A data-driven modeling method for machine tool power consumption

Abstract

The present invention provides a data-driven machine tool power consumption modeling method, comprising the following steps: selecting a machine tool as a source machine tool, obtaining a source data set T of the power consumption characteristics of the source machine tool through several experiments, src ; Obtain the target dataset T of the target machine tool power consumption characteristics through several experiments tar ; The source dataset T is transferred through transfer learning src To the target dataset T tar Migrate and obtain the migrated source dataset T s ' rc ; Using the target dataset T tar and the migrated source dataset T′ src Train the regression model; obtain the power consumption model of the target machine tool, which solves the problem that the existing machine tool power consumption modeling method is only applicable to the acquisition of a single machine tool power consumption model, and a large number of repetitive calibration experiments are required to obtain the power consumption model of each machine tool in the production line or workshop, resulting in low modeling efficiency. The power consumption model of the target machine tool can be quickly established through a small number of experiments, which greatly improves the modeling efficiency while ensuring the prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of machine tool power consumption model modeling, and in particular to a data-driven machine tool power consumption model modeling method. Background Art

[0002] As modern manufacturing places increasingly higher demands on energy efficiency, machine tool energy consumption has received widespread attention in recent years. The power consumption characteristics of machine tools directly affect the energy consumption of machine tool processing. The power consumption characteristics of different types of machine tools are completely different. Even for machine tools of the same model, their power consumption characteristics will vary.

[0003] Patent 201910013924.8 discloses a method for modeling the power consumption model of a machine tool milling process based on cutting force. This method divides the energy consumption of the machine tool milling process into no-load power consumption, cutting power consumption and additional load power consumption, and obtains the total power consumption model of the machine tool milling process through experiments. This modeling method is only suitable for obtaining the power consumption model of a single machine tool. However, if you want to obtain the power consumption model of each machine tool in a production line or workshop, you need to conduct a large number of repetitive calibration experiments, and the modeling efficiency is low. Summary of the Invention

[0004] The present invention discloses a data-driven machine tool power consumption model modeling method, which solves the problem that the existing machine tool power consumption model modeling method is only suitable for obtaining the power consumption model of a single machine tool, and a large number of repetitive calibration experiments are required when obtaining the power consumption model of each machine tool in a production line or workshop, resulting in low modeling efficiency. The power consumption model of the target machine tool can be quickly established through a small number of experiments, which greatly improves the modeling efficiency while ensuring the prediction accuracy.

[0005] To achieve the above object, the technical solution of the present invention is specifically implemented as follows:

[0006] The present invention discloses a data-driven method for modeling a machine tool power consumption model, comprising the following steps:

[0007] A machine tool is selected as the source machine tool, and a source data set T of the power consumption characteristics of the source machine tool is obtained through several experiments. src :

[0008] T src ={(x i src ,y i src )},i=1,2,....,n (1)

[0009] Among them, x i src Represents the processing parameters of the source machine tool, y i src represents the power consumption data of the source machine tool, and n represents the size of the source data set;

[0010] The target dataset T of the target machine tool power consumption characteristics is obtained through several experiments tar :

[0011] T tar ={(x i tar ,y i tar )},i=1,2,....m (2)

[0012] Among them, x i tar Indicates the processing parameters of the target machine tool, y i tar represents the power consumption data of the target machine tool, and m represents the size of the target data set;

[0013] The source dataset T is transferred through transfer learning src To the target dataset T tar Migrate and obtain the migrated source dataset T′ src ;

[0014] Using the target dataset T tar and the migrated source dataset T′ src Train a regression model;

[0015] Obtain the power consumption model of the target machine tool.

[0016] Furthermore, the source dataset T src To the target dataset T tar The migration steps include:

[0017] Affine transformation is used for domain adaptation so that the source dataset T src With the target dataset T tar The difference between them is the smallest, and the objective function is expressed as:

[0018] min|T src_1 WT tar_1 | (3)

[0019] in, W is the affine transformation matrix, n is the source data set T src The size of m is the target dataset T tar size;

[0020] Using the source dataset T src Training to obtain processing parameter space x i and power consumption data space y i The mapping relationship f src :

[0021] T src_1Can be replaced by T src_2 , T src_2 It can be expressed as:

[0022]

[0023] Among them, f src (x1 tar ) is the target processing parameter data x i tar Through the mapping relationship f src Obtained target power consumption data;

[0024] Then the objective function (3) is expressed as:

[0025] min|T src_2 WT tar_1 | (5)

[0026] Due to T src_2 With T tar_1 With a uniform column vector, equation (5) can be expressed as follows:

[0027]

[0028] Where H is the migration transformation matrix;

[0029] Solve the migration transformation matrix H:

[0030] H=(A T A) -1 A T B (7)

[0031] in,

[0032] The source dataset T after migration s ' rc Expressed as:

[0033] T′ src =[x src ,T src_1 ·H] (8).

[0034] Furthermore, using the target dataset T tar and the migrated source dataset T′ src The steps to train a regression model are:

[0035] The migrated source dataset T′ src and target dataset T tar To perform the merge:

[0036] T=T′ src ∪T tar (9)

[0037] Wherein, formula (9) is expressed as:

[0038] T={(x i ,y i )} (10)

[0039] in,

[0040] Furthermore, the adaptive weighting algorithm is used to increase the target data set T tar The weight of .

[0041] Beneficial technical effects:

[0042] The present invention discloses a data-driven machine tool power consumption modeling method, comprising the following steps: selecting a machine tool as a source machine tool, obtaining a source data set T of the power consumption characteristics of the source machine tool through several experiments, src ; Obtain the target dataset T of the target machine tool power consumption characteristics through several experiments tar ; The source dataset T is transferred through transfer learning src To the target dataset T tar Migrate and obtain the migrated source dataset T s ' rc ; Using the target dataset T tar and the migrated source dataset T′ src Train the regression model; obtain the power consumption model of the target machine tool, which solves the problem that the existing machine tool power consumption modeling method is only applicable to the acquisition of a single machine tool power consumption model, and a large number of repetitive calibration experiments are required to obtain the power consumption model of each machine tool in the production line or workshop, resulting in low modeling efficiency. The power consumption model of the target machine tool can be quickly established through a small number of experiments, which greatly improves the modeling efficiency while ensuring the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for describing the embodiments.

[0044] Figure 1 This is a flowchart of the steps of a data-driven machine tool power consumption modeling method described in the present invention. DETAILED DESCRIPTION

[0045] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0046] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0047] The present invention discloses a data-driven machine tool power consumption modeling method, see Figure 1 , specifically including the following steps:

[0048] S1: Select a machine tool as the source machine tool and obtain the source data set T of the power consumption characteristics of the source machine tool through several experiments src :

[0049] T src ={(x i src ,y i src )},i=1,2,....,n (1)

[0050] Among them, x i src Represents the processing parameters of the source machine tool, y i src represents the power consumption data of the source machine tool, and n represents the size of the source data set;

[0051] In this embodiment, taking the drilling of high-temperature alloy GH4169 as an example, the tool is a twist drill with a diameter of 10 mm, and two different machine tools are selected: CNC machining center YHVT850Z and CNC machining center VMC800. The high-temperature alloy GH4169 material is drilled on the YHVT850Z and VMC800 CNC machining centers respectively.

[0052] According to S1, drilling calibration experiments (31 groups) were carried out on the source machine tool (YHVT850Z), and the power consumption data of the source machine tool (YHVT850Z) were obtained as the source data set.

[0053] S2: Obtain the target dataset T of the target machine tool power consumption characteristics through several experiments tar :

[0054] T tar ={(x i tar ,y i tar )},i=1,2,....m (2)

[0055] Among them, x i tar Indicates the processing parameters of the target machine tool, y i tar represents the power consumption data of the target machine tool, and m represents the size of the target data set;

[0056] A small number of drilling calibration experiments (8 groups) were conducted on the target machine tool (VMC800), and the power consumption data of the target machine tool (VMC800) was obtained as the target data set.

[0057] S3: Transfer the source dataset T through transfer learning src To the target dataset T tar Migrate and obtain the migrated source dataset T′ src ;

[0058] Specifically, the source dataset T src To the target dataset T tar The migration steps include:

[0059] Affine transformation is used for domain adaptation so that the source dataset T src With the target dataset T tar The difference between them is the smallest, and the objective function is expressed as:

[0060] min|T src_1 WT tar_1 | (3)

[0061] in, W is the affine transformation matrix, n is the source data set T src The size of m is the target dataset T tar size;

[0062] Using the source dataset T src Training to obtain processing parameter space x i and power consumption data space y i The mapping relationship f src :

[0063] T src_1 Can be replaced by T src_2 , T src_2 It can be expressed as:

[0064]

[0065] Among them, f src (x1 tar ) is the target processing parameter data x i tar Through the mapping relationship f src Obtained target power consumption data;

[0066] Then the objective function (3) is expressed as:

[0067] min|T src_2 WT tar_1 | (5)

[0068] Due to T src_2 With T tar_1 With a uniform column vector, equation (5) can be expressed as follows:

[0069]

[0070] Where H is the migration transformation matrix;

[0071] Solve the migration transformation matrix H:

[0072] H=(A T A) -1 A T B (7)

[0073] in,

[0074] The source dataset T′ after migration src Expressed as:

[0075] T′ src =[x src ,T src_1 .H] (8).

[0076] S4: Using the target dataset T tar and the migrated source dataset T′ src Train a regression model;

[0077] Specifically, using the target dataset T tar and the migrated source dataset T′ src The steps to train a regression model are:

[0078] The migrated source dataset T′ src and target dataset T tar To perform the merge:

[0079] T=T′ src ∪T tar (9)

[0080] Wherein, formula (9) is expressed as:

[0081] T={(x i ,y i )} (10)

[0082] in,

[0083] S5: Obtain the power consumption model of the target machine tool.

[0084] The model established by the method of the present invention is compared with the model of the experimental calibration method. The comparison of the model expressions is shown in Table 1:

[0085] Table 1 Comparison of model expressions between the present invention method and the experimental calibration method

[0086]

[0087] Among them, P cuting is the cutting power, and s is the spindle speed.

[0088] It can be seen from Table 1 that the model expression obtained by the method for establishing the machine tool power consumption model disclosed in the invention is basically the same as the machine tool power consumption model expression obtained by the experimental calibration method.

[0089] The comparison of the number of experiments and time required to establish the model by the method of the present invention and the number of experiments and time required to establish the model by the experimental calibration method is shown in Table 2:

[0090] Table 2 Comparison of the number of experiments and time required to establish the model using the method of the present invention and the number of experiments and time required to establish the model using the experimental calibration method

[0091]

[0092] It can be seen from Table 2 that, through the method for establishing a machine tool power consumption model disclosed in the invention, the number of experiments required to establish a machine tool power consumption model is much smaller than the number of experiments required to establish a machine tool power consumption model by the experimental calibration method. As a result, the time required to establish a machine tool power consumption model by the method for establishing a machine tool power consumption model disclosed in the invention is also much smaller than the time required to establish a machine tool power consumption model by the experimental calibration method. The time required for the method for establishing a machine tool power consumption model disclosed in the invention is nearly 75% shorter than that of the method for establishing a machine tool power consumption model by the experimental calibration method, and the error with the actual power consumption of the machine tool measured is less than 2%.

[0093] The present invention discloses a data-driven machine tool power consumption modeling method for rapidly establishing and accurately evaluating the power consumption model of CNC machine tools. The prediction accuracy is high, and the error with the actual power consumption of the machine tool measured is less than 2%. The prediction accuracy is comparable to that of the experimental calibration method, while greatly reducing the amount of experiments. The proposed method shortens the experimental time by nearly 75% and improves the modeling efficiency. Therefore, it has great application value in practical work.

[0094] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0095] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0096] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0097] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0099] The above embodiments are merely descriptions of preferred implementations of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineers and technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A data-driven machine tool power consumption modeling method, characterized in that: The following steps are involved: A machine tool is selected as the source machine tool, and a source data set T of the power consumption characteristics of the source machine tool is obtained through several experiments. src : T src ={(x i src ,y i src )}i=1,2,....,n (1) Among them, x i src Represents the processing parameters of the source machine tool, y i src represents the power consumption data of the source machine tool, and n represents the size of the source data set; The target dataset T of the target machine tool power consumption characteristics is obtained through several experiments tar : T tar ={(x i tar ,y i tar )}i=1,2,....m (2) Among them, x i tar Indicates the processing parameters of the target machine tool, y i tar Indicates the power consumption data of the target machine tool, m represents the size of the target dataset; The source dataset T is transferred through transfer learning src To the target dataset T tar Migrate and obtain the migrated source dataset T′ src ; Using the target dataset T tar and the migrated source dataset T′ src Train a regression model; Obtain the power consumption model of the target machine tool; Source dataset T src To the target dataset T tar The migration steps include: Affine transformation is used for domain adaptation so that the source dataset T src With the target dataset T tar The difference between them is the smallest, and the objective function is expressed as: min|T src_1 ·W-T tar_1 | (3) in, W is the affine transformation matrix, n is the source data set T src The size of m is the target dataset T tar size; Using the source dataset T src Training to obtain processing parameter space x i and power consumption data space y i The mapping relationship f src : T src_1 Can be replaced by T src_2 , T src_2 It can be expressed as: Among them, f src (x1 tar ) is the target processing parameter data x i tar Through the mapping relationship f src Obtained target power consumption data; Then the objective function (3) is expressed as: min|T src_2 ·W-T tar_1 | (5)。 2. The data-driven machine tool power consumption modeling method according to claim 1, characterized in that: because T src_2 With T tar_1 With a uniform column vector, equation (5) can be expressed as follows: Where H is the migration transformation matrix; Solve the migration transformation matrix H: H=(A T A) -1 A T B (7) in, The source dataset T′ after migration src Expressed as: T′ src =[x src ,T src_1 ·H] (8)。 3. The data-driven machine tool power consumption modeling method according to claim 2, characterized in that: Using the target dataset T tar and the migrated source dataset T′ src The steps to train a regression model are: The migrated source dataset T′ src and target dataset T tar To perform the merge: T=T′ src AT tar (9) Wherein, formula (9) is expressed as: T={(x i ,y i )} (10) in, 4. The data-driven machine tool power consumption modeling method according to claim 3, characterized in that: Using adaptive weighting algorithm to increase the target data set T tar The weight of .

Images