Motion-perceived walnut shell cracking process regulation and control method and system

Abstract

The invention discloses an action-perceived walnut shell cracking process regulation and control method and system, and relates to the technical field of walnut processing, and the technical scheme is characterized by comprising the following steps: selecting a dynamic regulation and control unit of an automatic walnut shell cracking production line, configuring a distributed action sensing array, and determining anchoring sensing nodes; collecting a multi-dimensional data set of the dynamic regulation and control unit in a historical batch period; dividing main transmission cooperative nodes and auxiliary transmission cooperative nodes of anchoring sensing nodes based on the sensing precision of the distributed motion sensing array and the equipment transmission connection relation, and drawing a node motion conduction topological graph; stress correlation factors for regulating and controlling the action response efficiency of the main transmission cooperation node and the auxiliary transmission cooperation node are extracted in combination with a multi-dimensional data set, and the effects are that the stress correlation factors are extracted, accurate matching of shell breaking parameters and walnut characteristics is achieved, the parameters are dynamically adjusted according to the shell thickness and hardness difference, the whole kernel rate is increased, and the shell breaking loss rate is reduced. And the product value is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of walnut processing technology, and more specifically, to a method and system for controlling the walnut shell cracking process based on motion sensing. Background Technology

[0002] Walnuts, as a nutritious type of nut, rely heavily on shelling and kernel extraction as a core step in their deep processing industry. The quality of shelling directly determines the product's added value and market competitiveness. With the acceleration of industrialization in agricultural product processing, traditional manual shelling methods, due to their low efficiency, high cost, and uneven shelling quality, can no longer meet the needs of large-scale production. Automated walnut shelling production lines have emerged and become the mainstream production model in the industry.

[0003] However, current automated walnut shelling equipment is mainly divided into three types: striking, extrusion, and shearing. Their core working logic involves using preset, fixed mechanical parameters, such as striking force, extrusion pressure, and feeding speed, to process walnuts in batches. To improve applicability, some equipment has a limited number of adjustable parameter levels. Operators manually select the corresponding level based on their experience and the approximate type of walnut. However, the inherent differences in the characteristics of walnuts present many challenges to the stable operation of this type of equipment. Within the same batch of walnuts, there are significant differences in shell thickness, shell curvature, and hardness. The average shell thickness difference between smooth-skinned and rough-skinned walnuts can reach 2 millimeters, and the hardness fluctuation of different parts of a single walnut often exceeds 15%. Fixed parameters are difficult to adapt to these differences. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method and system for controlling the cracking process of walnut shells by motion sensing.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for controlling the walnut shell cracking process based on motion perception, the method comprising the following steps: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes, and collect multidimensional datasets of the dynamic control unit in historical batch cycles. Based on the sensing accuracy and device transmission connection relationship of the distributed motion sensor array, the main transmission cooperative node and the auxiliary transmission cooperative node of the anchor sensing node are divided, and the node motion transmission topology diagram is drawn. By combining multidimensional datasets, stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node are extracted; Input the data on the motion impact intensity fluctuation, response lag time and stress feedback deviation of the anchored sensing nodes in the historical batch cycle into the motion response prediction model to output the motion execution prediction result of the current batch cycle. Based on the action execution prediction results, determine whether the anchored sensing node triggers the shell breaking action warning; when the warning is triggered, retrieve the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle, and calculate the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle in combination with the node action transmission topology diagram. By integrating stress correlation factors with real-time motion sensing data of the current batch cycle, the motion control coverage area of ​​the anchored sensing node is determined. The risk assessment results are obtained by quantitatively evaluating the risk of failure of the shell-breaking action coordination within the action control coverage area based on the action deviation risk value, and dynamic adaptation of the rupture action control command is generated based on the risk assessment results.

[0006] Preferably, the multidimensional dataset includes fracture action stress time series data, walnut shell three-dimensional morphology data, and shell breakage integrity verification data.

[0007] Preferably, the stress correlation factor for regulating the action response efficiency of the main drive coordination node and the auxiliary drive coordination node is extracted by combining multidimensional datasets, specifically including the following steps: The multidimensional dataset is aligned and integrated according to the hatching operation time sequence to obtain a three-dimensional feature matrix dataset; A cross-dimensional feature fusion algorithm is used to determine the influence weight of each parameter in the three-dimensional feature matrix dataset on the node action response efficiency, and the core influencing parameters are selected by filtering the multidimensional dataset from the influence weights. The correlation coefficient between the variation range of core influencing parameters and the deviation value of node action response efficiency within different historical batch periods is calculated, and integrated to form a stress correlation factor for regulating node action response efficiency.

[0008] Preferably, the motion impact intensity fluctuation, response lag time, and stress feedback deviation data of the anchored sensing nodes within the historical batch cycle are input into the motion response prediction model to output the motion execution prediction result for the current batch cycle. This specifically includes the following steps: The full time-series data is obtained by extracting the impact intensity, response lag time and stress feedback signal of the anchored sensing nodes within the historical batch cycle, and simultaneously collecting environmental interference data on the equipment power supply stability and mechanical lubrication status of the corresponding cycle. The motion-related data is obtained by calculating the fluctuation coefficient of the motion impact intensity, the abnormality rate of the response lag time, and the deviation coefficient of the stress feedback signal from the full time series data. The motion-related data and the variation amplitude of the environmental disturbance data are coupled to obtain the motion-environment correlation factor. The motion environment correlation factors are input into the motion response prediction model to output the motion execution prediction result of the anchored sensing node in the current batch cycle. The motion execution prediction result includes the motion impact intensity value, response timeliness value, and stress feedback accuracy value.

[0009] Preferably, when an early warning is triggered, the system retrieves the action coordination error data of the main drive coordination node and the auxiliary drive coordination node within the historical cycle, and calculates the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node within the current batch cycle by combining the node action transmission topology diagram. This specifically includes the following steps: Collect action coordination error data of main drive coordination nodes and auxiliary drive coordination nodes, fatigue loss data of node components, and historical drive topology adjustment logs within the historical period, and integrate the fatigue loss data of node components with the historical drive topology adjustment logs into a basic dataset of coordination effects. The characteristic coefficient of the coordination deviation is obtained by calculating the fluctuation rate ratio based on the variation range of the coordination error data in different historical batches and the coordination influence basic dataset. The coordination deviation characteristic coefficient is calibrated and optimized by combining the weight ratio of the transmission link of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated by using the optimized coordination deviation characteristic coefficient.

[0010] Preferably, the coordination deviation characteristic coefficient is calibrated and optimized by combining the transmission link weight ratio of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated by using the optimized coordination deviation characteristic coefficient. Specifically, the following steps are included: Collect real-time operating parameters, component wear data, and material distribution density data of the main drive coordination node and auxiliary drive coordination node within the current batch cycle; The transmission link weight ratio of each node is determined based on the node action transmission topology diagram, and the matching deviation characteristic coefficient is weighted and calibrated by combining real-time operating parameters, component wear data and material distribution density data. The calibrated matching deviation characteristic coefficients are input into the risk probability calculation model to output the action deviation risk values ​​of the main drive coordination node and the auxiliary drive coordination node within the current batch cycle.

[0011] Preferably, the motion control coverage area of ​​the anchored sensing node is determined by fusing stress correlation factors with real-time motion sensing data of the current batch cycle, specifically including the following steps: Collect real-time motion parameters and sensor signal transmission intensity data of anchored sensing nodes and associated collaborative nodes within the current batch cycle, and integrate them to form a real-time motion sensing dataset. The stress correlation factor is coupled with the real-time motion sensing dataset to obtain the node motion influence weight matrix. The weight matrix of node actions is divided into regions, and the effective coverage area of ​​the action control of anchored sensing nodes is determined as the action control coverage area.

[0012] Preferably, the risk assessment result is obtained by quantitatively assessing the risk of failure of the shell-breaking action coordination within the action control coverage area based on the action deviation risk value, specifically including the following steps: Within the action control coverage area, several risk assessment sub-regions are divided, and the functional importance weights and action failure transmission coefficients of nodes within each risk assessment sub-region are determined. Input the action deviation risk value, functional importance weight, and action failure transmission coefficient into the risk quantification assessment model to output the overall shell-breaking action synergistic failure risk assessment result within the action control coverage area.

[0013] Preferably, the dynamic adaptation of the rupture action control command is generated based on the risk assessment results, specifically including the following steps: A set of risk thresholds is preset. The risk level is determined by comparing the risk assessment results with the set of risk thresholds one by one. The set of risk thresholds includes a first risk threshold, a second risk threshold, and a third risk threshold. By matching risk levels with differentiated action control strategies, rupture action control instructions are generated, which include action parameter correction schemes, execution timing optimization paths, and emergency action backup schemes.

[0014] A motion-sensing walnut shell cracking process control system includes: Data acquisition module: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes to collect multidimensional datasets of the dynamic control unit in historical batch cycles. Node partitioning module: Based on the sensing accuracy of the distributed motion sensor array and the relationship between the device transmission connection, the main transmission collaborative nodes and auxiliary transmission collaborative nodes of the anchor sensing nodes are partitioned, and the node motion transmission topology diagram is drawn. Stress factor extraction module: Combines multidimensional datasets to extract stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node; Action prediction module: Input the action impact intensity fluctuation, response lag time and stress feedback deviation data of the anchored sensing nodes in the historical batch cycle into the action response prediction model and output the action execution prediction result of the current batch cycle; The calculation module sets a threshold for qualified action response and determines whether the anchored sensing node triggers a shell-breaking action warning based on the action execution prediction result. When the warning is triggered, it retrieves the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle and calculates the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle by combining the node action transmission topology diagram. Control domain determination module: Integrates stress correlation factors with real-time motion sensing data of the current batch cycle to determine the motion control coverage domain of the anchor sensing node; Dynamic control module: Based on the action deviation risk value, the risk of failure of the shell breaking action coordination within the action control coverage area is quantitatively assessed to obtain the risk assessment result. Based on the risk assessment result, a dynamically adapted rupture action control command is generated and issued for execution.

[0015] Compared with existing technologies, this invention has the following advantages: By collecting stress time series and walnut shell three-dimensional morphology data through a distributed sensor array, stress correlation factors are extracted to achieve precise matching between shell breaking parameters and walnut characteristics. Parameters are dynamically adjusted according to differences in shell thickness and hardness to improve the whole kernel rate, reduce shell breaking loss rate, and significantly enhance product value. Impact intensity fluctuations are avoided in advance through a motion response prediction model, and dynamic control commands are directly applied to the equipment. By combining node topology diagrams and historical data to calculate deviation risk values, control is completed before component wear, avoiding hard wear. Equipment maintenance costs are reduced and lifespan is extended. Through data self-learning, it can quickly adapt to different varieties of walnuts, exhibiting strong adaptability. Attached Figure Description

[0016] Figure 1 A schematic diagram illustrating the steps of a motion-sensing method for controlling the cracking process of walnut shells proposed in this invention; Figure 2 This invention presents a schematic diagram of a motion-sensing walnut shell cracking process control system. Detailed Implementation

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0020] Reference Figures 1-2 As shown.

[0021] This embodiment further illustrates the motion-sensing method and system for controlling the cracking process of walnut shells proposed in this invention.

[0022] A method for controlling the walnut shell cracking process based on motion perception, the method comprising the following steps: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes, and collect multidimensional datasets of the dynamic control unit in historical batch cycles. Based on the sensing accuracy and device transmission connection relationship of the distributed motion sensor array, the main transmission cooperative node and the auxiliary transmission cooperative node of the anchor sensing node are divided, and the node motion transmission topology diagram is drawn. The dynamic control unit includes a roller mechanism responsible for conveying walnuts, a squeezing roller or striking component for performing the shell-breaking operation, and a motor and transmission gear set driving these components. To capture the operating status of the dynamic control unit in real time, a distributed motion sensor array needs to be configured. Appropriate sensing devices are selected based on the motion characteristics of each control unit. For example, pressure sensors are installed at key points where the squeezing roller contacts the walnut, vibration sensors are placed at the motor output shaft, and speed sensors are set at the transmission connection of the conveying roller. The distributed motion sensor array enables the collection of multi-dimensional data on the shell-breaking process. Anchored sensing nodes can filter out key points reflecting the core motion status from the sensor array. For example, the squeezing roller pressure sensor can directly acquire shell-breaking force data, and the motor vibration sensor can provide real-time feedback on the stability of power output. The dynamic control unit collects multi-dimensional datasets from historical batch cycles, including not only direct sensing data from each anchored node, such as pressure values, vibration frequency, and speed parameters, but also physical characteristic data of the corresponding batch of walnuts, such as walnut diameter, shell thickness, hardness, and final shell-breaking effect data, such as whole kernel rate and degree of shell breakage. This collectively constitutes the basic database for subsequent analysis and control of the technical solution, providing data support for precise control.

[0023] The functional positioning and collaborative relationships of each anchored sensing node are clearly defined, based on the differences in sensing accuracy of the distributed motion sensor array and the physical logic of the equipment transmission connections. Due to differences in installation location and measurement principles, the sensing devices of different anchored sensing nodes have clearly defined power transmission paths for each component through their transmission connections. For example, the main motor directly drives the extrusion roller through a reduction gear set, while the power for the conveyor roller is provided by the main motor through a belt drive. This transmission relationship determines that the sensing node corresponding to the extrusion roller has a direct power connection with the main motor node, while the conveyor roller node is at a branch point in the power transmission, thus defining the power transmission path for each anchored sensing node. The nodes are divided into main drive collaborative nodes and auxiliary drive collaborative nodes. To intuitively present the collaborative relationship, a node action transmission topology diagram needs to be drawn. The node action transmission topology diagram clearly marks the location distribution of each main drive collaborative node and auxiliary drive collaborative node in a graphical way, uses lines to connect and represent the transmission path of the transmission power, marks the sensing parameter type and accuracy range of each node, and also indicates the action sequence relationship between nodes. Through the node action transmission topology diagram, the functional division of each node, the power transmission logic and action collaboration sequence are clear at a glance, providing an intuitive visual basis for subsequent analysis of node action deviation and optimization of collaborative relationship.

[0024] By combining multidimensional datasets, stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node are extracted; Input the data on the motion impact intensity fluctuation, response lag time and stress feedback deviation of the anchored sensing nodes in the historical batch cycle into the motion response prediction model to output the motion execution prediction result of the current batch cycle. Based on the action execution prediction results, determine whether the anchored sensing node triggers the shell breaking action warning; when the warning is triggered, retrieve the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle, and calculate the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle in combination with the node action transmission topology diagram. By integrating stress correlation factors with real-time motion sensing data of the current batch cycle, the motion control coverage area of ​​the anchored sensing node is determined. The risk assessment results are obtained by quantitatively evaluating the risk of failure of the shell-breaking action coordination within the action control coverage area based on the action deviation risk value, and dynamic adaptation of the rupture action control command is generated based on the risk assessment results.

[0025] The multidimensional dataset includes fracture stress time series data, walnut shell three-dimensional morphology data, and shell breakage integrity verification data.

[0026] The multidimensional dataset specifically covers the time series data of fracture action stress, which is used to assist in the action response efficiency of transmission nodes. For example, in a walnut processing production line, the main transmission coordination node corresponds to the hydraulic power node that controls the hammer-type shell-breaking mechanism, and the auxiliary transmission coordination node corresponds to the indexing plate rotation node that conveys walnuts. By judging the peak value change of hydraulic impact stress in the fracture action stress time series data, and combining it with the shell thickness distribution characteristics in the three-dimensional morphology data of walnut shells, it was found that when the average thickness of the walnut shell increases by 1 mm, the hydraulic impact stress needs to be increased by 8% to ensure complete shell breaking. At the same time, the indexing plate rotation speed needs to be reduced by 5% to match the hammering rhythm. The rules extracted from the data correlation form the stress correlation factor for regulating the action response efficiency of the main and auxiliary transmission nodes, providing a data basis for subsequent regulation.

[0027] The stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node are extracted by combining multidimensional datasets, specifically including the following steps: The multidimensional dataset is aligned and integrated according to the hatching operation time sequence to obtain a three-dimensional feature matrix dataset; A cross-dimensional feature fusion algorithm is used to determine the influence weight of each parameter in the three-dimensional feature matrix dataset on the node action response efficiency, and the core influencing parameters are selected by filtering the multidimensional dataset from the influence weights. The correlation coefficient between the variation range of core influencing parameters and the deviation value of node action response efficiency within different historical batch periods is calculated, and integrated to form a stress correlation factor for regulating node action response efficiency.

[0028] The multidimensional dataset is aligned and integrated according to the shell-breaking operation sequence to form a three-dimensional feature matrix dataset. The shell-breaking operation sequence covers the complete process of walnuts entering the workstation, positioning, clamping, and shell-breaking detection. Each time sequence node corresponds to the synchronous acquisition of three types of data. For example, in the 10th batch of production, walnut numbered W202405 enters the workstation at the 3rd second. At this time, the positioning sensor records its three-dimensional morphological data: shell thickness 5.2 mm and curvature 8.3. The clamping is completed at the 4th second, the extrusion mechanism starts to operate, and the stress sensor collects stress time sequence data in real time: stress 28 MPa at the 4.1st second, peak stress 35 MPa at the 4.2nd second, and so on. The pressure drops to 12 MPa in 3 seconds; the shell breaks in 5 seconds, and the detection system outputs shell breakage integrity verification data with a kernel integrity rate of 100%. The three types of data for each walnut are matched one-to-one with the timestamp, and the data types of the batch number, time sequence node, and data type are used as the three-dimensional dimension to construct a three-dimensional feature matrix dataset. Assuming that the batch number is 1 to 50 on the X-axis, the time sequence node is 1 to 10 on the Y-axis corresponding to each stage of shell breaking, and the data type is 1 for stress time sequence data, 2 for three-dimensional morphology data, and 3 for shell breakage integrity data on the Z-axis, then X10Y4Z1 in the matrix represents the stress data of the cracking action at the 4th time sequence node of the 10th batch.

[0029] The response efficiency of a node action is usually measured by the action completion time and response accuracy. The response efficiency of the main drive collaborative node, such as the power node of the extrusion mechanism, is reflected in the time from receiving the command to reaching the target stress. The response efficiency of the auxiliary drive collaborative node, such as the conveying and positioning node, is reflected in the time from the walnut being in place to the completion of clamping. The cross-dimensional feature fusion algorithm first standardizes the data in the three-dimensional feature matrix to eliminate the influence of different dimensions. Then, it mines the correlation features of data in different dimensions through the feature interaction layer. Taking the main drive node of the extrusion mechanism as an example, it judges the correlation between the peak stress rise rate in the stress time series data of Z-axis 1, the shell thickness and shell curvature in the three-dimensional topography data of Z-axis 2, and the node response efficiency index and the time to reach the target stress.

[0030] When calculating the influence weights, a feature importance assessment method based on random forest is adopted. If the influence weight of shell thickness parameter is 0.32, peak stress parameter weight is 0.28, rise rate weight is 0.21, shell curvature weight is 0.08, and half-shell ratio weight in shell breakage integrity data is 0.11, the shell thickness peak stress rise rate with the highest weight is determined as the core influence parameter.

[0031] Calculate the correlation coefficient between the variation range of the core influencing parameters and the deviation value of the nodal action response efficiency within different historical batch periods, and integrate them to form a stress correlation factor; clarify the key indicators: 1. Variation range of the core influencing parameters: the average shell thickness of the 50 historical batches of walnuts was 4.8 mm, and the average shell thickness of the 51st batch was 5.4 mm, so the variation range of the shell thickness of this batch, i.e., the variation range of the core influencing parameters, is 12.5%; 2. Deviation value of nodal action response efficiency. Taking shell thickness as an example, if the standard response efficiency of the main drive node of the extrusion mechanism is 0.1 seconds to reach the target stress, and the average value of the 51st batch is 0.12 seconds, then the deviation of the node action response efficiency is 20%. Using the Pearson correlation coefficient to calculate the correlation coefficient between the two, the correlation coefficient between the shell thickness change and the node action response efficiency deviation is calculated to be 0.87, the correlation coefficient between the peak stress change and the main drive deviation is 0.79, and the correlation coefficient between the rise rate change and the main drive deviation is 0.72. The correlation coefficients are all close to 1, indicating that the core parameters and the response efficiency deviation are strongly positively correlated. Combining these correlation coefficients with the corresponding core parameters and incorporating the interaction between parameters, a stress correlation factor is finally formed. The quantified stress correlation factor provides a precise parameter basis for subsequent action control.

[0032] The data on the impact intensity fluctuations, response lag times, and stress feedback deviations of the anchored sensing nodes within historical batch cycles are input into the action response prediction model to output the action execution prediction results for the current batch cycle. Specifically, this includes the following steps: The full time-series data is obtained by extracting the impact intensity, response lag time and stress feedback signal of the anchored sensing nodes within the historical batch cycle, and simultaneously collecting environmental interference data on the equipment power supply stability and mechanical lubrication status of the corresponding cycle. The motion-related data is obtained by calculating the fluctuation coefficient of the motion impact intensity, the abnormality rate of the response lag time, and the deviation coefficient of the stress feedback signal from the full time series data. The motion-related data and the variation amplitude of the environmental disturbance data are coupled to obtain the motion-environment correlation factor. The motion environment correlation factors are input into the motion response prediction model to output the motion execution prediction result of the anchored sensing node in the current batch cycle. The motion execution prediction result includes the motion impact intensity value, response timeliness value, and stress feedback accuracy value.

[0033] The core data for anchoring the sensing nodes includes impact intensity, response lag time, and stress feedback signal. Full-volume time-series data and environmental interference data are extracted according to the historical batch cycle time sequence. Taking production data from batches 1 to 20 as an example, for the anchoring node of the main drive of the striking hammer, the impact intensity is collected by a pressure sensor at a frequency of 10 milliseconds / time, obtaining continuous time-series data within each batch. For example, in batch 15, the impact intensity data from second 3 to second 3.1 are 22 MPa, 28 MPa, 35 MPa, 32 MPa, and 25 MPa, respectively. MPa; The response lag time, i.e. the time interval from the control system issuing the striking command to the striking hammer contacting the walnut, was recorded synchronously with the displacement sensor and the timestamp. The response lag times for each striking action in the 15th batch were 0.08 seconds, 0.09 seconds, 0.08 seconds, 0.10 seconds, and 0.09 seconds, respectively; The stress feedback signal was collected by the stress feedback sensor. The difference between this signal and the actual impact intensity directly reflects the sensing accuracy. The corresponding data for the 15th batch were 21 MPa, 29 MPa, 34 MPa, 31 MPa, and 26 MPa.

[0034] Environmental interference data for each batch is collected synchronously. This data indirectly affects the response of nodes by influencing their operating status. It mainly includes the stability of the equipment power supply and the mechanical lubrication status. The stability of the equipment power supply is collected by a voltage sensor, recording the voltage fluctuation value within each batch. The mechanical lubrication status is determined by a combination of a lubricating oil viscosity meter and a wear sensor, and is represented by a lubrication status coefficient, with a value ranging from 0 to 1, where a higher value indicates a better status. The core action time sequence data of each batch is matched one-to-one with the corresponding environmental interference data according to the batch number and timestamp, forming a basic dataset integrating action and environment, providing comprehensive data support for subsequent feature extraction.

[0035] The calculation of motion-related data and the construction of motion-environment correlation factors; through feature mining of full-volume time-series data and combining the changes in environmental interference data, scattered data are transformed into correlation factors with model input value; the motion-related data calculated based on full-volume time-series data includes: the fluctuation coefficient of motion impact intensity, the anomaly rate of response lag time, and the deviation coefficient of stress feedback signal.

[0036] The impact intensity fluctuation coefficient is used to quantify the stability of impact force. For example, in batch 15, the standard deviation of the impact intensity time series data is 2.61 MPa, the mean of the impact intensity is 28.4 MPa, and the fluctuation coefficient CV is 9.2%. The response lag time anomaly rate is used to measure the stability of response timeliness. Assuming that the standard range of the response lag time of this production line is 0.07 seconds to 0.09 seconds, the number of anomalies in the response lag time exceeding the standard range in batch 15 is 1, i.e., 0.10 seconds. The total number of actions is 5, so the anomaly rate is 20%.

[0037] The deviation coefficient of the stress feedback signal is used to evaluate the accuracy of the feedback signal, and the calculation formula is as follows: Where Sfeedback is the stress feedback signal value of the i-th action, Sactual is the actual impact intensity value of the i-th action, and DC is the deviation coefficient. For example, in the data corresponding to the 15th batch, the deviation coefficient DC is 3.5%. The obtained action-related data and the variation amplitude of environmental interference data are coupled and calculated to obtain the action-environment correlation factor. The coupling calculation adopts a weighted summation method. Assuming that the weight of action-related data is 0.7 and the weight of environmental interference data is 0.3, the formula is: F=0.7×(a×CV+b×AR+c×DC)+0.3×(d×UV+e×LV), where a, b, and c are the internal weights of action-related data, d and e are the internal weights of environmental interference data, and a+b+c=1, d+e=1. After training, a=0.4, b=0.3, c=0.3, d=0.5, e=0.5, CV is the fluctuation coefficient, AR is the anomaly rate, DC is the deviation coefficient, UV is the voltage fluctuation coefficient, and LV is the lubrication coefficient variation rate. Substituting the data from the 15th batch, the action correlation component is 0.7065%; the final action environment correlation factor F is 8.2175%. The action response prediction model is trained by using the action environment correlation factors of historical batches and the corresponding actual action execution results, so that the action response prediction model has the ability to reverse the action execution state through factors. The input of the action response prediction model is the action environment correlation factor obtained from the previous data collection and calculation of the current batch, and the output is the action execution prediction result of the current batch, including the action impact intensity value, response timeliness value, and stress feedback accuracy value.

[0038] When an alert is triggered, the system retrieves the motion coordination error data of the main drive coordination node and the auxiliary drive coordination node from the historical period, and calculates the motion deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch period by combining the node motion transmission topology diagram. The specific steps include: Collect action coordination error data of main drive coordination nodes and auxiliary drive coordination nodes, fatigue loss data of node components, and historical drive topology adjustment logs within the historical period, and integrate the fatigue loss data of node components with the historical drive topology adjustment logs into a basic dataset of coordination effects. The characteristic coefficient of the coordination deviation is obtained by calculating the fluctuation rate ratio based on the variation range of the coordination error data in different historical batches and the coordination influence basic dataset. The coordination deviation characteristic coefficient is calibrated and optimized by combining the weight ratio of the transmission link of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated by using the optimized coordination deviation characteristic coefficient.

[0039] Comprehensive historical data on the coordination accuracy of the main drive and auxiliary drive nodes were collected, including error data that directly reflects the coordination state and loss and adjustment data that indirectly affect the stability of the coordination. First, the coordination error data within the historical period was collected, and the coordination deviation values ​​of key timing sequences were recorded batch by batch. Taking the production data of batches 1 to 30 as an example, for the coordination between the extrusion roller and the indexing plate, the time difference between the indexing plate completing the positioning signal and the extrusion roller starting to move, and the time difference between the extrusion roller completing the movement and the indexing plate starting to rotate were recorded. The coordination error data of batch 10 were 0.12 seconds from positioning to extrusion and 0.07 seconds from extrusion to rotation, while those of batch 20 were 0.15 seconds and 0.09 seconds, respectively.

[0040] Collect fatigue wear data of node components. The main drive coordination node needs to record the cumulative number of rotations of the extrusion roller shaft, the number of start-stops of the geared motor, and the bearing temperature change trend. The auxiliary drive node needs to record the number of meshings of the indexing plate gear and the peak torque fluctuation of the servo motor. Finally, collect historical transmission topology adjustment logs. The historical transmission topology adjustment logs record the transmission link adjustment information caused by equipment maintenance and process optimization, including adjustment time, adjustment location, transmission ratio parameters before and after adjustment, and adjustment reasons.

[0041] The fatigue loss data of each batch of node components is linked and integrated with the corresponding historical transmission topology adjustment logs by batch number to form a basic dataset of fit influence. The core is to establish the correlation link between component status, topology structure, and fit effect, so as to provide data support for subsequent analysis of the causes of fit error changes.

[0042] By quantifying the variation pattern of action coordination error data in historical batches and the correlation between the coordination influence on basic data, the characteristic coefficient of coordination deviation is obtained. The ratio of the variation amplitude and fluctuation rate of action coordination error data is clarified. Taking the time difference from positioning to extrusion as an example, the standard time is 0.1 seconds, and it is 0.12 seconds in the 10th batch, with the variation amplitude of action coordination error being 20%.

[0043] The basic dataset of the combined effects needs to be converted into a quantifiable effect intensity value of batch i, taking into account the dual effects of component fatigue and topology adjustment. The component fatigue effect is calculated as the proportion of cumulative wear to rated life. For example, the rated rotational life of the extrusion roller shaft is 100,000 times, and the 10th batch has a cumulative rotation of 12,000 times, with a fatigue effect accounting for 12%; the rated meshing life of the indexing plate gear is 80,000 times, and the 10th batch has a cumulative rotation of 12,000 times, with a fatigue effect accounting for 15%. The maximum value of the two is taken as the component fatigue effect value of 15%. The topology adjustment effect is determined according to the adjustment range.

[0044] The fluctuation rate ratio can reflect the change in fit error caused by the change in unit influence intensity. Taking the 10th and 20th batches as the analysis objects, the ratio for batches 10-20 is 3.97. The fit deviation characteristic coefficient, calculated by taking the 5th and 10th batches, is 2.03 for batches 5-10. This indicates that the fit error of the main and auxiliary nodes of the production line is relatively sensitive to the change in influence intensity. An increase in unit influence intensity will cause a 3-fold increase in error amplitude.

[0045] The node motion transmission topology diagram clearly marks the transmission link structure, power transmission path, and weight ratio of the main drive collaborative nodes and auxiliary drive collaborative nodes. The weight ratio is determined according to the core position of the node in the transmission link. The main drive collaborative nodes usually have a higher weight ratio because they directly determine the shell breaking action. Taking the production line topology diagram as an example, the transmission link includes the main link of the main motor, gearbox, and extrusion roller, and the auxiliary link of the servo motor, coupling, and indexing plate. The weight of each node is determined by dynamic analysis, and the sum of the weight ratios of each node is 1.

[0046] Calibration optimization requires adjusting the characteristic coefficient K of the coordination deviation based on the actual transmission status of the main and auxiliary nodes in the current batch, combined with the topology weights. The calibration formula is Kcalibration = K × (W_main × α + W_auxiliary × β), where W_main is the total weight of the main transmission node (0.7), W_auxiliary is the total weight of the auxiliary transmission node (0.3), α is the current state coefficient of the main transmission node (normal is 1, slight abnormality is 1.1, severe abnormality is 1.3), and β is the current state coefficient of the auxiliary transmission node (the value of which follows the same rule as α). Assuming that the current batch is found to have a slight abnormal noise in the main transmission gearbox through real-time monitoring (state coefficient α is 1.1), and the auxiliary transmission node is operating normally (β is 1), then Kcalibration is 3.21. The calibrated coefficient is more in line with the current equipment operating status.

[0047] The calculation of the motion deviation risk value needs to be combined with the predicted impact intensity value of the current batch. This value is calculated based on the real-time collected component fatigue data. The current extrusion roller has accumulated 30,000 rotations, with fatigue accounting for 30%. The topology status has not been adjusted recently, and the topology impact value is 4.2%. The predicted impact intensity value is 22.26%. When the motion deviation risk value is 23.3, the risk value is usually set as 0-30 for low risk, 30-60 for medium risk, and above 60 for high risk. The current risk value of 23.3 indicates that there is a slight motion deviation risk. It is necessary to adjust the lubrication status of the main drive reduction gearbox or slightly modify the transmission ratio to avoid the risk from escalating.

[0048] From in-depth mining of historical data to precise calibration of the current state, a complete risk quantification logic is formed. Through the correlation analysis of action coordination errors and influencing factors, and combined with the topological structure to clarify the source of risk, the final quantified risk value provides a precise basis for the generation of subsequent control instructions, which not only solves the problem of minor deviations, but also avoids excessive control that affects production efficiency.

[0049] The coordination deviation characteristic coefficient is calibrated and optimized by combining the transmission link weight ratio of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated using the optimized coordination deviation characteristic coefficient. The specific steps include: Collect real-time operating parameters, component wear data, and material distribution density data of the main drive coordination node and auxiliary drive coordination node within the current batch cycle; The transmission link weight ratio of each node is determined based on the node action transmission topology diagram, and the matching deviation characteristic coefficient is weighted and calibrated by combining real-time operating parameters, component wear data and material distribution density data. The calibrated matching deviation characteristic coefficients are input into the risk probability calculation model to output the action deviation risk values ​​of the main drive coordination node and the auxiliary drive coordination node within the current batch cycle.

[0050] Comprehensive core data for the current batch is collected, including real-time operating parameters, component wear data, and material distribution density data. For each type of data, key indicators are selected based on the collaborative characteristics of the main drive collaborative node and the auxiliary drive collaborative node. Real-time operating parameters represent the collaborative action status of the main drive collaborative node and the auxiliary drive collaborative node. For the main drive collaborative node, the actual value of the extrusion roller motor speed is collected (rated value 300 rpm), the actual value of the gearbox output torque is collected (rated value 50 Nm), and the actual value of the extrusion action response time is collected (rated value 0.1 seconds). For the auxiliary drive collaborative node, the actual value of the indexing plate servo motor speed is collected (rated value 60 rpm), and the actual value of the positioning response time is collected (rated value 0.05 seconds), thus directly reflecting the real-time stability of the node action.

[0051] The component wear data indicates the fatigue state of the core components. The main drive coordination node records the cumulative number of rotations of the extrusion roller shaft and the measured value of the lubricating oil quality in the gearbox; the auxiliary drive coordination node records the cumulative number of meshings of the indexing plate gear and the wear of the coupling. Component wear directly affects the accuracy of motion execution and is an important reference for the calibration of the characteristic coefficient of the matching deviation. The material distribution density data reflects the load characteristics of the current processing scenario. It is collected by infrared sensors in the production line material channel to calculate the number of walnuts per unit length. Changes in material density will cause load fluctuations in the main and auxiliary nodes, thus affecting the accuracy of motion matching.

[0052] The node motion transmission topology diagram determines the transmission link weight ratio of each node through dynamics and production experience. The total weight of the main transmission coordination node is higher than that of the auxiliary transmission coordination node, and the weight of the core execution component is higher than that of the power transmission component. The sum of the weight ratios of each node is 1, thereby ensuring that the calibration process can highlight the influence of the core components.

[0053] Weighted calibration requires first converting the collected real-time data into standardized state coefficients, and then combining the topology weights and coordination deviation characteristic coefficients to calculate the calibration value. The calculation of state coefficients requires setting rules for different data types. The weighted calibration formula is Kcalibration = K original × (W main × P main + W auxiliary × P auxiliary), where K original is the coordination deviation characteristic coefficient. Assuming K original is 3, W main is the total weight of the main drive 0.6, and W auxiliary is the total weight of the auxiliary drive 0.4, then K calibration is 0.72. The calibrated coefficients are more in line with the actual operating status of the main and auxiliary nodes in the current batch, avoiding deviations caused by relying solely on historical data.

[0054] The calibrated fit deviation characteristic coefficients are input into the risk probability calculation model, which outputs the action deviation risk value. The risk probability calculation model adopts an architecture that combines logistic regression with historical risk data. The risk probability calculation model has been trained with a large number of historical batches of K calibration values ​​and actual action deviation occurrence rates to establish a quantitative correlation between the two. The risk value calculation first outputs the probability of action deviation occurrence through the risk probability calculation model, and then combines it with the severity coefficient of the risk impact factor to obtain the final action deviation risk value. The severity coefficient is calculated based on the deviation of material hardness and ambient temperature.

[0055] After inputting the weighted and calibrated data into the risk probability calculation model, the probability of motion deviation is output by querying the risk probability corresponding to the similarity coefficient in historical data and combining it with the current risk impact factor. The production line is set with a risk value of 0 to 25 as low risk, 25 to 50 as medium risk, and above 50 as high risk. If the calculated final motion deviation risk value is 22.33, it indicates that there is a slight motion deviation risk in the main drive coordination node and the auxiliary drive coordination node, but it will not seriously affect the shell breaking quality. The control strategy can focus on slightly adjusting the speed of the main drive motor and the positioning timing of the auxiliary drive, without the need for machine shutdown maintenance.

[0056] Real-time data acquisition ensures the timeliness of calibration, topological weighting highlights the impact of core nodes, and a well-trained model enables quantitative risk assessment. This forms a complete logical chain from data to conclusions, avoiding the limitations of historical data and enhancing the comprehensiveness of risk assessment through multi-factor fusion. It provides a scientific basis for precise control of the production line and effectively balances production efficiency and shell-breaking quality.

[0057] By fusing stress correlation factors with real-time motion sensing data from the current batch cycle, the motion control coverage area of ​​the anchored sensing node is determined, specifically including the following steps: Collect real-time motion parameters and sensor signal transmission intensity data of anchored sensing nodes and associated collaborative nodes within the current batch cycle, and integrate them to form a real-time motion sensing dataset. The stress correlation factor is coupled with the real-time motion sensing dataset to obtain the node motion influence weight matrix. The weight matrix of node actions is divided into regions, and the effective coverage area of ​​the action control of anchored sensing nodes is determined as the action control coverage area.

[0058] The determination of the motion control coverage area is achieved through three steps: data acquisition, matrix coupling, and region division. The data is collected from anchored sensing nodes and associated collaborative nodes. Real-time motion parameters and sensor signal transmission intensity data are used to construct a real-time motion sensing dataset. The real-time motion parameters directly reflect the node operating status. For the main drive anchored nodes, the data collection should include the current batch average hydraulic rod extension speed of 0.8 m / s and the historical best of 0.6 m / s, and the average impact force of 32 MPa and the historical adaptation value of 28 MPa. For the auxiliary drive associated nodes, the data collection should include the average roller speed of 45 r / min and the standard adaptation speed of 40 r / min, and the average feeding positioning accuracy of ±0.3 mm and the standard value of ±0.2 mm.

[0059] The sensor signal transmission strength data ensures the reliability of the sensing. The signal transmission rate of each node sensor and the control system is collected. The current hammer sensor is 1.2Mbps, and the standard is ≥1.0Mbps. The signal packet loss rate of the current feed roller sensor is 0.3%, and the standard is ≤0.5%. This reflects whether the signal transmission will interfere with the accurate feedback of the action parameters. The data is integrated according to the structure of node number, parameter type and collection timestamp to form a real-time action sensing dataset of core parameters, which provides a basis for subsequent coupling calculation with stress correlation factors.

[0060] The stress correlation factors are extracted into a three-dimensional matrix F, with dimensions of main transmission parameters, auxiliary transmission parameters, and coordination coefficients. Among them, F11 is the matching coefficient between impact force and shell thickness (0.85), F12 is the coordination coefficient between feeding speed and impact frequency (0.78), F21 is the correlation coefficient between hydraulic pressure and impact accuracy (0.91), and F22 is the matching coefficient between feeding positioning and impact timing (0.72). The real-time motion sensing dataset is transformed into a real-time state matrix X, with the same dimensions as F. The elements in X are the deviation rates between the current parameters and the standard values. For example, the impact force deviation rate of X11 is 0.1429, the feeding speed deviation rate of X12 is 0.125, the hydraulic speed deviation rate of X21 is 0.3333, and the positioning accuracy deviation rate of X22 is 0.5.

[0061] Matrix coupling operations employ Hadamard product combined with weighted summation; the final calculated node action influence weight matrix is ​​W11 = 0.4649, W12 = 0.4345, W21 = 0.5660, and W22 = 0.5840. Higher weight values ​​indicate a greater impact of the node parameter on the action effect, requiring more control. Based on the node action transmission topology diagram, the nodes corresponding to the node action influence weight matrix are divided into four regions according to their physical location and functional association: Region 1 corresponds to W11, the main node of the hammer hydraulic mechanism; Region 2 corresponds to the main node of the feed roller drive motor; and Region 3 corresponds to W11, the main node of the feed roller drive motor. Node W12 corresponds to the hydraulic mechanism sensor node in region three, W21 corresponds to the hydraulic mechanism sensor node in region four, and W22 corresponds to the feeding and positioning sensor node in region four. A weight threshold of 0.45 is set, and regions with weight values ​​higher than the threshold (0.4649, 0.5660, and 0.5840) are selected. These three regions and the signal interaction links between them are integrated to determine the motion control coverage area, thereby clarifying the core range of control. The hydraulic mechanism of the hammer and the feeding and positioning sensor node are the focus, and the hydraulic sensor nodes are synchronously associated to avoid wasting control resources in low-impact areas.

[0062] The risk assessment results are obtained by quantitatively evaluating the risk of failure of coordinated breaking action within the action control coverage area based on the action deviation risk value, specifically including the following steps: Within the action control coverage area, several risk assessment sub-regions are divided, and the functional importance weights and action failure transmission coefficients of nodes within each risk assessment sub-region are determined. Input the action deviation risk value, functional importance weight, and action failure transmission coefficient into the risk quantification assessment model to output the overall shell-breaking action synergistic failure risk assessment result within the action control coverage area.

[0063] Within the motion control coverage area, risk assessment sub-regions are divided, and the functional importance weight and motion failure transmission coefficient of each sub-region are determined. The division of risk assessment sub-regions combines the node motion transmission topology diagram and equipment physical layout, following the principles of strong functional correlation and clear risk impact boundaries. For the control coverage area of ​​the production line, it can be divided into four core sub-regions: Sub-region 1 is the core area of ​​the main drive of the extrusion roller, including the extrusion roller motor, gearbox and direct drive components, providing the power for shell breaking; Sub-region 2 is the pressure sensing and monitoring area, composed of pressure sensors and signal preprocessing modules, responsible for real-time feedback of the shell breaking stress state; Sub-region 3 is the feed belt auxiliary drive area, covering the feed belt drive motor and speed controller, which dominates the conveying and positioning rhythm of walnuts; Sub-region 4 is the signal coordination control area, including the signal interaction modules and controllers of the main and auxiliary nodes, undertaking the key function of motion timing synchronization.

[0064] Functional importance weights measure the degree of influence of each sub-region on the synergistic effect of the shell-breaking action. They are determined by combining the irreplaceable function of each region with its contribution to the final shell-breaking quality. A higher weight indicates a greater disruption to the synergistic effect due to functional failure in that region. Based on historical production data, the functional importance weights of each sub-region can be determined as follows: Sub-region one, as the power core, directly determines the shell-breaking force and stability, and has the highest functional importance weight, set at 0.4; Sub-region four is responsible for the timing synchronization of the main and auxiliary actions, and synergistic failures often originate from here, with a weight set at 0.3; Sub-region two's sensor data provides the basis for control, with a weight set at 0.2; Sub-region three's feeding rhythm, although affecting synergy, can be quickly corrected through speed adjustment, with a weight set at 0.1. The sum of the weights of the four sub-regions is 1, ensuring the rigor of the evaluation logic.

[0065] The motion failure propagation coefficient reflects the range and degree of impact of motion failure in one sub-region on the motion stability of other sub-regions. The larger the coefficient value, the stronger the failure propagation capability and the wider the risk diffusion range. This coefficient needs to be determined based on the transmission link relationship in the node motion propagation topology diagram and the analysis of historical failure cases. The main drive failure in sub-region one will directly lead to signal coordination disorder in sub-region four, and at the same time affect the stress feedback accuracy of sub-region two. The motion failure propagation coefficient is set to 0.9. The signal coordination failure in sub-region four will simultaneously interfere with the motion timing of sub-region one and sub-region three. The propagation coefficient is set to 0.8. The sensor failure in sub-region two only affects the control accuracy and has little impact on the motion execution of other regions. The propagation coefficient is set to 0.3. The feeding failure in sub-region three mainly affects its own motion and has limited interference with the main drive and sensing system. The propagation coefficient is set to 0.4.

[0066] After completing the sub-region division and parameter determination, the overall collaborative failure risk assessment result is output by integrating the action deviation risk value, functional importance weight, and action failure transmission coefficient through a risk quantification assessment model. The risk quantification assessment model adopts a weighted fusion algorithm, first calculating the local collaborative failure risk value of each sub-region, and then integrating it with the functional importance weight to form the overall risk assessment result, ensuring that the assessment reflects both local risk differences and overall collaborative characteristics.

[0067] Define the core input parameters, among which the action deviation risk value is a quantitative indicator obtained through calculation. Assuming that the action deviation risk value calculated for the current batch is 28, this value reflects the basic risk level of the overall action deviation of the main and auxiliary nodes. The functional importance weight and action failure transmission coefficient of each sub-region are denoted as W1 to W4 and K1 to K4, respectively. At the same time, the real-time failure probability parameter P of the sub-region needs to be introduced. This parameter is calculated by the deviation between the real-time operating parameters and the rated parameters of the equipment in each sub-region, reflecting the possibility of action failure in each region under the current state.

[0068] The risk quantification assessment model is calculated in two steps. First, the local collaborative failure risk value Ri of each sub-region is calculated. Then, the basic risk is combined with the characteristics of the sub-region to obtain the local risk quantification value of each region. Substituting the parameters of the current batch, the local risk values ​​of each sub-region are calculated: R1 for sub-region one is 0.504; R2 for sub-region two is 0.0504; R3 for sub-region three is 0.0224; and R4 for sub-region four is 0.4032. The local collaborative failure risk values ​​of each sub-region are weighted and summed to obtain the overall shell-breaking action collaborative failure risk assessment result of 0.98 within the action control coverage area. Based on the risk level classification standard of this production line, the overall collaborative failure risk assessment result of 0 to 1 is extremely low risk, 1 to 3 is low risk, 3 to 5 is medium risk, and above 5 is high risk. The current assessment result of 0.98 indicates that although there is a certain basic risk of action deviation, the possibility of collaborative failure in each sub-region is extremely low, and the overall collaborative state of the shell-breaking action is stable. Based on this assessment, subsequent control instructions can focus on fine-tuning the signal coordination frequency of sub-region four and simultaneously optimizing the motor speed fluctuations in sub-region one. This eliminates the need for large-scale shutdowns and maintenance, ensuring shell-breaking quality while minimizing the impact on production efficiency. By using a multi-parameter fusion model to obtain the overall risk results, a complete assessment logic from local to overall and from qualitative to quantitative assessments is formed. This accurately identifies key risk points and provides a quantitative basis for subsequent control, significantly improving the scientific rigor and efficiency of walnut shell cracking control.

[0069] Based on the risk assessment results, dynamically adapted rupture action control instructions are generated, specifically including the following steps: A set of risk thresholds is preset. The risk level is determined by comparing the risk assessment results with the set of risk thresholds one by one. The set of risk thresholds includes a first risk threshold, a second risk threshold, and a third risk threshold. By matching risk levels with differentiated action control strategies, rupture action control instructions are generated, which include action parameter correction schemes, execution timing optimization paths, and emergency action backup schemes.

[0070] A risk threshold set is preset and risk level determination is completed. Through scientific threshold division, the control priority corresponding to different risk levels is clarified. The setting of the risk threshold set needs to combine the correlation between risk assessment results and shell breakage quality loss rate in historical production data, and also refer to equipment maintenance costs and production efficiency requirements. Finally, a three-level threshold system including a first risk threshold, a second risk threshold, and a third risk threshold is determined. Based on the actual operation data of the production line, the first risk threshold is set to 1, corresponding to a shell breakage quality loss rate of less than 0.5%, which is considered extremely low risk; the second risk threshold is set to 3, corresponding to a quality loss rate of 1% to 3%, which is considered low risk; the third risk threshold is set to 5, corresponding to a quality loss rate of 3% to 5%, which is considered medium risk; when the risk assessment result is higher than 5, the quality loss rate exceeds 5%, which is considered high risk.

[0071] Risk level determination is achieved by comparing the risk assessment results of the current batch's collaborative failure with the risk threshold set one by one. The determination logic is clear and operable. Different risk levels correspond to different control and response mechanisms. Level 1 risk does not require active control, but only requires enhanced real-time monitoring; Level 2 risk requires online parameter fine-tuning; Level 3 risk requires parameter correction and timing optimization; Level 4 risk requires the activation of emergency plans and suspension of some production processes to ensure that the risk is controllable.

[0072] By matching risk levels with differentiated action control strategies, three core elements of rupture control instructions are generated, ensuring that control strategies are matched with risk levels and risk sources to avoid over-control or under-control. The formulation of differentiated action control strategies needs to combine the causes of each risk level and determine the core targets and adjustment ranges of control based on the risk contribution of different sub-regions.

[0073] For Unit 1, which is at Level 1 risk, the risk assessment result of 0.8 is lower than the first risk threshold. The main source of risk is the slight speed fluctuation of the feed belt in Sub-region 3. The corresponding control strategy is primarily monitoring with fine-tuning as a secondary measure. In the generated control instructions, the action parameter correction scheme only requires a ±1% fine-tuning of the feed belt drive motor speed. Specifically, the current feed belt speed of 40 r / min is corrected to 39.8 r / min, offsetting the impact of speed fluctuations through this subtle adjustment. The execution timing optimization path does not require changing the core collaborative timing of the main and auxiliary nodes; it only increases the signal interaction detection frequency from 100 ms / time to 50 ms / time, enhancing real-time perception of timing deviations. The emergency action backup scheme does not need to be activated; it only requires presetting the trigger condition for automatically triggering Level 2 control when the speed fluctuation exceeds 2%. The characteristic of this type of instruction is that it does not affect production continuity and has extremely low control costs.

[0074] For Unit 2, a level 2 risk assessment result 2.2, the risk source is the slight signal deviation of the pressure sensor in sub-region 2 and the timing lag of the signal coordination in sub-region 4. The corresponding control strategy is precise parameter correction and local timing optimization. The action parameter correction scheme is as follows: 1. The pressure setting value of the extrusion roller, combined with the stress correlation factor and real-time sensor data, is currently set at 32MPa. Due to the sensor signal deviation, the actual pressure is too high, so it is corrected to 30.5MPa. The correction range is calculated by the deviation coefficient × stress correlation factor. The formula is: Correction value = current setting value × (1 - signal deviation coefficient). The signal deviation coefficient is obtained by comparing the real-time sensor data with the standard pressure calibration value. The current value is 0.0468, so the correction value is 30.5MPa. 2. The start and stop delay time of the feeding belt is corrected from the current 0.08 seconds to 0.07 seconds to ensure the synchronization of walnut positioning and extrusion action.

[0075] For Unit 3, a risk level of 3, the risk assessment result is 4.5. The risk sources are the power fluctuation of the extrusion roller motor in sub-region 1 and the transmission delay of the signal coordination module in sub-region 4. The corresponding control strategy is multi-parameter collaborative correction + timing reconstruction and emergency start. The timing optimization path needs to reconstruct the coordination logic of the main and auxiliary nodes, adjusting the original serial timing to parallel preprocessing timing. Specifically, the feeding belt conveys walnuts, synchronously starts the extrusion roller preheating and pressure calibration, the walnut positioning is completed, and the extrusion action is immediately triggered. Through parallel preprocessing, the overall coordination time is shortened from 0.18 seconds to 0.12 seconds, solving the signal transmission delay problem. The emergency action backup plan needs to be activated immediately to reduce the production load of the unit to 80% and use the backup pressure sensor to replace the equipment with the largest deviation. If the risk is still not reduced after replacement, the temporary shutdown procedure of the unit is triggered to avoid batch shelling quality problems. The control intensity is moderate. The moderate risk is solved through multi-parameter collaboration and timing reconstruction, and an emergency fallback mechanism is also provided.

[0076] For Unit 4, which is at risk level 4, the risk assessment result exceeds 5. The risk sources are mostly severe wear and tear on the main drive motor of Sub-area 1 or failure of the signal coordination module of Sub-area 4. The corresponding control command is emergency loss prevention. The action parameter correction plan requires immediately reducing the pressure of the extrusion roller to the safe value of 25MPa and stopping the feeding belt. The execution timing optimization path is paused, and the equipment safety lock program is initiated. The emergency action backup plan is fully activated, including switching to the backup main drive motor, notifying maintenance personnel for on-site inspection, and transferring the walnut material of this unit to other units for processing. At the same time, the quality traceability system is activated to conduct a comprehensive inspection of the walnuts already processed in this unit, thereby ensuring the effectiveness of the control command and enabling the system to dynamically adjust the control strategy based on actual feedback, forming a continuously optimized control capability.

[0077] By defining risk levels through risk thresholds, ensuring precise control through differentiated strategies, and guaranteeing execution effectiveness through closed-loop feedback, a complete logical chain from risk assessment to instruction implementation is formed. This avoids the rigidity of traditional fixed parameter control and balances control effectiveness with production efficiency through tiered response, providing reliable technical support for the stable operation of the walnut shell cracking production line.

[0078] A motion-sensing walnut shell cracking process control system includes: Data acquisition module: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes to collect multidimensional datasets of the dynamic control unit in historical batch cycles. Node partitioning module: Based on the sensing accuracy of the distributed motion sensor array and the relationship between the device transmission connection, the main transmission collaborative nodes and auxiliary transmission collaborative nodes of the anchor sensing nodes are partitioned, and the node motion transmission topology diagram is drawn. Stress factor extraction module: Combines multidimensional datasets to extract stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node; Action prediction module: Input the action impact intensity fluctuation, response lag time and stress feedback deviation data of the anchored sensing nodes in the historical batch cycle into the action response prediction model and output the action execution prediction result of the current batch cycle; The calculation module sets a threshold for qualified action response and determines whether the anchored sensing node triggers a shell-breaking action warning based on the action execution prediction result. When the warning is triggered, it retrieves the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle and calculates the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle by combining the node action transmission topology diagram. Control domain determination module: Integrates stress correlation factors with real-time motion sensing data of the current batch cycle to determine the motion control coverage domain of the anchor sensing node; Dynamic control module: Based on the action deviation risk value, the risk of failure of the shell breaking action coordination within the action control coverage area is quantitatively assessed to obtain the risk assessment result. Based on the risk assessment result, a dynamically adapted rupture action control command is generated and issued for execution.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for controlling the cracking process of walnut shells using motion sensing, characterized in that, The method includes the following steps: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes, and collect multidimensional datasets of the dynamic control unit in historical batch cycles. Based on the sensing accuracy and device transmission connection relationship of the distributed motion sensor array, the main transmission cooperative node and the auxiliary transmission cooperative node of the anchor sensing node are divided, and the node motion transmission topology diagram is drawn. By combining multidimensional datasets, stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node are extracted; Input the data on the motion impact intensity fluctuation, response lag time and stress feedback deviation of the anchored sensing nodes in the historical batch cycle into the motion response prediction model to output the motion execution prediction result of the current batch cycle. Based on the action execution prediction results, determine whether the anchored sensing node triggers the shell breaking action warning; when the warning is triggered, retrieve the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle, and calculate the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle in combination with the node action transmission topology diagram. By integrating stress correlation factors with real-time motion sensing data of the current batch cycle, the motion control coverage area of ​​the anchored sensing node is determined. The risk assessment results are obtained by quantitatively evaluating the risk of failure of the shell-breaking action coordination within the action control coverage area based on the action deviation risk value, and dynamic adaptation of the rupture action control command is generated based on the risk assessment results.

2. The method for controlling the walnut shell cracking process based on motion sensing according to claim 1, characterized in that, The multidimensional dataset includes fracture stress time series data, walnut shell three-dimensional morphology data, and shell breakage integrity verification data.

3. The method for controlling the walnut shell cracking process based on motion sensing according to claim 2, characterized in that, The stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node are extracted by combining multidimensional datasets, specifically including the following steps: The multidimensional dataset is aligned and integrated according to the hatching operation time sequence to obtain a three-dimensional feature matrix dataset; A cross-dimensional feature fusion algorithm is used to determine the influence weight of each parameter in the three-dimensional feature matrix dataset on the node action response efficiency, and the core influencing parameters are selected by filtering the multidimensional dataset from the influence weights. The correlation coefficient between the variation range of core influencing parameters and the deviation value of node action response efficiency within different historical batch periods is calculated, and integrated to form a stress correlation factor for regulating node action response efficiency.

4. The method for controlling the walnut shell cracking process based on motion sensing according to claim 3, characterized in that, The data on the impact intensity fluctuations, response lag times, and stress feedback deviations of the anchored sensing nodes within historical batch cycles are input into the action response prediction model to output the action execution prediction results for the current batch cycle. Specifically, this includes the following steps: The full time-series data is obtained by extracting the impact intensity, response lag time and stress feedback signal of the anchored sensing nodes within the historical batch cycle, and simultaneously collecting environmental interference data on the equipment power supply stability and mechanical lubrication status of the corresponding cycle. The motion-related data is obtained by calculating the fluctuation coefficient of the motion impact intensity, the abnormality rate of the response lag time, and the deviation coefficient of the stress feedback signal from the full time series data. The motion-related data and the variation amplitude of the environmental disturbance data are coupled to obtain the motion-environment correlation factor. The motion environment correlation factors are input into the motion response prediction model to output the motion execution prediction result of the anchored sensing node in the current batch cycle. The motion execution prediction result includes the motion impact intensity value, response timeliness value, and stress feedback accuracy value.

5. The method for controlling the walnut shell cracking process based on motion sensing according to claim 4, characterized in that, When an alert is triggered, the system retrieves the motion coordination error data of the main drive coordination node and the auxiliary drive coordination node from the historical period, and calculates the motion deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch period by combining the node motion transmission topology diagram. The specific steps include: Collect action coordination error data of main drive coordination nodes and auxiliary drive coordination nodes, fatigue loss data of node components, and historical drive topology adjustment logs within the historical period, and integrate the fatigue loss data of node components with the historical drive topology adjustment logs into a basic dataset of coordination effects. The characteristic coefficient of the coordination deviation is obtained by calculating the fluctuation rate ratio based on the variation range of the coordination error data in different historical batches and the coordination influence basic dataset. The coordination deviation characteristic coefficient is calibrated and optimized by combining the weight ratio of the transmission link of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated by using the optimized coordination deviation characteristic coefficient.

6. The method for controlling the walnut shell cracking process based on motion sensing according to claim 5, characterized in that, The coordination deviation characteristic coefficient is calibrated and optimized by combining the transmission link weight ratio of each node in the node action transmission topology diagram. The action deviation risk value of the main transmission coordination node and the auxiliary transmission coordination node in the current batch cycle is calculated using the optimized coordination deviation characteristic coefficient. The specific steps include: Collect real-time operating parameters, component wear data, and material distribution density data of the main drive coordination node and auxiliary drive coordination node within the current batch cycle; The transmission link weight ratio of each node is determined based on the node action transmission topology diagram, and the matching deviation characteristic coefficient is weighted and calibrated by combining real-time operating parameters, component wear data and material distribution density data. The calibrated matching deviation characteristic coefficients are input into the risk probability calculation model to output the action deviation risk values ​​of the main drive coordination node and the auxiliary drive coordination node within the current batch cycle.

7. The method for controlling the walnut shell cracking process based on motion sensing according to claim 6, characterized in that, By fusing stress correlation factors with real-time motion sensing data from the current batch cycle, the motion control coverage area of ​​the anchored sensing node is determined, specifically including the following steps: Collect real-time motion parameters and sensor signal transmission intensity data of anchored sensing nodes and associated collaborative nodes within the current batch cycle, and integrate them to form a real-time motion sensing dataset. The stress correlation factor is coupled with the real-time motion sensing dataset to obtain the node motion influence weight matrix. The weight matrix of node actions is divided into regions, and the effective coverage range of the action control of anchored sensing nodes is determined as the action control coverage area.

8. The method for controlling the walnut shell cracking process based on motion sensing according to claim 7, characterized in that, The risk assessment results are obtained by quantitatively evaluating the risk of failure of coordinated breaking action within the action control coverage area based on the action deviation risk value, which specifically includes the following steps: Within the action control coverage area, several risk assessment sub-regions are divided, and the functional importance weights and action failure transmission coefficients of nodes within each risk assessment sub-region are determined. Input the action deviation risk value, functional importance weight, and action failure transmission coefficient into the risk quantification assessment model to output the overall shell-breaking action synergistic failure risk assessment result within the action control coverage area.

9. The method for controlling the walnut shell cracking process based on motion sensing according to claim 8, characterized in that, Based on the risk assessment results, dynamically adapted rupture action control instructions are generated, specifically including the following steps: A set of risk thresholds is preset. The risk level is determined by comparing the risk assessment results with the set of risk thresholds one by one. The set of risk thresholds includes a first risk threshold, a second risk threshold, and a third risk threshold. By matching risk levels with differentiated action control strategies, rupture action control instructions are generated, which include action parameter correction schemes, execution timing optimization paths, and emergency action backup schemes.

10. A motion-sensing walnut shell cracking process control system, applied to the motion-sensing walnut shell cracking process control method described in any one of claims 1-9, characterized in that, include: Data acquisition module: Select the dynamic control unit of the automated walnut shell cracking production line, configure a distributed motion sensor array and establish anchor sensing nodes to collect multidimensional datasets of the dynamic control unit in historical batch cycles. Node partitioning module: Based on the sensing accuracy of the distributed motion sensor array and the relationship between the device transmission connection, the main transmission collaborative nodes and auxiliary transmission collaborative nodes of the anchor sensing nodes are partitioned, and the node motion transmission topology diagram is drawn. Stress factor extraction module: Combines multidimensional datasets to extract stress correlation factors that regulate the action response efficiency of the main drive coordinated node and the auxiliary drive coordinated node; Action prediction module: Input the action impact intensity fluctuation, response lag time and stress feedback deviation data of the anchored sensing nodes in the historical batch cycle into the action response prediction model and output the action execution prediction result of the current batch cycle; Accounting module: Sets the qualified threshold for action response, and determines whether the anchored sensing node triggers the shell breaking action warning based on the action execution prediction result; When an early warning is triggered, retrieve the action coordination error data of the main drive coordination node and the auxiliary drive coordination node in the historical cycle, and calculate the action deviation risk value of the main drive coordination node and the auxiliary drive coordination node in the current batch cycle by combining the node action transmission topology diagram. Control domain determination module: Integrates stress correlation factors with real-time motion sensing data of the current batch cycle to determine the motion control coverage domain of the anchor sensing node; Dynamic control module: Based on the action deviation risk value, the risk of failure of the shell breaking action coordination within the action control coverage area is quantitatively assessed to obtain the risk assessment result. Based on the risk assessment result, a dynamically adapted rupture action control command is generated and issued for execution.

Images