Optimization method of drag head operating parameters based on dredging volume of trailing suction hopper dredger

Abstract

This invention discloses a method for optimizing the operating parameters of a trailing suction hopper dredger based on the dredging volume of a trailing suction hopper dredger. The method involves dividing the trailing suction hopper dredger into standard dredging volume time periods, obtaining the dredging volume fluctuation interval and the maximum dredging volume fluctuation interval under the drag head multimodal parameters, obtaining the maximum dredging volume group, and calculating a maximum constraint condition quartile diagram. Furthermore, the method uses the drag head pulling force of the trailing suction hopper dredger to match the preset time nodes of the drag head multimodal parameters with the preset time nodes of the maximum parameters. The method then determines the constraints of the paired preset time nodes of the drag head multimodal parameters and the maximum parameter preset time nodes, and determines the range of the drag head operating parameters. This method reduces the need for manual intervention and improves the continuity and safety of operations.

Description

Technical Field

[0001] The present invention relates to the field of parameter optimization of a trailing suction hopper dredger, and in particular to a method for optimizing the operating parameters of a drag head based on the dredging volume of the trailing suction hopper dredger. Background Art

[0002] This method for optimizing the operating parameters of a trailing suction hopper dredger (TSD) based on the dredging volume aims to improve dredging efficiency and quality by rationally adjusting the various operating parameters of the TSD. Trailing suction hopper dredgers are widely used in projects such as river dredging, port construction, and waterway desilting. Their primary task is to capture and inhale sediment using the TSD. The proper configuration of parameters such as the TSD's operating angle, grab speed, operating depth, and inclination directly impacts the amount of sediment excavated and the effectiveness of the operation. Optimizing these operating parameters to improve efficiency and reduce energy consumption has become crucial for dredger design and operation. With the continuous advancement of science and technology, a growing number of studies are focusing on improving dredging efficiency through multi-parameter optimization, driving dredging technology towards greater efficiency and environmental friendliness.

[0003] While existing methods for optimizing drag head operating parameters have improved dredger efficiency to a certain extent, their optimization process often focuses solely on a single operating parameter and fails to fully consider the synergistic effects between these parameters. Most methods rely solely on adjusting a single drag head parameter (such as the operating angle or grab speed) for optimization, neglecting the multimodal optimization of multiple parameters, including the drag head's operating angle, grab speed, depth, and inclination. In reality, these drag head parameters are not independent and interact in complex ways, meaning that single-parameter optimization often fails to fully realize the drag head's maximum dredging potential. For example, changes in the drag head's angle directly affect the grab's entry into the water, while the grab's speed is limited by the operating depth and inclination angle. The synergistic effect of these two factors is key to optimizing dredging efficiency.

[0004] While existing optimization methods have extensively studied single parameters, they often fail to consider the integrated optimization of the drag head's operating parameters, resulting in unsatisfactory results in complex operating environments. Optimization of the drag head's operating angle and grab speed often fails to account for both operating depth and inclination, overlooking the impact of environmental variations on excavation efficiency. For example, in environments with complex sediment layers or highly variable water flows, single-parameter optimization struggles to adapt to dynamic environmental changes, resulting in low excavation efficiency or unstable operations. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for optimizing the operating parameters of a drag head based on the dredging volume of a trailing suction hopper dredger. The method calculates two excavation objects separately for each preset time node, then compares the constraints based on the respective calculation results. Based on the range of the drag head operating parameters of the constraint comparison results, it is determined whether the preset time node has an error of more than 30%, and the volume collected at the preset time node is discarded in advance to reduce the error of subsequent calculations, thereby ensuring the accuracy of the final analysis results.

[0006] The embodiment of the present invention is implemented as follows: a method for optimizing operating parameters of a drag suction hopper dredger based on the dredging volume of the drag suction hopper dredger comprises the following steps: dividing the drag suction hopper dredger into standard dredging volume periods, obtaining a dredging volume fluctuation interval and a maximum dredging volume fluctuation interval under the drag suction hopper dredger multimodal parameters, wherein the dredging volume fluctuation interval under the drag suction hopper dredger multimodal parameters is used to preset the floating range of the drag suction hopper dredger multimodal parameters for the dredging volume per unit time of different drag suction hopper dredger structures, and the maximum dredging volume fluctuation interval is used to preset the floating range of the maximum dredging volume per unit time of different drag suction hopper dredger structures; obtaining all the preset time nodes of the drag suction hopper dredger multimodal parameters within the dredging volume fluctuation interval under the drag suction hopper dredger multimodal parameters, determining the excavation volume of all the preset time nodes of the drag suction hopper dredger multimodal parameters, forming a drag head multimodal parameter excavation volume group, and calculating the constraint conditions of each excavation volume in the drag head multimodal parameter excavation volume group, and obtaining the drag head multimodal parameters. Constraint quartile diagram; obtain all maximum parameter preset time nodes within the maximum dredging volume fluctuation range, and determine the excavation volume of all maximum parameter preset time nodes, form a maximum excavation volume group, and calculate the constraints of each excavation volume in the maximum excavation volume group to obtain a maximum constraint quartile diagram; use the drag head pulling force of the trailing suction dredger to match the drag head multimodal parameter preset time node and the maximum parameter preset time node, determine the respective constraints of the paired drag head multimodal parameter preset time node and the maximum parameter preset time node, and judge the drag head operation parameter range; if the constraint condition value of the drag head multimodal parameter preset time node in the drag head multimodal parameter constraint quartile diagram is the same as the constraint condition value of the maximum parameter preset time node in the maximum constraint quartile diagram, then the drag head multimodal parameter preset time node and the maximum parameter preset time node are used as the volume drag head operation parameter optimization management time points.

[0007] In an optional embodiment, the preset time node of the rake head multimodal parameter of the volume rake head operation parameter optimization point is calibrated as the target preset time node of the rake head multimodal parameter, and the rake head multimodal parameter unit time discrete curve is established by using all the rake head multimodal parameter preset time nodes. The rake head multimodal parameter unit time discrete curve is used to determine the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node, and the similarity of all the rake head multimodal parameter correlation preset points and the volume excavated by the rake head multimodal parameter target preset time node is compared. According to the comparison result, the rake head operation parameter range is used to determine whether the rake head multimodal parameter target is to be used. The preset time node is used as the time point for optimizing the operation parameters of the volume rake head; the maximum parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the maximum value target preset time node, and the maximum value unit time discrete curve is established using all the maximum parameter preset time nodes; the maximum value unit time discrete curve is used to determine the maximum value correlation preset point of the maximum value target preset time node; all the maximum value correlation preset points are compared with the volume excavated by the maximum value target preset time node for similarity, and according to the comparison result rake head operation parameter range, it is determined whether the maximum value target preset time node is used as the time point for optimizing the operation parameters of the volume rake head.

[0008] In an optional embodiment, the similarity comparison includes the following steps: determining a specific time period for collecting the excavation volume, wherein the specific time period for collecting includes excavation of dredging volume with different rake head structures, excavation of dredging volume with different wind speeds, and excavation of dredging volume with different sediment properties; determining the weights between the various influencing factors in the specific time period for collecting, calculating the weight factors of the weights of the various influencing factors in the specific time period for collecting the corresponding correlation preset point and the weight factors of the weights of the various influencing factors in the specific time period for collecting the corresponding target preset time node, obtaining the degree of correlation of all weight factors, and using the degree of correlation to calculate the similarity comparison result between the corresponding correlation preset point and the corresponding target preset time node.

[0009] Before obtaining the correlation degree of all weight factors, it also includes encoding all weight factors to obtain different numbers;

[0010] In an optional embodiment, after determining the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node using the rake head multimodal parameter unit time discrete curve, the following steps are also included: obtaining the Spearman correlation coefficient between the rake head multimodal parameter target preset time node and the rake head multimodal parameter correlation preset point; and discarding the rake head multimodal parameter correlation preset point whose Spearman correlation coefficient exceeds the rake head multimodal parameter preset range.

[0011] In an optional embodiment, discarding the preset points of the rake head multimodal parameter correlation whose Spearman correlation coefficient exceeds the preset range of the rake head multimodal parameter also includes the following steps: using the unit time discrete curve of the rake head multimodal parameter to determine the number of nodes S between the discarded preset points of the rake head multimodal parameter correlation and the target preset time node of the rake head multimodal parameter; when the S value exceeds the preset range, the discarded preset points of the rake head multimodal parameter correlation are added as the basis for comparison with the target preset time node of the rake head multimodal parameter.

[0012] In an optional embodiment, the added drag head multimodal parameter correlation preset point is assigned to the dredging drag head operation parameter correction coefficient, and the dredging drag head operation parameter correction coefficient is used as the calculation basis for substituting the drag head multimodal parameter correlation preset point for similarity comparison.

[0013] The optimization method of drag head operation parameters based on the dredging volume of the trailing suction hopper dredger includes:

[0014] a dredging standard volume period division unit, which is used to divide the dredging standard volume period of the trailing suction hopper dredger into periods of standard volume, and obtain the dredging volume fluctuation interval and the maximum dredging volume fluctuation interval under the drag head multimodal parameters, wherein the dredging volume fluctuation interval under the drag head multimodal parameters is used to preset the floating range of the drag head multimodal parameters of the dredging volume per unit time for different drag head structures of the trailing suction hopper dredger, and the maximum dredging volume fluctuation interval is used to preset the floating range of the maximum dredging volume per unit time for different drag head structures of the trailing suction hopper dredger;

[0015] A dredger drag head multimodal parameter constraint condition calculation unit is used to obtain all drag head multimodal parameter preset time nodes within the dredging volume fluctuation range under the drag head multimodal parameters, determine the excavation volume at all the drag head multimodal parameter preset time nodes, form a drag head multimodal parameter excavation volume group, calculate the constraint conditions of each excavation volume in the drag head multimodal parameter excavation volume group, and obtain a drag head multimodal parameter constraint condition quartile diagram;

[0016] A dredger maximum constraint condition calculation unit is used to obtain all maximum parameter preset time nodes within the maximum dredging volume fluctuation range, determine the excavation volume at all maximum parameter preset time nodes, form a maximum excavation volume group, and calculate the constraint conditions of each excavation volume in the maximum excavation volume group to obtain a maximum constraint condition quartile diagram;

[0017] The drag head pulling force matching unit is used to match the drag head multi-modal parameter preset time node and the maximum parameter preset time node using the drag head pulling force of the trailing suction hopper dredger, determine the respective constraint conditions of the paired drag head multi-modal parameter preset time node and the maximum parameter preset time node, and judge the drag head operation parameter range;

[0018] The rake head operation parameter range judgment unit is used for rake head operation parameter range judgment: if the constraint condition value of the rake head multimodal parameter preset time node in the rake head multimodal parameter constraint condition quartile diagram is the same as the constraint condition value of the maximum parameter preset time node in the maximum value constraint condition quartile diagram, then the rake head multimodal parameter preset time node and the maximum parameter preset time node are used as the volume rake head operation parameter optimization management time point.

[0019] In an optional embodiment, it also includes a maximum value rake head operation parameter range judgment unit, which is used to calibrate the rake head multimodal parameter preset time node of the volume rake head operation parameter optimization point as the rake head multimodal parameter target preset time node, use all the rake head multimodal parameter preset time nodes to establish a rake head multimodal parameter unit time discrete curve, use the rake head multimodal parameter unit time discrete curve to determine the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node, compare all the rake head multimodal parameter correlation preset points with the volume excavated by the rake head multimodal parameter target preset time node for similarity, and judge whether the rake head operation parameter range is within the range according to the comparison result. The target preset time node of the rake head multimodal parameter is used as the time point for optimizing and managing the volume rake head operation parameters; the maximum parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the maximum value target preset time node, and all the maximum parameter preset time nodes are used to establish a maximum value unit time discrete curve; the maximum value unit time discrete curve is used to determine the maximum value correlation preset point of the maximum value target preset time node; all the maximum value correlation preset points are compared with the volume excavated by the maximum value target preset time node for similarity, and according to the comparison result rake head operation parameter range, it is determined whether the maximum value target preset time node is used as the time point for optimizing and managing the volume rake head operation parameters.

[0020] Beneficial effects

[0021] This paper proposes a method for optimizing the operating parameters of a trailing suction hopper dredger (TSSD) based on the dredging volume. By systematically analyzing different parameter combinations, the method accurately sets the fluctuation range of each operating parameter, ensuring that different slurry head configurations achieve optimal dredging efficiency under various operating conditions. Specifically, by presetting the dredging volume fluctuation range and the maximum fluctuation range, the slurry head's performance is optimized in different operating scenarios, maximizing dredging efficiency. This method incorporates quartile analysis to deeply explore the interactions and influence coefficients between parameters, further optimizing their synergy. Comprehensive consideration of factors such as the slurry head's operating angle, grab speed, operating depth, and inclination effectively avoids inefficiencies or operational instability caused by adjusting a single factor. The visualization of the quartile plot clearly displays the constraints and optimal operating ranges of each parameter, providing intuitive decision-making for operators. Furthermore, the parameter presetting and tension matching mechanisms incorporated into the method significantly enhance the automation level of the operation, reduce the need for manual intervention, and improve operational continuity and safety. Overall, this comprehensive optimization method not only improves the operating efficiency of the trailing suction hopper dredger, but also provides a feasible technical framework for future automated and intelligent operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 is a flow chart of the steps of the method of the present invention;

[0024] Figure 2 This is a diagram of the operating unit composition of the method of the present invention. DETAILED DESCRIPTION

[0025] It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of this application can be combined with each other. The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] like Figure 1 As shown, the method for optimizing the drag head operating parameters based on the dredging volume of a trailing suction hopper dredger provided in this embodiment includes the following steps:

[0027] A1: Divide the trailing suction hopper dredger into standard dredging volume periods, and obtain the dredging volume fluctuation interval and the maximum dredging volume fluctuation interval under the drag head multimodal parameters, wherein the dredging volume fluctuation interval under the drag head multimodal parameters is used to preset the floating range of the drag head multimodal parameters of the dredging volume per unit time of different drag head structures of the trailing suction hopper dredger, and the maximum dredging volume fluctuation interval is used to preset the floating range of the maximum dredging volume per unit time of different drag head structures of the trailing suction hopper dredger; this step indicates determining the different drag head structures of the trailing suction hopper dredger. The trailing suction hopper dredger is primarily equipped with two excavation dredging volumes: a maximum floating range of dredging volumes per unit time for different drag head structures and a floating range of multimodal parameters for dredging volumes per unit time for different drag head structures. These are achieved at different preset time nodes arranged in the trailing suction hopper dredger. At each preset time node, simultaneous excavation of the maximum floating range of dredging volumes per unit time for different drag head structures and the floating range of multimodal parameters for dredging volumes per unit time for different drag head structures is performed at the corresponding point. Considering that the probability of normal or abnormal excavation occurring simultaneously at the same preset time node for both the maximum floating range of dredging volumes per unit time for different drag head structures and the floating range of multimodal parameters for dredging volumes per unit time for different drag head structures is higher, and the probability of only one of them being abnormal is lower, the excavation volumes of the two dredging volumes corresponding to each preset time node are separately extracted and compared longitudinally. This can further determine whether abnormal excavation is likely at that preset time node within the scope of the drag head operation parameters. The standard dredging volume period of the trailing suction hopper dredger is divided into the dredging volume fluctuation range under the multi-modal parameters of the drag head and the maximum dredging volume fluctuation range in order to facilitate subsequent longitudinal comparison.

[0028] A2: Obtain all preset time nodes of the drag head multimodal parameters within the dredging volume fluctuation range under the drag head multimodal parameters, determine the excavation volume at all these preset time nodes, form a drag head multimodal parameter excavation volume group, and calculate the constraints of each excavation volume within this drag head multimodal parameter excavation volume group to obtain a quartile plot of the drag head multimodal parameter constraints. This step uses the constraints of the excavation volume to achieve the aforementioned longitudinal comparison. Since this step does not determine whether there is a possibility of abnormal excavation at the preset time node, it is necessary to present and analyze the excavation volume at all preset time nodes of the drag head multimodal parameters. Similarly, the same treatment measures are applied to the maximum parameter preset time node.

[0029] A3: Obtain all maximum parameter preset time nodes within the maximum dredging volume fluctuation range, and determine the excavation volumes of all the maximum parameter preset time nodes to form a maximum excavation volume group, and calculate the constraints of each excavation volume in the maximum excavation volume group to obtain a maximum constraint condition quartile map; present and analyze the excavation volumes of all maximum parameter preset time nodes.

[0030] By presenting the quartile diagram of the multimodal parameter constraints of the rake head and the quartile diagram of the maximum value constraints, it is possible to find a comparison between the excavated volumes of the two dredging volumes at the same preset time node and the excavated volumes of the two dredging volumes at other preset time nodes. If the excavated volume constraint curves of the two dredging volumes at the same preset time node are the same, it means that the excavated volumes of the two dredging volumes are accurate or correct from the perspective of the preset time node.

[0031] A4: Use the drag head pulling force of the trailing suction hopper dredger to match the drag head multimodal parameter preset time node and the maximum parameter preset time node, determine the respective constraints of the paired drag head multimodal parameter preset time node and the maximum parameter preset time node, and judge the drag head operation parameter range; the drag head pulling force of the trailing suction hopper dredger here mainly refers to the situation that belongs to the same configuration and management entity. In practice, the floating range of the drag head multimodal parameters of the dredging volume per unit time of different drag head structures and the floating range of the maximum value of the dredging volume per unit time of different drag head structures may be excavated at different excavation sites, but the different excavation sites are all supervised and responsible by the same entity. It is necessary to combine the corresponding separate drag head multimodal parameter preset time node and the maximum parameter preset time node to perform the above-mentioned longitudinal comparison, especially to judge the drag head operation parameter range according to their respective constraints.

[0032] A5: If the constraint condition value of the rake head multimodal parameter preset time node in the rake head multimodal parameter constraint condition quartile diagram is the same as the constraint condition value of the maximum parameter preset time node in the maximum value constraint condition quartile diagram, then the rake head multimodal parameter preset time node and the maximum parameter preset time node are used as the volume rake head operation parameter optimization management time points; this step indicates that if the rake head multimodal parameter preset time node and the maximum parameter preset time node are matched to the same excavation point.

[0033] Through the above technical solution, only the excavation volume of the excavation points with high possibility of abnormality is discarded from the perspective of preliminary screening, and the retained excavation points may also have abnormal excavation situations, but compared with the method of applying the averaged excavation volume of all excavation points, the error is smaller and the volume acquisition is more reliable. In some embodiments, in order to further analyze the retained excavation points, in order to further achieve the purpose of judging whether the excavation point is abnormal within the range of the rake head operating parameters. The analysis method provided in this embodiment also includes step A6 and its sub-steps. A6 is a step of finding the correlation preset point by establishing a unit time discrete curve for similarity comparison, so as to achieve further screening of the rake head multimodal parameter preset time node and the maximum parameter preset time node,

[0034] Specifically: the preset time node of the rake head multimodal parameter of the volume rake head operation parameter optimization point is calibrated as the target preset time node of the rake head multimodal parameter, all the rake head multimodal parameter preset time nodes are used to establish a rake head multimodal parameter unit time discrete curve, the rake head multimodal parameter unit time discrete curve is used to determine the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node, all the rake head multimodal parameter correlation preset points are compared with the volume excavated by the rake head multimodal parameter target preset time node for similarity, and according to the comparison result, the rake head operation parameter range is judged whether the rake head multimodal parameter target preset time node is used. The node between the two nodes is used as the time point for optimizing the operation parameters of the volume rake head; the maximum parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the maximum value target preset time node, and the maximum value unit time discrete curve is established using all the maximum parameter preset time nodes; the maximum value unit time discrete curve is used to determine the maximum value correlation preset point of the maximum value target preset time node; all the maximum value correlation preset points are compared with the volume excavated by the maximum value target preset time node for similarity, and according to the comparison result rake head operation parameter range, it is determined whether the maximum value target preset time node is used as the time point for optimizing the operation parameters of the volume rake head.

[0035] Step A6 indicates that the corresponding preset time node obtained in step A5, which can be used as the optimization point of the volumetric scraper head operating parameters, may have two situations in which the excavation and dredging volumes are normal or abnormal at the same time. In order to further identify, a further longitudinal comparison method is adopted. Taking the preset time node of the scraper head multimodal parameter of the volumetric scraper head operating parameter optimization point as an example, it is calibrated as the target preset time node of the scraper head multimodal parameter, and the unit time discrete curve relationship is established between the target preset time node of the scraper head multimodal parameter and the other preset time nodes of the scraper head multimodal parameter to obtain the unit time discrete curve of the scraper head multimodal parameter. The unit time discrete curve of the scraper head multimodal parameter is used to find the other preset time nodes of the scraper head multimodal parameter that are directly related to the target preset time node of the scraper head multimodal parameter and use them as the preset point of the correlation of the scraper head multimodal parameter. This or different scraper head multimodal parameter preset time nodes are used as the correlation point of the scraper head multimodal parameter. The similarity comparison between the excavation volume per unit time of different drag head structures at the modal parameter correlation preset point and the floating range of the drag head multimodal parameters of different drag head structures at the target preset time node of the drag head multimodal parameters is performed. The volume comparison is performed using the principle of direct correlation (i.e., directly connected in the actual underground environment) between two excavation points. The principle of connectivity, where the floating range of the drag head multimodal parameters of different drag head structures theoretically remains the same, further determines whether the volume collection at the target preset time node of the drag head multimodal parameters is normal. The target preset time node for the maximum value is processed similarly and will not be further elaborated here.

[0036] By performing the above-mentioned volume longitudinal comparison of the target preset time node of the rake head multimodal parameter or the maximum value target preset time node separately or performing the volume longitudinal comparison of the two simultaneously, it is possible to further identify whether the mining of the target preset time node of the rake head multimodal parameter or the maximum value target preset time node is abnormal, so as to judge whether the rake head operation parameter range can be used as the volume rake head operation parameter optimization management time point. For example, if it is found that the volume similarity between the target preset time node of the rake head multimodal parameter and the volume preset points of different rake head multimodal parameter correlations is low, the mining volume of the rake head multimodal parameter target preset time node is excluded to ensure that the volume source of the subsequent analysis has further reliability.

[0037] In this embodiment, the similarity comparison mainly performs differential analysis through smaller or more detailed indicators, and the similarity comparison includes the following steps:

[0038] B1: Determine the specific time period for collecting the excavation volume, wherein the specific collection period includes the excavation of dredging volume with different scraper head structures, the excavation of dredging volume with different wind speeds, and the excavation of dredging volume with different sediment properties; this step represents the method of determining the excavation volume obtained at the corresponding correlation preset point (the scraper head multimodal parameter correlation preset point or the maximum correlation preset point) and the corresponding target preset time node (the scraper head multimodal parameter target preset time node or the maximum target preset time node), wherein the correlation methods are mainly three types, the excavation of dredging volume with different scraper head structures, the excavation of dredging volume with different wind speeds, and the excavation of dredging volume with different sediment properties.

[0039] Then, step B2 proceeds: determining the weights of each influencing factor within the specific collection period, and calculating the weight factor between the weights of each influencing factor within the specific collection period at the corresponding preset correlation point and the weight factor of each influencing factor within the specific collection period at the corresponding preset target time node. This step compares the mining volume influencing factors between the corresponding preset correlation point and the corresponding preset target time node, further identifying the specific similarity mining indicator comparison. This utilizes the distance between the corresponding volumes to express the degree of similarity, thereby further determining the rake head operating parameter range judgment result for whether there is abnormal mining at the corresponding preset target time node. This leads to step B4: obtaining the correlation degree of all weight factors, and using this correlation degree to calculate the similarity comparison result between the corresponding preset correlation point and the corresponding preset target time node. This step involves observing and analyzing all weight factors, selecting the median based on the concentration, and then using this median selected by the concentration to determine the similarity of the rake head operating parameter range.

[0040] On the basis of the above technical solution, the following steps are also included before obtaining the correlation degree of all weight factors:

[0041] B3: Encode all weight factors and obtain different numbers;

[0042] The influencing factors of the different numbers are analyzed by graph neural network to obtain the mutual influence coefficient of the working efficiency of the multimodal parameters of the rake head, including:

[0043] 1. Encode weight factors and build graph structure. First, we need to encode all influencing factors. Assume there are four parameters: x 1: Working angle of rake head, x 2: Rake head grab speed, x 3: Depth of rake head, x4: Rake head tilt. These factors can be encoded numerically, such as normalized values or vectors, to generate feature vectors for each node. Each influencing factor becomes a node, and the relationships between these factors form graph edges. Edge weights can be calculated from historical data to represent the mutual influence between different nodes.

[0044] 2. Graph Neural Network Model Formula. The core idea of Graph Neural Networks (GNNs) is to update node states using information about their neighbors. In GNNs, we transfer and update node features through graph convolution operations. The following is a detailed formula description:

[0045] 2.1 Graph Construction, Assuming Graph G By node set V and edge sets E composition: ,Each node represents an influencing factor (working angle, grab speed, depth, inclination). E Represents the connection relationship between nodes, assuming that all nodes are connected to each other (complete graph).

[0046] 2.2 Node characteristics, each node v i (For example, the working angle of the rake head x 1) Both have an initial eigenvector h i (o) , representing a node v i The initial state of (usually a normalized numerical feature). For example: h 1 (o) = x 1 (eigenvector of the rake head working angle), h 2 (o) = x 2 (eigenvector of the drag head grab speed), h 3 (o) = x 3 (eigenvector of rake head depth), h 4 (o) = x 4 (eigenvector of rake head inclination).

[0047] 2.3 Graph Convolution Update Formula. The basic operation of graph neural networks is graph convolution. The key to graph convolution is to weight the neighbor information of a node and pass it to the node. The formula for the graph convolution operation is as follows:

[0048]

[0049] in: h i(k) :node v i In the k The feature vector of the layer. N( i ):node v i The set of neighbor nodes. W (k) : No. k The weight matrix of the layer is used to learn the transformation of node features. b (k) : No. k The bias term of the layer. σ (·): Activation function, commonly used are ReLU or Sigmoid.

[0050] The meaning of the formula is: each node i Aggregate its neighbor nodes j The features and weight matrix W (k) After the transformation, the features are updated. By stacking multiple graph convolutional layers, the complex relationships between nodes can be gradually captured.

[0051] 2.4 Edge Weights A ij Representation node v i and nodes v j The degree of influence between them is usually obtained from training data or expert experience. We can use an adjacency matrix A To represent the structure of the graph, A ij Representation node v i and v j The strength of the connection between them.

[0052] If the graph is complete (every node is connected to every other node), then the adjacency matrix A The non-zero elements in represent the connection strength, which may be a constant or calculated based on influencing factors.

[0053] 2.5 Multi-layer graph neural network, in order to better capture the high-order relationships between nodes, we usually use a multi-layer graph neural network. The node update formula for each layer is:

[0054]

[0055] in: H (k) It is k All node feature matrices of the layer, Ais the adjacency matrix of the graph (indicating the connectivity of edges), W (k) It is k The weight matrix of the layer, B (k) It is k The bias term of the layer.

[0056] The feature updates at each layer will cause nodes to propagate and aggregate information in the graph, gradually gaining an understanding of the coefficients that affect work efficiency.

[0057] 2.6 Output Layer: Through multi-layer graph neural network learning, we will eventually obtain the feature vector of each node (that is, each influencing factor). To obtain the final influence coefficient, we can use an output layer to map the node features to the influence coefficient of work efficiency:

[0058]

[0059] in: y i is a node v i The final influence coefficient of a certain influencing factor on work efficiency, h i (K) is through K Node after layer graph convolution v i characteristics. W out is the weight matrix of the output layer.

[0060] 3. Training process, loss function: To train this model, we can use a loss function to measure the difference between the work efficiency impact coefficient predicted by the model and the true value. For example, we can use the mean square error (MSE) loss function:

[0061]

[0062] - Optimization algorithm: Use common optimization algorithms, such as Adam optimizer, to update the weight matrix W (k) and W out .

[0063] This graph neural network model encodes influencing factors, constructs a graph structure, and uses a multi-layer graph convolutional network to learn the interactions between nodes. Ultimately, the model is able to predict the impact of each parameter (such as the drag head operating angle and grab bucket speed) on dredging efficiency, thereby supporting dredger operation optimization.

[0064] In actual similarity comparison calculations, there may be situations where the basic environments between adjacent nodes are different, especially the different rake head structures, the unit time dredging volume, and the floating range of the rake head multimodal parameters. In a geographical environment with certain faults or subsidence between the two nodes, the similarity of the excavation volume cannot be calculated by the connectivity principle due to the different rake head structures, the unit time dredging volume, and the floating range of the rake head multimodal parameters. At this time, the following steps are required, namely, after determining the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node using the rake head multimodal parameter unit time discrete curve, the following steps are also included: obtaining the Spearman correlation coefficient between the rake head multimodal parameter target preset time node and the rake head multimodal parameter correlation preset point; discarding the rake head multimodal parameter correlation preset point whose Spearman correlation coefficient exceeds the rake head multimodal parameter preset range. This step means discarding the rake head multimodal parameter correlation preset points with relatively large settlement between the rake head multimodal parameter correlation preset points and the rake head multimodal parameter target preset time nodes, and not using them as the basis for subsequent similarity comparison to ensure the rationality of the volume analysis. The rake head multimodal parameter preset range is predetermined and can be sufficiently small while ensuring the calculation accuracy.

[0065] On the basis of the above technical solution, taking into account that once all the preset points of the rake head multimodal parameter correlation exceed the preset range of the rake head multimodal parameter, there will be a situation where the sample of the preset points of the rake head multimodal parameter correlation is too small, which is not conducive to comprehensively obtaining a more reasonable volume similarity comparison result. That is, the preset points of the rake head multimodal parameter correlation that may cause a larger error in the excavation volume due to settlement can be discarded, and the rest can be supplemented. Then, in the process of discarding the preset points of the rake head multimodal parameter correlation whose Spearman correlation coefficient exceeds the preset range of the rake head multimodal parameter, the following steps are also included:

[0066] The unit time discrete curve of the rake head multimodal parameter is used to determine the number of nodes S between the discarded rake head multimodal parameter correlation preset point and the rake head multimodal parameter target preset time node. When the S value is less than or equal to the preset range, the discarded rake head multimodal parameter correlation preset point is added as the basis for comparison with the rake head multimodal parameter target preset time node.

[0067] This step represents the addition of rake head multimodal parameter correlation preset points for nodes that are far away but have reasonable settlement, while the rake head multimodal parameter correlation preset points for nodes that are close but have serious relative settlement are discarded. The preset range is also predetermined and can be sufficiently large while ensuring that the calculation meets the limit. Based on this solution, in order to further obtain a more accurate basis for volume calculation, the added rake head multimodal parameter correlation preset points are assigned a dredging rake head operating parameter correction coefficient. This dredging rake head operating parameter correction coefficient serves as the calculation basis for substituting this rake head multimodal parameter correlation preset point for similarity comparison. That is, the dredging rake head operating parameter correction coefficient is calculated based on reasonable settlement (for example, relative settlement caused by terrain difference), substituted as a weight and assigned to the weight factor, so that when performing similarity comparison calculations, a more reliable and reasonable basis for volume calculation is provided.

[0068] This embodiment also provides a method for optimizing the drag head operating parameters based on the dredging volume of the trailing suction hopper dredger, such as Figure 2 As shown, for example, each dredging volume unit can be divided, or two or more dredging volumes can be integrated into one processing unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software dredging volume unit. It should be noted that the division of units in the present invention is schematic and is only a logical dredging volume division. There may be other division methods in actual implementation. For example, in the case of dividing each dredging volume unit according to each dredging volume, it includes a dredging standard volume time period division unit, a dredger drag head multimodal parameter constraint condition calculation unit, a dredger maximum value constraint condition calculation unit, a drag head pulling force matching unit, and a drag head operating parameter range judgment unit. The dredging volume of each unit is explained below.

[0069] The dredging standard volume period division unit is used to divide the dredging standard volume period of the trailing suction hopper dredger, and obtain the dredging volume fluctuation interval and the maximum dredging volume fluctuation interval under the drag head multimodal parameters, wherein the dredging volume fluctuation interval under the drag head multimodal parameters is used to preset the floating range of the drag head multimodal parameters of the dredging volume per unit time of different drag head structures of the trailing suction hopper dredger, and the maximum dredging volume fluctuation interval is used to preset the floating range of the maximum dredging volume per unit time of different drag head structures of the trailing suction hopper dredger. The dredger drag head multimodal parameter constraint condition calculation unit is used to obtain all the drag head multimodal parameter preset time nodes within the dredging volume fluctuation range under the drag head multimodal parameter, and determine the excavation volume of all the drag head multimodal parameter preset time nodes to form a drag head multimodal parameter excavation volume group, and calculate the constraint conditions of each excavation volume in the drag head multimodal parameter excavation volume group to obtain the drag head multimodal parameter constraint condition quartile map; the dredger maximum value constraint condition calculation unit is used to obtain the All maximum parameter preset time nodes within the maximum dredging volume fluctuation range are determined, and the excavation volumes of all the maximum parameter preset time nodes are determined to form a maximum excavation volume group, and the constraint conditions of each excavation volume in the maximum excavation volume group are calculated to obtain a maximum constraint condition quartile map; a drag head pulling force matching unit is used to match the drag head multimodal parameter preset time node and the maximum parameter preset time node by using the drag head pulling force of the trailing suction dredger, determine the respective constraint conditions of the paired drag head multimodal parameter preset time node and the maximum parameter preset time node, and perform drag head operation parameter range judgment; a drag head operation parameter range judgment unit is used to use the drag head multimodal parameter preset time node and the maximum parameter preset time node as the volume drag head operation parameter optimization management time point if the constraint condition value of the drag head multimodal parameter preset time node in the drag head multimodal parameter constraint condition quartile map is the same as the constraint condition value of the maximum parameter preset time node in the maximum constraint condition quartile map.

[0070] In some embodiments, the method for optimizing the operating parameters of the drag head based on the dredging volume of the trailing suction hopper dredger further includes a maximum drag head operating parameter range judgment unit, which is used to calibrate the drag head multimodal parameter preset time node of the volume drag head operating parameter optimization point as the drag head multimodal parameter target preset time node, use all the drag head multimodal parameter preset time nodes to establish a drag head multimodal parameter unit time discrete curve, use the drag head multimodal parameter unit time discrete curve to determine the drag head multimodal parameter correlation preset point of the drag head multimodal parameter target preset time node, compare all the drag head multimodal parameter correlation preset points with the volume excavated by the drag head multimodal parameter target preset time node for similarity, and rake the volume according to the comparison result. The head operation parameter range determines whether to use the target preset time node of the rake head multimodal parameter as the volume rake head operation parameter optimization management time point; the maximum parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the maximum value target preset time node, and all the maximum parameter preset time nodes are used to establish a maximum value unit time discrete curve; the maximum value unit time discrete curve is used to determine the maximum value correlation preset point of the maximum value target preset time node; all the maximum value correlation preset points are compared with the volume excavated by the maximum value target preset time node for similarity, and according to the comparison result, the rake head operation parameter range determines whether to use the maximum value target preset time node as the volume rake head operation parameter optimization management time point.

[0071] In some embodiments, the maximum value scraper head operation parameter range judgment unit is also used to determine the specific time period for collecting the excavation volume, wherein the specific collection period includes the excavation of dredging volume with different scraper head structures, the excavation of dredging volume with different wind speeds, and the excavation of dredging volume with different sediment properties; determine the weights between the various influencing factors in the specific collection period, calculate the weight factors of the weights of the various influencing factors in the specific collection period at the corresponding correlation preset point and the weight factors of the weights of the various influencing factors in the specific collection period at the corresponding target preset time node; and obtain the Spearman correlation coefficient between the target preset time node of the scraper head multimodal parameter and the scraper head multimodal parameter correlation preset point; discard the scraper head multimodal parameter correlation preset points whose Spearman correlation coefficients exceed the scraper head multimodal parameter preset range.

[0072] Obviously, those skilled in the art may make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations.

Claims

1. A method for optimizing the drag head operating parameters based on the dredging volume of a trailing suction hopper dredger, characterized in that: The method includes: A1: Divide the trailing suction hopper dredger into standard dredging volume periods, and obtain the dredging volume fluctuation range and the maximum dredging volume fluctuation range under the multi-modal parameters of the drag head; A2: Obtaining all preset time nodes of the drag head multimodal parameters within the dredging volume fluctuation range under the drag head multimodal parameters, determining the excavation volumes at all the preset time nodes of the drag head multimodal parameters, forming a drag head multimodal parameter excavation volume group, and calculating the constraint conditions of each excavation volume in the drag head multimodal parameter excavation volume group to obtain a quartile diagram of the drag head multimodal parameter constraint conditions; A3: Obtain all maximum parameter preset time nodes within the maximum dredging volume fluctuation range, determine the excavation volumes at all the maximum parameter preset time nodes, form a maximum excavation volume group, and calculate the constraint conditions of each excavation volume in the maximum excavation volume group to obtain a maximum constraint condition quartile map; A4: Using the drag head pulling force of the trailing suction hopper dredger to match the drag head multimodal parameter preset time node with the maximum parameter preset time node, determining respective constraint conditions for the paired drag head multimodal parameter preset time node and the maximum parameter preset time node, and determining the drag head operating parameter range; A5: If the constraint condition value of the rake head multimodal parameter preset time node in the rake head multimodal parameter constraint condition quartile diagram is the same as the constraint condition value of the maximum parameter preset time node in the maximum value constraint condition quartile diagram, then the rake head multimodal parameter preset time node and the maximum parameter preset time node are used as the volume rake head operation parameter optimization management time point; The drag head multimodal parameters include drag head working angle, drag head grab speed, drag head depth, and drag head inclination; The rake head multimodal parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the rake head multimodal parameter target preset time node, all the rake head multimodal parameter preset time nodes are used to establish a rake head multimodal parameter unit time discrete curve, the rake head multimodal parameter unit time discrete curve is used to determine the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node, all the rake head multimodal parameter correlation preset points are compared with the volume excavated by the rake head multimodal parameter target preset time node for similarity, and according to the comparison result rake head operation parameter range, it is determined whether the rake head multimodal parameter target preset time node is used as the volume rake head operation parameter optimization management time point; The maximum parameter preset time node of the volume rake head operation parameter optimization point is calibrated as the maximum value target preset time node, and all the maximum parameter preset time nodes are used to establish a maximum value unit time discrete curve; the maximum value unit time discrete curve is used to determine the maximum value correlation preset point of the maximum value target preset time node; all the maximum value correlation preset points are compared with the volume excavated by the maximum value target preset time node for similarity, and according to the comparison result rake head operation parameter range, it is determined whether the maximum value target preset time node should be used as the volume rake head operation parameter optimization management time point.

2. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 1, characterized in that: The dredging volume fluctuation interval under the drag head multimodal parameters is used to preset the floating range of the drag head multimodal parameters of the dredging volume per unit time of different drag head structures of the trailing suction hopper dredger, and the maximum dredging volume fluctuation interval is used to preset the floating range of the maximum value of the dredging volume per unit time of different drag head structures of the trailing suction hopper dredger.

3. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 1, characterized in that: The similarity comparison comprises the following steps: Determine a specific time period for collecting the excavation volume, wherein the specific time period for collecting includes excavating the dredging volume with different rake head structures, excavating the dredging volume with different wind speeds, and excavating the dredging volume with different sediment properties; determine the weights between the various influencing factors in the specific time period for collecting, and calculate the weight factors of the weights of the various influencing factors in the specific time period for collecting the corresponding correlation preset point and the weight factors of the various influencing factors in the specific time period for collecting the corresponding target preset time node; obtain the degree of correlation of all weight factors, and use the degree of correlation to calculate the similarity comparison result between the corresponding correlation preset point and the corresponding target preset time node.

4. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 3, characterized in that: The step of obtaining the correlation degree of all weight factors also includes the following steps: All weight factors are encoded to obtain different numbers. The influencing factors in the different numbers are analyzed by graph neural network to obtain the mutual influence coefficient of the working efficiency of the multimodal parameters of the rake head, including: Assume there are four parameters: x 1 represents the working angle of the rake head, x 2 indicates the speed of the grab bucket. x 3 represents the depth of the rake head. x 4 represents the inclination of the rake head. The encoding of the weight factor is processed numerically to generate the feature vector of each node. Each influencing factor becomes a node, and based on the relationship between these influencing factors, the edges of the graph are constructed. The weight of the edge is calculated based on historical data to represent the mutual influence between different nodes. Graph construction, assuming graph G By node set V and edge sets E composition: , each node represents an influencing factor, E Represents the connection relationship between nodes, assuming that all nodes are connected to each other; Node characteristics, each node v i Each has an initial eigenvector h i (o) , representing a node v i The initial state of The graph convolution update formula weights the neighbor information of a node and passes it to the node. The formula for the graph convolution operation is as follows: ； in: h i (k) express v i In the k The feature vector of the layer, N( i ) represents a node v i The set of neighbor nodes of W (k) Indicates the k The weight matrix of the layer is used to learn the transformation of node features, b (k) Indicates the k The bias term of the layer, σ represents the activation function; Edge weight, edge weight A ij Representation node v i and nodes v j The degree of influence between them is expressed through an adjacency matrix A To represent the structure of the graph, A ij Representation node v i and v j The strength of the connection between Multi-layer graph neural network, uses a multi-layer graph neural network to capture high-order relationships, and the point update formula is: ； in: H (k) It is k All node feature matrices of the layer, A is the adjacency matrix of the graph, W (k) It is k The weight matrix of the layer, B (k) It is k The bias term of the layer; The output layer obtains the feature vector of each node through the learning of the multi-layer graph neural network, and maps the node features to the impact coefficient of work efficiency through an output layer: ； in: y i is a node v i The final impact coefficient on work efficiency, h i (K) is through K Node after layer graph convolution v i Features, W out is the weight matrix of the output layer; During the training process, a loss function is used to measure the difference between the work efficiency impact coefficient predicted by the model and the actual value.

5. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 1, characterized in that: After determining the rake head multimodal parameter correlation preset point of the rake head multimodal parameter target preset time node using the rake head multimodal parameter unit time discrete curve, the method further includes the following steps: Obtaining the Spearman correlation coefficient between the target preset time node of the rake head multimodal parameter and the preset point of the rake head multimodal parameter correlation; discarding the preset point of the rake head multimodal parameter correlation whose Spearman correlation coefficient exceeds the preset range of the rake head multimodal parameter.

6. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 5, characterized in that: The method of discarding the preset point of the rake head multimodal parameter correlation whose Spearman correlation coefficient exceeds the preset range of the rake head multimodal parameter also includes the following steps: using the unit time discrete curve of the rake head multimodal parameter to determine the number of nodes between the discarded preset point of the rake head multimodal parameter correlation and the target preset time node of the rake head multimodal parameter; when the value of the number of nodes exceeds the preset range, the discarded preset point of the rake head multimodal parameter correlation is added as the basis for comparison with the target preset time node of the rake head multimodal parameter.

7. The method for optimizing drag head operating parameters based on the dredging volume of a trailing suction hopper dredger according to claim 6, characterized in that: The added drag head multimodal parameter correlation preset point is assigned to the dredging drag head operation parameter correction coefficient, and the dredging drag head operation parameter correction coefficient serves as the calculation basis for substituting the drag head multimodal parameter correlation preset point for similarity comparison.

Images