Construction method of ship quality model

Abstract

The invention belongs to the technical field of ship manufacturing, and provides a ship quality model construction method, which comprises the steps of dividing a ship construction process into a plurality of manufacturing links; obtaining the quality problem of each manufacturing link; based on the quality problem, a key data set of a manufacturing link is obtained, and the key data set comprises process key data u, production process key data p and quality key data q; and constructing a quality model Q (q) = f (u, p). According to the method, production data accumulated in production is utilized to construct a data-driven ship quality model, so that quality data of a manufacturing link can be predicted in advance, data support is provided for actual construction, and a worker can make an analysis decision according to a prediction result; through quantitative description of process key data, production process key data and quality key data, construction and optimization of a quality model are supported, dependence on working experience is avoided, the accuracy and reliability of quality control are improved, and digital development of ship quality control is facilitated.

Description

Technical Field

[0001] The present application relates to the field of shipbuilding technology, and in particular to a method for constructing a ship quality model. Background Art

[0002] Shipbuilding is a complex and sophisticated industrial field. Quality control plays a vital role in the shipbuilding process and is the cornerstone for ensuring the final quality of the ship and ensuring navigation safety. As a large-scale means of transportation, ships have complex structures and numerous systems. From the laying of the keel of the hull to the completion of the superstructure, every component and every process is related to the overall performance and reliability of the ship. Once quality problems occur, it can lead to low ship operation efficiency, increase maintenance and time costs, or even cause safety accidents and threaten the lives of crew members.

[0003] The quality control methods in the traditional shipbuilding process mainly rely on manual inspection and analytical decision-making, and need to rely on the experience of quality management personnel to inspect, evaluate and make decisions on each link of shipbuilding. This control mode based on manual experience has many limitations. On the one hand, the efficiency of manual inspection is relatively low, and the scale of shipbuilding is large, involving many parts and complicated processes. Quality management personnel need to spend a lot of time and experience to check each detail one by one. It is difficult to complete a comprehensive and systematic quality assessment in a short time, resulting in a lag in the discovery and handling of quality problems. On the other hand, the accuracy of manual inspection is easily affected by many factors, and the experience level of quality management personnel is uneven. Different personnel may have different understandings and grasps of quality standards, which makes the accuracy of manual inspection poor, and there is a certain degree of subjectivity and uncertainty. Summary of the invention

[0004] In view of the shortcomings of the prior art described above, the purpose of the present application is to provide a method for constructing a ship quality model to solve the problems of low efficiency and poor accuracy of manual ship quality control in the prior art.

[0005] To achieve the above-mentioned and other related purposes, the present application provides a method for constructing a ship quality model, comprising the following steps:

[0006] Divide the shipbuilding process into several manufacturing stages;

[0007] Obtain the quality issues of each manufacturing link;

[0008] Based on the quality problems in the manufacturing link, a key data set of the manufacturing link is obtained, wherein the key data set includes process key data u, production process key data p and quality key data q;

[0009] Construct a quality model Q(q) = f(u, p) based on the key data set.

[0010] Optionally, dividing the shipbuilding process into several manufacturing links includes the following steps:

[0011] Obtain the process flow data set of the shipbuilding process;

[0012] Based on the process flow data set, divide the shipbuilding process into several manufacturing links;

[0013] Among them, several of the manufacturing links at least include a pretreatment link, a cutting link, a processing and forming link, a small block assembly manufacturing link, and a medium block assembly manufacturing link.

[0014] Optionally, obtaining the key data set of the manufacturing link includes the following steps:

[0015] Obtain the process flow information of the manufacturing link;

[0016] Based on the process flow information, divide the manufacturing link into several manufacturing sub-links;

[0017] Based on the quality problems of the manufacturing link, obtain the sub-link parameter data of the manufacturing sub-link;

[0018] Respectively extract process key data, production process key data, and quality key data from the sub-link parameter data.

[0019] Optionally, the process key data includes part parameter data, the production process key data includes process parameter data, and the quality key data includes deviation parameter data.

[0020] Optionally, the steps of constructing a quality model according to the key data set include:

[0021] Establish an expression of the quality model Q(q) = f(u, p) = a + b(u + p) + c(u + p) 2 ; where a, b, and c are the coefficients of the quality model expression;

[0022] Based on the key data set, fit the expression of the quality model to calculate the coefficients in the expression of the quality model.

[0023] Optionally, the steps of constructing a quality model according to the key data set include:

[0024] Establish a machine training model;

[0025] Using the process key data and the production process key data as input data, and using the quality key data as output data, train the machine learning model;

[0026] Construct the quality model according to the trained machine learning model.

[0027] Optionally, after constructing the quality model according to the key data set, it further includes:

[0028] Detect the constructed quality model to determine whether the quality model is qualified.

[0029] Optionally, determining whether the quality model is qualified includes the following steps:

[0030] Obtain the confidence threshold;

[0031] Calculate the confidence of the constructed quality model;

[0032] When the confidence of the quality model is less than the confidence threshold, determine that the quality model is unqualified.

[0033] Optionally, determining whether the quality model is qualified includes the following steps:

[0034] Based on the quality problems in the manufacturing process, obtain the actual production data;

[0035] According to the actual production data, use the quality model to calculate the quality prediction value;

[0036] Calculate the difference between the quality prediction value and the actual quality value in the actual production data, and record it as the quality deviation value;

[0037] Obtain the quality deviation threshold;

[0038] When the quality deviation value is greater than the quality threshold, determine that the quality model is unqualified.

[0039] Optionally, after determining whether the quality model is qualified, it further includes the following steps:

[0040] When it is determined that the quality model is unqualified, based on the quality problems in the manufacturing process, obtain the additional key data set of the manufacturing process;

[0041] Based on the additional key data set, adjust the constructed quality model.

[0042] As described above, compared with the prior art, the method for constructing a ship quality model provided by this application has at least the following beneficial effects:

[0043] In this application, a large amount of production data accumulated during production is utilized to construct a data-driven ship quality model, which can predict the quality data in the manufacturing process in advance, thereby providing data support for actual construction. Staff can analyze and make decisions based on the prediction results to determine whether it is necessary to adjust process parameters and the production process; by quantitatively describing the key process data, key production process data, and key quality data, the specific data ranges of the key data sets are clarified and quantified, realizing the construction and optimization of the quality model, getting rid of the dependence on work experience, reducing the possible errors in quality analysis and decision-making, helping to improve the accuracy and reliability of quality control, and facilitating the development of ship quality control towards the direction of digitization and refinement; the influence of other non-critical factors is ignored during the construction process, reducing the difficulty of constructing the quality model, improving the timeliness and work efficiency of quality control operations, and reducing the safety risks of ships. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0045] Figure 1 It shows a schematic flowchart of the method for constructing a ship quality model provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] To make the technical objectives, technical solutions, and technical effects of the present application clearer, the technical solutions in the present application will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations.

[0047] Therefore, the following detailed description of the embodiments of the present application is not intended to limit the scope of the present application to be protected, but only represents the selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0048] In the description of the present application, it should be noted that the descriptions with reference to terms such as "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0049] In view of the problems in the prior art, such as the low efficiency and poor accuracy of manual inspection in the ship quality control method mainly based on manual inspection and analysis decision-making, this embodiment provides a method for constructing a ship quality model.

[0050] Specifically, referring to Figure 1 , the method for constructing the ship quality model provided by the present application includes steps S1 to S4, specifically including:

[0051] Step S1: Divide the shipbuilding process into several manufacturing links;

[0052] Step S2: Obtain the quality problems of each of the manufacturing links;

[0053] Step S3: Based on the quality problems of the manufacturing links, obtain the key data sets of the manufacturing links, where the key data sets include process key data u, production process key data p, and quality key data q;

[0054] Step S4: Construct a quality model Q(q) = f(u, p) according to the key data sets.

[0055] The following will detail the method for constructing the ship quality model of this embodiment. It should be noted that the above order does not strictly represent the step order of the method for constructing the ship quality model protected by the present application, and those skilled in the art can adjust it according to actual needs.

[0056] First, execute step S1 to divide the shipbuilding process into several manufacturing links.

[0057] In this embodiment, executing step S1 includes the following steps: Obtain the process flow data set of the shipbuilding process; Based on the process flow data set, divide the shipbuilding process into several manufacturing links. Optionally, the shipbuilding process may include, for example, a pretreatment link, a cutting link, a processing and forming link, a small assembly manufacturing link, a medium assembly manufacturing link, and other related manufacturing links.

[0058] Next, execute step S2 to obtain the quality problems of each manufacturing link.

[0059] In the shipbuilding process, each manufacturing link has a different quality formation mechanism due to differences in manufacturing processes, and thus has its specific quality problems. When the manufacturing link and the quality problems faced are different, the constructed quality model is also different. By dividing the shipbuilding process into several manufacturing links, it is possible to clarify the usage link of the established quality model, and it is also convenient to clarify the quality problems to be faced and solved subsequently, so as to construct a specific quality model to ensure that the constructed quality model can be applied to the actual quality control process.

[0060] In an alternative embodiment, performing step S2 includes the following steps: obtaining the past production data of the shipbuilding process; extracting the quality problems of the manufacturing links from the past production data. Among them, the past production data includes past process parameter data, past structural parameter data, past process requirement data, and other relevant data in the past shipbuilding process, such as past user feedback data, etc.

[0061] Further, several manufacturing links include, for example, the cutting link. The quality problems of the cutting link may include, for example: the cutting size deviation is too large. Specifically, for example, the process requirement range of the size deviation in the cutting link is -1.5 mm to 1.5 mm, and the actual deviation range of the size deviation in actual processing is -3.2 mm to 2.3 mm. It can be seen that the actual deviation range exceeds the process requirement range; therefore, the usage link of the quality model is the cutting link, the quality problem to be solved by the quality model is that the cutting size deviation is too large, and based on the quality model, the range for solving the quality problem is to control the deviation range in actual processing within -1.5 mm to 1.5 mm.

[0062] Further, several manufacturing links include, for example, the sub-assembly manufacturing link. The quality problems of the sub-assembly manufacturing link may include, for example, that the levelness deviation of the sub-assembly welding is too large. Specifically, for example, the process requirement range of the welding levelness deviation in the sub-assembly manufacturing link is -0.5 mm to 1.5 mm, and the range of the sub-assembly welding levelness deviation in actual processing is -3.2 mm to 2.3 mm. It can be seen that the actual range of the sub-assembly welding levelness deviation exceeds the process requirement range; therefore, the usage link of the constructed quality model is the sub-assembly manufacturing link, the quality problem to be solved by the quality model is that the levelness deviation of the sub-assembly welding is too large, and based on the quality model, the method for solving the quality problem is to control the range of the sub-assembly welding levelness deviation in actual processing within -0.5 mm to 1.5 mm.

[0063] Next, perform step S3. Based on the quality problems of the manufacturing link, obtain the key data set of the manufacturing link. Among them, the key data set includes process key data u, production process key data p, and quality key data q.

[0064] Specifically, based on the quality problems to be solved, analyze the manufacturing process of the corresponding manufacturing link and the formation mechanism of the quality problems, and determine the key factors affecting the quality problems of this manufacturing link. Optionally, the steps for obtaining the key data set of the manufacturing link in step S3 include the following steps: obtain the process flow information of the manufacturing link; based on the process flow information, divide the manufacturing link into several manufacturing sub-links; based on the quality problems of the manufacturing link, obtain the sub-link parameter data of the manufacturing sub-links; from the sub-link parameter data, extract process key data, production process key data, and quality key data respectively. Further, the process key data includes, for example, part parameter data, the production process key data includes, for example, process parameter data, and the quality key data includes, for example, deviation parameter data.

[0065] Among them, the process key data, production process key data, and quality key data in the sub-link parameter data are the key factors affecting the quality problems of the corresponding manufacturing link in the manufacturing sub-links, and the set of key factors of each sub-link forms a key data set. Optionally, the sub-link parameter data may also include non-critical data, and the non-critical data is a non-critical factor that has a negligible impact or no impact on the quality problems of the manufacturing link, and the set of non-critical data of each manufacturing sub-link forms a non-critical data set.

[0066] In an alternative embodiment, several manufacturing links include, for example, a cutting link, and the manufacturing sub-links of the cutting link include a steel plate loading sub-link, a positioning sub-link, a trial cutting sub-link, a marking sub-link, and a cutting sub-link, etc. Analyze and sort out the above-mentioned various manufacturing sub-links to clarify the formation mechanism of their quality problems, that is, sort out the factors affecting the quality problems of each manufacturing sub-link. Based on the quality problems of the manufacturing link, the obtained sub-link parameter data may include, for example, the positioning accuracy of the positioning sub-link, the marking accuracy of the marking sub-link, the cutting speed of the cutting sub-link, and other related data. From the sub-link parameter data, the extracted process key data includes, for example, part parameter data such as steel plate material, steel plate thickness, and part size, and other appropriate related data; the extracted production process key data includes, for example, process parameter data such as cutting speed and cutting power, and other appropriate related data; the extracted quality key data includes, for example, deviation parameter data such as part size deviation, and other appropriate related data.

[0067] In an alternative embodiment, several manufacturing processes, for example, include the sub-assembly manufacturing process. The manufacturing sub-processes of the sub-assembly manufacturing process, for example, include the part loading sub-process, the assembly sub-process, the grinding sub-process, the welding sub-process, etc. Analyze and sort out the above-mentioned various manufacturing sub-processes to clarify the formation mechanism of their quality problems, that is, sort out the factors affecting quality problems in each manufacturing sub-process. Based on the quality problems of the manufacturing process, the obtained sub-process parameter data may include, for example, the assembly accuracy and spot welding sequence of the assembly sub-process, the environmental parameters (temperature, humidity, wind speed, etc.) and welding machine parameters (current, voltage, etc.) of the welding sub-process, and other related data. From the sub-process parameter data, the extracted process key data includes, for example, the assembly sequence, spot welding sequence, part material, groove size, etc.; the extracted key production process data includes, for example, environmental parameters (temperature, humidity, wind speed, etc.) and welding machine parameters (current, voltage, etc.); the extracted key quality data includes, for example, the pre-welding inspection data of the sub-assembly and the post-welding inspection data of the sub-assembly.

[0068] Finally, perform step S4. According to the key data set, construct the quality model Q(q) = f(u, p). Select appropriate mathematical models and mathematical tools to describe the quality problems to construct the quality model.

[0069] In an alternative embodiment, performing step S4 includes the following steps: establishing an expression of the quality model Q(q) = f(u, p) = a + b(u + p) + c(u + p) 2 ; where a, b, and c are the coefficients of the expression of the quality model; based on the key data set, fit the expression of the quality model to calculate the coefficients in the expression of the quality model. Among them, the process key data u may include, for example, multiple sets of data such as u1, u2,... u m and so on, and the key production process data p may include, for example, multiple sets of data such as p1, p2,... p r and so on, and the key quality data q may include, for example, multiple sets of data such as q1,..., q n and so on.

[0070] In an alternative embodiment, performing step S4 includes the following steps: establishing a machine training model; using the process key data and the key production process data as input data, and using the key quality data as output data to train the machine training model; based on the trained machine training model, obtain the quality model; specifically, obtain the parameters in the quality model through the machine training model to obtain the quality model.

[0071] In an alternative embodiment, before constructing the quality model, the following steps are further included: simplifying the assumptions for the construction process of the quality model. Specifically, by assuming that the non-critical factors affecting the quality problems in the manufacturing process are ideal situations, when constructing the quality model, the influence of non-critical factors can be ignored, and only the critical factors need to be considered, which simplifies the construction process of the quality model, reduces the construction difficulty of the quality model, and improves work efficiency. Optionally, the sub-process parameter data includes non-critical data, and the non-critical data in the sub-process parameter data can be standardized to simplify the assumptions for the construction process of the quality model.

[0072] Furthermore, several manufacturing processes, such as the cutting process, can assume that the environmental parameters of the cutting process are ideal situations to ignore the influence of environmental factors on the cutting process, and assume that the working conditions of the cutting equipment are stable and reliable ideal situations to ensure the accuracy of cutting data and precision. Other appropriate assumptions can also be made according to actual needs to ignore the influence of non-critical factors on quality problems.

[0073] Furthermore, several manufacturing processes, such as the sub-assembly production process, can assume that the welding equipment, welding materials, and welding process are all stable and reliable ideal situations to ensure the accuracy of the acquired data, so as to ignore the influence of non-critical factors such as different construction workers and welding machine voltage drops on quality problems.

[0074] Since there are many influencing factors for the formation of quality problems, and certain assumption simplifications are made when constructing the quality model, there must be a certain deviation between the quality prediction value and the measured value. Therefore, in the actual ship production process, one or more manufacturing processes can be used as objects, combined with the process data in the production process and the production process data collected, and the constructed quality model can be used for prediction to obtain the quality data predicted before production, and this quality data can be compared and analyzed with the quality data after production to determine whether the constructed quality model is accurate and reliable.

[0075] In an alternative embodiment, after constructing the quality model according to the key data set, the following is further included: detecting the constructed quality model to determine whether the quality model is qualified.

[0076] Furthermore, determining whether the quality model is qualified may include the following steps: obtaining a confidence threshold; calculating the confidence of the constructed quality model; when the confidence of the quality model is less than the confidence threshold, determining that the quality model is unqualified. Among them, the confidence threshold can be set according to actual needs. For example, the confidence threshold can be set to 95%, and for the key manufacturing processes of certain ships, the confidence threshold can also be set to 99%.

[0077] Further, determining whether the quality model is qualified may include the following steps: Based on the quality problems in the manufacturing process, obtain the actual production data; According to the actual production data, use the quality model to calculate the quality prediction value; Calculate the difference between the quality prediction value and the actual quality value in the actual production data, and record it as the quality deviation value; Obtain the quality deviation threshold; When the quality deviation value is greater than the quality deviation threshold, determine that the quality model is unqualified. Among them, the actual production data includes actual process key data, actual production key data, and actual quality value, which respectively correspond to process key data, production process key data, and quality key data. There is a mathematical relationship between the actual process key data, actual production key data, and actual quality value that is approximately the same as or the same as the quality model. Based on the constructed quality model, the quality prediction value can be calculated according to the actual process key data and actual production key data. The quality deviation threshold can be set according to actual needs. For example, the quality deviation threshold is set to 0.5 mm, and the confidence level of the quality model is 96%, that is, there is a 96% probability that the quality deviation value between the quality prediction value and the actual quality value of this quality model falls within the range of 0.5 mm or less.

[0078] In an alternative embodiment, after determining whether the quality model is qualified, the following steps are further included: When it is determined that the quality model is unqualified, based on the quality problems in the manufacturing process, obtain the additional key data set of the manufacturing process; Based on the additional key data set, adjust the constructed quality model. Among them, the additional key data set is used to expand the data volume of the key data set. According to the increased additional key data set, the relevant parameters in the constructed quality model can be adjusted to optimize the quality model, or the method of constructing the quality model (such as adjusting the mathematical model for constructing the quality model) can also be adjusted to optimize the quality model, so that the prediction result of the quality model is more accurate and reliable.

[0079] As described above, in the method for constructing the ship quality model of this embodiment, a large amount of production data accumulated in production is used to construct a data-driven ship quality model, which can predict the quality data in the manufacturing process in advance, thereby providing data support for actual construction. Staff can analyze and make decisions based on the prediction results to determine whether it is necessary to adjust the process parameters and production process; By quantitatively describing the process key data, production process key data, and quality key data, the specific data range of the key data set is clarified and quantified, which can not only support the construction and optimization of the quality model, but also get rid of the dependence on work experience, reduce the possible errors in making quality analysis decisions, help improve the accuracy and reliability of quality control, and facilitate the development of ship quality control towards digitalization and refinement; Moreover, the influence of other non-critical factors is ignored during the construction process, reducing the difficulty of constructing the quality model, improving the timeliness and work efficiency of quality control operations, and reducing the safety risks of ships.

[0080] The above embodiments are only illustrative of the principles and effects of the present application and are not intended to limit the present application. Any person familiar with this technology can modify, change or combine the above embodiments without departing from the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed in the present application should still be covered by the claims of the present application.

Claims

1. A method for constructing a ship quality model, characterized in that: The following steps are involved: Divide the shipbuilding process into several manufacturing stages; Obtain the quality issues of each manufacturing link; Based on the quality problems in the manufacturing link, a key data set of the manufacturing link is obtained, wherein the key data set includes process key data u, production process key data p and quality key data q; According to the key data set, a quality model Q(q)=f(u,p) is constructed.

2. The method for constructing a ship quality model according to claim 1, characterized in that: The shipbuilding process is divided into several manufacturing links, including the following steps: Acquire a process flow data set of the ship construction process; Based on the process flow data set, the shipbuilding process is divided into several manufacturing links; Among them, some of the manufacturing links at least include a pre-processing link, a cutting link, a processing and forming link, a small group assembly production link, and a medium assembly production link.

3. The method for constructing a ship quality model according to claim 1, characterized in that: The step of obtaining the key data set of the manufacturing process comprises the following steps: Obtaining process flow information of the manufacturing link; Based on the process flow information, the manufacturing process is divided into a number of manufacturing sub-processes; Based on the quality problem of the manufacturing link, obtaining sub-link parameter data of the manufacturing sub-link; From the sub-link parameter data, process key data, production process key data and quality key data are extracted respectively.

4. The method for constructing a ship quality model according to claim 3, characterized in that: The key process data includes part parameter data, the key production process data includes process parameter data, and the key quality data includes deviation parameter data.

5. The method for constructing a ship quality model according to claim 1, characterized in that: Based on the key data sets, the steps of building a quality model include: The expression of the quality model is established: Q(q)=f(u,p)=a+b(u+p)+c(u+p) 2 ; Wherein, a, b, c are the coefficients of the quality model expression; Based on the key data set, the expression of the quality model is fitted to calculate the coefficients in the expression of the quality model.

6. The method for constructing a ship quality model according to claim 1, characterized in that: Based on the key data sets, the steps of building a quality model include: Building machine training models; Training the machine training model using the process key data and the production process key data as input data and the quality key data as output data; The quality model is constructed according to the trained machine training model.

7. The method for constructing a ship quality model according to claim 1, characterized in that: After building the quality model based on the key data set, it also includes: The constructed quality model is tested to determine whether the quality model is qualified.

8. The method for constructing a ship quality model according to claim 7, characterized in that: Determining whether the quality model is qualified includes the following steps: Get the confidence threshold; Calculating the confidence of the constructed quality model; When the confidence of the quality model is less than the confidence threshold, the quality model is determined to be unqualified.

9. The method for constructing a ship quality model according to claim 7, characterized in that: Determining whether the quality model is qualified includes the following steps: Based on the quality issues in the manufacturing process, obtain actual production data; According to the actual production data, the quality prediction value is calculated by using the quality model; Calculate the difference between the quality prediction value and the actual quality value in the actual production data, and record it as the quality deviation value; Get the quality deviation threshold; When the quality deviation value is greater than the quality threshold, the quality model is determined to be unqualified.

10. The method for constructing a ship quality model according to claim 8 or 9, characterized in that: After determining whether the quality model is qualified, the following steps are also included: When it is determined that the quality model is unqualified, based on the quality problem of the manufacturing link, obtaining an additional key data set of the manufacturing link; Based on the additional key data sets, the constructed quality model is adjusted.

Images