CN111859777A - Method for calculating intelligent excavation technological parameters of trailing suction ship - Google Patents
Method for calculating intelligent excavation technological parameters of trailing suction ship Download PDFInfo
- Publication number
- CN111859777A CN111859777A CN202010464705.4A CN202010464705A CN111859777A CN 111859777 A CN111859777 A CN 111859777A CN 202010464705 A CN202010464705 A CN 202010464705A CN 111859777 A CN111859777 A CN 111859777A
- Authority
- CN
- China
- Prior art keywords
- data
- construction
- angle
- density
- drag
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Excavating Of Shafts Or Tunnels (AREA)
Abstract
A method for calculating intelligent excavation technological parameters of a trailing suction ship specifically comprises the following steps: collecting a daily report and historical data; counting the construction data range; generating learning sample data; calculating a neural network; and selecting process parameters. According to the method, a model of slurry density and flow rate is built by four factors including the angle of a drag head, the speed of navigation to the ground, the pressure of a wave compensator, the angle of a drag pipe and the like, the yield maximization is further solved, a stable and efficient construction process parameter combination is solved by the yield model in turn, the efficient dredging construction of the drag suction ship is guided, and the domestic blank of a technology for intelligently searching for the optimal construction parameter combination is filled.
Description
Technical Field
The invention relates to the technical field of calculation of excavation parameters of a trailing suction hopper dredger, in particular to a method for calculating intelligent excavation technological parameters of a trailing suction hopper dredger.
Background
In recent years, along with the rapid development of an automatic integration technology in the dredging industry, particularly, a data acquisition and monitoring system of a newly-built trailing suction hopper dredger is mature day by day, and a plurality of sensors of the trailing suction hopper dredger realize the functions of data acquisition, transmission and local storage; meanwhile, in recent two years, with the successful application of commercial satellite broadband technology in the mobile trailing suction hopper dredger, the historical and real-time data of the trailing suction hopper dredger of a dredging company can be remotely transmitted and stored on a shore-based terminal server, so that construction enterprises have a large amount of dredging data resources. However, the dredging data which has been accumulated and is continuously increasing is only staying at the level of providing data reference or simple data statistical analysis, and the construction data depth analysis, research and utilization level is not really realized, and the treasure of the dredging construction data assets is not really discovered and is not successfully applied to the production practice process. Through research on key technologies of the method for intelligently mining the technological parameter database by the trailing suction dredger, the relevance of construction technological parameters can be analyzed on line in real time, valuable construction parameter combinations can be found in time, and the method can be applied to dredging construction in time. The change requirements are timely provided according to the construction boundary conditions, the current situation of manual test lag is changed, the effective utilization rate of construction efficiency and time is improved, and energy consumption is reduced.
In conclusion, the dredging enterprises in China are still blank in the aspect of the key technology research on the optimization of the process parameters of the trailing suction dredger, the importance of dredging construction data assets is gradually shown along with the development of the dredging industry in the direction of refinement and intellectualization, and the key technology research on the optimization of the process parameters of the dredging construction of the trailing suction dredger is the development trend of the dredging industry in the future.
To solve the above problems, a series of improvements have been made.
Disclosure of Invention
The invention aims to provide a method for calculating intelligent excavation process parameters of a trailing suction vessel, which overcomes the defects and shortcomings in the prior art.
A method for calculating intelligent excavation technological parameters of a trailing suction ship comprises the following steps:
step 1: collecting a daily report and historical data;
step 2: counting the construction data range, discarding the data in the non-construction time period, wherein the counted construction data is as follows: the speed of the ground, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe, the depth of a suction port of a dredge pump, the flow rate of silt and the density of silt;
and step 3: generating learning sample data, setting the learning sample data step by step according to the data type in the step 2, grouping according to the construction data range and the step, filtering the data according to a set threshold, reversely calculating the density of the rake head, and finally screening the data;
And 4, step 4: calculating a neural network, inputting the sample data in the step 3 into a neural network model, and inputting layer signals as follows: the speed of the ground is controlled, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a suction port of a dredge pump are controlled, the hidden layers are 4 hidden layers, and the signals of an output layer are flow speed and density;
and 5: selecting technological parameters, selecting a neural network with small error to extract data for a scheme, counting, dividing sections and extracting the data, and finally selecting the data from a construction parameter combination.
Further, in the step 2, the calculation method of the non-construction time period is as follows:
pi＝ni/nzong，
in the formula, nzongIs the total number of points during construction, niNumber of points for each section, piFor each segment probability size, pi>The 70% range is considered to be within the valid construction data range.
Further, the step 3 comprises the following steps,
step 3.1: extracting a signal: the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a mud pump suction port are respectively set to be 0.1, 0.5, 2 and 5 in steps of the ground speed, the pressure of the compensator, the angle of the drag head, the angle of the drag pipe and the depth of the mud pump suction port, and the grouping number is obtained by the construction data range and the steps;
Step 3.2: and data filtering, wherein the first layer filters construction time period data, the second layer filters construction data within a specified construction time period, the ground speed is required to be greater than 1kn, the compensator stroke is required to be greater than 0.1m, and the third layer filters effective construction data, wherein the effective construction data is required to be between the two layers and does not contain a set value: mud density (1.025,1.6) and flow rate (4.5, 6);
step 3.3: performing reverse pushing, namely reversely calculating the density of the drag head silt, wherein the density in the construction data is the density measured by the flowmeter at the time point, because the construction density has time delay benefit, obtaining a density advance time according to the formula drag pipe length/silt speed, and reversely calculating the drag head silt density according to the time point;
step 3.4: in the screening, the number of valid points in any sample data must be greater than or equal to 5 to be considered as valid sample data.
Further, the step 4 comprises the following steps,
step 4.1: carrying out normalization processing on the construction data for min-max standardization, wherein max is the maximum value of the sample data, and min is the minimum value of the sample data;
step 4.2: the input layer is 4 nodes: speed of a ship or plane, wave compensator pressure, rake head angle, lower rake pipe angle, dredge pump suction inlet degree of depth, the output layer is 2 nodes: the flow rate and the density are set, the number of hidden layers, the number of neural nodes and errors are set, wherein the number of hidden layers is 4, the mean square error and the iteration of each node are adjusted and are sequentially reduced to ten-thousandth to two, when each set error is reached, the system can store the neural network under the error, after the calculation of the neural network is completed, the neural networks with different accuracies are selected to calculate comparison data, statistical comparison and selection statistical data and learning data errors are counted, and the neural network with small errors is selected.
Further, the step 5 comprises the following steps,
step 5.1: and (3) reversely screening data, wherein the construction data is subjected to dimensionality reduction: firstly, reducing the dimension of the depth of a mud pump suction port by taking 5 meters as steps, and dividing the depth into two groups of mud pump suction port depth data;
step 5.2: each group of pumping depth has 12 combinations, and the 12 combinations can perform dimensionality reduction on 4-dimensional signals: forming signals with 5 dimensions by adding 2 groups of suction port depths of the dredge pump to the ground speed, the pressure of the wave compensator, the rake head angle and the lower rake pipe angle;
step 5.3: screening and extracting data, respectively counting the maximum point number of a group of schemes in the combination and the number of schemes with the point number of a high point area exceeding 7.5 percent of the total point number for 2 groups of pumping depth data, and selecting one group of data which meets the requirement most;
step 5.4: counting the upper limit value and the lower limit value of the mud density of one group of scheme data, taking the range of the area as a basis, counting 80% of mud density points in the area again, and counting the construction data;
step 5.5: judging the number of the counted high points of the slurry density again, wherein the number of the counted high points of the slurry density is more than 80% of the total number of the points, the longitudinal and transverse rows are larger than the square of the number of the points, the construction parameters are finally generated according to the above basis, and the construction parameter combinations are named according to the project names and the soil property parameters and are used for guiding the construction state of the ship.
The invention has the beneficial effects that:
the invention guides the construction operation on the ship by building the drag head-output model, and efficiently controls the operation of equipment on the ship.
The method is characterized in that a model of slurry density and flow rate is built by the aid of four factors including drag head angle, ground speed, pressure of a wave compensator, drag pipe angle and the like, yield maximization is further solved, stable and efficient construction process parameter combinations are solved by the aid of the yield model, efficient dredging construction of the drag suction ship is guided finally, and the domestic blank of a technology for intelligently finding optimal construction parameter combinations is filled.
Description of the drawings:
FIG. 1 is a flow chart of the calculation of the present invention.
Fig. 2 is a schematic diagram of step 4.1.
Fig. 3 is a schematic diagram of step 4.2.
Fig. 4 is a schematic diagram of step 5.2.
Fig. 5 is a schematic diagram of step 5.5.
Detailed Description
The present invention is further illustrated by the following examples, which are intended to be purely exemplary of the invention and are not intended to limit the scope of the invention.
Example 1
FIG. 1 is a flow chart of the calculation of the present invention. Fig. 2 is a schematic diagram of step 4.1. Fig. 3 is a schematic diagram of step 4.2. Fig. 4 is a schematic diagram of step 5.2. Fig. 5 is a schematic diagram of step 5.5.
As shown in fig. 1-5, a method for calculating intelligent excavation process parameters of a trailing suction vessel comprises the following steps:
Step 1: collecting a daily report and historical data;
step 2: counting the construction data range, discarding the data in the non-construction time period, wherein the counted construction data is as follows: the speed of the ground, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe, the depth of a suction port of a dredge pump, the flow rate of silt and the density of silt;
and step 3: generating learning sample data, setting the learning sample data step by step according to the data type in the step 2, grouping according to the construction data range and the step, filtering the data according to a set threshold, reversely calculating the density of the rake head, and finally screening the data;
and 4, step 4: calculating a neural network, inputting the sample data in the step 3 into a neural network model, and inputting layer signals as follows: the speed of the ground is controlled, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a suction port of a dredge pump are controlled, the hidden layers are 4 hidden layers, and the signals of an output layer are flow speed and density;
and 5: selecting technological parameters, selecting a neural network with small error to extract data for a scheme, counting, dividing sections and extracting the data, and finally selecting the data from a construction parameter combination.
In the step 2, the calculation method of the non-construction time period comprises the following steps:
pi＝ni/nzong，
in the formula, n zongIs the total number of points during construction, niNumber of points for each section, piFor each segment probability size, pi>The 70% range is considered to be within the valid construction data range.
In the step 3, the method comprises the following steps,
step 3.1: extracting a signal: the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a mud pump suction port are respectively set to be 0.1, 0.5, 2 and 5 in steps of the ground speed, the pressure of the wave compensator, the angle of the drag head, the angle of the drag pipe and the depth of the mud pump suction port, and the grouping number is obtained by the range and the steps of construction data;
step 3.2: and data filtering, wherein the first layer filters construction time period data, the second layer filters construction data within a specified construction time period, the ground speed is required to be greater than 1kn, the compensator stroke is required to be greater than 0.1m, and the third layer filters effective construction data, wherein the effective construction data is required to be between the two layers and does not contain a set value: mud density (1.025,1.6) and flow rate (4.5, 6);
step 3.3: performing reverse pushing, namely reversely calculating the density of the drag head, wherein the density in the construction data is the density measured by the flowmeter at the time point, because the construction density has time delay benefit, obtaining a density advance time according to the formula drag pipe length/silt speed, and reversely calculating the silt density of the drag head according to the time point;
Step 3.4: in the screening, the number of valid points in any sample data must be greater than or equal to 5 to be considered as valid sample data.
In the step 4, the method comprises the following steps,
step 4.1: carrying out normalization processing on the construction data for min-max standardization, wherein max is the maximum value of the sample data, and min is the minimum value of the sample data;
step 4.2: the input layer is 4 nodes: speed of a ship or plane, wave compensator pressure, rake head angle, lower rake pipe angle, dredge pump suction inlet degree of depth, the output layer is 2 nodes: the flow rate and the density are set, the number of hidden layers, the number of neural nodes and errors are set, wherein the number of hidden layers is 4, the mean square error and the iteration of each node are adjusted and are sequentially reduced to ten-thousandth to two, when each set error is reached, the system can store the neural network under the error, after the calculation of the neural network is completed, the neural networks with different precisions are selected to calculate comparison data, statistical comparison and selection statistical data and learning data errors are counted, and the neural network with small errors is selected.
In the step 5, the method comprises the following steps,
step 5.1: and (3) reversely screening data, wherein the construction data is subjected to dimensionality reduction: firstly, reducing the dimension of the depth of a mud pump suction port by taking 5 meters as steps, and dividing the depth into two groups of mud pump suction port depth data;
Step 5.2: every group dredge pump suction inlet depth divides 12 combinations, and 12 combinations can be to 4 dimension signals dimensionality reduction: forming signals with 5 dimensions by adding 2 groups of suction port depths of the dredge pump to the ground speed, the pressure of the wave compensator, the rake head angle and the lower rake pipe angle;
step 5.3: screening and extracting data, respectively counting the maximum point number of a group of schemes in the combination and the number of schemes with the point number of a high point area exceeding 7.5 percent of the total point number for 2 groups of mud pump suction port depth data, and selecting a group of data which meets the requirement most;
step 5.4: counting the upper limit value and the lower limit value of the mud density of one group of scheme data, taking the range of the area as a basis, counting 80% of mud density points in the area again, and counting the construction data;
step 5.5: judging and counting the number of high points of the slurry density again, wherein the number of points exceeds 80% of the total number of points, the longitudinal and transverse rows are larger than the square of the number of points, the construction parameters are finally generated according to the above basis, the project names and the soil property parameters are named as construction parameter combinations, and the parameter combinations are used for guiding the construction state of the ship.
At present, the traditional algorithm is limited by the technology, the influence of at most 2 parameters on the yield can be analyzed, and the traditional algorithm can only be displayed in a curve form and cannot describe the corresponding relation. Due to the small number of samples and the immaturity of the algorithm, the generalization and convergence of the calculation are poor. Moreover, the drag suction dredging process is a very complicated process, the data shock is very large, the convergence is very poor, but there are some basic rules. Finally, some variables that affect production are interactive, dependent factors such as speed over ground, heave compensator pressure, drag head angle.
The process of the invention is as follows:
step 1: collecting daily reports and historical data, wherein the data are from a ship production information data center, the daily reports need to be downloaded on a daily report interface, the historical data need to be downloaded on a data downloading page, and the total construction data is more than 1000 million;
step 2: counting the construction data range, discarding the data in the non-construction time period, wherein the counted construction data is as follows: the speed of the ground, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe, the depth of a suction port of a dredge pump, the flow rate of silt and the density of silt. The method for calculating the non-construction time period comprises the following steps:
pi＝ni/nzong，
In the formula, nzongIs the total number of points during construction, niNumber of points for each section, piFor each segment probability size, pi>The 70% range is considered to be within the valid construction data range.
And step 3: generating learning sample data, setting the learning sample data step by step according to the data type in the step 2, grouping according to the construction data range and the step, filtering the data according to a set threshold, reversely calculating the density of the rake head, and finally screening the data.
Step 3.1: extracting a signal: the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a mud pump suction port are respectively set to be 0.1, 0.5, 2 and 5 in steps of the ground speed, the pressure of the wave compensator, the angle of the drag head, the angle of the drag pipe and the depth of the mud pump suction port, the grouping number is obtained by the construction data range and the steps, and the data statistics is as follows;
| signal name | Minimum value (inclusive) | Maximum value (none) | Stepping operation | Number of packets |
| Speed of flight to ground | 0.500 | 3.500 | 0.100 | 30 |
| Pressure of compensator | 26.000 | 30.000 | 0.500 | 8 |
| Rake head angle | -30.000 | 0.000 | 2.000 | 15 |
| Angle of lower harrow tube | 27.000 | 37.000 | 2.000 | 5 |
| Depth of suction port of dredge pump | 20.000 | 30.000 | 5.000 | 2 |
Step 3.2: and data filtering, wherein the first layer filters construction time period data, the second layer filters construction data within a specified construction time period, the ground speed is required to be greater than 1kn, the compensator stroke is required to be greater than 0.1m, and the third layer filters effective construction data, wherein the effective construction data is required to be between the two layers and does not contain a set value: mud density (1.025,1.6) and flow rate (4.5, 6);
Step 3.3: performing reverse pushing, namely reversely calculating the density of the drag head mud, wherein the density in the construction data is the density measured by the flowmeter at the time point, because the construction density has time delay benefit, obtaining a density advance time according to the formula drag pipe length/silt speed, and reversely calculating the drag head silt density according to the time point;
step 3.4: in the screening, the number of valid points in any sample data must be greater than or equal to 5 to be considered as valid sample data.
And 4, step 4: calculating a neural network, inputting the sample data in the step 3 into a neural network model, and inputting layer signals as follows: the speed of the ground is controlled, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a suction port of a dredge pump are controlled, the hidden layers are 4 hidden layers, and the signals of an output layer are flow speed and density;
step 4.1: carrying out normalization processing on the construction data for min-max standardization, wherein max is the maximum value of the sample data, and min is the minimum value of the sample data;
the basic principle is as follows: data normalization is a data preprocessing method, namely, data to be processed is limited within a required certain range through a certain algorithm, and for the purpose of more convenient later data processing and ensuring that convergence is accelerated when a neural network operates, the data is generally limited between 0 and 1. For example, for singular sample data, which is a sample vector that is particularly large or small relative to other input samples, the data causes increased network training time and may cause the network to fail to converge, so the data set for which there is singular sample data for the training samples is preferably normalized before training. Expressed by the formula:
The formula: wherein X is an input layer signal, such as instantaneous construction data of the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe, the depth of a suction port of a dredge pump and the like, max is the maximum value of the data in the construction time period, and min is the minimum value of the data in the construction time period;
as shown in fig. 2, the whole process is:
wherein the selection of the activation function:
step 4.2: the input layer is 4 nodes: speed of a ship or plane, wave compensator pressure, rake head angle, lower rake pipe angle, dredge pump suction inlet degree of depth, the output layer is 2 nodes: the flow rate and the density are set, the number of hidden layers, the number of neural nodes and errors are set, wherein the number of hidden layers is 4, the mean square error and the iteration of each node are adjusted and are sequentially reduced to ten-thousandth to two, when each set error is reached, the system can store the neural network under the error, after the calculation of the neural network is completed, the neural networks with different accuracies are selected to calculate comparison data, statistical comparison and selection statistical data and learning data errors are counted, and the neural network with small errors is selected.
As shown in fig. 3, the left four parameters are five boxes, and one more box is prepared for adding more construction parameter factor data later. When the neural network is used for calculation, the input layer is 4 nodes, the output layer is 2 nodes, the total number of the nodes is more than 1000 and ten thousand samples, and the number of each layer of the nodes of the neural network is 4 multiplied by 32. There are 128 total weights and offsets for 4 x 32, with various crossovers. The algorithm adopts 4 hidden layers. The number of hidden layers, the number of neural nodes and errors are set. Each data record is trained until over 1000 ten thousand records. And (3) adjusting the mean square error and iteration of each node every time of training, sequentially reducing the mean square error and iteration to ten-thousandth to two, and when each set error is reached, storing the neural network under the error by the system. The general error is set to 6 different values. That is, the software will store 6 neural networks, and select the appropriate neural network when the scheme is to be extracted. After the neural network calculation is finished, selecting the neural networks with different precisions to calculate the comparison data, counting the errors of the comparison and selection statistical data and the learning data, and selecting the neural network with small error.
And 5: selecting technological parameters, selecting a neural network with small error to extract data for a scheme, counting, dividing sections and extracting the data, and finally selecting the data from a construction parameter combination.
Step 5.1: and (3) reversely screening data, wherein the construction data is subjected to dimensionality reduction: firstly, the depth of the mud pump suction port is reduced, the dimension is reduced by taking 5 meters as steps, and the depth data of the mud pump suction port is divided into two groups.
Step 5.2: every group dredge pump suction inlet depth divides 12 combinations, and 12 combinations can be to 4 dimension signals dimensionality reduction: the speed of the ground, the pressure of the wave compensator, the angle of the drag head and the angle of the drag pipe are added with 2 groups of suction port depths of the dredge pump to form signals with 5 dimensions. The method comprises the following steps: speed of flight to the ground, heave compensator pressure, drag head angle, drag tube angle, speed of flight to the ground, drag head angle, heave compensator pressure, drag tube angle, speed of flight to the ground, drag tube angle, heave compensator pressure, drag head angle, speed of flight to the ground, drag tube angle, heave compensator pressure, drag tube angle, speed of flight to the ground, drag head angle, heave compensator pressure, speed of flight to the ground, drag head angle, drag tube angle, speed of flight to the ground, wave compensator pressure, drag tube angle, drag head angle, heave compensator pressure, speed of flight to the ground, drag tube angle, speed of flight to the ground, heave compensator pressure, drag head angle, drag tube angle, speed of flight to the ground, drag head angle, drag tube angle, heave compensator pressure, speed of flight to the ground, drag tube angle, drag, Drag head angle, drag pipe angle, drag head angle, speed to ground, heave compensator pressure. The above 12 combinations can perform dimensionality reduction on 4-dimensional signals: the dimension of signals of 5 dimensions can be reduced by the aid of the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a suction port of a dredge pump which is divided into 2 groups.
The correspondence between the combination numbers and the signals is as follows: the speed to the ground is 1, the pressure of the heave compensator is 2, the angle of the drag head is 3, the angle of the drag pipe is 4, and the depth of the suction port of the dredge pump is 0. As shown in fig. 4, the back two dimensions correspond to XY coordinates of the surface, and the XY coordinates are reversed without affecting the surface representation, so that the back two dimensions do not need to be combined. The 12 sets may not be needed for the extraction of the solution, because the combined surface representations of the first two-dimensional exchange may not be the same, and the solution should be extracted consistently. The correspondence table is as follows:
step 5.3: screening and extracting data, respectively counting the maximum point number of a group of schemes in the combination and the number of schemes with the point number of a high point area exceeding 7.5 percent of the total point number for 2 groups of mud pump suction port depth data, and selecting a group of data which meets the requirement most;
step 5.4: counting the upper limit value and the lower limit value of the mud density of one group of scheme data, taking the range of the area as a basis, counting 80% of the mud density number in the area again, and counting the construction data;
step 5.5: judging and counting the number of high points of the slurry density again, wherein the number of points exceeds 80% of the total number of points, the longitudinal and transverse rows are larger than the square of the number of points, the construction parameters are finally generated according to the above basis, the project names and the soil property parameters are named as construction parameter combinations, and the parameter combinations are used for guiding the construction state of the ship. The principle is shown in fig. 5, wherein a: the total number of mud density points in the high point area; a: the number of points in the horizontal direction in the high point area; b: the number of points in the vertical direction in the high point area; t: total number of points of curved surface, B: percent high point constraint, C: high point constraint points (the total number of mud density points (i.e., a) in the high point region must be greater than or equal to this value); d: the number of horizontal and vertical direction constraint points (the number of horizontal and vertical direction constraint points must be equal to or greater than this value),
In this embodiment, B is set to 7.5%, where a is 36 and D is 6, that is, 6 points are required for the horizontal axis and 6 points are required for the vertical axis, or the horizontal axis of the vertical axis is easily formed into a curved surface region, which is convenient for the subsequent output result to be embodied.
While the present invention has been described with reference to the specific embodiments, the present invention is not limited thereto, and various changes may be made without departing from the spirit of the present invention.
Claims (5)
1. A method for calculating intelligent excavation technological parameters of a trailing suction ship is characterized by comprising the following steps:
step 1: collecting a daily report and historical data;
step 2: counting the construction data range, discarding the data in the non-construction time period, wherein the counted construction data is as follows: the speed of the ground, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe, the depth of a suction port of a dredge pump, the flow rate of silt and the density of silt;
and step 3: generating learning sample data, setting the learning sample data step by step according to the data type in the step 2, grouping according to the construction data range and the step, filtering the data according to a set threshold, reversely calculating the density of the rake head, and finally screening the data;
and 4, step 4: calculating a neural network, inputting the sample data in the step 3 into a neural network model, and inputting layer signals as follows: the speed of the ground is controlled, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a suction port of a dredge pump are controlled, the hidden layers are 4 hidden layers, and the signals of an output layer are flow speed and density;
And 5: selecting technological parameters, selecting a neural network with small error to extract data for a scheme, counting, dividing sections and extracting the data, and finally selecting the data from a construction parameter combination.
2. The method for calculating the intelligent excavation process parameters of the trailing suction vessel as claimed in claim 1, wherein the method comprises the following steps: in the step 2, the calculation method of the non-construction time period comprises the following steps:
pi＝ni/nzong，
in the formula, nzongIs the total number of points during construction, niNumber of points for each section, piFor each segment probability size, pi>The 70% range is considered to be within the valid construction data range.
3. The method for calculating the intelligent excavation process parameters of the trailing suction vessel as claimed in claim 1, wherein the method comprises the following steps: in the step 3, the method comprises the following steps,
step 3.1: extracting a signal: the ground speed, the pressure of a wave compensator, the angle of a drag head, the angle of a drag pipe and the depth of a mud pump suction port are respectively set to be 0.1, 0.5, 2 and 5 in steps of the ground speed, the pressure of the wave compensator, the angle of the drag head, the angle of the drag pipe and the depth of the mud pump suction port, and the grouping number is obtained by the range and the steps of construction data;
step 3.2: and data filtering, wherein the first layer filters construction time period data, the second layer filters construction data within a specified construction time period, the ground speed is required to be greater than 1kn, the compensator stroke is required to be greater than 0.1m, and the third layer filters effective construction data, wherein the effective construction data is required to be between the two layers and does not contain a set value: mud density (1.025,1.6) and flow rate (4.5, 6);
Step 3.3: performing reverse pushing, namely reversely calculating the density of the drag head silt, wherein the density in the construction data is the density measured by the flowmeter at the time point, because the construction density has time delay benefit, obtaining a density advance time according to the formula drag pipe length/silt speed, and reversely calculating the drag head silt density according to the time point;
step 3.4: in the screening, the number of valid points in any sample data must be greater than or equal to 5 to be considered as valid sample data.
4. The method for calculating the intelligent excavation process parameters of the trailing suction vessel as claimed in claim 1, wherein the method comprises the following steps: in the step 4, the method comprises the following steps,
step 4.1: carrying out normalization processing on the construction data for min-max standardization, wherein max is the maximum value of the sample data, and min is the minimum value of the sample data;
step 4.2: the input layer is 4 nodes: speed of a ship or plane, wave compensator pressure, rake head angle, lower rake pipe angle, dredge pump suction inlet degree of depth, the output layer is 2 nodes: the flow rate and the density are set, the number of hidden layers, the number of neural nodes and errors are set, wherein the number of hidden layers is 4, the mean square error and the iteration of each node are adjusted and are sequentially reduced to ten-thousandth to two, when each set error is reached, the system can store the neural network under the error, after the calculation of the neural network is completed, the neural networks with different accuracies are selected to calculate comparison data, statistical comparison and selection statistical data and learning data errors are counted, and the neural network with small errors is selected.
5. The method for calculating the intelligent excavation process parameters of the trailing suction vessel as claimed in claim 1, wherein the method comprises the following steps: in the step 5, the method comprises the following steps,
step 5.1: and (3) reversely screening data, wherein the construction data is subjected to dimensionality reduction: firstly, reducing the dimension of the depth of a mud pump suction port by taking 5 meters as steps, and dividing the depth into two groups of mud pump suction port depth data;
step 5.2: every group dredge pump suction inlet depth divides 12 combinations, and 12 combinations can be to 4 dimension signals dimensionality reduction: forming signals with 5 dimensions by adding 2 groups of suction port depths of the dredge pump to the ground speed, the pressure of the wave compensator, the rake head angle and the lower rake pipe angle;
step 5.3: screening and extracting data, respectively counting the maximum point number of a group of schemes in the combination and the number of schemes with the point number of a high point area exceeding 7.5 percent of the total point number for 2 groups of mud pump suction port depth data, and selecting a group of data which meets the requirement most;
step 5.4: counting the upper limit value and the lower limit value of the mud density of one group of scheme data, taking the range of the area as a basis, counting 80% of the mud density number in the area again, and counting the construction data;
Step 5.5: judging and counting the number of high points of the slurry density again, wherein the number of points exceeds 80% of the total number of points, the longitudinal and transverse rows are larger than the square of the number of points, the construction parameters are finally generated according to the above basis, the project names and the soil property parameters are named as construction parameter combinations, and the parameter combinations are used for guiding the construction state of the ship.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010464705.4A CN111859777A (en) | 2020-05-28 | 2020-05-28 | Method for calculating intelligent excavation technological parameters of trailing suction ship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010464705.4A CN111859777A (en) | 2020-05-28 | 2020-05-28 | Method for calculating intelligent excavation technological parameters of trailing suction ship |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111859777A true CN111859777A (en) | 2020-10-30 |
Family
ID=72985824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010464705.4A Withdrawn CN111859777A (en) | 2020-05-28 | 2020-05-28 | Method for calculating intelligent excavation technological parameters of trailing suction ship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111859777A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114991243A (en) * | 2022-06-23 | 2022-09-02 | 中交广州航道局有限公司 | Construction method for clay bow blowing of trailing suction ship |
| CN116290203A (en) * | 2023-01-12 | 2023-06-23 | 中港疏浚有限公司 | Dredging construction parameter optimization model method based on neural network |
-
2020
- 2020-05-28 CN CN202010464705.4A patent/CN111859777A/en not_active Withdrawn
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114991243A (en) * | 2022-06-23 | 2022-09-02 | 中交广州航道局有限公司 | Construction method for clay bow blowing of trailing suction ship |
| CN114991243B (en) * | 2022-06-23 | 2023-09-29 | 中交广州航道局有限公司 | Construction method for clay bow blowing of trailing suction hopper |
| CN116290203A (en) * | 2023-01-12 | 2023-06-23 | 中港疏浚有限公司 | Dredging construction parameter optimization model method based on neural network |
| CN116290203B (en) * | 2023-01-12 | 2023-10-03 | 中港疏浚有限公司 | Dredging construction parameter optimization model method based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104615862B (en) | The method that high water cut oil field based on evolution algorithm determines well location | |
| CN111859777A (en) | Method for calculating intelligent excavation technological parameters of trailing suction ship | |
| CN105118090A (en) | Adaptive point-cloud filtering method for complex terrain structure | |
| CN107965301B (en) | Quantitative evaluation method for water injection process of digital oil field | |
| CN108549962B (en) | Wind power prediction method based on historical segmented sequence search and time sequence sparsification | |
| CN103591942B (en) | Method for quickly simplifying and automatically optimizing sound velocity profile based on maximum offset of sound velocity | |
| US11687693B2 (en) | Method for evaluating fracture connectivity and optimizing fracture parameters based on complex network theory | |
| CN105045091A (en) | Dredging process intelligent decision analysis method based on fuzzy neural control system | |
| CN112329979A (en) | Ultra-short-term wind power prediction method based on self-adaptive depth residual error network | |
| CN110826790A (en) | Intelligent prediction method for construction productivity of cutter suction dredger | |
| CN115860197A (en) | Data-driven coal bed gas yield prediction method and system | |
| CN112307410A (en) | Seawater temperature and salinity information time sequence prediction method based on shipborne CTD measurement data | |
| CN111460737A (en) | Intelligent settlement prediction method and system for slurry air pressure balance shield | |
| CN113486556B (en) | Improved efficient automatic history fitting method for oil and gas reservoir | |
| CN113156492B (en) | Real-time intelligent early warning method applied to TBM tunnel rockburst disasters | |
| CN105512383A (en) | Dredging process regulation and control parameter screening method based on BP neural network | |
| CN112862946B (en) | Gray rock core image three-dimensional reconstruction method for generating countermeasure network based on cascade condition | |
| AU2021104794A4 (en) | Stage-wise stochastic inversion identification method of aquifer structure based on deep learning | |
| CN111488974A (en) | Deep learning neural network-based ocean wind energy downscaling method | |
| CN116070768A (en) | Short-term wind power prediction method based on data reconstruction and TCN-BiLSTM | |
| CN113988382A (en) | Oil field recovery rate prediction method based on ensemble learning | |
| CN115618480A (en) | Dredger energy consumption optimization method and system based on genetic algorithm | |
| CN114169217A (en) | Method for intelligently adding foam drainage agent based on data driving and edge cloud coordination to solve water blockage of gas well | |
| CN111749675A (en) | Stratum drillability prediction method and system based on cascade model algorithm | |
| Zhao et al. | Flow velocity prediction method based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201030 |