A multi-source data fusion job control method and system

Abstract

The application relates to a multi-source data fusion operation control method and system, and relates to the technical field of port automation and intelligence. The method comprises the following steps: acquiring multi-source operation state data of ship unloader grab operation, extracting a feature vector, and generating an operation state matrix; based on a dynamics model of the ship unloader grab and the operation state matrix, a standard state equation of the ship unloader grab containing state variables and control variables is constructed; under the dynamic relationship defined by the standard state equation, a multi-objective function containing a path optimization target and an energy consumption control target is constructed, and a solving constraint condition of the multi-objective function is set; a non-dominated sorting genetic algorithm is used to optimize and solve the multi-objective function under the solving constraint condition, and an optimized control result of the grab operation is generated; and the optimized control result is issued to an executing mechanism of the ship unloader to control at least one ship unloader to execute the grab operation. By using the application, the automation level and operation efficiency of the ship unloader grab operation can be improved.

Description

Technical Field

[0001] This application relates to the field of port automation and intelligent technology, and in particular to an operation control method and system for multi-source data fusion. Background Technology

[0002] As the core loading and unloading equipment in bulk cargo terminals, the operation efficiency and energy consumption level of ship unloaders directly affect the port's operating costs and carbon emissions. Traditional ship unloader operation control systems mostly rely on manual operation or simple automation logic. The grab bucket path planning lacks comprehensive consideration of material distribution, equipment status and environmental disturbances, often resulting in problems such as long empty strokes, long operation cycles and high energy consumption. Summary of the Invention

[0003] Therefore, it is necessary to provide a multi-source data fusion operation control method and system to address the aforementioned technical problems.

[0004] Firstly, this application provides a job control method for multi-source data fusion, the method comprising: Acquire multi-source operation status data of the unloader grab bucket operation, extract feature vectors from the multi-source operation status data, and generate an operation status matrix. The multi-source operation status data includes at least two of the following: material characteristic data, operation environment data, equipment status data, and operation task data. Based on the dynamic model of the unloader grab and the operation state matrix, a standard state equation containing state variables and control variables is constructed for the unloader grab. Under the dynamic relationship defined by the standard state equation, a multi-objective function containing path optimization objective and energy consumption control objective is constructed, and the solution constraints of the multi-objective function are set. The multi-objective function is optimized under the given constraints using a non-dominated sorting genetic algorithm to generate optimized control results for the grab bucket operation. The optimized control results are sent to the actuators of the ship unloader to control at least one ship unloader to perform grab bucket operations. The actuators include at least a traveling mechanism, a hoisting mechanism, and a luffing mechanism.

[0005] Optionally, the feature vectors of the multi-source job status data are extracted to generate a job status matrix, including: The multi-source operation status data is cleaned and standardized; Feature vectors of the multi-source job status data are extracted using a convolutional neural network; The feature vectors are weighted by combining an attention mechanism to obtain a weighted feature vector. The weighted feature vectors at multiple time points are combined to generate the job state matrix.

[0006] Optionally, under the dynamic relationship defined by the standard state equations, a multi-objective function is constructed, comprising path optimization and energy consumption control objectives, and the solution constraints of the multi-objective function are set, including: Under the dynamic coupling relationship between state variables and control variables defined by the standard state equation, a path performance objective function based on the state trajectory and an energy consumption performance objective function based on control variables and state variables are constructed respectively. The solution constraint conditions of the multi-objective function, which includes the constraints of the standard state equation, kinematic boundary constraints and equipment operation boundary constraints, are established.

[0007] Optionally, the path performance objective function includes an operation time function and a path smoothness function, wherein the path smoothness is calculated by the curvature of the grab's motion trajectory.

[0008] Optionally, the energy consumption performance objective function is a unit loading and unloading volume energy consumption function, which includes the motor efficiency curve and corresponding terms of the operation stage. The operation stage includes at least an unloaded rising stage, a fully loaded falling stage, and a horizontal movement stage.

[0009] Optionally, the method further includes: The real-time execution data of the grab bucket operation of at least one ship unloader is obtained and compared with the optimized control results; When the deviation between the real-time execution data and the optimized control result is detected to be greater than a preset threshold, the non-dominated sorting genetic algorithm is used again to optimize the multi-objective function under the solution constraints to generate a new optimized control result for the grab operation.

[0010] Optionally, the method further includes: Based on the aforementioned operational status matrix, distributed model predictive control is used to divide the operational area, allocate task priorities, and generate grab paths for each ship unloader in order to avoid spatial and temporal conflicts among multiple ship unloaders.

[0011] Secondly, this application also provides a multi-source data fusion operation control system, the system comprising: A multi-source data acquisition unit is used to acquire multi-source operation status data of the unloader grab bucket operation, extract feature vectors from the multi-source operation status data, and generate an operation status matrix. The multi-source operation status data includes at least two of the following: material characteristic data, operation environment data, equipment status data, and operation task data. The state equation construction unit is used to construct the standard state equation containing state variables and control variables of the unloader grab bucket based on the dynamic model of the unloader grab bucket and the operation state matrix. The objective function setting unit is used to construct a multi-objective function containing path optimization objective and energy consumption control objective under the dynamic relationship defined by the standard state equation, and to set the solution constraints of the multi-objective function; The control result solving unit is used to optimize the multi-objective function under the solved constraints using a non-dominated sorting genetic algorithm to generate optimized control results for the grab bucket operation. The control result execution unit is used to send the optimized control result to the execution mechanism of the ship unloader, and control at least one ship unloader to perform grab bucket operation. The execution mechanism includes at least a traveling mechanism, a hoisting mechanism and a luffing mechanism.

[0012] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in the first aspect.

[0013] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0014] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0015] This application employs a multi-source data fusion-based operation control method. By fusing and unifying multi-source operation status data, the ship unloader's operation control system can simultaneously consider multiple factors such as material characteristics, operating environment, equipment operating status, and operation tasks. This improves the overall perception capability and data representation completeness of the ship unloader's grab bucket operation status. By constructing a standard state equation based on a dynamic model, a clear dynamic constraint relationship is established between the grab bucket's motion state and the control input, providing a unified mathematical model foundation for operation path planning and energy consumption control. Furthermore, by establishing a multi-objective function that includes path optimization and energy consumption control objectives, and... The non-dominated sorting genetic algorithm is used for optimization, which can coordinate the optimization of operation efficiency and equipment energy consumption under the condition of satisfying equipment operation constraints, thereby obtaining an optimized control result that takes into account both operation efficiency and energy consumption level. At the same time, by directly sending the optimized control result to the actuator of the ship unloader, the optimization result can be transformed into specific equipment control commands, realizing coordinated control of grab bucket travel, lifting and luffing movements. This can improve the automation level and operation efficiency of the grab bucket operation of the ship unloader, reduce overall energy consumption while ensuring stable equipment operation, and enhance the adaptability of the ship unloader operation control system in complex operating environments. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a multi-source data fusion job control method in one embodiment; Figure 2 This is a schematic diagram of the framework structure of a multi-source data fusion operation control system in one embodiment; Figure 3 This is a schematic diagram of the framework structure of a multi-source data fusion operation control system in one embodiment; Figure 4 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] This application provides a multi-source data fusion-based operation control method, which can be applied to the operation control system of a ship unloader at a port bulk cargo terminal. Ship unloader operation control systems are typically used for the transfer of bulk materials such as coal, ore, and grain from transport vessels to the terminal yard or conveyor system. In actual operation scenarios, the ship unloader grab needs to frequently reciprocate between the ship hold and the unloading point. Its movement trajectory, operation rhythm, and equipment operating status are closely related to unloading efficiency, equipment energy consumption, and operational safety. Simultaneously, due to the complex structure of the ship hold and the significant impact of wind speed and material conditions on the operating environment, the grab is prone to problems such as unreasonable path design, excessive energy consumption, or significant equipment impact during operation. Therefore, it is necessary to perform coordinated optimization control of the ship unloader grab's movement trajectory and equipment energy consumption while ensuring unloading efficiency and operational safety, in order to improve the overall operating efficiency of the unloading operation and reduce equipment energy consumption.

[0019] The ship unloader can include a traveling mechanism, a luffing mechanism, a hoisting mechanism, and a grab bucket device, used to grab and transport materials between the ship hold and the unloading area. The ship unloader's operating system can collect and process real-time data on equipment operating status, environmental parameters, and material characteristics during the unloading process, and generate corresponding grab bucket motion control strategies based on the processing results. During actual operation, multi-source data such as grab bucket position, speed, motor operating parameters, material weight, and ambient wind speed can be collected and then fused and analyzed to obtain the comprehensive operating status of the current unloading operation. After obtaining the comprehensive operating status, the grab bucket's motion trajectory and motor output power can be adjusted according to preset path optimization and energy consumption control strategies, thereby achieving coordinated optimization control of the grab bucket's operating path and equipment energy consumption.

[0020] The following combination Figure 1 The multi-source data fusion job control method provided in the embodiments of this application will be described in detail, wherein the uppercase and lowercase forms of the same letter refer to different contents, and may specifically include the following steps: Step S1: Obtain multi-source operation status data of the unloader grab bucket operation, extract the feature vector of the multi-source operation status data, and generate the operation status matrix.

[0021] The multi-source operational status data includes at least two of the following: material characteristic data, operational environment data, equipment status data, and operational task data. Material characteristic data may include parameters such as material density, humidity, particle size distribution, or actual load weight of the grab bucket; operational environment data may include environmental parameters such as wind speed, wind direction, temperature, or dust concentration; equipment status data may include operating parameters such as grab bucket spatial coordinates, motor speed, motor current, wire rope tension, or luffing angle; and operational task data may include task information such as loading and unloading volume, target unloading location, or operational priority.

[0022] In implementation, multi-source operational status data can be used to reflect the comprehensive operating status of the unloader's grab bucket during the current operation. These data sources can include material characteristic data, operating environment data, equipment status data, and task data. By unifying the representation of data from different sources, a status description can be formed for subsequent path optimization and energy consumption control model calculations. In actual implementation, this multi-source operational status data can be collected through various sensors and system interfaces installed on the unloader. For example, the grab bucket weight can be obtained through a weight sensor installed in the lifting mechanism, the grab bucket spatial position can be obtained through an encoder or positioning sensor, the ambient wind speed can be obtained through a wind speed sensor in the dock area, and the motor operating status can be collected through a frequency converter or controller.

[0023] After acquiring multi-source job status data, various data types can be synchronized according to timestamps, and different data types can be organized into a unified data vector format. Let the multi-source job status data collected at a certain moment be:

[0024] in, This represents the i-th state parameter.

[0025] To facilitate subsequent calculations, the multi-source job status data can be converted into feature vector form:

[0026] in, Let represent the j-th feature extracted from the original state data, or it can be represented as . .

[0027] After continuously collecting data at multiple time points, the feature vectors from each time point can be combined to form a job status matrix:

[0028] Where k represents the number of sampling times and q represents the number of feature dimensions.

[0029] Using the above method, operational data from different sources and of different types can be uniformly represented as an operational status matrix, thereby providing input data for subsequent grab path optimization and energy consumption coordinated control.

[0030] In one embodiment, the process of generating the job state matrix in S1 can be as follows: cleaning and standardizing the multi-source job state data, extracting the feature vectors of the multi-source job state data through a convolutional neural network, assigning weights to the feature vectors using an attention mechanism to obtain weighted feature vectors, and combining the weighted feature vectors from multiple time points to generate the job state matrix.

[0031] In practice, multi-source operational status data may contain noise, outliers, or dimensional differences during actual data collection. Therefore, data preprocessing is necessary. Specifically, the 3σ criterion can be used to remove outliers from continuous data, linear interpolation can be used to fill missing values ​​in continuous data, the mode can be used to fill missing values ​​in discrete data, and data standardization can map each data point to the [0,1] interval. First, the original data is cleaned using outlier detection methods, such as using the mean and standard deviation to identify outliers. When a data point meets the outlier detection criteria... At that time, among them for abbreviation, The mean of all data. The standard deviation of all data points can be used to identify outliers and remove or correct them. Subsequently, data from different dimensions can be standardized to bring them into a uniform numerical range. For example, normalization can be used.

[0032] in, This represents the standardized data. , The maximum and minimum values ​​of this feature are identified. After data preprocessing, the standardized multi-source job status data can be organized into a multi-channel input matrix according to the time series, so that a convolutional neural network can be used to extract features from the multi-source job status data. The convolutional neural network performs local feature learning on the input data through convolution operations, and includes three convolutional layers (with kernel sizes of 5, 3, and 3, and channel numbers of 64, 128, and 256 respectively), the ReLU activation function, and max pooling downsampling; the output is a low-dimensional feature vector (dimension 128) for each data source.

[0033] Its convolution operation can be represented as:

[0034] in, For input features, For convolution kernel weights, For bias terms, This represents the convolution operation. This is the activation function.

[0035] Convolution operations can be used to extract representative feature vectors from multi-source job status data. Furthermore, to improve the effectiveness of different types of features in representing job status, an attention mechanism can be introduced to assign weights to each feature vector. In the attention mechanism, the feature vectors are mapped to the query matrix Q, the key matrix K, and the value matrix V, respectively, and their calculation formulas are as follows: ,in, , , These are the query mapping weight matrix, key mapping weight matrix, and value mapping weight matrix, respectively, which are obtained through neural network training. Then, the correlation weights between features are calculated using a scaled dot product attention mechanism.

[0036] Let A represent the dimension of the key vector, and let A represent the attention weight matrix, whose elements are... This represents the importance of the i-th feature to the j-th feature. In scalar form, the importance weight of each feature can be calculated:

[0037] in, This represents the attention score for the i-th feature. This represents the normalized weight coefficients.

[0038] Finally, the weighted feature vector can be obtained: By combining the weighted feature vectors from consecutive sampling times, the final job state matrix can be formed:

[0039] The above methods can effectively integrate multi-source operational status data, enabling the generated operational status matrix to more accurately reflect the comprehensive operational status of the current unloader grab bucket operation, thereby providing a more reliable data foundation for subsequent path optimization and energy consumption collaborative control models.

[0040] Step S2: Based on the dynamic model and operational state matrix of the unloader grab, construct the standard state equations of the unloader grab, which include state variables and control variables.

[0041] During implementation, the grab bucket of the ship unloader is affected by various factors such as motor driving force, gravity, frictional resistance, wind load, and equipment structural parameters during operations such as material handling, lifting, horizontal movement, and unloading. Therefore, a dynamic model of the grab bucket's motion can be established to describe its motion law in space. To enable unified solutions for subsequent path optimization and energy consumption control models, the dynamic relationship can be further transformed into a standard state equation containing state variables and control variables. State variables describe the system's operating state at any given time and typically include physical quantities such as the grab bucket's position and velocity. Control variables describe the system's adjustable inputs, such as motor output force or driving force. The operational state matrix reflects the impact of the current operating environment, equipment status, and task conditions on the system parameters.

[0042] Specifically, consider the unloader's grab bucket as a moving body of mass m, whose motion obeys Newton's second law. Let the grab bucket's position in three-dimensional space be: , in, These are the horizontal position, the horizontal vertical position, and the vertical height, respectively.

[0043] The dynamic model of the grab's motion can be expressed as:

[0044] in, For motor driving force, For wind-borne disturbance force, For frictional resistance, For the gravity of the grab bucket.

[0045] In practical applications, external disturbances are related to factors such as equipment operating status and environmental conditions. Therefore, the system parameters can be corrected using the operating state matrix S generated in step S1, for example: ,in The wind load influence coefficient is determined by the operational state matrix. Let the ambient wind speed be . Similarly, frictional resistance can also be expressed as: ,in, The coefficient of friction is determined by the condition of the equipment.

[0046] To facilitate subsequent optimization, the dynamic model can be converted into a state-space form. The state variables are defined as follows: ,Right now ,in, For the spatial position of the grab, The grab speed. The control variable is defined as: This refers to the control force generated by the grab bucket drive motor.

[0047] Based on the above dynamic relationships, the second-order equations of motion can be transformed into first-order state-space equations: Specifically, it can be expanded as follows:

[0048] Further, write it in a unified form:

[0049] in, , representing the combined disturbance force determined by the operational state matrix.

[0050] By using the above method, we can obtain the standard state equation describing the dynamic characteristics of the grab bucket's motion. This equation reflects the dynamic relationship between state variables, control variables, and the operational state matrix.

[0051] Considering that different operating conditions will change the system dynamic parameters during actual operation, the model parameters can be adaptively corrected using the operating state matrix S, such as:

[0052]

[0053] in, For the basic quality of the grab bucket, The influence of material weight. The friction parameters of the equipment under standard operating conditions are given. This is a friction correction term used to describe the effect of different operating conditions on the friction coefficient.

[0054] Step S3: Under the dynamic relationship defined by the standard state equation, construct a multi-objective function that includes path optimization objective and energy consumption control objective, and set the solution constraints for the multi-objective function.

[0055] Specifically, under the dynamic coupling relationship between state variables and control variables defined by the standard state equation, a path performance objective function based on the state trajectory and an energy consumption performance objective function based on control variables and state variables are constructed respectively. Furthermore, the solution constraints of the multi-objective function, which includes constraints of the standard state equation, kinematic boundary constraints, and equipment operation boundary constraints, are established.

[0056] In implementation, the grab bucket dynamics state-space equations established in step S2 are as follows: This describes the dynamic relationship between the grab's motion state and the control input. Therefore, when performing path optimization and energy consumption control, it can be ensured that the optimization variables satisfy this dynamic relationship constraint.

[0057] Based on this, a multi-objective optimization model is constructed: ,in, The objective function for path performance is... The objective function is energy consumption performance, and the solution constraints for multiple objective functions are established simultaneously:

[0058] in, Represents the feasible region of the state variable; This represents the feasible region of the control variables. In the actual solution process, this constraint can be further included to ensure that the optimization results conform to the dynamic laws of the equipment, such as standard state equation constraints, kinematic boundary constraints (e.g., limits on the grab's speed and acceleration), and equipment operation boundary constraints (e.g., safety distance limits between the grab and surrounding equipment and structures). In this way, a unified multi-objective optimization solution model can be formed.

[0059] In this embodiment, the multi-objective function is decomposed into a path optimization sub-model and an energy consumption control sub-model. The path performance objective function can be expressed as: , The energy consumption performance objective function for the formation of the grab's motion trajectory can be expressed as: The overall optimization problem can be expressed as: And at least satisfy the above-mentioned standard state equation constraints, kinematic boundary constraints and equipment operation boundary constraints. The equipment operation boundary constraints may include the safe distance between the grab bucket and the adjacent unloader, the distance between the grab bucket and the ship's bulkhead, and the grab bucket tilt angle constraints. The range of the grab bucket tilt angle can be determined according to the material's moisture content to prevent material spillage during wet material operations.

[0060] Furthermore, the path performance objective function includes an operation time function and a path smoothness function, with path smoothness calculated using the curvature of the grab bucket's trajectory. Parallel to this, the energy consumption performance objective function is a unit loading / unloading energy consumption function, which includes the motor efficiency curve and corresponding terms for the operation stages. The operation stages at least include an unloaded ascent stage, a fully loaded descent stage, and a horizontal movement stage.

[0061] In implementation, the path optimization sub-model aims to minimize operation time and maximize path smoothness, and its objective function can be expressed as: ,in This indicates the operation time required for the grab bucket to complete one loading and unloading cycle. This indicates path smoothness. Path smoothness can be calculated from the curvature change of the grab trajectory:

[0062] Where L represents the length of the grab's trajectory, and s represents the trajectory arc length parameter. This represents the trajectory curvature, and the formula for calculating trajectory curvature is:

[0063] in, The first derivative of the trajectory is represented. , It represents the second derivative of the trajectory.

[0064] In actual calculations, a continuous trajectory function can be obtained by performing cubic spline interpolation on discrete trajectory points, and the curvature value can be obtained by differentiating this function. This method can avoid too many sharp turns or sudden speed changes in the grab bucket trajectory, thereby improving the operational stability of the equipment.

[0065] The energy consumption control sub-model aims to minimize energy consumption per unit of loading and unloading volume, and its objective function is expressed as:

[0066] Where E represents the total energy consumption of the equipment during the operation cycle, and Q represents the amount of material loaded and unloaded.

[0067] Equipment energy consumption can be expressed as:

[0068] Among them, P This represents the motor output power, and T represents the operating time.

[0069] The output power of a motor can be determined by its efficiency curve.

[0070] in, This indicates the motor's output torque. Indicates the angular velocity of the motor. The motor efficiency curve is represented. The motor efficiency curve is calibrated through motor bench testing and stored in the device's local database.

[0071] In actual control, the motor output power can be dynamically adjusted according to different operational stages. During the unloaded lifting stage, the grab bucket is relatively light, so the output torque of the lifting motor can be appropriately reduced to utilize gravity-assisted movement, thereby reducing energy consumption. During the fully loaded descent stage, the energy feedback braking device can be activated to convert the potential energy generated during the grab bucket's descent into electrical energy and feed it back to the power grid. During the horizontal movement stage, an S-shaped speed curve control strategy can be adopted to reduce the impact during start-up and stopping, thereby reducing energy consumption. Through these methods, energy consumption per unit loading and unloading volume can be reduced while ensuring operational efficiency.

[0072] Step S4: The non-dominated sorting genetic algorithm is used to optimize the multi-objective function under the constraints, and the optimized control results of the grab bucket operation are generated.

[0073] In implementation, to solve the multi-objective optimization model established in step S3, the Non-dominated Sorting Genetic Algorithm II (NSGA-II) can be used to optimize the multi-objective function. This algorithm simulates a natural evolutionary process, iteratively searching for path planning and control variables while satisfying system constraints, thereby obtaining a set of Pareto optimal solutions. The multi-objective optimization problem can be expressed as:

[0074] Under the constraints established in S3, the NSGA-II algorithm can be used to search for the optimization variables and obtain a set of Pareto optimal solutions that satisfy different objective trade-offs. The specific steps are as follows: First, generate an initial population randomly: Where N represents the population size, Let represent the i-th individual solution, which includes path trajectory parameters and control variables. In this embodiment, the population size is set to N=100.

[0075] Then, the population is ordered by non-dominated hierarchy, dividing individuals into multiple Pareto fronts. If solution A is not inferior to solution B on all objective functions, and is superior to solution B on at least one objective function, then solution A is said to dominate solution B. Based on this dominance relationship, the population can be divided into multiple non-dominated tiers: ,in, This is the first Pareto front. Let be the solution set of the r-th layer.

[0076] To maintain the uniformity of solution distribution, the crowding distance can be calculated for individuals within the same Pareto level:

[0077] in, Let M represent the crowding distance of the i-th individual, and M represent the number of objective functions. Let represent the nth objective function. Individuals with larger crowding distances are preferentially retained to ensure the diversity of the solution set.

[0078] Next, genetic operations such as tournament selection, crossover, and mutation are used to generate a new generation population. The crossover probability is set to... The mutation probability is set to New offspring populations are generated through crossover and mutation. Then, the parent population and the offspring population are merged to obtain: The algorithm then re-sorts and recalculates the non-dominated population, selecting the top N individuals to form a new generation. In this embodiment, the maximum number of iterations can be set to 100. The algorithm terminates when the maximum number of iterations is reached or the optimization result converges.

[0079] Thus, after iterative calculations, a set of Pareto optimal solutions can be obtained: h is the total number of solutions in the set, where each solution... A set of grab control parameters is provided, including the optimal grab trajectory, hoisting motor output power curve, horizontal movement speed curve, and operation stage control strategies. Ultimately, the optimal control scheme can be selected from the Pareto solution set based on operational requirements, and this optimized control result is output to the unloader's operation control system, thereby achieving intelligent optimized control of the grab operation process.

[0080] Step S5: The optimized control results are sent to the actuators of the unloader to control at least one unloader to perform grab bucket operations. The actuators include at least a traveling mechanism, a hoisting mechanism, and a luffing mechanism.

[0081] In implementation, after optimizing the multi-objective function using a non-dominated sorting genetic algorithm in step S4, the optimized control result of the grab bucket operation can be obtained. The optimized control result is essentially a set of optimal control variables satisfying the constraints of the standard state equations, which can be expressed as:

[0082] And the corresponding optimal state trajectory:

[0083] in, This represents the optimized output torque sequence of the hoisting motor. This represents the sequence of grab bucket operating speeds. This represents a sequence of variable angles.

[0084] In the actual control process, the ship unloader operation control system first performs time discretization processing on the optimized continuous control variables to form a discrete control sequence under the control period: ,in This indicates the control cycle number. Within each control cycle, the ship unloader operation control system adjusts its operation based on the current system state. With the target state The deviation generates an execution control signal, which is then sent to the corresponding actuator via the equipment control bus. Upon receiving the control signal, the actuator drives the mechanical structure to complete the corresponding action, thereby gradually bringing the system state closer to the optimized trajectory. During the grab's spatial movement, the grab's spatial position is determined by the coordinated movement of the traveling mechanism, the hoisting mechanism, and the luffing mechanism.

[0085] In the horizontal direction of motion, the traveling mechanism drives the unloader to move along the track via a traveling motor, and its position update relationship can be expressed as: ,in This indicates the travel speed. The ship unloader's operation control system calculates the target speed based on the optimized path. It is used as the speed control input for the walking motor, thereby enabling the grab to track the path on the horizontal plane.

[0086] In the vertical direction, the hoisting mechanism uses a hoisting motor to drive a winch system to raise and lower the grab bucket. The driving force of the hoisting motor is determined by its output torque. The decision, and its driving relationship, can be expressed as: ,in The moment of inertia of the motor. The angular velocity of the motor. This refers to the load torque. The lifting speed of the grab bucket can be controlled by adjusting the motor's output torque. This allows for changes in the height of the grab bucket. .

[0087] During the adjustment of the horizontal working range, the luffing mechanism changes the working radius of the grab bucket by controlling the luffing angle of the boom. The angle change relationship is as follows: ,in The variable angular velocity is used. The ship unloader's operation control system calculates the target angle based on the optimized trajectory. It also drives the luffing motor to adjust the boom angle, thereby changing the working range of the grab bucket in the horizontal plane.

[0088] Through the above control process, the optimized control results can be transformed into specific drive commands for the unloader's actuators. Under the scheduling of the unloader's operation control system, the traveling mechanism, lifting mechanism, and luffing mechanism are driven to move in coordination, thereby completing the grab, lifting, horizontal movement, and unloading operations, and realizing the automation and optimized control of the unloader's grab operation process.

[0089] In one implementation, during the process of controlling the unloader to perform grab bucket operations in step S5, the unloader operation control system can further monitor the execution status of the grab bucket operations in real time to achieve dynamic correction of the optimized control results. That is, it acquires the real-time execution data of the grab bucket operations of at least one unloader and compares it with the optimized control results. When it is detected that the deviation between the real-time execution data and the optimized control results is greater than a preset threshold, the optimized control results of the grab bucket operations are re-solved and generated.

[0090] Specifically, the ship unloader operation control system can continuously acquire real-time execution data of the grab bucket operation through sensors and status acquisition modules installed on the ship unloader. The real-time execution data includes at least the ship unloader's travel position, lifting height, luffing angle, grab bucket opening and closing status, and actual unloading volume per unit time. The aforementioned real-time execution data is input to the status evaluation module of the ship unloader operation control system, and forms a real-time status vector corresponding to the time variable t according to a preset sampling period.

[0091] The ship unloader operation control system can compare and analyze the real-time state vector with the optimized control result generated in step S4. The optimized control result is represented by a control command vector, which describes the target control state of the ship unloader at time t. The ship unloader operation control system can calculate the deviation between the actual operating state and the target control state based on real-time execution data, and obtain a comprehensive deviation index through a preset deviation evaluation function. When the comprehensive deviation index is less than or equal to a preset threshold, the ship unloader operation control system can continue to drive the ship unloader to perform grab bucket operations according to the current optimized control result; when the comprehensive deviation index is detected to be greater than the preset threshold, the ship unloader operation control system can determine that there is a significant deviation between the current operating state and the optimized control result.

[0092] Upon detecting a deviation exceeding a threshold, the ship unloader's operation control system can input the latest collected real-time execution data as an updated equipment status into the optimization solution module to re-solve the grab bucket operation control strategy. Based on the updated equipment status parameters, remaining tasks, and current operating environment information, the optimization solution module recalculates the control command vector, generating a new optimized control result for the grab bucket operation. This new result is then redistributed to the ship unloader's actuators for dynamic adjustments to subsequent grab bucket operations. Through this method, the ship unloader can continuously adjust its control strategy based on real-time operating conditions, achieving adaptive optimization control of the grab bucket operation and improving the stability and overall efficiency of the unloading operation.

[0093] In one implementation, after generating the optimized control results for the unloader's grab bucket operation in step S4, the unloader operation control system can further perform collaborative planning of the unloader's grab bucket operation area based on the operation state matrix to avoid spatial and temporal conflicts between multiple unloaders during operation. Specifically, based on the operation state matrix, distributed model predictive control is used to divide the operation area, allocate task priorities, and generate grab bucket paths for each unloader to avoid spatial and temporal conflicts between multiple unloaders.

[0094] Specifically, the ship unloader operation control system can first comprehensively analyze the spatial distribution of materials to be unloaded in the hold, the current position of each ship unloader, and the current operation progress based on the operation status matrix obtained and constructed in steps S1 to S3. Then, it uses a distributed model predictive control algorithm to divide the overall operation area. By predicting the operating status of the ship unloaders over several future control cycles, the ship unloader operation control system can divide the hold operation space into multiple sub-regions. Based on the remaining cargo volume in each sub-region, the accessibility of the ship unloaders, and the current equipment load, it can allocate corresponding operation areas and task priorities to each ship unloader, thus forming a regional allocation scheme for multi-ship unloader collaborative operation.

[0095] After the area division is completed, the ship unloader operation control system can predict and plan the grab bucket operation path according to the corresponding operation area and task priority of each ship unloader. Specifically, the ship unloader operation control system can use a distributed model predictive control framework to perform rolling optimization calculations on the motion state of each ship unloader in the future control time domain, and generate the corresponding grab bucket motion path, so that each ship unloader can complete the grab, lifting, rotating and unloading operations within its assigned area.

[0096] To further improve the real-time performance of path planning, the aforementioned calculation process can be completed in an edge computing module. This edge computing module can employ an industrial-grade edge gateway deployed within the ship unloader's control cabinet, with a sampling frequency set to 50Hz to continuously collect the ship unloader's actual operating data. This actual operating data includes at least the actual operating current of the motor and the real-time spatial position parameters of the grab bucket, and is input in real-time to the model predictive control module to update the equipment status.

[0097] During path planning and execution, the ship unloader operation control system can continuously monitor the ship unloader's operating deviations by comparing and analyzing real-time acquired data with predicted states. When the deviation between the actual motor current and the predicted current exceeds ±5%, or the deviation between the actual spatial position of the grab bucket and the planned position exceeds ±0.2m, the ship unloader operation control system can determine that there is a significant difference between the current execution state and the predicted state. Based on the updated equipment state, it will re-execute the model predictive control solution, thereby dynamically adjusting the grab bucket path to ensure that the grab bucket's motion trajectory is consistent with the optimized control results.

[0098] Through the above methods, during the division of the work area and the generation of the grab bucket path, the ship unloader operation control system can comprehensively consider the operational coordination relationship between multiple ship unloaders under the distributed model predictive control framework, realize the dynamic optimization planning of the grab bucket operation path, thereby effectively avoiding conflicts between multiple ship unloaders in terms of work space and work time, and improving the safety and efficiency of the ship unloading operation process.

[0099] This application employs a multi-source data fusion-based operation control method. By fusing and unifying multi-source operation status data, it enables the unloader to simultaneously consider multiple factors such as material characteristics, operating environment, equipment operating status, and operation tasks during operation. This improves the overall perception of the unloader's grab bucket's operating status and the completeness of data representation. By constructing a standard state equation based on a dynamic model, a clear dynamic constraint relationship is established between the grab bucket's motion state and the control input, providing a unified mathematical model foundation for operation path planning and energy consumption control. Furthermore, by establishing a multi-objective function that includes path optimization and energy consumption control objectives, and... The non-dominated sorting genetic algorithm is used for optimization, which can coordinate the optimization of operation efficiency and equipment energy consumption under the condition of satisfying equipment operation constraints, thereby obtaining an optimized control result that takes into account both operation efficiency and energy consumption level. At the same time, by directly sending the optimized control result to the actuator of the ship unloader, the optimization result can be transformed into specific equipment control commands, realizing coordinated control of grab bucket travel, lifting and luffing movements, thereby improving the automation level and operation efficiency of the grab bucket operation of the ship unloader, reducing overall energy consumption while ensuring stable equipment operation, and enhancing the adaptability of the ship unloader operation control system in complex operating environments.

[0100] The following description uses a typical embodiment as an example, specifically including the following contents: At a large bulk cargo terminal, a bridge-type grab unloader with a rated lifting capacity of 40 tons is equipped with the unloading operation method described in this application; the unloader's operation control system first acquires real-time information: the density sensor measures the coal density to be 1.3 t / m³. 3 The humidity sensor showed a moisture content of 8%, and the particle size analyzer identified an average particle size of 25mm. The wind speed and direction sensor detected a wind speed of 6m / s and a wind direction perpendicular to the track. The motor speed, current, wire rope tension, and grab bucket GPS coordinates were uploaded in real time. Meanwhile, the task of this operation was to unload 5,000 tons, with a priority of "high".

[0101] The collected data is processed by the edge computing module: abnormal current data are removed using the 3σ criterion, missing values ​​are filled by linear interpolation, and all data are normalized to the [0,1] interval; subsequently, a one-dimensional convolutional neural network extracts features from each data source, and dynamically weights them using an attention mechanism to generate a fused operation status matrix that comprehensively represents the current operation environment and equipment status.

[0102] Based on this operational state matrix, the ship unloader's operational control system constructs a joint optimization model for path and energy consumption: the path optimization sub-model aims to shorten operation time and improve trajectory smoothness, constraining the grab bucket's maximum speed to 3.2 m / s and maximum acceleration to 0.7 m / s². 2 The grab bucket tilt angle is set to no more than 20° according to the material moisture content; the energy consumption control sub-model aims to reduce energy consumption per unit loading and unloading volume, and dynamically adjusts the motor output power in different stages such as unloaded ascent, full-load descent, and horizontal movement based on the pre-calibrated motor efficiency curve; a non-dominated sorting genetic algorithm is used to solve the multi-objective optimization problem, and outputs the optimal grab bucket motion trajectory coordinate sequence and motor power control curve.

[0103] Control commands are issued to the three main mechanisms of traveling, hoisting, and luffing for execution. The edge computing module monitors the actual motor current and grab position at a frequency of 50Hz. When the current deviation exceeds ±5% or the position deviation exceeds ±0.2m (such as trajectory deviation caused by wind load disturbance), the re-optimization process is immediately triggered to re-execute data fusion and model solving to ensure control accuracy.

[0104] In multi-machine operation scenarios, the ship unloader operation control system collaborates with other ship unloaders through distributed model predictive control: dynamically divide the operation area according to the energy consumption level of each machine to ensure that the energy consumption difference rate is controlled within ±8%; when the spatial distance between two grab buckets is less than 2 meters and the time difference is less than 1 second, automatically adjust the operation sequence or plan an avoidance path to avoid collision risks.

[0105] In this embodiment, the ship unloader operation control system operates stably, with an increase in operation efficiency of about 18%, a reduction in unit energy consumption of about 12%, and a decrease in path conflict rate of more than 90%, verifying the effectiveness and advancement of the present invention under actual working conditions.

[0106] In summary, this unloading operation method, through in-depth processing and analysis of multi-source heterogeneous data using intelligent algorithms, achieves grab bucket path optimization and effective energy consumption control, significantly improving operational efficiency and energy utilization. It not only dynamically responds to abnormal situations but also supports efficient collaborative operations among multiple devices, significantly reducing operational risks and equipment conflicts, demonstrating its enormous potential and application value in improving the intelligence level of port loading and unloading operations.

[0107] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0108] Based on the same inventive concept, such as Figure 2 As shown in the illustration, this application also provides a multi-source data fusion operation control system 200, the system comprising: The multi-source data acquisition unit 201 is used to acquire multi-source operation status data of the unloader grab bucket operation, extract feature vectors from the multi-source operation status data, and generate an operation status matrix. The multi-source operation status data includes at least two of the following: material characteristic data, operation environment data, equipment status data, and operation task data. The state equation construction unit 202 is used to construct the standard state equation containing state variables and control variables of the unloader grab bucket based on the dynamic model of the unloader grab bucket and the operation state matrix. The objective function setting unit 203 is used to construct a multi-objective function containing path optimization objective and energy consumption control objective under the dynamic relationship defined by the standard state equation, and to set the solution constraints of the multi-objective function; The control result solving unit 204 is used to optimize the multi-objective function under the solving constraints using a non-dominated sorting genetic algorithm to generate optimized control results for the grab bucket operation. The control result execution unit 205 is used to send the optimized control result to the execution mechanism of the unloader, and control at least one unloader to perform grab bucket operation. The execution mechanism includes at least a traveling mechanism, a hoisting mechanism and a luffing mechanism.

[0109] In one embodiment, the multi-source data acquisition unit 201 is specifically used for: The multi-source job status data is cleaned and standardized, and feature vectors of the multi-source job status data are extracted by a convolutional neural network. Weights are then assigned using an attention mechanism to generate a job status matrix.

[0110] In one embodiment, the objective function setting unit 203 is specifically used for Under the dynamic coupling relationship between state variables and control variables defined by the standard state equation, a path performance objective function based on the state trajectory and an energy consumption performance objective function based on control variables and state variables are constructed respectively. The solution constraint conditions of the multi-objective function, which includes the constraints of the standard state equation, kinematic boundary constraints and equipment operation boundary constraints, are established.

[0111] In one embodiment, the path performance objective function includes an operation time function and a path smoothness function, wherein the path smoothness is calculated by the rate of change of curvature of the grab's motion trajectory.

[0112] In one embodiment, the energy consumption performance objective function is a unit loading and unloading volume energy consumption function, which includes a motor efficiency curve and corresponding terms for the operation stage. The operation stage includes at least an unloaded ascent stage, a fully loaded descent stage, and a horizontal movement stage.

[0113] In one embodiment, the control result solving unit 204 is further configured to: The real-time execution data of the grab bucket operation of at least one ship unloader is obtained and compared with the optimized control results; When the deviation between the real-time execution data and the optimized control result is detected to be greater than a preset threshold, the optimized control result of the grab bucket operation is re-solved and generated.

[0114] In one embodiment, such as Figure 3 As shown, the operation control system 200 also includes a multi-machine collaborative control unit 206, used for: Based on the operation status matrix, distributed model predictive control is used to divide the operation area, allocate task priorities, and generate grab bucket paths for each ship unloader in order to avoid spatial and temporal conflicts among multiple ship unloaders.

[0115] In one embodiment, a computer device is provided, the internal structure of which can be shown as follows: Figure 4As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for data exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a multi-source data fusion job control method.

[0116] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0117] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0118] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0119] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0120] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited thereto.

[0121] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0122] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A job control method based on multi-source data fusion, characterized in that, The method includes: Acquire multi-source operation status data of the unloader grab bucket operation, extract feature vectors from the multi-source operation status data, and generate an operation status matrix. The multi-source operation status data includes at least two of the following: material characteristic data, operation environment data, equipment status data, and operation task data. Based on the dynamic model of the unloader grab bucket and the operation state matrix, a standard state equation containing state variables and control variables is constructed for the unloader grab bucket. The system parameters of the standard state equation are corrected by the operation state matrix. The system parameters include wind load influence coefficient, friction correction term and material weight influence. Under the dynamic relationship defined by the standard state equation, a multi-objective function is constructed, which includes path optimization objective and energy consumption control objective. The path optimization objective includes minimizing operation time and maximizing path smoothness, and the energy consumption control objective includes minimizing energy consumption per unit loading and unloading volume. The solution constraints of the multi-objective function are set. The multi-objective function is optimized under the given constraints using a non-dominated sorting genetic algorithm to generate optimized control results for the grab bucket operation. The optimized control results are sent to the actuators of the ship unloader to control at least one ship unloader to perform grab bucket operations. The actuators include at least a traveling mechanism, a hoisting mechanism, and a luffing mechanism.

2. The method according to claim 1, characterized in that, Extracting feature vectors from the multi-source job status data to generate a job status matrix includes: The multi-source operation status data is cleaned and standardized; Feature vectors of the multi-source job status data are extracted using a convolutional neural network; The feature vectors are weighted by combining an attention mechanism to obtain a weighted feature vector. The weighted feature vectors at multiple time points are combined to generate the job state matrix.

3. The method according to claim 1, characterized in that, Under the dynamic relationship defined by the standard state equations, a multi-objective function is constructed, comprising path optimization and energy consumption control objectives. The solution constraints for this multi-objective function are then defined, including: Under the dynamic coupling relationship between state variables and control variables defined by the standard state equation, a path performance objective function based on the state trajectory and an energy consumption performance objective function based on control variables and state variables are constructed respectively. The solution constraints of the multi-objective function, which includes standard state equation constraints, kinematic boundary constraints and equipment operation boundary constraints, are established.

4. The method according to claim 3, characterized in that, The path performance objective function includes an operation time function and a path smoothness function, with path smoothness calculated by the curvature of the grab's motion trajectory.

5. The method according to claim 3, characterized in that, The energy consumption performance objective function is a unit loading and unloading volume energy consumption function, which includes the motor efficiency curve and corresponding terms of the operation stage. The operation stage includes at least an unloaded rising stage, a fully loaded falling stage, and a horizontal movement stage.

6. The method according to claim 1, characterized in that, The method further includes: The real-time execution data of the grab bucket operation of at least one ship unloader is obtained and compared with the optimized control results; When the deviation between the real-time execution data and the optimized control result is detected to be greater than a preset threshold, the non-dominated sorting genetic algorithm is used again to optimize the multi-objective function under the solution constraints to generate a new optimized control result for the grab operation.

7. The method according to claim 1, characterized in that, The method further includes: Based on the aforementioned operational status matrix, distributed model predictive control is used to divide the operational area, allocate task priorities, and generate grab paths for each ship unloader in order to avoid spatial and temporal conflicts among multiple ship unloaders.

8. A multi-source data fusion operation control system, characterized in that, The system includes: A multi-source data acquisition unit is used to acquire multi-source operation status data of the unloader grab bucket operation, extract feature vectors from the multi-source operation status data, and generate an operation status matrix. The multi-source operation status data includes at least two of the following: material characteristic data, operation environment data, equipment status data, and operation task data. The state equation construction unit is used to construct a standard state equation containing state variables and control variables for the unloader grab bucket based on the dynamic model of the unloader grab bucket and the operation state matrix. The system parameters of the standard state equation are corrected by the operation state matrix. The system parameters include wind load influence coefficient, friction correction term and material weight influence amount. The objective function setting unit is used to construct a multi-objective function containing path optimization objective and energy consumption control objective under the dynamic relationship defined by the standard state equation. The path optimization objective includes minimizing operation time and maximizing path smoothness, and the energy consumption control objective includes minimizing energy consumption per unit loading and unloading volume. The unit also sets the solution constraints for the multi-objective function. The control result solving unit is used to optimize the multi-objective function under the solved constraints using a non-dominated sorting genetic algorithm to generate optimized control results for the grab bucket operation. The control result execution unit is used to send the optimized control result to the execution mechanism of the ship unloader, and control at least one ship unloader to perform grab bucket operation. The execution mechanism includes at least a traveling mechanism, a hoisting mechanism and a luffing mechanism.

9. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the multi-source data fusion operation control method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium has a computer program stored thereon, which, when executed by a processor, implements the multi-source data fusion operation control method according to any one of claims 1 to 7.

Images