Part type selection method of ship pipeline system

Abstract

The invention provides a part type selection method for a ship pipeline system. The part type selection method comprises the steps that S1, ship parts are classified according to functions; s2, respectively extracting feature data of the ship parts according to different classifications; s3, the databases of different types are trained, different machine learning models are obtained, and a part specification prediction model is selected from the multiple machine learning models; and S4, calling the corresponding specification prediction model according to the category of the ship part, determining the model of the part based on the corresponding specification prediction model, and completing the model selection of the part. Historical data of the ship three-dimensional model are trained and learned through the AI technology so as to determine model selection preferences and special requirements of different types of ship parts. The model selection of the ship parts is performed through the specification prediction model, so that the model selection time of the ship parts can be greatly shortened, the model selection efficiency of the ship parts is improved, the model selection deviation caused by insufficient experience and knowledge of designers is reduced, the accuracy of the model selection of the ship parts is improved, and popularization and application are facilitated.

Description

Technical Field

[0001] The present application relates to the technical field of ship design, and in particular to a method for selecting parts for a ship piping system. Background Art

[0002] In the design of a ship's piping system, parts selection is a crucial link, which is directly related to the performance, safety and cost-effectiveness of the entire system. In the actual design process, quality problems caused by the wrong selection of piping parts are common, and the main reasons are reflected in the following aspects:

[0003] 1. There are many types of parts and a huge parts library

[0004] The number of parts and models of piping systems is increasing, forming a huge and complex component library. Designers need to spend a lot of time looking for parts from it, which not only increases time costs but also increases the possibility of errors.

[0005] 2. Different series of ships have different application preferences

[0006] Different series of ships have different requirements for piping systems, and designers need to fully consider these application preferences. However, this difference in application preferences makes the selection process more complicated, requiring designers to have deep professional knowledge and rich project experience.

[0007] 3. Selection specifications are difficult to organize into itemized content, which makes it difficult for designers to quickly and accurately understand and apply the specifications during the selection process, increasing the risk of errors.

[0008] In summary, it is necessary to provide an improved technical solution to address the above-mentioned deficiencies in the prior art. Summary of the invention

[0009] The purpose of an embodiment of the present application is to provide a method for selecting parts for a ship piping system, which can complete the selection of parts in a ship piping system based on an AI algorithm.

[0010] The present application embodiment specifically provides a method for selecting parts for a ship piping system, comprising the following steps:

[0011] S1. Classify ship parts according to their functions;

[0012] S2, extracting characteristic data of ship parts according to different classifications;

[0013] S3, training different categories of databases respectively to obtain different machine learning models, and selecting one or more of the multiple machine learning models as a category of part specification prediction models;

[0014] S4. Retrieve a corresponding specification prediction model according to the category of the ship parts, determine the model of the parts based on the corresponding specification prediction model, and complete the selection of the parts.

[0015] In one practicable manner, in step S1, the parts can be divided into at least flanges, joints, reducers, elbows, tees, welding sockets, and gaskets according to their functions.

[0016] In one practicable manner, in step S1 , the parts may be classified according to the ship type, or the parts may be classified according to the ship type and function.

[0017] In one practicable manner, in step S2, at least the following contents are included:

[0018] S21, extracting structured data of different ship parts based on the three-dimensional model;

[0019] S22, splitting the data set of ship parts so that each category of ship parts forms a data set;

[0020] S23, performing data cleaning on each of the data sets, including at least completing missing feature data, and correcting or deleting erroneous feature data;

[0021] S24. Establishing a database based on the corrected data set.

[0022] In an implementable manner, the structured data includes at least category data, feature data, and label data; and the three-dimensional model includes at least a pipeline model and an equipment model.

[0023] In an implementable manner, the category data is determined by the classification information in step S1; the feature data at least includes ship type, region, system, pipe diameter, wall thickness, material, and pressure level; and the label data at least includes component model.

[0024] In one feasible manner, step S3 includes at least the following contents: pre-training the database using different machine learning algorithms according to the problem type, data quantity or data distribution of the database, and selecting a machine learning algorithm as the designated algorithm for the database according to the pre-training results.

[0025] In one practicable manner, the machine learning algorithms in pre-training include at least: decision tree, random forest, extreme gradient boosting tree, and lightweight gradient boosting machine.

[0026] In one feasible method, the data set is split into a training data set and a test data set; the training data set is trained using different machine learning algorithms to obtain different selection machine learning models; based on the test data set, evaluation indicators of multiple selection machine learning models are obtained respectively, and the machine learning model corresponding to the optimal indicator is selected from multiple evaluation indicators as a category of part specification prediction model.

[0027] In an practicable manner, in step S4, at least the following contents are included: determining the category of ship parts according to the ship design content, and extracting feature data corresponding to the category. Selecting a corresponding part specification prediction model according to the category. Inputting the feature data based on the part specification prediction model to obtain a predicted part model.

[0028] Compared with the prior art, the beneficial effects of this application are:

[0029] In the technical solution of this application, the historical data of the three-dimensional model of the ship is trained and learned through AI technologies such as machine learning algorithms to determine the selection preferences and special requirements of different types of ship parts. The selection of ship parts through the specification prediction model provided by this application can greatly shorten the time for selecting ship parts, improve the efficiency of selecting ship parts, reduce the selection deviation caused by the designer's lack of experience and knowledge, improve the accuracy of ship part selection, and facilitate promotion and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of a method for selecting parts for a ship piping system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. These embodiments are only used to illustrate the present invention, but not to limit the present invention.

[0032] In the description of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0033] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0034] Furthermore, in the description of the present invention, unless otherwise specified, “plurality” means two or more.

[0035] See also Figure 1 The present application provides a method for selecting parts for a ship piping system, comprising the following steps:

[0036] S1. Classify parts according to their functions;

[0037] It should be noted that parts can be divided into at least flanges, joints, reducers, elbows, tees, welding seats and gaskets according to their functions.

[0038] It should also be noted that, in step S1, the parts may be classified according to the ship type, or classified according to the ship type and function, so as to ensure the accuracy of subsequent selection.

[0039] S2, extracting characteristic data of ship parts according to different classifications;

[0040] In one practicable manner, in step S2, at least the following contents are included:

[0041] S21, extracting structured data of different ship parts based on the three-dimensional model;

[0042] It should be noted that the structured data at least includes category data, feature data, and label data. The three-dimensional model includes a pipeline model and an equipment model.

[0043] Specifically, as shown in Table 1, Table 1 is a header detail of structured data. The category data is determined according to the classification information of the parts in step S1. The feature data at least includes ship type, region, system, pipe diameter, wall thickness, material, and pressure level. The label data at least includes component model.

[0044] Table 1

[0045] OID category Ship Type Diameter area system Wall thickness Material Pressure level Part number

[0046] S22, splitting the data set of ship parts so that each category of ship parts forms a data set;

[0047] S23, performing data cleaning on each of the data sets, including at least completing missing feature data, and correcting or deleting erroneous feature data;

[0048] Specifically, if the pressure level data of a component is missing, it can be supplemented according to the pipeline system in which it is located; if the historical three-dimensional model has model interference or does not comply with existing design rules, the relevant component models shall be corrected.

[0049] S24, establishing a database based on the corrected data set, that is, saving the cleaned data into the database.

[0050] S3. Train different categories of databases respectively to obtain different machine learning models, and select one or more of the multiple machine learning models as a category of part specification prediction models.

[0051] In one practicable manner, in step S3, at least the following contents are included:

[0052] According to the problem type, data quantity or data distribution of the database, different machine learning algorithms are used to pre-train the database, and a machine learning algorithm is selected as the designated algorithm for the database based on the pre-training results.

[0053] In one practicable manner, the machine learning algorithms in pre-training include at least: decision tree, random forest, extreme gradient boosting tree, and lightweight gradient boosting machine.

[0054] In an practicable manner, step S3 also includes the following contents: extracting a data set of a corresponding category according to the part category, and splitting the data set into a training data set and a test data set. The training data set is trained using different machine learning algorithms to obtain different selection machine learning models. Based on the test data set, evaluation indicators of multiple selection machine learning models are obtained respectively, and the optimal indicator is selected from the multiple evaluation indicators to determine the optimal selection machine learning model.

[0055] Specifically, the flange selection data set is split into a flange selection training data set and a flange selection test data set; the flange selection training data set is taken, and at least multiple machine learning models including decision tree, random forest, extreme gradient boosting tree, and lightweight gradient boosting machine are used for training, and a flange selection machine learning model based on decision tree, a flange selection machine learning model based on random forest, and a flange selection machine learning model based on extreme gradient boosting tree are obtained respectively; the flange selection test data set is taken to obtain evaluation indicators of the above multiple machine learning models, and the machine learning model is selected based on the evaluation indicators.

[0056] It should be noted that the accuracy rate can be used as the evaluation indicator, and the machine learning model with the highest accuracy rate is selected for use.

[0057] In one feasible method, in order to enhance the prediction effect of the machine learning model, different data sets are trained separately to generate multiple dedicated machine learning models suitable for different categories of components, and the machine learning models are uniquely named, such as the flange specification prediction model.

[0058] S4. Retrieve a corresponding specification prediction model according to the category of the ship parts, determine the model of the parts based on the corresponding specification prediction model, and complete the selection of the parts.

[0059] Specifically, in step S4, at least the following contents are included:

[0060] The category of ship parts is determined according to the ship design content, and the characteristic data corresponding to the category is extracted. The corresponding part specification prediction model is selected according to the category. The characteristic data is input based on the part specification prediction model to obtain the predicted part model.

[0061] In an practicable manner, after the characteristic data is input, at least three candidate models are output for subsequent further selection.

[0062] In an practicable manner, after step S4, the method further includes: determining the predicted part model.

[0063] Specifically, by adding judgment rules, such as whether the material matches, etc., a preliminary correctness judgment is made on the model output by the part specification prediction model, and the component model and judgment results are fed back to the designer to facilitate further judgment by the designer.

[0064] In summary, this application uses AI technologies such as machine learning algorithms to train and learn historical data of ship 3D models to determine the selection preferences and special requirements of different types of ship parts. The selection of ship parts through the specification prediction model provided by this application can greatly shorten the time for selecting ship parts, improve the efficiency of selecting ship parts, reduce the selection deviation caused by the designer's lack of experience and knowledge, improve the accuracy of selecting ship parts, and facilitate promotion and use.

[0065] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.

Claims

1. A method for selecting parts for a ship piping system, characterized in that: The following steps are involved: S1. Classify ship parts according to their functions; S2, extracting characteristic data of ship parts according to different classifications; S3, training different categories of databases respectively to obtain different machine learning models, and selecting one or more of the multiple machine learning models as a category of part specification prediction models; S4. Retrieve a corresponding specification prediction model according to the category of the ship parts, determine the model of the parts based on the corresponding specification prediction model, and complete the selection of the parts.

2. The method for selecting parts for a ship piping system according to claim 1, characterized in that: In step S1, parts can be divided into at least flanges, joints, reducers, elbows, tees, welding sockets, and gaskets according to their functions.

3. The method for selecting parts for a ship piping system according to claim 1, characterized in that: In step S1, the parts may be classified according to the ship type, or the parts may be classified according to the ship type and function.

4. The method for selecting parts for a ship piping system according to claim 1, characterized in that: In step S2, at least the following contents are included: S21, extracting structured data of different ship parts based on the three-dimensional model; S22, splitting the data set of ship parts so that each category of ship parts forms a data set; S23, performing data cleaning on each of the data sets, including at least completing missing feature data, and correcting or deleting erroneous feature data; S24. Establishing a database based on the corrected data set.

5. The method for selecting parts for a ship piping system according to claim 4, characterized in that: The structured data at least includes category data, feature data, and label data; the three-dimensional model at least includes a pipeline model and an equipment model.

6. The method for selecting parts for a ship piping system according to claim 5, characterized in that: The category data is determined by the classification information in step S1; the feature data at least includes ship type, region, system, pipe diameter, wall thickness, material, and pressure level; and the label data at least includes component model.

7. The method for selecting parts for a ship piping system according to claim 1, characterized in that: In step S3, at least the following contents are included: according to the problem type, data quantity or data distribution of the database, different machine learning algorithms are used to pre-train the database, and a machine learning algorithm is selected as the designated algorithm for the database according to the pre-training results.

8. The method for selecting parts for a ship piping system according to claim 7, characterized in that: The machine learning algorithms in pre-training include at least: decision tree, random forest, extreme gradient boosting tree, and lightweight gradient boosting machine.

9. The method for selecting parts for a ship piping system according to claim 7, characterized in that: The data set is split into a training data set and a test data set; the training data set is trained using different machine learning algorithms to obtain different selection machine learning models; based on the test data set, evaluation indicators of multiple selection machine learning models are obtained respectively, and the machine learning model corresponding to the optimal indicator is selected from the multiple evaluation indicators as a category of part specification prediction model.

10. The method for selecting parts for a ship piping system according to claim 1, characterized in that: In step S4, at least the following contents are included: determining the category of ship parts according to the ship design content, and extracting feature data corresponding to the category; A corresponding part specification prediction model is selected according to the category; and the feature data is input based on the part specification prediction model to obtain a predicted part model.

Images