A big data-based production comprehensive monitoring system

Abstract

The present application belongs to the technical field of production supervision, and particularly relates to a production comprehensive supervision system based on big data. The present application is based on the constructed multi-dimensional data acquisition module, state time sequence prediction module, dynamic strategy generation module, closed-loop control execution module and digital twin and visualization module, accurately predicts the risks of the target material conveying equipment from multiple dimensions such as material, equipment health and energy consumption, analyzes the optimal dynamic control parameter sequence, thereby realizes the dynamic speed regulation, feed control and other strategies, avoids material congestion, improves the conveying capacity per unit time, reduces the energy consumption of equipment operation, and guarantees the production continuity. Furthermore, the pre-rehearsal verification of the optimal dynamic control parameter sequence of the target material conveying equipment is beneficial to reduce the failure rate and secondary risk, improve the production safety, and is beneficial to adapt to complex dynamic production conditions.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of production supervision, and particularly relates to a production comprehensive supervision system based on big data. BACKGROUND

[0002] In modern industrial production, material conveying equipment is the core hub connecting various production links, and undertakes the task of transferring raw materials, semi-finished products and finished products, and its running state directly affects the continuity, efficiency and safety of the production process. Effective supervision of material conveying equipment is a key link to ensure the timely execution of production plans, reduce operating costs and reduce safety accidents.

[0003] The material conveying equipment supervision method in the prior art has basically met the use requirements, but it still has certain deficiencies: (1) The prior art can only respond passively or simply to the state, and cannot foresee complex dynamic conditions such as material flow fluctuations and potential equipment failures, thereby causing frequent congestion and shutdown problems.

[0004] (2) The prior art uses fixed parameters to control the material conveying equipment, ignoring multi-dimensional detection and analysis of the material conveying equipment, so that it cannot be dynamically adjusted according to the real-time state, which may result in high energy consumption or low conveying efficiency.

[0005] (3) The prior art ignores simulation and prediction of the material conveying equipment before adjustment, which may result in poor parameter adaptability due to lack of verification of the adjustment strategy, which is not conducive to reducing invalid energy consumption and equipment wear and tear, and cannot prolong the service life of the equipment and reduce operating costs. SUMMARY

[0006] In view of this, in order to solve the problems raised in the background art, a production comprehensive supervision system based on big data is provided.

[0007] The purpose of the application can be achieved by the following technical solutions: The application provides a production comprehensive supervision system based on big data, which comprises a multi-dimensional data acquisition module, a state time sequence prediction module, a dynamic strategy generation module, a closed-loop control execution module and a digital twin and visualization module.

[0008] The multi-dimensional data acquisition module is connected with the state time sequence prediction module, the state time sequence prediction module is connected with the dynamic strategy generation module, the dynamic strategy generation module is connected with the closed-loop control execution module, and the digital twin and visualization module is bidirectionally connected with the state time sequence prediction module and the dynamic strategy generation module, respectively.

[0009] The multi-dimensional data acquisition module acquires real-time original running data sets of the target material conveying equipment.

[0010] The state-time prediction module preprocesses the original running dataset and imports it into a pre-trained state-time prediction model, outputting a multi-dimensional state-time prediction vector for the target material conveying equipment within a preset future time window.

[0011] The dynamic strategy generation module determines whether there is operational risk in the multi-dimensional state time-series prediction vector by running risk identification rules. When there is operational risk, it automatically generates the optimal dynamic control parameter sequence through multi-objective collaborative optimization decision logic.

[0012] The closed-loop control execution module converts the optimized dynamic control parameter sequence into control commands for the corresponding actuators and issues them. At the same time, it feeds back the adjusted actual operating data to the multi-dimensional data acquisition module, forming a continuously iteratively optimized monitoring and control closed loop.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention accurately predicts the risks of the target material conveying equipment from multiple dimensions such as material, equipment health and energy consumption, and analyzes them to obtain the optimal dynamic control parameter sequence, thereby realizing strategies such as dynamic speed regulation and feeding control, avoiding material congestion, increasing the conveying capacity per unit time, reducing equipment operating energy consumption, and ensuring production continuity.

[0014] 2. This invention uses digital twin and visualization modules to pre-test and verify the optimal dynamic control parameter sequence of the target material conveying equipment, which helps to reduce the failure rate and secondary risks, improve production safety, and help it adapt to complex dynamic production conditions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1

[0019] Please see Figure 1 As shown, the present invention provides a comprehensive production monitoring system based on big data, with the following specific module distribution: multi-dimensional data acquisition module, state timing prediction module, dynamic strategy generation module, and closed-loop control execution module.

[0020] The multidimensional data acquisition module is connected to the state timing prediction module, the state timing prediction module is connected to the dynamic strategy generation module, and the dynamic strategy generation module is connected to the closed-loop control execution module.

[0021] The multi-dimensional data acquisition module collects the original operating dataset of the target material conveying equipment in real time.

[0022] In a preferred feasible example of the present invention, the original operating dataset includes physical status data from the device's own sensors, device operating condition data from the control unit interface, and scheduling instruction data from the upper-level production execution unit.

[0023] In one specific example, the physical state data includes vibration spectrum data of key load-bearing points of the equipment, surface temperature data of the drive motor and gearbox, volume and morphological distribution data of the material on the conveyor belt surface, and real-time rotational speed data of each drive roller.

[0024] It should be further explained that the specific methods for acquiring the vibration spectrum data of the key load-bearing points of the equipment, the surface temperature data of the drive motor and gearbox, the volume and morphological distribution data of the material on the conveyor belt surface, and the real-time rotational speed data of each drive roller are as follows: Vibration spectrum data of the key load-bearing points of the equipment is acquired through an accelerometer; surface temperature data of the drive motor and gearbox is acquired through a temperature sensor; volume and morphological distribution data of the material on the conveyor belt surface is acquired through an industrial camera or LiDAR; and real-time rotational speed data of each drive roller is acquired through an encoder. The accelerometer, temperature sensor, industrial camera or LiDAR, and encoder mentioned above are relatively mature technologies in the prior art and will not be elaborated upon here.

[0025] The equipment operating data includes the current, voltage, and torque data of each drive motor, as well as the tension setting value of each section of the conveyor belt.

[0026] It should be further explained that the specific method for obtaining the current, voltage, and torque data of each drive motor and the tension setting value of each section of the conveyor belt is as follows: the current, voltage, and torque data of each drive motor and the tension setting value of each section of the conveyor belt are directly read from the programmable logic controller (PLC).

[0027] The scheduling instruction data includes the batch, type, target path, and planned start and stop times of the materials to be transported.

[0028] It should be further explained that the specific method for obtaining the batch, type, target path, and planned start and stop time information of the materials to be transported is to directly obtain the batch, type, target path, and planned start and stop time information of the materials to be transported from the warehouse management platform (WMS).

[0029] The state-time prediction module preprocesses the original running dataset and imports it into a pre-trained state-time prediction model, outputting a multi-dimensional state-time prediction vector for the target material conveying equipment within a preset future time window.

[0030] It should be noted that the preprocessing process performed by the state time series prediction module after receiving the original running dataset includes: (1) Data cleaning and alignment: missing values ​​in the original running dataset are filled by time series nearest neighbor interpolation, outliers are identified and removed by isolated forest algorithm, and the timestamps of multi-source heterogeneous data are aligned based on a unified minimum sampling period to form a standardized time series data matrix.

[0031] (2) Feature engineering construction: extract time-domain features and frequency-domain features from the standardized time-series data matrix; the time-domain features include mean, variance, peak value and kurtosis, and the frequency-domain features are obtained by performing fast Fourier transform on the vibration spectrum data; the extracted features are concatenated with the original scheduling instruction data to form the final feature vector of the original running dataset input to the state time-series prediction model.

[0032] It should be further explained that the state temporal prediction model is a sequence-to-sequence model with an encoder-decoder architecture; wherein, the encoder part adopts a multi-layer stacked long short-term memory network (LSTM) structure to learn and compress the historical temporal dependencies of the input feature vector and generate a fixed-length context vector; the decoder part also adopts a multi-layer stacked long short-term memory network structure and combines an attention mechanism to gradually generate the multi-dimensional state temporal prediction vector based on the context vector.

[0033] In a preferred feasible example of the present invention, the specific method of the output target material conveying equipment within a future preset time window for multidimensional state time-series prediction vector includes: the state time-series prediction model performs feature extraction and time-series correlation learning on the input original operating dataset; the model's bottom layer encodes different types of data in the original operating dataset separately; then, the fusion layer captures the correlation relationship; then, the recurrent layer learns the long-term dependencies in historical data; and finally, the output layer calculates for the three target sub-vectors respectively and integrates them into the final output according to the time dimension and the target dimension, thereby generating the multidimensional state time-series prediction vector.

[0034] One specific example is the frequency domain characteristics of vibration spectra or the discrete characteristics of scheduling instructions.

[0035] Such relationships include, for example, the timing correlation between heavy-load batches and the increase in motor current in scheduling instructions.

[0036] It should be further explained that the time dimension is to ensure that the time step of the three sub-vectors is consistent with the preset time window, for example, in the next 30 minutes, there is one time point every 5 minutes, for a total of 6 time steps.

[0037] The target dimension is the vector structure formation. For example, the final vector shape can be "[preset time window length, total dimension of sub-vectors]".

[0038] For example, [6,15] represents 6 time steps, which may include 5 flow parameters, 7 health parameters, and 3 energy consumption parameters.

[0039] In a preferred feasible example of the present invention, the specific training process of the state time series prediction model includes: extracting a large amount of historical original operation datasets of target material conveying equipment and their corresponding adjusted actual operation state data from the database as training samples.

[0040] A state-time prediction model is built using a deep learning network. The historical raw running dataset is processed and used as the model input. The corresponding adjusted actual running state data is used as the label. The loss function is set as the mean squared error (MSE) between the multidimensional state-time prediction vector and the true label. The Adam optimizer is used, and the network weights and biases in the encoder and decoder are iteratively updated through the backpropagation algorithm until the value of the loss function converges to below a preset threshold or reaches a preset number of training epochs. The model training is then completed, and the state-time prediction model is obtained.

[0041] For example, the convergence threshold of the loss function is set to 0.005, that is, the mean square error between the multidimensional state time-series prediction vector and the true label is 0.005. The preset number of training cycles is 100 rounds. If the convergence is not achieved within 100 rounds, the training is terminated.

[0042] In a preferred feasible example of the present invention, the multidimensional state time-series prediction vector includes at least one prediction sub-vector regarding the material flow distribution of the target material conveying equipment, one prediction sub-vector regarding the health status of the key components of the target material conveying equipment, and one prediction sub-vector regarding the total energy consumption of the target material conveying equipment.

[0043] The predicted sub-vectors for the material flow distribution include flow rate indicators, spatial distribution indicators, and dynamic change indicators.

[0044] It should be noted that the flow rate index reflects the flow rate of the material being transported and is an important parameter for evaluating the material transport volume.

[0045] The spatial distribution index reflects the spatial distribution of materials on the conveying equipment, which helps to determine whether the materials are evenly distributed and avoid local congestion.

[0046] The dynamic change index shows the dynamic trend of material flow over time and can provide early warning of sudden increases or decreases in flow.

[0047] The predictive subvectors for the health status of the key components include performance degradation indicators, failure risk indicators, and health scores.

[0048] It should be noted that the performance degradation index is used to measure the degree of performance degradation of key components over time, and can detect the downward trend of component performance in advance.

[0049] The failure risk index assesses the likelihood of critical components failing, providing a basis for failure prevention.

[0050] The health score is used to evaluate the overall health status of key components in a quantitative way, making it easy to intuitively understand the health level of the components.

[0051] The predictor vector for total energy consumption includes total energy consumption, energy consumption distribution, and energy consumption trend.

[0052] The dynamic strategy generation module determines whether there is operational risk in the multi-dimensional state time-series prediction vector by running risk identification rules. When there is operational risk, it automatically generates the optimal dynamic control parameter sequence through multi-objective collaborative optimization decision logic.

[0053] In a preferred feasible example of the present invention, the operational risks include material congestion risk, equipment failure risk, and energy consumption deviation risk.

[0054] The specific content of the operational risk identification rule is as follows: the flow rate index, spatial distribution index and dynamic change index in the predicted sub-vector of the material flow distribution of the target material conveying equipment are compared with the corresponding preset safety thresholds. If any one of them is not within the range of the corresponding preset safety threshold, it is determined that the target material conveying equipment has a risk of material congestion.

[0055] The performance degradation index, failure risk index and health score in the prediction sub-vector of the health status of the key components of the target material conveying equipment are compared with the corresponding preset health baseline. If any of them does not fall within the range of the corresponding preset health baseline, the target material conveying equipment is determined to have equipment failure risk.

[0056] The total energy consumption, energy consumption distribution, and energy consumption trend in the predicted sub-vector of the total energy consumption of the target material conveying equipment are compared with the theoretical optimal energy consumption benchmark for the corresponding current operating condition. If any of these items deviates from the theoretical optimal energy consumption benchmark for the corresponding current operating condition, it is determined that the target material conveying equipment has a risk of energy consumption deviation.

[0057] In a preferred feasible example of the present invention, the specific content of the multi-objective collaborative optimization decision logic includes: A1, objective function construction: constructing an objective function that includes three sub-objectives: maximizing transmission throughput, minimizing equipment energy consumption, and minimizing overall equipment loss, and assigning a weighted sum objective function with dynamically adjustable weight coefficients to each sub-objective.

[0058] It should be noted that the significance of constructing a target that includes maximizing conveying throughput, minimizing equipment energy consumption, and minimizing overall equipment loss is as follows: Maximizing conveying throughput aims to increase the amount of material conveyed per unit time, ensuring the efficient progress of the production process and meeting the production plan's requirements for material conveying efficiency. Minimizing equipment energy consumption focuses on reducing energy consumption during equipment operation, aligning with the concept of energy conservation and emission reduction, while also reducing energy costs in production. Minimizing overall equipment loss emphasizes reducing wear and tear, aging, and other losses of various equipment components, extending equipment lifespan, and reducing maintenance and replacement costs.

[0059] It should be further explained that the significance of assigning a dynamically adjustable weight coefficient to the weighted sum objective function for each sub-objective is that the dynamic adjustment of the weight coefficients can adapt to different production scenarios and demand priorities. For example, during peak production periods, the weight of "maximizing transport throughput" may be increased to prioritize timely material delivery; while during periods of energy shortage, the weight of "minimizing equipment energy consumption" may be increased to prioritize energy consumption control. By integrating the three sub-objectives into a comprehensive objective function through weighted summation, multi-objective synergistic optimization is achieved, avoiding the performance degradation in other aspects that may result from optimizing a single objective, and ensuring that the optimal equipment control strategy is found at the overall level.

[0060] A2. Parameter optimization: When there is an operational risk in the target material conveying equipment, the objective function is used as the optimization target. Within the preset feasible domain of control parameters, the particle swarm optimization algorithm is used to perform global optimization calculation to obtain a set of dynamic control parameters that make the objective function value optimal, and this set is recorded as the optimal dynamic control parameter sequence.

[0061] It should be noted that the particle swarm optimization algorithm is a global optimization algorithm based on swarm intelligence, simulating the cooperative behavior of birds foraging in flocks. In the algorithm, each "particle" represents a potential solution. By tracking the individual optimal solution and the swarm optimal solution, it continuously adjusts its position and velocity, iteratively searching in the solution space, and ultimately approximating the global optimum. Its characteristics include fast convergence speed, simple parameter settings, and applicability to multi-dimensional and multi-constraint optimization problems.

[0062] It should be further explained that the specific process of global optimization calculation using particle swarm optimization algorithm includes: (1) Initializing particle swarm: Within the preset feasible domain of control parameters, a certain number of particles are randomly generated. Each particle corresponds to a set of dynamic control parameter sequences, such as motor speed and torque limit of each section, as the initial potential solution.

[0063] (2) Calculate the fitness value: Substitute the sequence of dynamic control parameters corresponding to each particle into the objective function and calculate its fitness value, which is the objective function value. The better the fitness value, the closer the set of parameters is to the optimal solution.

[0064] (3) Update individual and swarm optimal solutions: Each particle tracks the parameters corresponding to its own historical best fitness value, i.e., the individual optimal solution, and the parameters corresponding to the optimal fitness value of the entire particle swarm, i.e., the swarm optimal solution.

[0065] (4) Iterative optimization of parameters: Based on the individual optimal solution and the group optimal solution, the particle adjusts its "position" (i.e., parameter value) and "velocity" (i.e., parameter adjustment range), as shown in the following formula: Velocity update: ,in For inertial weights, As a learning factor, It is a random number. For the optimal position of an individual, The optimal position for the group. For "location". Location update: .

[0066] (5) Termination of iteration: When the number of iterations reaches the preset value or the fitness value converges and no longer changes significantly, the iteration stops. At this time, the dynamic control parameter sequence corresponding to the population optimal solution is the optimal dynamic control parameter sequence.

[0067] For example, inertia weight The learning factor has a value between 0.6 and 0.8. Both are set to 2.0, the number of particles is set to 30, and the number of iterations is set to 50.

[0068] In a preferred feasible example of the present invention, the optimized dynamic control parameter sequence is specifically manifested as a set of cooperative instructions bound to the time axis for different segment drive units.

[0069] The collaborative instruction set includes: conveyor belt speed adjustment parameters, feed rate control parameters, distribution correction parameters, load adjustment parameters, tension dynamic correction parameters, cooling and lubrication control parameters, operation mode switching parameters, collaborative operation parameters, and regeneration energy recovery parameters.

[0070] It should be noted that the conveyor belt speed adjustment parameter is the target speed value of the material conveying equipment for each time period within the future preset time window, for example, 0-10 minutes: 1.2m / s; 10-30 minutes: 0.9m / s. By reducing or increasing the speed, overload can be avoided or conveying efficiency can be improved.

[0071] The feed rate control parameter is the opening percentage sequence of the feed valve or feeder of the target material conveying equipment, for example, reducing from 80% to 60% within the next 5 minutes to control the material input rate to match the conveying capacity.

[0072] The distribution correction parameter refers to the angle and position adjustment value of the material guiding device in the target material conveying equipment. For example, the left guide plate of the conveyor belt is deflected by 3° to correct the material unbalanced load and avoid local wear or deviation.

[0073] The load adjustment parameter is the output power limit value of the drive motor of the target material conveying equipment. For example, the maximum power should not exceed 25kW in the next 20 minutes to avoid overload and overheating of the motor.

[0074] The tension dynamic correction parameter is the pressure adjustment sequence of the conveyor belt tensioning cylinder of the target material conveying equipment. For example, it is gradually reduced from the current 1.0MPa to 0.8MPa, with a reduction of 0.1MPa every 5 minutes, to alleviate the wear of the rollers caused by excessive tension.

[0075] The cooling and lubrication control parameters are the rotation speed of the motor cooling fan of the target material conveying equipment, such as maintaining high-speed operation for the next 15 minutes, and the start-stop interval of the bearing lubrication pump, such as starting once every 10 minutes for 30 seconds, to control the temperature of key components within a safe range.

[0076] The operating mode switching parameters refer to the switching sequence of the target material conveying equipment between high-efficiency mode and energy-saving mode. For example, during off-peak hours (20-30 minutes), the equipment switches to energy-saving mode and the motor frequency drops to 40Hz.

[0077] The coordinated operation parameters are the start and stop linkage commands for the multi-section conveyor belt of the target material conveying equipment. For example, when the flow rate of section A is less than 0.5 m³ / h, section B is delayed in starting for 5 minutes to avoid idling and energy consumption.

[0078] The regenerative energy recovery parameter is the activation threshold of the braking energy recovery device of the target material conveying equipment. For example, when the conveyor belt deceleration rate exceeds 0.2m / s², energy recovery is activated to improve energy utilization.

[0079] This invention accurately predicts the risks of target material conveying equipment from multiple dimensions such as material, equipment health, and energy consumption, and analyzes them to obtain the optimal dynamic control parameter sequence. This enables strategies such as dynamic speed regulation and feed control, avoiding material congestion, increasing the conveying capacity per unit time, reducing equipment operating energy consumption, and ensuring production continuity.

[0080] The closed-loop control execution module converts the optimized dynamic control parameter sequence into control commands for the corresponding actuators and issues them to drive the target material conveying equipment to adjust its operating state. At the same time, it feeds back the adjusted actual operating data to the multi-dimensional data acquisition module, forming a continuously iteratively optimized monitoring and control closed loop.

[0081] In a preferred feasible example of the present invention, the specific method of converting the optimized dynamic control parameter sequence into control commands for the corresponding actuator and issuing them includes: classifying the optimized dynamic control parameter sequence according to the control objectives to obtain the dynamic control parameters of each control objective to which the target material conveying equipment belongs, establishing the association between them and the corresponding actuators based on the database, and converting the dynamic control parameters of each control objective to which the target material conveying equipment belongs into specific values ​​that can be recognized by the actuators according to the physical characteristics of the target material conveying equipment.

[0082] It should be noted that the control objectives include material flow and distribution control objectives, critical component health protection control objectives, and energy consumption optimization and coordination control objectives.

[0083] Among them, the dynamic control parameters for the material flow and distribution control objectives include conveyor belt speed adjustment parameters, feed rate control parameters, and distribution correction parameters.

[0084] The dynamic control parameters for the health protection control objectives of critical components include load adjustment parameters, tension dynamic correction parameters, and cooling and lubrication control parameters.

[0085] The dynamic control parameters for the energy consumption optimization and coordinated control objectives include operating mode switching parameters, cooperative operation parameters, and regenerative energy recovery parameters.

[0086] For example, "speed adjustment amount 0.5m / s" is converted into a motor frequency command of 45Hz, and "tension +5%" is converted into a cylinder pressure command of 1.2MPa, taking into account the actuator's range limitation.

[0087] The specific numerical values ​​that the actuator can recognize are converted into the corresponding format according to the actuator communication protocol before the instruction is issued.

[0088] For example, the actuator communication protocol includes, but is not limited to, Modbus, Profinet, and CANopen.

[0089] The corresponding formats include, but are not limited to, hexadecimal codes, binary signals, and pulse signals.

[0090] In a specific example, “45Hz” is encoded as register address 0x0001 and value 0x002D via the Modbus protocol.

[0091] In a preferred feasible example of the present invention, the system further includes a digital twin and visualization module, which is bidirectionally connected to the state timing prediction module and the dynamic strategy generation module.

[0092] The digital twin and visualization module specifically includes: B1, constructing a three-dimensional visualized digital twin based on the geometric model, physical properties, and kinematic logic of the target material conveying equipment.

[0093] It should be noted that the geometric model, physical properties, and kinematic logic of the target material conveying equipment are all obtained by real-time scanning and recording using a deployed laser 3D scanner.

[0094] It should be further noted that the construction of the three-dimensional visualization digital twin of the target material conveying equipment is a relatively mature technology in the existing field, and will not be elaborated further here.

[0095] B2. The original running dataset, the multidimensional state time-series prediction vector, and the optimal dynamic control parameter sequence are mapped in real time onto the three-dimensional visual digital twin of the target material conveying equipment to present the current status, future trends, and control strategies to be executed of the equipment in a visual manner.

[0096] B3. Before sending the control commands of the actuator corresponding to the dynamic control parameter sequence to the closed-loop control execution module, a virtual operation simulation is performed in the three-dimensional visualization digital twin environment of the target material conveying equipment to evaluate the effects and potential secondary risks. When the simulation results meet the preset safety and efficiency standards, the control commands are authorized to be sent to the closed-loop control execution module.

[0097] This invention uses digital twins and visualization modules to pre-test and verify the optimal dynamic control parameter sequence of the target material conveying equipment, which helps to reduce the failure rate and secondary risks, improve production safety, and help it adapt to complex dynamic production conditions.

[0098] Example 2

[0099] In a second embodiment of the present invention, in conjunction with the above-described comprehensive production supervision system based on big data, the present invention provides the following technical solution: a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-described comprehensive production supervision system based on big data.

[0100] Those skilled in the art will understand that the data in the flowchart, or logic and / or steps otherwise described herein, for example, can be considered as a sequenced data table of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0101] More specific examples of readable media (a non-exhaustive list of data) include: electrical connections (electronic devices) with one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0102] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0103] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A big data-based production integrated supervision system, characterized in that: The method comprises the following steps: a multi-dimensional data acquisition module acquires real-time original operation data sets of a target material conveying device; a state time series prediction module imports the original operation data sets after preprocessing into a pre-trained state time series prediction model, and outputs a multi-dimensional state time series prediction vector of the target material conveying device in a future preset time window; a dynamic strategy generation module determines whether there is an operation risk in the multi-dimensional state time series prediction vector through an operation risk identification rule, and automatically generates an optimal dynamic control parameter sequence through multi-objective collaborative optimization decision logic when there is an operation risk; a closed-loop control execution module converts the optimal dynamic control parameter sequence into control instructions for corresponding actuators and issues the control instructions, and simultaneously feeds back the adjusted actual operation data to the multi-dimensional data acquisition module to form a continuously iteratively optimized supervisory control closed loop; the original operation data sets include physical state data from device body sensors, device working condition data from control unit interfaces, and scheduling instruction data from upper-layer production execution units; the specific manner of outputting the multi-dimensional state time series prediction vector of the target material conveying device in the future preset time window comprises: the state time series prediction model performs feature extraction and time series association learning on the input original operation data sets, the model bottom layer encodes different types of data in the original operation data sets, then captures the association relationship through a fusion layer, then learns the long-term dependence in the historical data through a recurrent layer, and finally integrates the three target sub-vectors into the final output according to the time dimension and the target dimension through the output layer to generate the multi-dimensional state time series prediction vector. 2.The production integrated monitoring system based on big data according to claim 1, characterized in that: the specific training process of the state time series prediction model comprises: extracting a large amount of historical original operation data sets of target material conveying devices and their corresponding adjusted actual operation state data from a database as training samples; building a state time series basic prediction model through a deep learning network, processing the historical original operation data sets as model input, taking the corresponding adjusted actual operation state data as labels, setting the loss function as the mean square error between the multi-dimensional state time series prediction vector and the true labels, using the Adam optimizer, and iteratively updating the network weights and biases in the encoder and the decoder through the back propagation algorithm until the value of the loss function converges to below a preset threshold or reaches a preset training period, completing the model training, and obtaining the state time series prediction model. 3.The production integrated monitoring system based on big data according to claim 2, characterized in that: the multi-dimensional state time series prediction vector contains at least one prediction sub-vector of material flow distribution of the target material conveying device, one prediction sub-vector of the health state of a key component of the target material conveying device, and one prediction sub-vector of the total energy consumption of the target material conveying device; the prediction sub-vector of the material flow distribution includes flow indicators, spatial distribution indicators, and dynamic change indicators; the prediction sub-vector of the health state of the key component includes performance degradation indicators, fault risk indicators, and health score; the prediction sub-vector of the total energy consumption includes total energy consumption, energy consumption distribution, and energy consumption trend.

4. The production integrated monitoring system based on big data according to claim 3, characterized in that: The operation risk includes material congestion risk, equipment failure risk, and energy consumption deviation risk. The specific content of the operation risk identification rule is that the flow index, spatial distribution index, and dynamic change index in the prediction sub-vector of the material flow distribution of the target material conveying equipment are compared with the corresponding preset safety threshold, and if any of them is not within the corresponding preset safety threshold range, it is determined that the target material conveying equipment has a material congestion risk. The performance degradation index, failure risk index, and health score in the prediction sub-vector of the health status of the key components of the target material conveying equipment are compared with the corresponding preset health baseline, and if any of them is not within the corresponding preset health baseline range, it is determined that the target material conveying equipment has an equipment failure risk. The total energy consumption, energy consumption distribution, and energy consumption trend in the prediction sub-vector of the total energy consumption of the target material conveying equipment are compared with the theoretical optimal energy consumption benchmark of the corresponding current working condition, and if any of them deviates from the theoretical optimal energy consumption benchmark of the corresponding current working condition, it is determined that the target material conveying equipment has an energy consumption deviation risk.

5. The production integrated monitoring system based on big data according to claim 4, characterized in that: The specific content of the multi-objective collaborative optimization decision logic includes: A1, target function construction: a weighted sum target function is constructed, which includes three sub-targets of maximum conveying throughput, minimum equipment energy consumption, and minimum equipment comprehensive loss, and each sub-target is assigned a dynamically adjustable weight coefficient; A2, parameter optimization: when the target material conveying equipment has an operation risk, the target function is used as the optimization objective, and a global optimization calculation is performed within the feasible region of the preset control parameters using a particle swarm optimization algorithm to obtain a sequence of dynamic control parameters that optimizes the target function value, which is denoted as the optimal dynamic control parameter sequence. 6.The production integrated monitoring system based on big data according to claim 5, wherein: The optimal dynamic control parameter sequence specifically represents a set of collaborative instruction sets for different section driving units bound to the time axis; The collaborative instruction set includes: conveying belt speed adjustment parameters, feed quantity control parameters, distribution correction parameters, load adjustment parameters, tension dynamic correction parameters, cooling and lubrication control parameters, operation mode switching parameters, collaborative operation parameters, and regeneration energy consumption recovery parameters.

7. The production integrated monitoring system based on big data according to claim 6, characterized in that: The specific way of converting the optimal dynamic control parameter sequence into control instructions for the corresponding actuators and issuing them includes: The optimal dynamic control parameter sequence is classified according to the control objectives to obtain dynamic control parameters for each control objective of the target material conveying equipment, and the association between the dynamic control parameters and the corresponding actuators is established based on a database. The dynamic control parameters for each control objective of the target material conveying equipment are converted into specific values recognizable by the actuators according to the physical characteristics of the target material conveying equipment; The specific values recognizable by the actuators are converted into the corresponding format according to the actuator communication protocol and then issued as instructions. 8.The production integrated monitoring system based on big data according to claim 7, wherein: The system also includes a digital twin and visualization module that is bidirectionally connected to the state time series prediction module and the dynamic strategy generation module; The digital twin and visualization module specifically includes: B1, based on the geometric model, physical properties, and kinematics logic of the target material conveying equipment, a three-dimensional visual digital twin is constructed. B2, mapping the original operation data set, the multi-dimensional state time series prediction vector and the optimized dynamic control parameter sequence to the target material conveying equipment three-dimensional visual digital twin in real time; B3, before the dynamic control parameter sequence corresponding to the control instruction of the actuator is issued to the closed-loop control execution module, virtual operation deduction is carried out in the target material conveying equipment three-dimensional visual digital twin environment, the effect and potential secondary risk generated are evaluated, and when the deduction result meets the preset safety and efficiency standard, the control instruction is authorized to be issued to the closed-loop control execution module.

Images