improved through inertial temperature rise, and the grinding energy consumption and efficiency are optimized.

Description

Technical Field

[0001] The present application relates to the technical field of cocoa grinding, and in particular to an automated cocoa grinding control system and method. Background Art

[0002] Cocoa grinding is the core process of chocolate and high-end food manufacturing. Its quality directly determines the product flavor, taste and market value, affecting the taste experience of billions of consumers around the world and the economic benefits of the industry.

[0003] However, existing grinding control mainly relies on preset fixed parameters or manual experience adjustment, which cannot adapt to the state changes of cocoa during the grinding process, and easily leads to irreversible loss of cocoa flavor, which not only reduces the product qualification rate, but also causes raw material waste and rising production costs. Summary of the Invention

[0004] The present application provides an automated cocoa grinding control system and method to solve the above technical problems.

[0005] In a first aspect, the present application provides an automated cocoa grinding control method, the method comprising: Obtain a cocoa initial state set and a machine default state set; based on the cocoa initial state set and the machine default state set, track cocoa particle state changes and determine a grinding transition start time node; track a temperature change performance parameter set during the grinding transition stage based on the grinding transition start time node; obtain a cocoa flavor target data set; based on the cocoa flavor target data set and the temperature change performance parameter set, analyze the cocoa target flavor temperature range requirements during the grinding transition stage, predict cocoa temperature change characteristics, and determine a machine dynamic control strategy; control the grinding machine to make corresponding dynamic adjustments based on the machine dynamic control strategy, and output an adjustment report.

[0006] This solution simultaneously collects raw material characteristics and the initial state of the equipment, monitors changes in the ground particles in real time, and automatically determines the entry into the fine grinding stage when a rapid decrease in particle size is detected. Secondly, the equipment temperature trend and material temperature fluctuations are continuously tracked during the fine grinding process. Furthermore, based on the target flavor standard and combined with temperature changes, the flavor formation path is predicted and the optimal equipment adjustment parameters are calculated. Finally, the adjustment instructions are transmitted in real time to the execution unit to dynamically control the equipment's operating status, while the temperature trajectory and flavor index deviations throughout the entire process are recorded to form a traceable process report. This process ensures that the grinding process always remains within the ideal reaction temperature range, thereby achieving the core benefits of increased flavor retention and significantly reduced unit energy consumption. It also has the ability to adapt to different raw material characteristics and ensure quality consistency.

[0007] Optionally, tracking the state changes of cocoa particles based on the cocoa initial state set and the machine default state set to determine the grinding transition start time node includes: The initial cocoa state set includes initial particle size and initial moisture content; Tracking the change in particle size ratio and fat precipitation status through the initial particle size and the machine default state set; Matching the particle size change with a preset transition starting particle ratio to determine a critical time interval for phase transition; generating an indirect determination factor based on the initial moisture content and the fat precipitation state; Based on the phase transition key time interval and according to the indirect determination factor, a phase transition start time point is determined, and the phase transition start time point is used as the grinding transition start time node.

[0008] Through this solution, based on the key time interval of stage conversion and the indirect judgment factor, the starting time point of stage conversion is determined, and the starting time point of stage conversion is used as the starting time node of grinding transition. By accurately capturing the synchronization point of cell wall rupture and fat release, the dissolution rate of target cocoa flavor substances is improved, and the degradation of flavor precursors due to overheating or insufficient release is avoided.

[0009] Optionally, generating an indirect determination factor based on the initial moisture content and the fat precipitation state includes: Tracking and determining real-time moisture content based on the initial moisture content and the machine default state set; determining a real-time fat precipitation amount according to the fat precipitation state; determining a time correction amount according to the real-time moisture content and the real-time fat precipitation amount; The real-time moisture content is positively correlated with the relative position of the phase transition start time point within the phase transition critical time interval, that is, the higher the real-time moisture content, the closer the phase transition start time point is to the supremum of the phase transition critical time interval; The real-time fat precipitation amount is negatively correlated with the relative position of the stage transition start time point within the stage transition critical time interval, that is, the lower the real-time fat precipitation amount, the closer the stage transition start time point is to the infimum of the stage transition critical time interval; The time correction amount is used as the indirect determination factor.

[0010] Through this solution, based on the dynamic correction of dual physical property parameters, the systematic deviation of the single particle ratio determination is eliminated, and the matching degree between the starting time point of the grinding transition and the actual physical state evolution is improved; insufficient release of flavor substances due to premature conversion or burnt smell caused by late conversion is avoided, thus ensuring the stability of the target flavor and guaranteeing flavor consistency.

[0011] Optionally, tracking the temperature change performance parameter set during the grinding transition stage according to the grinding transition start time node includes: According to a preset monitoring unit time period, the grinding transition stage is divided into a plurality of monitoring sub-intervals with the grinding transition start time node as a starting reference; According to the monitoring subintervals, obtaining grinding temperature data within each monitoring subinterval and extracting temperature fluctuation characteristics; Based on the temperature fluctuation characteristics, a temperature feature subset is screened to generate the temperature change performance parameter set.

[0012] Through this solution, the monitoring sub-intervals are monitored and the temperature fluctuation characteristics are extracted to increase the monitoring frequency, thereby accelerating the response speed and effectively avoiding the decomposition of flavor substances caused by instantaneous overheating. At the same time, feature screening reduces the amount of input data, improving the efficiency of control strategy generation while avoiding control command oscillation caused by overfitting. By focusing on key fluctuation characteristics to predict the inertial temperature rise trend, the deceleration command is triggered in advance to enhance the robustness of flavor control and ensure that the temperature is strictly within the target flavor range.

[0013] Optionally, the cocoa initial state set further includes cocoa heat capacity, and the machine default state set further includes machine heat capacity. The acquiring of the cocoa flavor target data set, analyzing the cocoa target flavor temperature range requirements during the grinding transition stage based on the cocoa flavor target data set and the temperature change performance parameter set, predicting cocoa temperature change characteristics, and determining a machine dynamic control strategy include: The cocoa flavor target data set includes floral and fruity aroma target data, nutty flavor target data, and caramel flavor target data; obtaining grinding requirement information, and extracting target flavor data corresponding to the grinding requirement information from a cocoa flavor target data set based on the grinding requirement information; Determining a flavor target temperature difference based on the temperature change performance parameter set and the temperature requirement of the target flavor data; Based on the temperature change performance parameter set and in combination with the cocoa heat capacity and the machine heat capacity, a dynamic temperature rise inertia prediction interval is generated; Acquiring the ambient temperature, analyzing the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature, and determining the inertia temperature rise peak value; Based on the inertia temperature rise range and the peak value of inertia temperature rise, the actual temperature rise inertia is predicted and the actual inertia temperature rise time series data is determined; Analyzing the actual inertia temperature rise time series data according to the inertia temperature rise peak value to determine the inertia temperature rise peak time and peak temperature duration; Based on the flavor target temperature difference, the inertia temperature rise peak value, the inertia temperature rise peak time, and the peak temperature duration, determining the safety shutdown / speed reduction recovery extension time; A dynamic control strategy for the machine is constructed based on the safety stop / speed reduction recovery extension time.

[0014] Through this solution, based on the prediction of flavor-sensitive intervals and thermal inertia peak values, the safe intervention time is accurately calculated and executed, achieving proactive intervention in advance and avoiding irreversible flavor deterioration caused by heat penetrating into cocoa butter. The heat capacity differences of cocoa beans from different origins are quantified and the prediction interval is dynamically corrected to effectively adapt to changes in material properties and solve the problem of excessive temperature rise after shutdown caused by ignoring thermal inertia in fixed parameter models. Based on real-time moisture content and ambient temperature, the dynamic temperature rise inertia prediction interval is analyzed to determine the inertia temperature rise peak value. Dual dynamic calibration of the inertia temperature rise peak value using ambient temperature and real-time moisture is introduced to compensate for the superimposed influence of temperature and humidity fluctuations on the grinding thermal effect, ensuring flavor consistency between batches and ultimately achieving precise and stable control of cocoa flavor.

[0015] Optionally, the step of generating a dynamic temperature rise inertia prediction interval based on the temperature change performance parameter set and combining the cocoa heat capacity and the machine heat capacity includes: Extracting the basic temperature rise rate and the current temperature according to the temperature change performance parameter set; generating a heat transfer efficiency factor based on the cocoa heat capacity and the machine heat capacity according to an inverse heat transfer algorithm; Based on the heat conduction efficiency factor, the basic temperature rise rate is corrected to determine the inertial temperature rise rate; Based on the current temperature and the inertia temperature rise rate, an inertia temperature rise simulation is performed by a Kalman filter algorithm to determine the dynamic temperature rise inertia prediction interval.

[0016] Through this solution, the basic temperature rise rate is corrected based on the heat conduction efficiency factor to determine the inertial temperature rise rate. Based on the inertial temperature rise rate, the Kalman filter algorithm is used to simulate the inertial temperature rise and determine the dynamic temperature rise inertia prediction range. A dynamic temperature rise prediction model that adapts to complex grinding environments is constructed, filling the technical gap in temperature inertia control. A safety threshold is also set based on the supremum of the prediction range to ensure that the temperature is always below the target flavor degradation critical point, reducing the risk of flavor loss.

[0017] Optionally, analyzing the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature to determine the inertia temperature rise peak value includes: The ambient temperature is positively correlated with the inertia temperature rise peak value. The higher the ambient temperature, the greater the increase in the supremum of the dynamic temperature rise inertia prediction interval. The real-time moisture content is negatively correlated with the inertia temperature rise peak value. The lower the real-time moisture content, the smaller the increase in the lower bound of the dynamic temperature rise inertia prediction interval caused by the real-time moisture content. Determining a first correction coefficient based on a positive correlation between the ambient temperature and the inertia temperature rise peak value; determining a second correction coefficient based on a negative correlation between the real-time moisture content and the inertia temperature rise peak value; Dynamically calibrating the inertia temperature rise range according to the first correction coefficient and the second correction coefficient to generate a corrected actual inertia temperature rise range; The upper limit of the corrected actual inertia temperature rise range is taken as the inertia temperature rise peak value.

[0018] Through this solution, the inertial temperature rise range is dynamically calibrated through the first correction coefficient and the second correction coefficient, generating the corrected actual inertial temperature rise range, thereby improving the temperature rise prediction accuracy; the temperature rise peak is precisely controlled within the target flavor temperature range, improving the retention rate of key volatile substances, and significantly improving the flavor consistency between batches; through preventive peak control, the frequency of emergency shutdowns is reduced, the thermal stress damage of equipment is reduced, the service life of grinding parts is extended, and production losses are reduced.

[0019] Optionally, controlling the machine to perform corresponding dynamic adjustments according to the machine dynamic control strategy and outputting an adjustment report includes: Based on the inertia temperature rise peak value and the safety shutdown / speed reduction recovery extension time, the machine shutdown time / speed reduction power is set; Acquiring the real-time overall temperature during the inertial temperature rise process, and when the real-time overall temperature reaches the flavor target temperature, tracking the subsequent deviation between the real-time overall temperature and the flavor target temperature; According to the deviation amplitude and a preset temperature deviation amplitude reference value, determining whether the deviation amplitude exceeds the preset temperature deviation amplitude reference value; If the deviation exceeds the preset temperature deviation reference value, dynamically increase / decrease the deceleration power according to the difference between the deviation and the preset temperature deviation reference value to determine the adjusted deceleration power; Dynamically adjusting the machine operating state according to the machine downtime / the adjusted speed reduction power to determine the adjusted machine operating state; According to the adjusted machine operating status, the corresponding adjustment report is output.

[0020] Through this solution, the speed reduction power is dynamically increased / decreased according to the difference between the deviation amplitude and the preset temperature deviation amplitude reference value, and the adjusted speed reduction power is determined. Combined with the safety shutdown / speed reduction recovery extension time, a dynamic buffer mechanism is established to avoid residual heat impacting the flavor critical point. At the same time, based on the closed-loop feedback of the deviation amplitude between the real-time overall temperature and the flavor target temperature, the speed reduction power is dynamically corrected, realizing the transition from open-loop preset to adaptive control, and completely resolving the core contradiction between control lag and response rigidity in traditional processes.

[0021] Optionally, the method further includes: Acquire historical grinding experimental data, analyze the mapping relationship between feed rate and temperature change based on the historical grinding experimental data, and determine the rate of influence of the feed process on the reference temperature in the inertial temperature rise stage; determining an inertia compensation temperature difference between the current temperature and a preset temperature based on the reference temperature impact rate and the inertia temperature rise peak value; According to the inertia compensation temperature difference, the feed rate and feed duration required in the inertia temperature rise stage are analyzed to construct a dynamic feeding strategy; According to the dynamic feeding strategy, the feeding state of the machine is adjusted, and the corresponding feeding state is written into the adjustment report.

[0022] Through this solution, the dynamic feeding strategy allows for the most efficient feeding in the low-temperature stage, and only the minimum necessary reduction is required when approaching the critical point, thereby shortening the single-batch grinding time and reducing grinding energy consumption; through the feedforward compensation mechanism, the maximum temperature deviation in the inertial temperature rise stage is controlled within an extremely small range, avoiding the decomposition of floral and fruity aroma substances or overload of caramel flavor due to overshoot, ensuring the accurate presentation of the target flavor.

[0023] In a second aspect, the present application provides an automated cocoa grinding control system, the system comprising: a time node module, configured to obtain a cocoa initial state set and a machine default state set, track cocoa particle state changes based on the cocoa initial state set and the machine default state set, and determine a grinding transition start time node; A temperature change module, configured to track a temperature change performance parameter set during a grinding transition phase according to the grinding transition start time node; a control strategy module for acquiring a cocoa flavor target data set, analyzing the cocoa flavor target temperature range requirements during the grinding transition phase based on the cocoa flavor target data set and the temperature variation performance parameter set, predicting cocoa temperature variation characteristics, and determining a machine dynamic control strategy; The execution output module is used to control the grinding machine to perform corresponding dynamic adjustments according to the machine dynamic control strategy and output an adjustment report. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0025] Figure 1 A schematic diagram of an application scenario provided in one embodiment of the present application; Figure 2 A flowchart of an automated cocoa grinding control method provided in one embodiment of the present application; Figure 3 A schematic structural diagram of an automated cocoa grinding control system provided in one embodiment of the present application; DETAILED DESCRIPTION To make the purpose, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0026] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document, unless otherwise specified, generally indicates an "or" relationship between the related objects.

[0027] The embodiments of the present application are described in further detail below with reference to the accompanying drawings.

[0028] Existing grinding control mainly relies on preset fixed parameters or manual experience adjustment, which cannot adapt to the changes in the state of cocoa during the grinding process, and easily leads to irreversible loss of cocoa flavor, which not only reduces the product qualification rate, but also causes raw material waste and rising production costs.

[0029] Based on this, the present application provides an automated cocoa grinding control system and method. The raw material characteristics and the initial status of the equipment are collected simultaneously, and the changes in the grinding particles are monitored in real time. When a rapid decrease in particle size is detected, it is automatically determined to enter the fine grinding stage; secondly, the temperature change trend of the equipment and the temperature fluctuation of the material are continuously tracked during the fine grinding process; then, according to the target flavor standard, the flavor formation path is predicted in combination with the temperature change, and the optimal equipment adjustment parameters are calculated; finally, the adjustment instructions are transmitted to the execution unit in real time to dynamically control the operation status of the equipment, and the temperature trajectory and flavor index deviation of the entire process are recorded at the same time to form a traceable process report. This process enables the grinding process to always maintain the ideal reaction temperature range, thereby achieving the core benefits of improved flavor substance retention and significantly reduced unit energy consumption, and has the ability to adapt to different raw material characteristics and ensure quality consistency.

[0030] Figure 1 This is a schematic diagram of an application scenario provided by this application. In the cocoa grinding process, the method provided by this application achieves the core benefits of increased flavor retention and significantly reduced unit energy consumption. It also has the ability to adapt to different raw material characteristics and ensure quality consistency.

[0031] Specifically, the method of the present application is applied to any server, which communicates with the raw material incoming quality inspection system and the equipment PLC system, and obtains the cocoa initial state set and cocoa flavor target data set provided by the raw material incoming quality inspection system and the temperature change performance parameter set provided by the equipment PLC system through the server, synchronously collects the raw material characteristics and the initial state of the equipment, monitors the changes in the grinding particles in real time, and automatically determines to enter the fine grinding stage when it is detected that the particle size is decreasing rapidly; secondly, continuously tracks the equipment temperature change trend and material temperature fluctuation during the fine grinding process; then, based on the target flavor standard, the flavor formation path is predicted in combination with the temperature change, and the optimal equipment adjustment parameters are calculated; finally, the adjustment instructions are transmitted to the execution unit in real time to dynamically control the equipment operation status, and the temperature trajectory and flavor index deviation of the entire process are recorded, and the resulting traceable process report is provided to the equipment management personnel.

[0032] For specific implementation methods, please refer to the following embodiments.

[0033] Figure 2 This is a flow chart of an automated cocoa grinding control method provided in one embodiment of the present application. The method of this embodiment can be applied to the server in the above scenario. Figure 2 As shown, the method includes: S201: Obtain a cocoa initial state set and a machine default state set, track cocoa particle state changes based on the cocoa initial state set and the machine default state set, and determine a grinding transition start time node.

[0034] The initial state set of cocoa may include the initial moisture content, particle size, and initial distribution of cocoa beans, and the data comes from the raw material incoming quality inspection system.

[0035] The default state set of the machine can be the initial speed of the grinder, the roller gap, the basic temperature control threshold, and the heat capacity, and the data comes from the equipment PLC system.

[0036] The grinding transition start time node can be a critical time point when the cocoa particles switch from the coarse grinding stage to the fine grinding stage, which is triggered by real-time monitoring when the particle size change rate reaches a preset threshold.

[0037] Specifically, traditional cocoa grinding relies on fixed timelines or manual experience to switch between grinding stages. This is unable to account for changes in the initial state of the cocoa beans and lacks precise control over the state of the cocoa particles. This leads to uncontrolled state transfer between the various grinding stages (coarse and fine grinding). For example, if the cocoa beans do not remain in the coarse grinding stage sufficiently, large particles entering the fine grinding stage will suddenly heat up and damage the flavor. This step tracks changes in particle size and fat extraction based on initial particle size and moisture content, combined with machine status. First, a key time period is identified by matching particle size changes to a preset ratio. A time correction is then calculated based on the real-time moisture content (higher values ​​are closer to the end of the time period) and fat extraction (lower values ​​are closer to the beginning of the time period). This correction is used to precisely determine the starting time within the critical time period. This ensures sufficient particle breakage, avoids localized overheating caused by hard particles entering the fine grinding stage, and lays a stable foundation for subsequent flavor control.

[0038] S202 : Tracking a temperature change performance parameter set during the grinding transition stage according to the grinding transition start time node.

[0039] The temperature change performance parameter set may include temperature rise rate, maximum regional temperature difference, and temperature stability coefficient, and the data comes from the embedded temperature control system of the grinding machine.

[0040] The grinding transition stage may be the complete grinding process interval starting from the grinding transition start time node until the target particle fineness is achieved.

[0041] Specifically, relying solely on a static topology table is insufficient to handle sudden temperature anomalies during grinding. Traditional methods can only be adjusted after the fact, and cannot predict how the temperature anomaly will spread along the grinding path. For example, a "hotspot" in the coarse grinding stage may quickly transmit to the sensitive fine grinding stage, causing large-scale flavor denaturation. This step begins at a determined starting time point and divides the grinding transition stage into multiple small time periods at fixed time intervals. Temperature data is collected in each small time period, and its fluctuation characteristics are analyzed. The important parts are then screened from these fluctuation characteristics and combined into a set of key parameters that can represent the temperature changes. This method accurately grasps the temperature changes during the transition stage and provides a more accurate temperature adjustment strategy for the subsequent preset cocoa target flavor.

[0042] S203: Obtain a cocoa flavor target dataset, analyze the cocoa flavor target temperature range requirements during the grinding transition phase based on the cocoa flavor target dataset and the temperature variation performance parameter set, predict cocoa temperature variation characteristics, and determine a machine dynamic control strategy.

[0043] The cocoa flavor target dataset can be the target flavor type of cocoa grinding, which includes the target acidity value, bitterness intensity, and aroma retention rate. The data comes from the raw material incoming quality inspection system.

[0044] Specifically, traditional cocoa grinding processes suffer from systemic flaws. The core issue lies in the excessive focus on fineness control, while completely ignoring the critical thermodynamic role of the transition from coarse to fine grinding. This not only mistakenly treats the transient temperature rise caused by crushing energy conversion and the inertial temperature rise after shutdown as purely interference sources, but also passively suppresses flavor formation pathways such as the Maillard reaction through forced cooling, leading to uncontrolled changes in flavor formation pathways. To address these issues, this step example provides a flavor control method based on parameter manipulation during the coarse-to-fine grinding transition. First, a mapping relationship is established between the target flavor (e.g., burnt bitterness intensity) and the transition temperature threshold. Real-time monitoring of material fineness and temperature changes accurately identifies the transition starting point. Upon entering the transition phase, dynamic power reduction or staged shutdown is implemented, leveraging the material's inertial temperature rise to raise the temperature to the target range, activating the production of burnt bitter flavor compounds while reducing inefficient energy consumption. This method transforms negative effects such as inertial temperature rise into controllable process variables, enabling the targeted output of specific flavors such as burnt bitterness. This improves sensory evaluation consistency, reduces energy consumption during the transition phase, and maintains a highly stable fineness acceptance rate.

[0045] S204: According to the machine dynamic control strategy, control the grinding machine to perform corresponding dynamic adjustments and output an adjustment report.

[0046] The adjustment report may include equipment parameter execution records (timestamp, actual speed, roller gap value), flavor index achievement (measured / target ratio), and energy consumption statistics (kWh / kg).

[0047] Specifically, traditional monitoring technology suffers from slow and imprecise adjustments, unable to match the complex outputs of dynamic networks. If, after identifying the risk transmission path and extent, targeted control instructions are not immediately generated and executed, the prediction becomes meaningless. Based on the analysis of peak inertial temperature rise, the system directly predicts and adjusts machine control parameters in real time. Using data visualization technology, it generates corresponding adjustment reports and provides them to equipment managers via human-computer interaction devices, such as the high-definition display screens on the grinding equipment, enabling them to accurately understand the real-time status of the grinding equipment. This report, used for process optimization, clearly documents events and responses, ensuring precise achievement of flavor targets.

[0048] This solution simultaneously collects raw material characteristics and the initial state of the equipment, monitors changes in the ground particles in real time, and automatically determines the entry into the fine grinding stage when a rapid decrease in particle size is detected. Secondly, the equipment temperature trend and material temperature fluctuations are continuously tracked during the fine grinding process. Furthermore, based on the target flavor standard and combined with temperature changes, the flavor formation path is predicted and the optimal equipment adjustment parameters are calculated. Finally, the adjustment instructions are transmitted in real time to the execution unit to dynamically control the equipment's operating status, while the temperature trajectory and flavor index deviations throughout the entire process are recorded to form a traceable process report. This process ensures that the grinding process always remains within the ideal reaction temperature range, thereby achieving the core benefits of increased flavor retention and significantly reduced unit energy consumption. It also has the ability to adapt to different raw material characteristics and ensure quality consistency.

[0049] In some embodiments, the initial state set of cocoa includes initial particle size and initial moisture content; the changes in particle coarseness ratio and fat precipitation status are tracked through the initial particle size and the machine default state set; the changes in particle coarseness are matched with the preset transition starting particle ratio to determine the key time interval for stage transition; based on the initial moisture content and fat precipitation status, an indirect determination factor is generated; based on the key time interval for stage transition, the starting time point of stage transition is determined according to the indirect determination factor, and the starting time point of stage transition is used as the starting time node of grinding transition.

[0050] The initial particle size can be the average particle size of the cocoa beans before grinding, reflecting the physical state of the raw materials.

[0051] The initial moisture content can be the moisture content of the cocoa beans, affecting the adhesion and fat release efficiency during grinding.

[0052] The change in the particle size ratio can describe the dynamic change in the mass ratio of coarse particles to fine particles during the grinding process.

[0053] The fat precipitation state can characterize the degree of release of cocoa butter from the cell wall and directly affect the dissolution efficiency of flavor substances.

[0054] The preset transition starting particle ratio can be an empirically determined critical ratio of coarse to fine particles, marking the timing of the transition to the grinding stage.

[0055] The critical time interval for stage transition can be the potential transition period from coarse grinding to fine grinding, which is delineated by the particle ratio change trend prediction model.

[0056] The indirect determination factor can be a correction parameter generated by comprehensively combining the moisture content and the fat precipitation state, which is used to calibrate the time node.

[0057] The starting time point of the stage transition can be the optimal switching opportunity of the grinding stages, which is dynamically determined by matching the particle ratio and the indirect determination factor.

[0058] Specifically, traditional cocoa grinding control has two major defects: First, process deviation caused by static thresholds: if the grinding stage is switched based solely on a preset fixed time threshold, it cannot adapt to the differences in physical properties of cocoa beans from different batches. For example, cocoa beans with a moisture content greater than a certain proportion have increased viscosity, and the coarse grinding time needs to be extended to fully break the cell walls. Premature switching to fine grinding will cause fat to be wrapped in unbroken cells, reducing the flavor extraction rate; second, the coupling relationship between key state parameters is ignored: when the particle fineness is insufficient, even if the grinding time is extended, the fat cannot be effectively released. When the moisture content is too high, the fat precipitation rate decreases nonlinearly with the increase in fineness. In response to the above problems and needs, this step first defines the state parameter tracking and interval: the initial particle size and initial moisture content are obtained in real time. Based on the default state set of the machine, a high-temperature resistant industrial camera is installed on the side wall of the grinding chamber to continuously monitor the changes in the particle coarseness and fineness ratio. When the proportion of coarse particles drops to the preset transition starting particle ratio for the first time, time point one is recorded, and the fat precipitation rate is detected synchronously by the viscosity sensor. When the rate reaches the threshold, time point two is recorded, and the union of time point one and time point two is taken to generate the stage transition key time interval [minimum time, maximum time]; then, an indirect determination factor is generated: within the key time interval, the real-time moisture content and real-time fat precipitation amount are collected once per cycle time, and the moisture deviation coefficient is calculated: Moisture deviation coefficient = (initial moisture content - real-time moisture content) ) / initial moisture content, calculate the fat precipitation correction amount: fat precipitation correction amount = real-time fat precipitation amount / theoretical maximum precipitation amount in the machine's default state, generate an indirect determination factor: indirect determination factor = moisture deviation coefficient × fat precipitation correction amount; finally, dynamically determine the time node: according to the value of the indirect determination factor, the system will dynamically adjust the starting time node of the phase transition: when the factor value is high, the starting time point will be biased towards the front end of the time interval; when the factor value is low, the starting time point will be close to the back end of the time interval; if the factor is in the middle range, the starting time point will be automatically set to the middle of the entire time interval. The three situations correspond to different interval positioning strategies to ensure that the selection of the time node is always logically related to the determination factor level.

[0059] Through this solution, based on the key time interval of stage conversion and the indirect judgment factor, the starting time point of stage conversion is determined, and the starting time point of stage conversion is used as the starting time node of grinding transition. By accurately capturing the synchronization point of cell wall rupture and fat release, the dissolution rate of target cocoa flavor substances is improved, and the degradation of flavor precursors due to overheating or insufficient release is avoided.

[0060] In some embodiments, the real-time moisture content is tracked and determined based on the initial moisture content and the machine default state set; the real-time fat precipitation amount is determined based on the fat precipitation state; the time correction amount is determined based on the real-time moisture content and the real-time fat precipitation amount; the real-time moisture content is positively correlated with the relative position of the phase transition start time point within the phase transition critical time interval, that is, the higher the real-time moisture content, the closer the phase transition start time point is to the upper bound of the phase transition critical time interval; the real-time fat precipitation amount is negatively correlated with the relative position of the phase transition start time point within the phase transition critical time interval, that is, the lower the real-time fat precipitation amount, the closer the phase transition start time point is to the lower bound of the phase transition critical time interval; the time correction amount is used as an indirect determination factor.

[0061] The real-time moisture content can be the real-time moisture content of the cocoa particles during the grinding process, reflecting the fluidity and heat conduction efficiency of the material.

[0062] The real-time fat release can be the real-time mass of cocoa butter released from solid particles during the grinding process, which affects the material viscosity and frictional heat generation.

[0063] The time correction amount may be a dynamic adjustment value of the starting time point within the critical time interval of the phase transition according to the synergistic effect of the moisture content and the fat precipitation amount.

[0064] The supremum / infimum can be the theoretical earliest time point (infimum) and latest time point (supremum) of the critical time interval of phase transition, constituting the boundary constraints of time correction.

[0065] Specifically, relying solely on particle ratio matching to determine the transition time point of the grinding stage has significant defects, because it cannot adapt to the dynamic changes in moisture content and fat precipitation and the antagonistic effect between the two (moisture inhibits fat precipitation, and fat blocks water volatilization) resulting in non-simple proportional increases and decreases. The indirect determination factor generates a time correction amount by quantifying the synergistic effect of real-time moisture content (positively correlated with the delayed transition time point) and real-time fat precipitation (negatively correlated with the advanced transition time point), and dynamically and accurately corrects the transition timing. It is a necessary compensation mechanism to overcome the disconnection between physical state evolution and theoretical models and ensure the effectiveness of subsequent temperature control. This step first uses the initial moisture content and the default parameters of the grinder to monitor the grinding chamber vapor pressure in real time through a humidity sensor, infer the moisture evaporation rate, and dynamically update the real-time moisture content. Then, based on the fat precipitation state, a fat precipitation amount-time accumulation model is established, and the real-time fat precipitation amount is fitted using sensor data. Finally, the time offset coefficient is calculated based on the real-time moisture content (moisture content↑→offset coefficient↑→time point approaches the upper bound), and the time offset coefficient is calculated based on the real-time fat precipitation amount (fat precipitation↓→offset coefficient↓→time point approaches the lower bound). The two coefficients are weighted and fused to generate a time correction value, and the starting time point of the stage transition is dynamically adjusted.

[0066] Through this solution, based on the dynamic correction of dual physical property parameters, the systematic deviation of the single particle ratio determination is eliminated, and the matching degree between the starting time point of the grinding transition and the actual physical state evolution is improved; insufficient release of flavor substances due to premature conversion or burnt smell caused by late conversion is avoided, thus ensuring the stability of the target flavor and guaranteeing flavor consistency.

[0067] In some embodiments, according to a preset monitoring unit time period, the grinding transition stage is divided into multiple monitoring sub-intervals with the grinding transition start time node as the starting reference; according to the monitoring sub-intervals, the grinding temperature data in each monitoring sub-interval is obtained, and the temperature fluctuation characteristics are extracted; based on the temperature fluctuation characteristics, the temperature feature subset is screened to generate a temperature change performance parameter set.

[0068] The preset monitoring unit time period may be a pre-set fixed time interval, which is a basic time unit for discrete monitoring of the grinding process.

[0069] The monitoring subintervals may be discrete time segments obtained by dividing the grinding transition phase into preset monitoring unit time periods.

[0070] The grinding temperature data may be a sequence of original temperature values ​​collected by a temperature sensor in each monitoring subinterval.

[0071] Temperature fluctuation characteristics can be statistical and morphological indicators extracted from grinding temperature data.

[0072] Specifically, during the automated cocoa grinding process, flavor formation is extremely sensitive to temperature, especially during the grinding transition stage, where changes in the fineness of cocoa particles trigger a nonlinear accumulation of frictional heat. Target flavors such as floral, fruity, and caramel have very low tolerance for temperature ranges. Using traditional temperature monitoring methods will result in serious defects such as data overload and feature flooding, delayed control response, and weakened flavor correlation. To address the above issues, this step, based on the preset starting time node of the grinding transition phase, first subdivides the transition phase into multiple continuous time monitoring subintervals according to preset fixed-duration unit monitoring cycles. Subsequently, within each monitoring subinterval, grinding temperature data (usually with millisecond timestamps) is synchronously collected in real time, and two types of key temperature fluctuation features are independently extracted for each subinterval: statistical features (including maximum value, minimum value, arithmetic mean, and variance), and dynamic features (referring to the absolute average of the temperature difference between adjacent sampling points, and the cumulative number of times the temperature value exceeds the preset threshold). Finally, the full high-dimensional feature matrix formed by the extraction is screened in two stages: the first stage is correlation filtering (only features with an absolute correlation coefficient that is closely related to the temperature requirement of the target flavor (e.g., 95%) are retained), and the second stage is redundancy filtering (using principal component analysis (PCA) to merge features with significant linear correlations and reduce dimensionality). Ultimately, a refined set of temperature variation performance parameters that can effectively characterize the key temperature behavior patterns during the grinding transition process is generated.

[0073] Through this solution, the monitoring sub-intervals are monitored and the temperature fluctuation characteristics are extracted to increase the monitoring frequency, thereby accelerating the response speed and effectively avoiding the decomposition of flavor substances caused by instantaneous overheating. At the same time, feature screening reduces the amount of input data, improving the efficiency of control strategy generation while avoiding control command oscillation caused by overfitting. By focusing on key fluctuation characteristics to predict the inertial temperature rise trend, the deceleration command is triggered in advance to enhance the robustness of flavor control and ensure that the temperature is strictly within the target flavor range.

[0074] In some embodiments, a cocoa flavor target data set includes floral and fruity aroma target data, nutty flavor target data, and caramel flavor target data; grinding requirement information is obtained, and target flavor data corresponding to the grinding requirement information in the cocoa flavor target data set is extracted based on the grinding requirement information; a flavor target temperature difference is determined based on a temperature change performance parameter set and the temperature requirement of the target flavor data; a dynamic temperature rise inertia prediction interval is generated based on the temperature change performance parameter set and the cocoa heat capacity and the machine heat capacity; the ambient temperature is obtained, and the dynamic temperature rise inertia prediction interval is analyzed based on the real-time moisture content and the ambient temperature to determine the inertia temperature rise peak value; the actual temperature rise inertia is predicted based on the inertia temperature rise interval and the inertia temperature rise peak value to determine the actual inertia temperature rise time series data; the actual inertia temperature rise time series data is analyzed based on the inertia temperature rise peak value to determine the inertia temperature rise peak time and the peak temperature duration; based on the flavor target temperature difference, the safety shutdown / speed reduction recovery extension time is determined based on the inertia temperature rise peak value, the inertia temperature rise peak time, and the peak temperature duration; and a machine dynamic control strategy is constructed based on the safety shutdown / speed reduction recovery extension time.

[0075] The grinding requirement information can be user-defined grinding task requirements (such as flavor preference priority), which is used to extract target flavor data from the flavor target dataset.

[0076] The flavor target temperature difference may be a deviation value between the ideal temperature range required by the target flavor data and the actual predicted temperature.

[0077] The dynamic temperature rise inertia prediction interval can be a set of parameters for predicting the temperature rise range caused by the thermal inertia of materials and machines during the grinding process, the heat capacity of the machine, and the temperature change performance.

[0078] The peak inertia temperature rise is the predicted maximum temperature that the grinding system may reach under the action of thermal inertia.

[0079] The safety shutdown / speed reduction recovery extension time can be the buffer time required for early shutdown or speed reduction to prevent temperature from exceeding the standard.

[0080] Specifically, traditional cocoa grinding control methods have serious flaws: they rely on post-feedback from temperature sensors, resulting in delayed control, and the heat has penetrated deep into the cocoa butter, causing irreversible deterioration of flavor components (such as the volatilization of floral and fruity aromas); at the same time, fixed parameter models cannot adapt to the significant differences in the heat capacity of cocoa beans from different origins (such as high-fat beans have strong heat storage properties), and ignoring thermal inertia causes the temperature rise to exceed the standard after shutdown; in addition, fluctuations in ambient temperature and humidity (such as warming in summer) will superimpose the grinding heat effect, and the system lacks a compensation mechanism, resulting in poor flavor consistency between batches. This step achieves precise temperature control by building a "flavor-temperature-inertia" collaborative prediction chain. First, the target flavor data is extracted: the grinding requirement information submitted by the user is read, and the corresponding nut flavor target data is matched from the cocoa flavor target data set to determine its temperature sensitive range; secondly, the flavor target temperature difference is calculated: based on the temperature change performance parameter set (current temperature rise trend, fluctuation frequency), the temperature range required by the target flavor data is compared to obtain the maximum allowable positive deviation value (flavor target temperature difference); then the dynamic temperature rise inertia is predicted: the cocoa heat capacity and the machine heat capacity are input, the heat conduction efficiency factor is calculated through the inverse heat conduction model, the monitored basic temperature rise rate is corrected, and the inertial temperature rise rate is obtained. Taking the current temperature as the starting point, the card The Mann filter simulates the inertial temperature rise path and outputs a dynamic temperature rise inertia prediction interval; the inertial temperature rise peak is recalibrated: the ambient temperature (workshop sensor) and real-time moisture content (online moisture meter) are collected, and the first correction coefficient (environment) and the second correction coefficient (moisture) are generated based on their positive / negative correlation with the temperature rise peak. The prediction interval boundary is dynamically adjusted, and the upper bound of the corrected interval is used as the inertial temperature rise peak; finally, a safety strategy is generated: combined with the flavor target temperature difference, the amplitude, peak time and duration of the inertial temperature rise peak exceeding the target interval are analyzed. If the peak is close to the upper threshold limit, the safety shutdown / speed reduction recovery extension time required for early intervention is calculated. Based on this extension time and the preset temperature warning threshold, a dynamic machine control strategy is constructed.

[0081] Through this solution, based on the prediction of flavor-sensitive intervals and thermal inertia peak values, the safe intervention time is accurately calculated and executed, achieving proactive intervention in advance and avoiding irreversible flavor deterioration caused by heat penetrating into cocoa butter. The heat capacity differences of cocoa beans from different origins are quantified and the prediction interval is dynamically corrected to effectively adapt to changes in material properties and solve the problem of excessive temperature rise after shutdown caused by ignoring thermal inertia in fixed parameter models. Based on real-time moisture content and ambient temperature, the dynamic temperature rise inertia prediction interval is analyzed to determine the inertia temperature rise peak value. Dual dynamic calibration of the inertia temperature rise peak value using ambient temperature and real-time moisture is introduced to compensate for the superimposed influence of temperature and humidity fluctuations on the grinding thermal effect, ensuring flavor consistency between batches and ultimately achieving precise and stable control of cocoa flavor.

[0082] In some embodiments, based on a set of temperature change performance parameters, a basic temperature rise rate and a current temperature are extracted; based on the cocoa heat capacity and the machine heat capacity, a heat conduction efficiency factor is generated according to an inverse heat conduction algorithm; based on the heat conduction efficiency factor, the basic temperature rise rate is corrected to determine the inertial temperature rise rate; based on the current temperature as a reference, an inertial temperature rise simulation is performed according to the inertial temperature rise rate through a Kalman filter algorithm to determine a dynamic temperature rise inertia prediction interval.

[0083] Cocoa heat capacity can be the amount of heat required to increase the temperature of unit mass of cocoa beans, reflecting the heat storage capacity of cocoa materials.

[0084] The machine heat capacity can be the amount of heat required to raise the temperature of the grinding equipment and characterizes the thermal inertia of the machine components.

[0085] The inverse heat conduction algorithm can be a numerical calculation method for inferring the heat source transfer efficiency by inverting the temperature field distribution.

[0086] The heat transfer efficiency factor can be a correction factor that quantifies the efficiency of heat transfer between the cocoa and the machine.

[0087] The basic temperature rise rate may be the theoretical temperature rise rate without considering thermal inertia.

[0088] The current temperature may be a real-time temperature value at the start of the grinding transition phase.

[0089] The inertia temperature rise rate may be a predicted value of the actual temperature rise rate after correcting the effect of thermal inertia.

[0090] The Kalman filter algorithm can be a thermal inertia dynamic modeling method based on state estimation. By converting the thermal capacity characteristics of the grinding system into state transfer constraints, the noise suppression and optimal estimation capabilities of the Kalman filter are used to perform high-precision time series prediction of the continuous temperature rise phenomenon caused by the heat conduction delay effect.

[0091] Specifically, the core problem of temperature control lag caused by thermal inertia in the cocoa grinding process is that in the traditional grinding process, when the temperature is detected to be close to the flavor target threshold, cooling measures (such as shutdown or speed reduction) are taken. However, due to the heat capacity effect of machine components and the cocoa itself, the temperature will continue to rise, causing the actual temperature to exceed the safe range, destroying the target flavor compounds (such as the volatilization of floral and fruity esters and excessive carbonization of caramel flavor precursors), and thus resulting in a double fatal defect: the static model ignores the effects of thermal inertia and water evaporation, and temperature overshoot causes irreversible degradation of flavor substances. This step first extracts parameters: the basic temperature rise rate (such as the sliding average of the temperature change rate over the past 30 seconds) and the current temperature at the start of the grinding transition are parsed from the temperature change performance parameter set. The heat conduction efficiency factor is then calculated: the inverse heat conduction algorithm is called, the cocoa heat capacity and the machine heat capacity are input, the heat flux distribution state in the grinding chamber is inverted, and the heat conduction efficiency factor is output. The inertial temperature rise rate is then corrected: the basic temperature rise rate is multiplied by the heat conduction efficiency factor to obtain the corrected inertial temperature rise rate. Finally, a Kalman filter simulation is performed: starting with the current temperature and the inertial temperature rise rate as the state variable, considering sensor noise, and the temperature value is extrapolated forward at a unit time step (such as 1 second). The Monte Carlo simulation is repeated several times, and the upper and lower temperature boundaries of the majority of the confidence interval are taken to form a dynamic temperature rise inertia prediction interval.

[0092] Through this solution, the basic temperature rise rate is corrected based on the heat conduction efficiency factor to determine the inertial temperature rise rate. Based on the inertial temperature rise rate, the Kalman filter algorithm is used to simulate the inertial temperature rise and determine the dynamic temperature rise inertia prediction range. A dynamic temperature rise prediction model that adapts to complex grinding environments is constructed, filling the technical gap in temperature inertia control. A safety threshold is also set based on the supremum of the prediction range to ensure that the temperature is always below the target flavor degradation critical point, reducing the risk of flavor loss.

[0093] In some embodiments, the ambient temperature is positively correlated with the peak value of the inertial temperature rise. The higher the ambient temperature, the greater the increase in the supremum of the dynamic temperature rise inertia prediction interval; the real-time moisture content is negatively correlated with the peak value of the inertial temperature rise. The lower the real-time moisture content, the smaller the increase in the infimum of the dynamic temperature rise inertia prediction interval; based on the positive correlation between the ambient temperature and the peak value of the inertial temperature rise, a first correction coefficient is determined; based on the negative correlation between the real-time moisture content and the peak value of the inertial temperature rise, a second correction coefficient is determined; based on the first correction coefficient and the second correction coefficient, the inertial temperature rise interval is dynamically calibrated to generate a corrected actual inertial temperature rise interval; the supremum of the corrected actual inertial temperature rise interval is used as the peak value of the inertial temperature rise.

[0094] The ambient temperature may be a real-time temperature value of the environment in which the cocoa grinding device is located.

[0095] The first correction coefficient may be a dynamic weight value that quantifies the positive impact of the ambient temperature on the temperature rise peak.

[0096] The second correction coefficient may be a dynamic weight value that quantifies the negative inhibitory effect of moisture content on the temperature rise peak.

[0097] Dynamic calibration can be the process of adaptively adjusting the prediction interval based on real-time parameters.

[0098] The corrected actual inertial temperature rise range may be the temperature fluctuation range calibrated by the ambient temperature and moisture content.

[0099] Specifically, the existing technology only considers a single factor, ambient temperature or moisture content, in isolation, and cannot accurately reflect the amplifying effect of the dual-factor coupling effect on temperature rise under extreme working conditions such as high-temperature drying, resulting in a high deviation between the static prediction interval and the actual range, thereby triggering frequent ineffective shutdowns. This step achieves real-time dual-channel calibration of the dynamic temperature-rise inertia prediction interval by simultaneously introducing a positive correlation correction coefficient for ambient temperature and a negative correlation correction coefficient for moisture content. Specifically, the associated parameters are first analyzed: ambient temperature sensor data is acquired in real time, the impact of ambient temperature on the peak is calculated based on a preset positive correlation mapping table, and the NIR moisture detection value is simultaneously read. The inhibitory effect of moisture content on temperature rise is evaluated based on a negative correlation model. Correction coefficients are then generated: the difference between ambient temperature and reference temperature is input into an exponential function to output the first correction coefficient, and the deviation between the real-time moisture content and the critical moisture value is input into a piecewise function to generate the second correction coefficient. The interval is then dynamically calibrated: the original temperature-rise prediction interval [minimum time, maximum time] is extracted, the first correction coefficient is applied to expand the upper bound of the interval: maximum time' = maximum time × (1 + first correction coefficient), and the second correction coefficient is applied to shrink the lower bound: minimum time' = minimum time × (1 - absolute value of the second correction coefficient). The corrected actual inertia temperature rise interval [minimum time', maximum time'] is output. Finally, the peak value is determined: maximum time' is directly taken as the inertia temperature rise peak value and incorporated into the machine dynamic control strategy.

[0100] Through this solution, the inertial temperature rise range is dynamically calibrated through the first correction coefficient and the second correction coefficient, generating the corrected actual inertial temperature rise range, thereby improving the temperature rise prediction accuracy; the temperature rise peak is precisely controlled within the target flavor temperature range, improving the retention rate of key volatile substances, and significantly improving the flavor consistency between batches; through preventive peak control, the frequency of emergency shutdowns is reduced, the thermal stress damage of equipment is reduced, the service life of grinding parts is extended, and production losses are reduced.

[0101] In some embodiments, based on the peak value of the inertial temperature rise, the machine downtime / speed reduction power is set according to the safety shutdown / speed reduction recovery extension time; the real-time overall temperature during the inertial temperature rise process is obtained, and when the real-time overall temperature reaches the flavor target temperature, the deviation amplitude between the subsequent real-time overall temperature and the flavor target temperature is tracked; based on the deviation amplitude and the preset temperature deviation amplitude reference value, it is determined whether the deviation amplitude exceeds the preset temperature deviation amplitude reference value; if the deviation amplitude exceeds the preset temperature deviation amplitude reference value, the speed reduction power is dynamically increased / decreased according to the difference between the deviation amplitude and the preset temperature deviation amplitude reference value to determine the adjusted speed reduction power; based on the machine downtime / adjusted speed reduction power, the machine operating status is dynamically adjusted accordingly to determine the adjusted machine operating status; based on the adjusted machine operating status, a corresponding adjustment report is output.

[0102] The machine downtime / speed reduction power can be the reserved time for maintaining the equipment status after the temperature rise is predicted to reach the peak in order to prevent temperature overshoot and damage to the flavor.

[0103] The real-time overall temperature may be the comprehensive temperature value of the contact surface between the cocoa material and the machine in the grinding chamber.

[0104] The flavor target temperature may be the median of an ideal temperature range required to satisfy a target flavor (eg, caramel flavor, floral or fruity aroma).

[0105] The preset temperature deviation amplitude reference value may be a maximum threshold value of allowable temperature fluctuations, exceeding which triggers power adjustment.

[0106] The adjusted speed reduction power may be the grinding operation power that is dynamically corrected according to the deviation amplitude.

[0107] The adjusted machine operating state may be an equipment operating mode after a shutdown or speed reduction operation is performed.

[0108] The adjustment report can be an output document that records the machine status adjustment process and parameters.

[0109] Specifically, traditional cocoa grinding control has serious defects. The inertial characteristics of mechanical friction heat cause the temperature to continue to rise after shutdown, resulting in temperature overshoot and destroying the target flavor. The fixed threshold control mode passively responds to temperature fluctuations, causing the equipment to start and stop frequently, which not only aggravates mechanical wear, but also causes uneven crystallization of cocoa butter due to sudden temperature changes, resulting in a decrease in flavor consistency. To address the above issues, we first match the safety shutdown / speed reduction recovery extension time based on the predicted inertial temperature rise peak, and set the initial machine shutdown time or speed reduction power value; during the grinding transition stage, we continuously monitor the real-time overall temperature. When it reaches the flavor target temperature, the system calculates the deviation between the real-time temperature and the target temperature at a certain time period. If the deviation exceeds the preset reference value (such as ±3°C) for multiple consecutive times, the speed reduction power is dynamically adjusted according to the difference between the excess value and the reference value. That is, for every 1°C exceeding the limit, the original speed reduction power is increased (when the temperature is too high) or decreased (when the temperature is too low) by a certain proportion to generate the adjusted speed reduction power; finally, the equipment operating status is controlled according to the power value or the preset shutdown time, and an adjustment report containing key parameters such as temperature extremes, deviation duration, and power adjustment ratio is output to form a closed-loop control.

[0110] Through this solution, the speed reduction power is dynamically increased / decreased according to the difference between the deviation amplitude and the preset temperature deviation amplitude reference value, and the adjusted speed reduction power is determined. Combined with the safety shutdown / speed reduction recovery extension time, a dynamic buffer mechanism is established to avoid residual heat impacting the flavor critical point. At the same time, based on the closed-loop feedback of the deviation amplitude between the real-time overall temperature and the flavor target temperature, the speed reduction power is dynamically corrected, realizing the transition from open-loop preset to adaptive control, and completely resolving the core contradiction between control lag and response rigidity in traditional processes.

[0111] In some embodiments, historical grinding experimental data is obtained, and based on the historical grinding experimental data, the mapping relationship between feed speed and temperature change is analyzed to determine the rate of influence of the base temperature of the feed process on the inertia temperature rise stage; based on the base temperature influence rate and the inertia temperature rise peak value, the inertia compensation temperature difference between the current temperature and the preset temperature is determined; based on the inertia compensation temperature difference, the feed speed and feed duration required in the inertia temperature rise stage are analyzed, and a dynamic feeding strategy is constructed; based on the dynamic feeding strategy, the corresponding feeding state of the machine is adjusted, and the corresponding feeding state is written into the adjustment report.

[0112] Historical grinding experiment data may refer to a complete process data set accumulated through multiple cocoa grinding experiments, including parameters such as feed rate, real-time temperature, particle fineness, ambient temperature and humidity, etc.

[0113] The reference temperature impact rate may be a temperature change rate characterized by a unit feed rate change.

[0114] The inertia temperature rise stage may refer to the stage in which the temperature continues to rise due to the thermal inertia of the machine and cocoa after shutdown / speed reduction.

[0115] The inertia compensation temperature difference may be the difference between the target flavor temperature upper limit and the actual predicted peak temperature.

[0116] A dynamic feeding strategy can be an adjustment scheme that includes feeding rate and duration.

[0117] The feeding status can be a set of real-time working parameters of the grinding machine feed port, including conveyor belt speed and material flow rate.

[0118] Specifically, the traditional fixed feeding strategy ignores the thermal inertia effect, which still leads to temperature overshoot after shutdown / speed reduction, causing the temperature to far exceed the preset upper limit of the cocoa flavor, causing flavor deterioration. The hysteresis control triggered by the temperature threshold cannot eliminate the accumulated residual heat, and the preset strategy is difficult to adapt to the dynamic changes of the material. That is, the reduction of initial moisture will increase the temperature rise rate, and the temperature runaway rate of high-fat varieties remains high when friction heat is generated due to the lubrication effect. This step establishes a quantitative mapping of feed rate to temperature rise rate through historical data modeling, actively compensates for thermal inertia before overshoot occurs, and achieves precise temperature control across working conditions. First, a benchmark temperature influence model is established: valid experimental data in the historical database is retrieved, and the fields [feed rate, initial temperature, temperature rise rate] are extracted. The feed rate-temperature rise rate relationship curve is fitted by the least squares method to obtain the benchmark temperature influence rate function; then the inertia compensation temperature difference is calculated: the current temperature and the upper limit of the target flavor temperature are monitored in real time, and the inertia temperature rise peak value is predicted based on the thermodynamic model. The inertia temperature rise peak value = current temperature + (machine heat capacity × friction coefficient) / (cocoa mass × specific heat capacity) is used to calculate the compensation temperature difference; then a dynamic feeding strategy is generated: The equipment uses the default feed speed as a benchmark and performs inverse proportional adjustment according to the size of the compensated temperature difference. That is, for every certain percentage increase in the temperature difference, the feed speed is reduced by a certain percentage of the basic value to ensure that thermal buffer space is reserved before the temperature approaches the critical point. Based on the current temperature rise rate and the compensated temperature difference value, the necessary duration of the speed reduction operation is calculated. The specific duration is determined by the difference between the target temperature and the current temperature / the benchmark temperature rise rate×the safety buffer coefficient. This coefficient avoids excessive extension and affects efficiency. Finally, the strategy is executed and a report is written: the strategy is sent to the grinder PLC through the OPC protocol, the frequency converter is adjusted to control the feed belt speed, the adjusted feed speed, duration, actual temperature rise curve and other parameters are recorded, a structured adjustment report is generated and stored in the database.

[0119] Through this solution, the dynamic feeding strategy allows for the most efficient feeding in the low-temperature stage, and only the minimum necessary reduction is required when approaching the critical point, thereby shortening the single-batch grinding time and reducing grinding energy consumption; through the feedforward compensation mechanism, the maximum temperature deviation in the inertial temperature rise stage is controlled within an extremely small range, avoiding the decomposition of floral and fruity aroma substances or overload of caramel flavor due to overshoot, ensuring the accurate presentation of the target flavor.

[0120] Figure 3 This is a schematic diagram of the structure of an automated cocoa grinding control system provided in one embodiment of the present application, as shown in FIG. Figure 3 As shown, an automated cocoa grinding control system 300 of this embodiment includes: a time node module 301 , a temperature change module 302 , a control strategy module 303 and an execution output module 304 .

[0121] A time node module 301 is configured to obtain a cocoa initial state set and a machine default state set, track cocoa particle state changes based on the cocoa initial state set and the machine default state set, and determine a grinding transition start time node; A temperature change module 302 is configured to track a temperature change performance parameter set during a grinding transition phase according to the grinding transition start time node; A control strategy module 303 is configured to obtain a cocoa flavor target dataset, analyze the cocoa flavor target temperature range requirements during the grinding transition phase based on the cocoa flavor target dataset and the temperature variation performance parameter set, predict cocoa temperature variation characteristics, and determine a dynamic control strategy for the machine; The execution output module 304 is used to control the grinding machine to perform corresponding dynamic adjustments according to the machine dynamic control strategy and output an adjustment report.

[0122] Optionally, the time node module 301 tracks the state changes of the cocoa particles based on the cocoa initial state set and the machine default state set, and determines the grinding transition start time node, specifically for: The initial cocoa state set includes initial particle size and initial moisture content; Tracking the change in particle size ratio and fat precipitation status through the initial particle size and the machine default state set; Matching the particle size change with a preset transition starting particle ratio to determine a critical time interval for phase transition; generating an indirect determination factor based on the initial moisture content and the fat precipitation state; Based on the phase transition key time interval and according to the indirect determination factor, a phase transition start time point is determined, and the phase transition start time point is used as the grinding transition start time node.

[0123] Optionally, when generating the indirect determination factor based on the initial moisture content and the fat precipitation state, the time node module 301 is specifically configured to: Tracking and determining real-time moisture content based on the initial moisture content and the machine default state set; determining a real-time fat precipitation amount according to the fat precipitation state; determining a time correction amount according to the real-time moisture content and the real-time fat precipitation amount; The real-time moisture content is positively correlated with the relative position of the phase transition start time point within the phase transition critical time interval, that is, the higher the real-time moisture content, the closer the phase transition start time point is to the supremum of the phase transition critical time interval; The real-time fat precipitation amount is negatively correlated with the relative position of the stage transition start time point within the stage transition critical time interval, that is, the lower the real-time fat precipitation amount, the closer the stage transition start time point is to the infimum of the stage transition critical time interval; The time correction amount is used as the indirect determination factor.

[0124] Optionally, the temperature change module 302 is specifically configured to: According to a preset monitoring unit time period, the grinding transition stage is divided into a plurality of monitoring sub-intervals with the grinding transition start time node as a starting reference; According to the monitoring subintervals, obtaining grinding temperature data within each monitoring subinterval and extracting temperature fluctuation characteristics; Based on the temperature fluctuation characteristics, a temperature feature subset is screened to generate the temperature change performance parameter set.

[0125] Optionally, the control strategy module 303, when analyzing the cocoa target flavor temperature range requirements during the grinding transition stage based on the cocoa flavor target data set and the temperature change performance parameter set, predicting the cocoa temperature change characteristics, and determining the machine dynamic control strategy, is specifically configured to: The cocoa flavor target data set includes floral and fruity aroma target data, nutty flavor target data, and caramel flavor target data; obtaining grinding requirement information, and extracting target flavor data corresponding to the grinding requirement information from a cocoa flavor target data set based on the grinding requirement information; Determining a flavor target temperature difference based on the temperature change performance parameter set and the temperature requirement of the target flavor data; Based on the temperature change performance parameter set and in combination with the cocoa heat capacity and the machine heat capacity, a dynamic temperature rise inertia prediction interval is generated; Acquiring the ambient temperature, analyzing the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature, and determining the inertia temperature rise peak value; Based on the inertia temperature rise range and the peak value of inertia temperature rise, the actual temperature rise inertia is predicted and the actual inertia temperature rise time series data is determined; Analyzing the actual inertia temperature rise time series data according to the inertia temperature rise peak value to determine the inertia temperature rise peak time and peak temperature duration; Based on the flavor target temperature difference, the inertia temperature rise peak value, the inertia temperature rise peak time, and the peak temperature duration, determining the safety shutdown / speed reduction recovery extension time; A dynamic control strategy for the machine is constructed based on the safety stop / speed reduction recovery extension time.

[0126] Optionally, when the control strategy module 303 generates the dynamic temperature rise inertia prediction interval based on the temperature change performance parameter set and in combination with the cocoa heat capacity and the machine heat capacity, it is specifically configured to: Extracting the basic temperature rise rate and the current temperature according to the temperature change performance parameter set; generating a heat transfer efficiency factor based on the cocoa heat capacity and the machine heat capacity according to an inverse heat transfer algorithm; Based on the heat conduction efficiency factor, the basic temperature rise rate is corrected to determine the inertial temperature rise rate; Based on the current temperature and the inertia temperature rise rate, an inertia temperature rise simulation is performed by a Kalman filter algorithm to determine the dynamic temperature rise inertia prediction interval.

[0127] Optionally, when the control strategy module 303 analyzes the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature to determine the inertia temperature rise peak value, it is specifically configured to: The ambient temperature is positively correlated with the inertia temperature rise peak value. The higher the ambient temperature, the greater the increase in the supremum of the dynamic temperature rise inertia prediction interval. The real-time moisture content is negatively correlated with the inertia temperature rise peak value. The lower the real-time moisture content, the smaller the increase in the lower bound of the dynamic temperature rise inertia prediction interval caused by the real-time moisture content. Determining a first correction coefficient based on a positive correlation between the ambient temperature and the inertia temperature rise peak value; determining a second correction coefficient based on a negative correlation between the real-time moisture content and the inertia temperature rise peak value; Dynamically calibrating the inertia temperature rise range according to the first correction coefficient and the second correction coefficient to generate a corrected actual inertia temperature rise range; The upper limit of the corrected actual inertia temperature rise range is taken as the inertia temperature rise peak value.

[0128] Optionally, the execution output module 304 is specifically configured to: Based on the inertia temperature rise peak value and the safety shutdown / speed reduction recovery extension time, the machine shutdown time / speed reduction power is set; Acquiring the real-time overall temperature during the inertial temperature rise process, and when the real-time overall temperature reaches the flavor target temperature, tracking the subsequent deviation between the real-time overall temperature and the flavor target temperature; According to the deviation amplitude and a preset temperature deviation amplitude reference value, determining whether the deviation amplitude exceeds the preset temperature deviation amplitude reference value; If the deviation exceeds the preset temperature deviation reference value, dynamically increase / decrease the deceleration power according to the difference between the deviation and the preset temperature deviation reference value to determine the adjusted deceleration power; Dynamically adjusting the machine operating state according to the machine downtime / the adjusted speed reduction power to determine the adjusted machine operating state; According to the adjusted machine operating status, the corresponding adjustment report is output.

[0129] Optionally, the system 300 further includes a feed control module 305, specifically configured to: Acquire historical grinding experimental data, analyze the mapping relationship between feed rate and temperature change based on the historical grinding experimental data, and determine the rate of influence of the feed process on the reference temperature in the inertial temperature rise stage; determining an inertia compensation temperature difference between the current temperature and a preset temperature based on the reference temperature impact rate and the inertia temperature rise peak value; According to the inertia compensation temperature difference, the feed rate and feed duration required in the inertia temperature rise stage are analyzed to construct a dynamic feeding strategy; According to the dynamic feeding strategy, the feeding state of the machine is adjusted, and the corresponding feeding state is written into the adjustment report.

[0130] The system of this embodiment can be used to execute the method of any of the above embodiments. Its implementation principles and technical effects are similar and will not be described in detail here.

Claims

1. An automated cocoa grinding control method, characterized in that: include: Obtaining a cocoa initial state set and a machine default state set, tracking cocoa particle state changes based on the cocoa initial state set and the machine default state set, and determining a grinding transition start time node; Tracking a temperature change performance parameter set during the grinding transition stage according to the grinding transition start time node; Obtaining a cocoa flavor target data set, analyzing the cocoa flavor target temperature range requirements during the grinding transition stage based on the cocoa flavor target data set and the temperature change performance parameter set, predicting cocoa temperature change characteristics, and determining a machine dynamic control strategy; According to the machine dynamic control strategy, the grinding machine is controlled to perform corresponding dynamic adjustments and an adjustment report is output.

2. The method according to claim 1, characterized in that Tracking the state changes of cocoa particles based on the cocoa initial state set and the machine default state set to determine the grinding transition start time node includes: The initial cocoa state set includes initial particle size and initial moisture content; Tracking the change in particle size ratio and fat precipitation status through the initial particle size and the machine default state set; Matching the change in the particle size ratio with a preset transition starting particle ratio to determine a critical time interval for phase transition; generating an indirect determination factor based on the initial moisture content and the fat precipitation state; Based on the phase transition key time interval and according to the indirect determination factor, a phase transition start time point is determined, and the phase transition start time point is used as the grinding transition start time node.

3. The method according to claim 2, characterized in that The generating of an indirect determination factor based on the initial moisture content and the fat precipitation state includes: Tracking and determining real-time moisture content based on the initial moisture content and the machine default state set; determining a real-time fat precipitation amount according to the fat precipitation state; determining a time correction amount according to the real-time moisture content and the real-time fat precipitation amount; The real-time moisture content is positively correlated with the relative position of the phase transition start time point within the phase transition critical time interval, that is, the higher the real-time moisture content, the closer the phase transition start time point is to the supremum of the phase transition critical time interval; The real-time fat precipitation amount is negatively correlated with the relative position of the stage transition start time point within the stage transition critical time interval, that is, the lower the real-time fat precipitation amount, the closer the stage transition start time point is to the infimum of the stage transition critical time interval; The time correction amount is used as the indirect determination factor.

4. The method according to claim 3, characterized in that The step of tracking the temperature change performance parameter set during the grinding transition stage according to the grinding transition start time node includes: According to a preset monitoring unit time period, the grinding transition stage is divided into a plurality of monitoring sub-intervals with the grinding transition start time node as a starting reference; According to the monitoring subintervals, obtaining grinding temperature data within each monitoring subinterval and extracting temperature fluctuation characteristics; Based on the temperature fluctuation characteristics, a temperature feature subset is screened to generate the temperature change performance parameter set.

5. The method according to claim 4, characterized in that The cocoa initial state set also includes cocoa heat capacity, and the machine default state set also includes machine heat capacity. The cocoa target flavor temperature range requirements during the grinding transition stage are analyzed based on the cocoa flavor target data set and the temperature change performance parameter set, cocoa temperature change characteristics are predicted, and a machine dynamic control strategy is determined, including: The cocoa flavor target data set includes floral and fruity aroma target data, nutty flavor target data, and caramel flavor target data; obtaining grinding requirement information, and extracting target flavor data corresponding to the grinding requirement information from a cocoa flavor target data set based on the grinding requirement information; Determining a flavor target temperature difference based on the temperature change performance parameter set and the temperature requirement of the target flavor data; Based on the temperature change performance parameter set and in combination with the cocoa heat capacity and the machine heat capacity, a dynamic temperature rise inertia prediction interval is generated; Acquiring the ambient temperature, analyzing the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature, and determining the inertia temperature rise peak value; Based on the inertia temperature rise range and the peak value of inertia temperature rise, the actual temperature rise inertia is predicted and the actual inertia temperature rise time series data is determined; Analyzing the actual inertia temperature rise time series data according to the inertia temperature rise peak value to determine the inertia temperature rise peak time and peak temperature duration; Based on the flavor target temperature difference, the inertia temperature rise peak value, the inertia temperature rise peak time, and the peak temperature duration, determining the safety shutdown / speed reduction recovery extension time; A dynamic control strategy for the machine is constructed based on the safety stop / speed reduction recovery extension time.

6. The method according to claim 5, characterized in that The method of generating a dynamic temperature rise inertia prediction interval based on the temperature change performance parameter set and combining the cocoa heat capacity and the machine heat capacity includes: Extracting the basic temperature rise rate and the current temperature according to the temperature change performance parameter set; generating a heat transfer efficiency factor based on the cocoa heat capacity and the machine heat capacity according to an inverse heat transfer algorithm; Based on the heat conduction efficiency factor, the basic temperature rise rate is corrected to determine the inertial temperature rise rate; Based on the current temperature and the inertia temperature rise rate, an inertia temperature rise simulation is performed by a Kalman filter algorithm to determine the dynamic temperature rise inertia prediction interval.

7. The method according to claim 5, characterized in that The step of analyzing the dynamic temperature rise inertia prediction interval based on the real-time moisture content and the ambient temperature to determine the inertia temperature rise peak value includes: The ambient temperature is positively correlated with the inertia temperature rise peak value. The higher the ambient temperature, the greater the increase in the supremum of the dynamic temperature rise inertia prediction interval. The real-time moisture content is negatively correlated with the inertia temperature rise peak value. The lower the real-time moisture content, the smaller the increase in the lower bound of the dynamic temperature rise inertia prediction interval caused by the real-time moisture content. Determining a first correction coefficient based on a positive correlation between the ambient temperature and the inertia temperature rise peak value; determining a second correction coefficient based on a negative correlation between the real-time moisture content and the inertia temperature rise peak value; Dynamically calibrating the inertia temperature rise range according to the first correction coefficient and the second correction coefficient to generate a corrected actual inertia temperature rise range; The upper limit of the corrected actual inertia temperature rise range is taken as the inertia temperature rise peak value.

8. The method according to claim 7, characterized in that According to the machine dynamic control strategy, controlling the machine to perform corresponding dynamic adjustments and outputting an adjustment report includes: Based on the inertia temperature rise peak value and the safety shutdown / speed reduction recovery extension time, the machine shutdown time / speed reduction power is set; Acquiring the real-time overall temperature during the inertial temperature rise process, and when the real-time overall temperature reaches the flavor target temperature, tracking the subsequent deviation between the real-time overall temperature and the flavor target temperature; According to the deviation amplitude and a preset temperature deviation amplitude reference value, determining whether the deviation amplitude exceeds the preset temperature deviation amplitude reference value; If the deviation exceeds the preset temperature deviation reference value, dynamically increase / decrease the deceleration power according to the difference between the deviation and the preset temperature deviation reference value to determine the adjusted deceleration power; Dynamically adjusting the machine operating state according to the machine downtime / the adjusted speed reduction power to determine the adjusted machine operating state; According to the adjusted machine operating status, the corresponding adjustment report is output.

9. The method according to claim 6, characterized in that The method further comprises: Acquire historical grinding experimental data, analyze the mapping relationship between feed rate and temperature change based on the historical grinding experimental data, and determine the rate of influence of the feed process on the reference temperature in the inertial temperature rise stage; determining an inertia compensation temperature difference between the current temperature and a preset temperature based on the reference temperature impact rate and the inertia temperature rise peak value; Analyzing the feed rate and feed duration required in the inertia temperature rise phase according to the inertia compensation temperature difference, and constructing a dynamic feeding strategy; According to the dynamic feeding strategy, the feeding state of the machine is adjusted, and the corresponding feeding state is written into the adjustment report.

10. An automated cocoa grinding control system, characterized in that: include: a time node module, configured to obtain a cocoa initial state set and a machine default state set, track cocoa particle state changes based on the cocoa initial state set and the machine default state set, and determine a grinding transition start time node; A temperature change module, configured to track a temperature change performance parameter set during a grinding transition phase according to the grinding transition start time node; a control strategy module for acquiring a cocoa flavor target data set, analyzing the cocoa flavor target temperature range requirements during the grinding transition phase based on the cocoa flavor target data set and the temperature variation performance parameter set, predicting cocoa temperature variation characteristics, and determining a machine dynamic control strategy; The execution output module is used to control the grinding machine to perform corresponding dynamic adjustments according to the machine dynamic control strategy and output an adjustment report.

Images