Process beat determination method, apparatus, computer device, and storage medium

Abstract

The application relates to a process rhythm determination method and device, computer equipment and a storage medium. The method comprises the following steps: based on production process data and production line body attribute data, a first data set is processed, the first data set comprises the duration of each action of a target line body in a target station and a target production model relative to the first action; when there is a missing value in the first data set, the first data set is converted into a first matrix through two-dimensional matrix transformation; the missing value in the first matrix is filled by using a K-nearest neighbor model and a chain equation model of a random forest, to obtain a corresponding filling value; the filling value is supplemented to the position of the corresponding missing value in the first data set, to obtain a second data set; and based on the second data set, the process rhythm of the target line body in the target station and the target production model is determined. The method can better fill the missing value and improve the accuracy of the process rhythm calculation result.

Description

Technical Field

[0001] This application relates to the field of industrial production line control technology, and in particular to a method, apparatus, computer equipment and storage medium for determining process cycle time. Background Technology

[0002] With the acceleration of the industrialization of flexible manufacturing, modern measurement technology, intelligent sensor technology and information technology have been widely applied to industries such as automobile manufacturing. A large amount of process data has been accumulated in the production process. However, in the actual data collection and transmission process, data is missing at certain points in time. In particular, when the data of key actions at the workstation is missing, there will be certain errors in the calculation of the process cycle.

[0003] Currently, there are many methods for calculating process cycle time, but they cannot provide accurate handling of missing data, such as the missing duration of the last action at the workstation or the missing start time of the action, which limits the accuracy of process cycle time determination. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, and storage medium for determining process cycle time that can complete missing data to improve the accuracy of process cycle time, in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides a method for determining process cycle time. The method includes:

[0006] Based on production process data and production line attribute data, a first dataset is obtained. The first dataset includes the duration of each action of the target line in the target workstation and target production vehicle process relative to the first action.

[0007] If there are missing values ​​in the first dataset, the first dataset is transformed into a two-dimensional matrix to obtain the first matrix;

[0008] The missing values ​​in the first matrix are filled using the K-nearest neighbor model and the chain equation model of random forest to obtain the corresponding filled values;

[0009] The missing values ​​are filled into the positions of the corresponding missing values ​​in the first dataset to obtain the second dataset;

[0010] Based on the second dataset, the process cycle time of the target production line at the target workstation and in the process of producing the target vehicle model is determined.

[0011] In one embodiment, the missing values ​​in the first matrix are filled using a K-nearest neighbor model and a chain equation model of random forest, resulting in the corresponding filled values, including:

[0012] The missing values ​​in the first matrix are filled using the K-nearest neighbor model to obtain the second matrix;

[0013] The missing values ​​in the first matrix are filled using a chain equation model of random forest to obtain the third matrix;

[0014] Based on the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix, the filling values ​​corresponding to the missing values ​​are obtained.

[0015] In one embodiment, based on the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix, the filling values ​​corresponding to the missing values ​​are obtained, including:

[0016] The missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix are weighted and summed to obtain the filling values ​​corresponding to the missing values.

[0017] In one embodiment, a first dataset is obtained based on production process data and production line attribute data, including:

[0018] Acquire production process data and production line attribute data of the production line; production process data includes production vehicle model, line unique identifier, workstation unique identifier, action group unique identifier, action unique identifier, and action time information; production line attribute data includes production line unique identifier, workstation unique identifier, action group unique identifier, and action unique identifier.

[0019] Based on the line information in the production process data and the unique identification codes of the line, workstation, action group, and action in the production line attribute data, the production process data and the production line attribute data are correlated to obtain a third dataset; the third dataset includes the duration of each action performed by each workstation of each line for each production model relative to the start time.

[0020] The first dataset is obtained by extracting the duration of each production action relative to the first action in the process of the target production line at the target workstation and the target production model from the third dataset.

[0021] In one embodiment, the production line attribute data also includes the line number, workstation number, action group number, and action number;

[0022] Obtain production line attribute data, including:

[0023] Obtain the inherent attribute data of the production line, which includes the line number, line unique identifier, workstation number, workstation unique identifier, action group unique identifier, original action number, and action unique identifier.

[0024] The inherent attribute data of the production line are grouped according to the line number and workstation number, and sorted according to the action group number and original action number in each workstation.

[0025] Based on the order of sorted actions, action sequence numbers are generated, and a mapping table between different workstations and actions of different production lines is constructed to obtain production line attribute data.

[0026] In one embodiment, the process cycle time includes the actual cycle time of the target station, the production cycle time of the target station, the start time of each action, the end time of each action, and the duration of each action; wherein, the actual cycle time of the station is the maximum duration of the last action in the target station relative to the first action; the production cycle time of the station is the difference between the start time of the first action of the target station and the start time of the first action of the next station.

[0027] Secondly, this application also provides a process cycle determination device. The device includes:

[0028] The first dataset acquisition module is used to process production process data and production line attribute data to obtain the first dataset. The first dataset includes the duration of each production action of the target line in the target workstation and target production vehicle process relative to the first action.

[0029] The first matrix acquisition module is used to convert the first dataset into a two-dimensional matrix to obtain the first matrix when there are missing values ​​in the first dataset. The value of each element in the first matrix is ​​the corresponding value in the first dataset.

[0030] The fill value determination module is used to fill the missing values ​​in the first matrix using the K-nearest neighbor model and the chain equation model of random forest to obtain the corresponding fill values;

[0031] The second dataset generation module is used to fill in the missing values ​​in the first dataset to obtain the second dataset.

[0032] The process cycle time determination module is used to calculate the process cycle time of the target production line at the target workstation and the target production model based on the second dataset.

[0033] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0034] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0035] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described method.

[0036] The aforementioned method, apparatus, computer equipment, storage medium, and computer program products for determining process cycle time, based on production process data and production line attribute data, process to obtain a first dataset. The first dataset includes the duration of each action of the target line at the target workstation and in the process of producing the target vehicle model, relative to the first action. When missing values ​​exist in the first dataset, it is transformed into a two-dimensional matrix to obtain a first matrix. The missing values ​​in the first matrix are filled using a K-nearest neighbor model and a chain equation model of random forest to obtain the corresponding filled values. The filled values ​​are then inserted into the positions of the corresponding missing values ​​in the first dataset to obtain a second dataset. Based on the second dataset, the process cycle time of the target line at the target workstation and in the process of producing the target vehicle model is determined. By fusing production process data with production line attribute data, a time-series dataset of actions at different workstations of different lines is constructed, providing support for subsequent missing data filling, thereby enabling better filling of missing values ​​and improving the accuracy of the process cycle time calculation results. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a method for determining process cycle time according to one embodiment;

[0038] Figure 2 This is a schematic diagram of the action mapping table for different workstations of line number 1 in one embodiment;

[0039] Figure 3 This is a flowchart illustrating the process of filling missing values ​​in the first matrix using a chain equation model of K-nearest neighbors and random forest, as an example, to obtain the corresponding filled values.

[0040] Figure 4 This is a flowchart illustrating how a first dataset is obtained by processing production process data and production line attribute data according to one embodiment.

[0041] Figure 5 This is a schematic diagram of a process for obtaining production line attribute data according to one embodiment;

[0042] Figure 6 This is a partial data diagram showing the duration (milliseconds) of each action of workstation 1 at production line number 1 in a production vehicle model number 100 relative to the first action, according to one embodiment.

[0043] Figure 7 A flowchart illustrating a method for determining process cycle time according to another embodiment;

[0044] Figure 8 This is a schematic diagram of the structure of an industrial cycle time determination device according to an embodiment;

[0045] Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0047] In one embodiment, such as Figure 1 As shown, a method for determining process cycle time is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0048] Step 102: Based on the production process data and production line attribute data, a first dataset is obtained. The first dataset includes the duration of each action of the target line in the target workstation and target production vehicle process relative to the first action.

[0049] Specifically, production process data can be data collected from the production line during the production process, or data obtained by further processing data collected from the production line. For example, it can include one or more of the following data: the target vehicle model produced by the production line, the line's unique identifier, the workstation's unique identifier, the action's unique identifier, the workstation's first action marker, the action's start time, the action's end time, and the action's duration.

[0050] Production line attribute data can be raw attribute data such as inherent attribute data of the production line and manually configured data, or it can be data obtained by further processing of raw attribute data. For example, it can be one or more of the following data: line number, line unique identifier, workstation number, workstation unique identifier, action group number, action group unique identifier, action number, and action unique identifier.

[0051] Based on the process cycle time to be determined, the target production line, target workstation, and target production vehicle model are identified. Then, based on the production process data and production line attribute data, a first dataset is obtained. The first dataset includes the duration of each action of the target production line at the target workstation and in the process of producing the target production vehicle model, relative to the first action.

[0052] Figure 2The example shows partial data of the duration (in milliseconds) of each action (action number 1-12) relative to the first action at workstation number 1 of line number 1 in production model number 100.

[0053] Step 104: When there are missing values ​​in the first dataset, the first dataset is transformed into a two-dimensional matrix to obtain the first matrix.

[0054] Specifically, if errors or omissions occur during the actual data collection and transmission process, data may be missing at certain points in time, resulting in missing values ​​in the first dataset.

[0055] In the case of missing data, the accuracy of process cycle calculation is limited. Therefore, the first dataset can be transformed into a two-dimensional matrix to fill in the missing values.

[0056] by Figure 2 Taking the first dataset as an example, the location of missing data is marked as nan, and the cell indicates that the corresponding data is missing.

[0057] Step 106: Use the K-nearest neighbor model and the chain equation model of random forest to fill the missing values ​​in the first matrix to obtain the corresponding filled values.

[0058] For example, the fill value can be a value derived from existing values ​​in the matrix, used to supplement missing data. Figure 2 Taking the first dataset as an example, the positions marked as NaN in the first dataset have empty values ​​in the corresponding positions of the first matrix. The padding values ​​can be calculated values ​​obtained from the empty positions in the first matrix. For example, the padding values ​​can represent the duration of missing actions or the start time of the first action caused during data acquisition and transmission.

[0059] Step 108: Fill in the missing values ​​in the first dataset to obtain the second dataset.

[0060] For example, a cell that is represented as nan in the first dataset is represented as the corresponding fill value in the second dataset.

[0061] Alternatively, the second matrix can be obtained by padding the first matrix with values, and then the second matrix can be converted into a second dataset.

[0062] Step 110: Based on the second dataset, determine the process cycle time of the target production line at the target workstation and in the process of producing the target vehicle model.

[0063] Specifically, the second dataset is the dataset obtained by filling in the missing values. That is, in the second dataset, the duration data of each action of the target line relative to the first action in the target station and the target production model process is not missing. Based on the complete duration of each action of the target line relative to the first action in the target station and the target production model process, the process cycle time of the target line in the target station and the target production model process can be determined.

[0064] In the aforementioned method for determining the cycle time, a first dataset is obtained based on production process data and production line attribute data. This first dataset includes the duration of each action of the target line at the target workstation and in the process of producing the target vehicle model, relative to the first action. If missing values ​​exist in the first dataset, it is transformed into a two-dimensional matrix to obtain a first matrix. The missing values ​​in the first matrix are filled using a K-nearest neighbor model and a chain equation model based on random forests, yielding the corresponding filled values. These filled values ​​are then inserted into the positions of the corresponding missing values ​​in the first dataset to obtain a second dataset. Based on the second dataset, the cycle time of the target line at the target workstation and in the process of producing the target vehicle model is determined. By fusing production process data with production line attribute data, a time-series dataset of actions at different workstations on different lines is constructed, providing support for subsequent missing data filling. This allows for better filling of missing values ​​and improves the accuracy of the cycle time calculation results.

[0065] In one embodiment, such as Figure 3 As shown, the missing values ​​in the first matrix are filled using the K-nearest neighbor model and the chain equation model of random forest, and the corresponding filled values ​​can include:

[0066] Step 302: Use the K-nearest neighbor model to fill in the missing values ​​in the first matrix to obtain the second matrix.

[0067] Specifically, based on a preset number of neighboring samples used to fill in missing values, the K-nearest neighbor model is used to fill in the missing values ​​in the first matrix Y, resulting in a second matrix determined by the K-nearest neighbor model. For example, the second matrix can be called the filling matrix Y_1. The value at the corresponding position of the missing value in the first matrix Y in the filling matrix Y_1 can be the first filling value for that missing value. For example, the number of neighboring samples can be 5. This application embodiment does not limit the numerical value of the number of neighboring samples.

[0068] Step 304: Use the chain equation model of random forest to fill the missing values ​​in the first matrix to obtain the third matrix.

[0069] Specifically, all action columns and the first action start time column of the base matrix Y are used as feature columns, and a new column is added as the category column. A chain equation model of random forest is constructed to fill the missing data in the feature columns of the base matrix Y, resulting in a third matrix. For example, the third matrix can be called the filling matrix Y_2; where the missing value in the first matrix Y corresponds to the value in the filling matrix Y_2, and can be the second filling value for that missing value.

[0070] Step 306: Based on the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix, obtain the filling values ​​corresponding to the missing values.

[0071] Specifically, based on preset conditions, the fill value corresponding to the missing value can be obtained according to the first fill value and the second fill value.

[0072] The process cycle determination method provided in this application uses both the K-nearest neighbor model and the chain equation model of random forest to calculate the filling value corresponding to the missing value, thereby improving the accuracy of the filling value calculation.

[0073] In one embodiment, obtaining the imputed value corresponding to the missing value based on the imputed values ​​in the second matrix and the imputed values ​​in the third matrix may include:

[0074] The missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix are weighted and summed to obtain the filling values ​​corresponding to the missing values.

[0075] For example, the fill values ​​obtained from the two methods are weighted and summed to obtain the final fill value. For example, fill value = first fill value * +Second fill value* ,in The weighting coefficient corresponding to the first fill value. The weighting coefficient corresponding to the second filler value. .

[0076] The process cycle determination method provided in this application uses both the K-nearest neighbor model and the chain equation model of random forest to calculate the filling value corresponding to the missing value, and performs a weighted sum of the filling values ​​calculated by the two methods to obtain the final filling value. This method considers both the characteristics of the neighboring data of the missing data and the nonlinear correlation between actions, and can better fill the missing value.

[0077] In one embodiment, such as Figure 4 As shown, the first dataset is obtained by processing production process data and production line attribute data, which may include the following steps:

[0078] Step 402: Obtain the production process data and production line attribute data of the production line. The production process data includes the vehicle model being produced, the line's unique identifier, the workstation's unique identifier, the action group's unique identifier, the action's unique identifier, and the action's timing information. The production line attribute data includes the production line's unique identifier, the workstation's unique identifier, the action group's unique identifier, and the action's unique identifier.

[0079] For example, production process data and production line attribute data can be obtained through a production line data service platform. In a specific application scenario, production process data and production line attribute data can be collected through a self-built intelligent manufacturing data service platform.

[0080] Step 404: Based on the line information in the production process data and the unique identification code of the line, the unique identification code of the workstation, the unique identification code of the action group, and the unique identification code of the action in the production line attribute data, the production process data and the production line attribute data are associated to obtain a third dataset; the third dataset includes the duration of each action performed by each workstation of each line for each production model relative to the start time.

[0081] For example, the production line attribute data includes the line number. and unique identifier Workstation number and unique identifier Action group number and unique identifier Action sequence number and unique identifier Attribute data, where Indicates the first A line body, , Indicates the first The first line body One workstation. , Indicates the first The first line body The first workstation A set of actions, , Indicates the first The first line body The first workstation The first action group One action, Production process data includes data for different vehicle models. Unique identification code for the line Unique identification code for workstation Unique Identifier Code for Action Group Action unique identifier The first action sign of the workstation ( (Indicates the first action), action start time. Action end time Duration of the action ,in .

[0082] Among them, the unique identifiers for the production line, workstation, action group, and action exist in both the production line attribute data and the production process data. By associating the production line attribute data and the production process data using these unique identifiers, a fused dataset can be obtained. This refers to the third dataset. Indicates the first A line body, Indicates the first One workstation. Indicates the first One action, Indicates vehicle model, Indicates the first The action starts at time [time missing]. The duration is long.

[0083] Step 406: Extract the duration of each production action of the target line body relative to the first action in the process of the target workstation and the target production model from the third dataset to obtain the first dataset.

[0084] Specifically, the third dataset includes the duration of multiple production actions relative to the first action in multiple production processes across multiple work lines and multiple vehicle models, and may also include the start time of the first action. The process cycle time is determined for each production action of the target work line at the target station and in the target vehicle model process; therefore, the first dataset required to determine the process cycle time is extracted from the third dataset.

[0085] For example, from the dataset Extracting the line body workstation ,action Models in production The relative duration of the first action is used to obtain the dataset. (i.e., the first dataset). Among them... Indicates the first action marker The number of times it occurs, and the time difference between adjacent first actions is greater than This refers to the number of cycles within a certain period of time. , Indicates the first The first line body Total number of actions per workstation This column indicates the start time of the first action. This indicates the end time of the final action. Indicates the first The first cycle The duration of each action relative to the first action. Indicates the first The start time of the first action and the end time of the last action in each cycle. These are adjustable parameters.

[0086] In the process cycle determination method provided in this application embodiment, the production process data and production line attribute data are integrated based on a unique identifier, which can accurately construct the time sequence dataset of each action of different workstations on different lines, providing support for subsequent filling of missing data.

[0087] In one embodiment, the production line attribute data also includes the line number, workstation number, action group number, and action number;

[0088] like Figure 5 As shown, obtaining production line attribute data may include the following steps:

[0089] Step 502: Obtain the inherent attribute data of the production line body. The inherent attribute data of the production line body includes the line body number, the line body unique identifier, the workstation number, the workstation unique identifier, the action group unique identifier, the original action number, and the action unique identifier.

[0090] The inherent attribute data can be data such as the number, sequence number, and identification code of the production line, workstation, action group, and action. For example, inherent attribute data can be obtained through data platforms such as production line data service platforms or enterprise-built intelligent manufacturing data service platforms.

[0091] Step 504: Group the inherent attribute data of the production line according to the line number and workstation number, and sort them based on the action group number and original action number in each workstation.

[0092] In a specific application scenario, such as Figure 6 As shown, the inherent attribute data of production line body number 1 is grouped and sorted. For example, the sequence numbers of the action groups and the action numbers of each workstation are sorted in ascending order to obtain the following results: Figure 6 The arrangement order is shown.

[0093] Step 506: Based on the order of the sorted actions, generate action sequence numbers, construct a mapping table between different workstations and actions of different production lines, and obtain production line attribute data.

[0094] Specifically, in Figure 6 In the example production line inherent attribute data, based on the ascending sorting of the action group number and action number of each workstation, new action numbers are generated sequentially, and a mapping table between different workstations and actions of different production lines is constructed to obtain the production line attribute data.

[0095] The process cycle determination method provided in this application provides an accurate basis for the association and fusion of production process data and production line attribute data by constructing action attribute codes and mapping tables for different workstations on different production lines.

[0096] In one embodiment, the process cycle time includes the actual cycle time of the target station, the production cycle time of the target station, the start time of each action, the end time of each action, and the duration of each action; wherein, the actual cycle time of the station is the maximum duration of the last action in the target station relative to the first action; the production cycle time of the station is the difference between the start time of the first action of the target station and the start time of the first action of the next station.

[0097] For example, calculate the first The production cycle of each workstation is corresponding to one cycle. , can be represented as ; Calculate the actual cycle time of the workstation , can be represented as , Indicates the first The duration of the last action of each cycle relative to the duration of the first action. Indicates the first The first cycle The duration of each action relative to the first action. Indicates the first The start time of the first action and the end time of the last action in each cycle.

[0098] To further illustrate the solutions of the embodiments of this application, a specific example is provided below, taking the application of this method to production line attribute data and production process data extracted from an intelligent manufacturing data service platform as an example. Figure 7 As shown, it includes the following steps:

[0099] Step 1: Extract the sequence number and unique identifier of the production line, workstation, action group, and action from the intelligent manufacturing data service platform.

[0100] Step 2: Extract the dataset with line number 1, group the attribute data according to line number and workstation number, and sort the action group number and action number of each workstation in ascending order to generate new numbers between different workstations and each action. Construct a mapping table between different workstations and different actions for different lines, as shown below. Figure 6 The mapping table shown.

[0101] Step 3: Extract vehicle model number 100 from the period from 00:00:00 on 2022 / 09 / 10 to 00:00:00 on 2022 / 09 / 17. Taking the production process data of each action for line number 1 and unique identifier 100001, and workstation number 1 and unique identifier 200001 as an example, this includes the first action marker of the workstation, the start time of each action, the end time of each action, and the duration of each action relative to the first action (milliseconds); where, the duration of each action relative to the first action = the start time of the first action of the workstation + (the end time of each action - the start time of each action);

[0102] Step 4: Associate the S2 and S3 data using the line number and unique identifier to obtain the fused dataset X.

[0103] Step 5: Extract dataset D from dataset X the duration of the 12 actions of line number 1 and workstation number 1 relative to the first action in production model number 100; for example, the duration (milliseconds) of each action of line number 1 and workstation number 1 in production model number 100 relative to the first action is shown below. Figure 2 The dataset shown is illustrated. The start time of the first action shown in the figure is Linux time, meaning it is calculated from the start time of January 1, 1970, 00:00, in milliseconds. For example, the start time of the first action for sample number 1 is 1662747181451, indicating that 1662747181451 milliseconds have elapsed since the start time of January 1, 1970, 00:00. The corresponding Windows time is 2022 / 09 / 10 02:13:01. The duration of action number 1 relative to the first action for sample number 1 is 1154, indicating that action number 1 ended 1154 milliseconds after starting from 1662747181451. This application embodiment does not limit the unit and format of time data.

[0104] Step 6: Obtain the number of rows with missing values ​​in dataset D, G=16.

[0105] Step 7: Based on dataset D, obtain a basic matrix Y, where the values ​​of each element of the basic matrix Y are the corresponding values ​​in dataset D, and the locations where there is missing data are marked as nan.

[0106] Step 8: Set the number of neighboring samples used to fill missing values ​​to 5, construct the K-nearest neighbor model to fill the missing data in dataset D, and obtain the filling matrix Y_1.

[0107] Step 9: Use all action columns and the first action start time column of dataset D as feature columns, add a new column as a category column, and construct a chain equation model of random forest to fill the missing data in the feature columns of dataset D to obtain the filling matrix Y_2.

[0108] Step 10: Sum the fill values ​​obtained from the two models using a weighted sum. For example, if the fill values ​​for sample number 5 (5th cycle) and action number 12 are 39246 and 38970 respectively, then the final fill value is 39246×0.5+38970×0.5=39108.

[0109] Step 11: Obtain the maximum duration of each action in the 5th cycle relative to the first action, which is 39108 milliseconds after filling, and is also the actual cycle time of the workstation 39108 / 1000 = 39.108 seconds; calculate the workstation production cycle time (1662747423468-1662747375528) / 1000 = 47.94 seconds; and so on, to obtain the duration of each missing action relative to the first action and the actual cycle time and production cycle time of the workstation for each cycle.

[0110] The process cycle determination method provided in this application can effectively encode the attribute information of each action at different workstations on different production lines and construct a mapping table. By fusing it with production process data, a time-series dataset of each action at different workstations on different production lines is constructed, providing support for subsequent missing data filling. Simultaneously, the method employs a K-nearest neighbor model and a chain equation model based on random forests, considering both the proximity characteristics of missing data and the nonlinear relationships between actions, thus enabling better filling of missing values.

[0111] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0112] Based on the same inventive concept, this application also provides a process cycle determination apparatus for implementing the process cycle determination method described above. The solution provided by this apparatus is similar to the solution described in the above method; therefore, the specific limitations in one or more process cycle determination apparatus embodiments provided below can be found in the limitations of the process cycle determination method described above, and will not be repeated here.

[0113] In one embodiment, such as Figure 8 As shown, a process cycle determination device 800 is provided, the device comprising:

[0114] The first dataset acquisition module 801 is used to process production process data and production line attribute data to obtain the first dataset. The first dataset includes the duration of each production action of the target line in the target workstation and target production vehicle process relative to the first action.

[0115] The first matrix acquisition module 802 is used to convert the first dataset into a two-dimensional matrix to obtain the first matrix when there are missing values ​​in the first dataset. The value of each element in the first matrix is ​​the corresponding value in the first dataset.

[0116] The filling value determination module 803 is used to fill the missing values ​​in the first matrix using the K-nearest neighbor model and the chain equation model of random forest to obtain the corresponding filling values;

[0117] The second dataset generation module 804 is used to fill in the missing values ​​in the first dataset to obtain the second dataset.

[0118] The process cycle time determination module 805 is used to calculate the process cycle time of the target production line at the target workstation and the target production model based on the second dataset.

[0119] In one embodiment, the fill value determination module includes:

[0120] The second matrix determination unit is used to fill in the missing values ​​in the first matrix using the K-nearest neighbor model to obtain the second matrix;

[0121] The third matrix determination unit is used to fill in the missing values ​​in the first matrix using the chain equation model of random forest to obtain the third matrix;

[0122] The filling value determination unit is used to obtain the filling value corresponding to the missing value based on the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix.

[0123] In one embodiment, the fill value determination unit includes:

[0124] The weighted summation sub-unit is used to perform a weighted summation of the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix to obtain the filling values ​​corresponding to the missing values.

[0125] In one embodiment, the first dataset acquisition module includes:

[0126] The production process data and production line attribute data acquisition unit is used to acquire the production process data and production line attribute data of the production line. The production process data includes the production vehicle model, the line's unique identification code, the workstation's unique identification code, the action group's unique identification code, the action's unique identification code, and the action time information. The production line attribute data includes the production line's unique identification code, the workstation's unique identification code, the action group's unique identification code, and the action's unique identification code.

[0127] The data association unit is used to associate the production process data and the production line attribute data based on the line information in the production process data and the line unique identifier, workstation unique identifier, action group unique identifier, and action unique identifier in the production line attribute data to obtain a third dataset; the third dataset includes the duration of each action performed by each workstation of each line for each production model relative to the start time;

[0128] The data extraction unit is used to extract the duration of each production action of the target line body relative to the first action in the process of the target workstation and the target production model from the third dataset, so as to obtain the first dataset.

[0129] In one embodiment, the production line attribute data also includes the line number, workstation number, action group number, and action number;

[0130] The production process data and production line attribute data acquisition unit includes:

[0131] The inherent attribute data acquisition subunit is used to acquire the inherent attribute data of the production line body. The inherent attribute data of the production line body includes the line body number, the line body unique identifier, the workstation number, the workstation unique identifier, the action group unique identifier, the original action number, and the action unique identifier.

[0132] The data sorting subunit is used to group the inherent attribute data of the production line according to the line number and the workstation number, and sort them based on the action group number and the original action number in each workstation.

[0133] The action sequence number determination sub-unit is used to generate action sequence numbers based on the order of sorted actions, construct a mapping table between different workstations and actions of different production lines, and obtain production line attribute data.

[0134] In one embodiment, the process cycle time includes the actual cycle time of the target station, the production cycle time of the target station, the start time of each action, the end time of each action, and the duration of each action; wherein, the actual cycle time of the station is the maximum duration of the last action in the target station relative to the first action; the production cycle time of the station is the difference between the start time of the first action of the target station and the start time of the first action of the next station.

[0135] Each module in the aforementioned process cycle determination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0136] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores production cycle time determination data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a production cycle time determination method.

[0137] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0138] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0139] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps of the method described above.

[0140] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0141] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0143] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining process cycle time, characterized in that, The method includes: Based on production process data and production line attribute data, a first dataset is obtained. The first dataset includes the duration of each action of the target line in the process of the target workstation and the target production model relative to the first action. If there are missing values ​​in the first dataset, the first dataset is transformed into a two-dimensional matrix to obtain the first matrix; The missing values ​​in the first matrix are filled using the K-nearest neighbor model and the chain equation model of random forest to obtain the corresponding filled values; the filled values ​​are used to represent the duration of the missing action or the start time of the first action caused during data acquisition and transmission. The imputation values ​​are inserted into the positions of the corresponding missing values ​​in the first dataset to obtain the second dataset; Based on the second dataset, the process cycle time of the target production line in the target workstation and the target production vehicle is determined; The process of filling missing values ​​in the first matrix using a chain equation model based on the K-nearest neighbors model and random forest to obtain the corresponding filled values ​​includes: The missing values ​​in the first matrix are filled using the K-nearest neighbor model to obtain the second matrix; The missing values ​​in the first matrix are filled using a chain equation model of random forest to obtain the third matrix; Based on the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix, the filling values ​​corresponding to the missing values ​​are obtained; The step of obtaining the filled values ​​corresponding to the missing values ​​based on the filled missing values ​​in the second matrix and the filled missing values ​​in the third matrix includes: The missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix are weighted and summed to obtain the filling values ​​corresponding to the missing values.

2. The method according to claim 1, characterized in that, The first dataset, obtained by processing production process data and production line attribute data, includes: Acquire production process data and production line attribute data of the production line; the production process data includes production vehicle model, line unique identifier, workstation unique identifier, action group unique identifier, action unique identifier, and action time information; the production line attribute data includes the line unique identifier, workstation unique identifier, action group unique identifier, and action unique identifier of the production line. Based on the line information in the production process data and the unique identification code of the line, unique identification code of the workstation, unique identification code of the action group, and unique identification code of the action in the production line attribute data, the production process data and the production line attribute data are associated to obtain a third dataset; the third dataset includes the duration of each action performed by each workstation of each line for each production model relative to the start time. The first dataset is obtained by extracting the duration of each production action of the target line relative to the first action in the process of the target workstation and the target production model from the third dataset.

3. The method according to claim 2, characterized in that, The production line attribute data also includes the line number, workstation number, action group number, and action number; The acquisition of production line attribute data includes: Obtain the inherent attribute data of the production line body, which includes the line body serial number, line body unique identifier, workstation serial number, workstation unique identifier, action group unique identifier, action original serial number, and action unique identifier. The inherent attribute data of the production line are grouped according to the line number and the workstation number, and sorted according to the action group number and the original action number in each workstation. Based on the order of the sorted actions, the action sequence number is generated, and a mapping table between different workstations and actions of different production lines is constructed to obtain the attribute data of the production line.

4. The method according to any one of claims 1 to 3, characterized in that, The process cycle time includes the actual cycle time of the target station, the production cycle time of the target station, the start time of each action, the end time of each action, and the duration of each action; wherein, the actual cycle time of the station is the maximum duration of the last action in the target station relative to the first action; the production cycle time of the station is the difference between the start time of the first action of the target station and the start time of the first action of the next station.

5. A process cycle time determination device, characterized in that, The device includes: The first dataset acquisition module is used to process production process data and production line attribute data to obtain a first dataset. The first dataset includes the duration of each production action of the target line in the process of the target workstation and the target production model relative to the first action. The first matrix acquisition module is used to convert the first dataset into a two-dimensional matrix to obtain a first matrix when there are missing values ​​in the first dataset. The value of each element in the first matrix is ​​the corresponding value in the first dataset. The fill value determination module is used to fill the missing values ​​in the first matrix using the K-nearest neighbor model and the chain equation model of random forest to obtain the corresponding fill value; the fill value is used to represent the duration of the missing action or the start time of the first action caused during data acquisition and transmission. The second dataset generation module is used to fill in the missing values ​​in the first dataset to obtain the second dataset. The process cycle time determination module is used to determine the process cycle time of the target production line in the target workstation and the target production vehicle based on the second dataset; The filling value determination module is further configured to fill the missing values ​​in the first matrix using a K-nearest neighbor model to obtain a second matrix; fill the missing values ​​in the first matrix using a random forest chain equation model to obtain a third matrix; and obtain the filling value corresponding to the missing value based on the filled missing values ​​in the second matrix and the filled missing values ​​in the third matrix. The filling value determination module is further configured to perform a weighted summation of the missing values ​​filled in the second matrix and the missing values ​​filled in the third matrix to obtain the filling value corresponding to the missing value.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

Images