Numerical control system thermal error intelligent compensation method based on multi-source temperature sensing

Abstract

The invention relates to the technical field of numerical control system control, in particular to a numerical control system thermal error intelligent compensation method based on multi-source temperature sensing. The method comprises the following steps: respectively training two prediction models according to collected data; the predicted instantaneous temperature residual output by the loss temperature residual prediction model can ensure the reference, and meanwhile, the problem of control lag cannot be generated; the machine tool background temperature prediction model is a low-frequency model with a sampling period as a time unit, planned injection energy and the open-loop thermal instability degree serve as driving characteristics of machine tool heat source physical attributes in the sampling period, and the predicted machine tool background temperature in the future period is predicted. And compensation control can be performed on the laser power and the machine tool feed amount according to prediction results of the two models. According to the method, through effective training and application of the two prediction models, the technical problems that transient overheating control lags behind, and a strong heat source in the load increasing process is not considered for machine tool thermal drift compensation are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of numerical control system control, in particular to a numerical control system thermal error intelligent compensation method based on multi-source temperature sensing. BACKGROUND

[0002] In the process of metal additive manufacturing, high-energy-density heat sources in the machine tool system induce two different time scale thermal errors, which jointly restrict the final machining accuracy. One is on the fast time scale of milliseconds to seconds, when the heat source scans to the sharp corners, thin walls and other heat dissipation limited areas of the part, the local heat quickly accumulates, which may cause transient thermal deformation such as overheating collapse of the workpiece. The second is on the slow time scale of minutes to hours, the overall dissipated heat of the additive process will continue to accumulate and conduct to the machine tool bed, column and other key structural components, causing uneven thermal expansion of the tool center point, and then causing macro positioning drift. The two error sources are completely different in cause and time scale, but are related to each other.

[0003] The prior art has limitations in dealing with such problems, focusing on single-scale error compensation, which is difficult to effectively deal with the complex working conditions of double-scale errors. For the control of transient overheating of the workpiece, it mainly depends on the lagging temperature feedback or fixed process parameters, which is difficult to prevent overheating caused by complex geometry and machining path. For the compensation of machine tool macro thermal drift, the current method generally considers the temperature field of the machine tool itself in isolation, ignoring the strong heat source of the additive process as the actual driving force for the thermal deformation of the machine tool. SUMMARY

[0004] In order to solve the technical problems of the prior art that the control of transient overheating is lagging and the compensation of machine tool thermal drift does not consider the strong heat source in the loading process, the purpose of the present application is to provide a numerical control system thermal error intelligent compensation method based on multi-source temperature sensing, and the technical scheme adopted is as follows: The present application provides a numerical control system thermal error intelligent compensation method based on multi-source temperature sensing, which comprises the following steps: According to the task information, obtain the static heat dissipation characteristics of the numerical control machine tool machining task; during the execution of the machining task, collect the instruction power and instantaneous molten pool temperature at each time point in the preset sampling period; collect temperature readings at different positions on the characteristic structural components of the numerical control machine tool, obtain the machine tool structure temperature feature vector, and obtain the machine tool background temperature; At each time point, obtain the machining temperature according to the machine tool background temperature and the instruction power, and obtain the instantaneous temperature residual at each time point according to the difference between the instantaneous molten pool temperature and the machining temperature; use the machine tool structure temperature feature vector, the instruction power and the static heat dissipation characteristics as input data, and use the instantaneous temperature residual at the next time point as output data to train the instantaneous temperature residual prediction model; According to the instruction power, the planned injection energy in each sampling period is counted, and the open-loop thermal instability degree is obtained according to the instantaneous temperature residual change characteristics in each sampling period; the instruction power, the planned injection energy and the open-loop thermal instability degree in the sampling period are taken as input data, and the machine tool background temperature in the next sampling period is taken as output data, so as to train the machine tool background temperature prediction model; The data in the real-time sampling period is processed by using the instantaneous temperature residual prediction model to obtain the predicted instantaneous temperature residual; the predicted machine tool background temperature residual is obtained according to the difference between the machine tool background temperature in the real-time sampling period and the predicted machine tool background temperature output by the machine tool background temperature prediction model; The laser power is adjusted according to the predicted instantaneous temperature residual, and the machine tool feeding amount is adjusted according to the predicted instantaneous temperature residual and the predicted machine tool background temperature residual.

[0005] Further, the static heat dissipation characteristics include geometric heat dissipation conditions and path repeated heating risks; the volume ratio of the solidified entity to the to-be-melted powder in the local area of each path point is obtained, and the average value of all volume ratios is taken as the geometric heat dissipation condition; the repeated heating times of each path point are counted, and the average value of the repeated heating times of all path points is taken as the path repeated heating risk.

[0006] Further, the machine tool background temperature is the average value of all elements in the machine tool structure temperature characteristic vector.

[0007] Further, the method for obtaining the processing temperature comprises: The instruction power is multiplied by a preset conversion coefficient to obtain an input temperature, and the sum of the input temperature and the machine tool background temperature is taken as the processing temperature.

[0008] Further, the method for obtaining the instantaneous temperature residual comprises: The difference between the instantaneous molten pool temperature and the processing temperature is taken as the instantaneous temperature residual.

[0009] Further, the open-loop thermal instability degree is the standard deviation of the instantaneous temperature residuals at all time points in the sampling period.

[0010] Further, the method for obtaining the predicted machine tool background temperature comprises: The predicted instantaneous temperature residual sequence of the real-time sampling period is obtained by using the instantaneous temperature residual prediction model, the predicted open-loop thermal instability degree is obtained according to the predicted instantaneous temperature residual sequence; the predicted open-loop thermal instability degree, the planned injection energy in the real-time sampling period and the instruction power in the real-time sampling period are taken as input data, and the predicted machine tool background temperature is output by the machine tool background temperature prediction model.

[0011] Further, the predicted machine tool background temperature residual is a difference between the predicted machine tool background temperature and the machine tool background temperature in a real-time sampling period.

[0012] Further, the adjusting the laser power according to the predicted instantaneous temperature residual comprises: Dividing the predicted instantaneous temperature residual by a preset conversion coefficient to obtain a compensation power; and subtracting the compensation power from a laser base power to obtain an adjusted laser power.

[0013] Further, the adjusting the machine tool feed amount according to the predicted instantaneous temperature residual and the predicted machine tool background temperature residual comprises: Taking a sum of the predicted instantaneous temperature residual and the predicted machine tool background temperature residual as a machine tool thermal error factor cumulative value; if the machine tool thermal error factor cumulative value is greater than a preset first threshold value and less than a preset second threshold value, a linearly descending method is used to lower the feed amount; and if the machine tool thermal error factor cumulative value is greater than or equal to the second threshold value, the feed amount is adjusted to a preset minimum value.

[0014] The present application has the following advantages: The present application respectively collects static heat dissipation characteristics before processing and multiple heat characteristics in the processing process, trains two prediction models according to the collected data, which are an instantaneous temperature residual prediction model and a machine tool background temperature prediction model. The loss temperature residual prediction model is a high-frequency model trained according to high-frequency data, so that the predicted instantaneous temperature residual output by the model can ensure the reference and will not cause control lag. In the training process, the model can learn the nonlinear difference of the overheating response caused by the same geometric path under different global thermal backgrounds, so as to obtain a prediction data that has taken into account the macro influence. The machine tool background temperature prediction model is a low-frequency model with a sampling period as a time unit, and the planned injection energy and the open-loop thermal instability degree are taken as the driving characteristics of the machine tool heat source physical properties in the sampling period. Therefore, the machine tool background temperature prediction model can learn the constitutive relationship between the additive process and the machine tool thermal deformation, so that the model can accurately predict the predicted machine tool background temperature in the future period. Then, the laser power and the machine tool feed amount can be timely and reasonably controlled according to the prediction results of the two models. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0016] Figure 1A flow chart of a multi-source temperature sensing based intelligent compensation method of thermal errors of a numerical control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined object of the application, the following describes in detail the specific implementation, structure, features and effects of a multi-source temperature sensing based intelligent compensation method of thermal errors of a numerical control system according to the present application, in combination with the accompanying drawings and preferred embodiments. Different "one embodiment" or "another embodiment" in the following description do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0019] The specific scheme of the multi-source temperature sensing based intelligent compensation method of thermal errors of a numerical control system provided by the present application is specifically described below in combination with the accompanying drawings.

[0020] Referring to Figure 1 , a flow chart of a multi-source temperature sensing based intelligent compensation method of thermal errors of a numerical control system according to an embodiment of the present application is shown, which includes the following steps. Step S1: Obtain the static heat dissipation characteristics of the machining task of the numerical control machine tool according to the task information; collect the instruction power and the instantaneous molten pool temperature at each time point in the preset sampling period during the execution of the machining task; collect temperature readings at different positions on the characteristic structural parts of the numerical control machine tool to obtain the machine tool structure temperature characteristic vector and the machine tool background temperature.

[0021] The embodiment of the present application aims to solve the problem of control lag by training an effective prediction model, so it is necessary to collect various information of the machining task of the numerical control machine tool, including historical information and real-time information, wherein the historical information is used for training the model, and the real-time information is used for input into the trained model for prediction output. In the embodiment of the present application, the sampling period is set to 1 second, and 100 time points are set under one sampling period. And for each machining task, the embodiment of the present application obtains two kinds of machining information, i.e. static characteristics and dynamic characteristics.

[0022] The static characteristics are the static heat dissipation characteristics obtained according to the task information, which are usually represented as the heat dissipation characteristics corresponding to the geometric characteristics of the workpiece or the machining path in the task design file, etc. The static heat dissipation characteristics can be used to represent the basic heat information of the machining task before the offline preprocessing stage of the machining task.

[0023] The dynamic feature is a feature collected in real time during the execution of the machining task, and specifically includes the instruction power and the instantaneous molten pool temperature at each time point in a sampling period, a machine tool structure temperature feature vector composed of temperature readings at different positions on the machine tool feature structure in the sampling period, and a machine tool background temperature estimated according to the machine tool structure temperature feature vector. The collection of the dynamic feature can determine the heat injected by the numerical control machine tool during the execution of the machining task, the heat fed back in real time, and the machine tool background temperature and other heat information. Analysis of the information can determine the heat generation and dissipation form of the machine tool during the execution of the task.

[0024] The machine tool feature structure can be set as a key structure such as a machine tool bed and a column, and details are not repeated and limited. The machine tool structure temperature feature vector is a vector composed of the temperatures collected at the positions. It should be noted that the machine tool feature structure should be consistent in all machining tasks, that is, the lengths of different machine tool structure temperature feature vectors are consistent, and the machine tool feature structures corresponding to the same position elements should be the same.

[0025] The static feature and the dynamic feature provide structured data for protecting the macro state, the dynamic process and the static attribute, and based on the data, the training analysis of the prediction model can be used. In the embodiment of the present application, for each machining task, in a sampling period of 1 second, an instruction power sequence and an instantaneous molten pool temperature sequence composed of 100 time points can be obtained; a machine tool structure temperature feature vector and a machine tool background temperature can be obtained; and a static heat dissipation feature of the machining task can be obtained.

[0026] Preferably, in the embodiment of the present application, the static heat dissipation feature includes a geometric heat dissipation condition and a path repeated heating risk; the volume ratio of the solidified entity to the to-be-melted powder in the local area of each path point is obtained, and the average value of all volume ratios is taken as the geometric heat dissipation condition; the number of repeated heating of each path point is counted, and the average value of the number of repeated heating of all path points is taken as the path repeated heating risk. The geometric heat dissipation condition can represent the inherent heat dissipation ability of the path point due to the local set shape (for example, sharp corner, thin wall) through the volume ratio; the repeated heating risk represents the risk of repeated injection of heat required by the path point position during the execution of the task. That is, the static heat dissipation feature in the embodiment of the present application is a two-dimensional vector.

[0027] It should be noted that the size of the local area of the path point can be determined according to the size of the component corresponding to the specific machining task, and details are not limited and repeated.

[0028] Preferably, in the embodiment of the present application, the machine tool background temperature is the average value of all elements in the machine tool structure temperature feature vector.

[0029] Step S2: at each time point, the processing temperature is obtained according to the machine tool background temperature and the instruction power, the instantaneous temperature residual at each time point is obtained according to the difference between the instantaneous molten pool temperature and the processing temperature; the machine tool structure temperature feature vector, the instruction power and the static heat dissipation feature are taken as input data, and the instantaneous temperature residual at the next time point is taken as output data, and the instantaneous temperature residual prediction model is trained.

[0030] After the information of the historical machining task is collected by step S1, the information needs to be further extracted to facilitate the training of the model.

[0031] First, the high-frequency instantaneous temperature residual prediction model is trained, that is, the model is analyzed and predicted in each time unit of the sampling period, and the output is the predicted instantaneous temperature residual at the future time point. Because the output purpose is to obtain the instantaneous temperature residual at the future time point to adjust the laser power of the numerical control machine tool, and then ensure the stability of the temperature. Therefore, the instantaneous temperature residual at each time point needs to be determined in the model training process, and then used as the label of the previous time point for training. For each time point of each completed machining task, the processing temperature is first obtained according to the machine tool background temperature and the instruction power, that is, the machine tool background temperature is the temperature basis, and the instruction power is the power parameter when the machine tool injects heat into the workpiece material, so the theoretical processing temperature can be obtained based on the two data. Further, the instantaneous temperature residual at each time point is obtained according to the difference between the instantaneous molten pool temperature and the processing temperature. That is, the instantaneous temperature residual represents the additional temperature rise at the current time point due to factors such as static heat dissipation characteristics, which exceeds the basic physical law. It should be noted that the additional temperature rise here can also be negative, that is, the temperature decreases, and the specific value of the instantaneous temperature residual can represent the heat generation and heat dissipation comparison in the actual machining process, so this feature can be used as the basis for adjusting the laser power in the numerical control machine tool.

[0032] For the instantaneous temperature residual prediction model, the machine tool structure temperature feature vector, the instruction power and the static heat dissipation feature are taken as input data, and the instantaneous temperature residual at the next time point is taken as output data for training.

[0033] In the embodiment of the present application, the instantaneous temperature residual prediction model adopts a long short-term memory network model. In order to realize the guidance of the macro state to the micro prediction, before processing each new data sequence, the model produces the initial hidden state and cell state of the model through an independently addressable affine transformation layer for the machine tool structure temperature feature vector corresponding to the data. Then the sequence containing the historical instantaneous temperature residual, the historical instruction power and the static heat dissipation features of the corresponding machining task is sent into the network model for training. The loss function used in the training is the mean square error function, that is, the training target is to minimize the mean square error between the output predicted instantaneous temperature residual sequence and the historical instantaneous temperature residual sequence at the corresponding time. The training method of the long short-term memory network is a technical means familiar to those skilled in the art, which will not be repeated and limited here.

[0034] Preferably, in the embodiment of the present application, the method for obtaining the machining temperature comprises: The instruction power is multiplied by a preset conversion coefficient to obtain an input temperature, and the sum of the input temperature and the machine tool background temperature is taken as the machining temperature. It should be noted that the preset conversion coefficient can be determined in advance by linear regression analysis on the machining data of a simple large block collection area, or can be consulted from the instruction manual of the laser or other prior knowledge, and the specific content will not be repeated and limited.

[0035] Preferably, in the embodiment of the present application, the difference between the instantaneous molten pool temperature and the machining temperature is taken as the instantaneous temperature residual. That is, the instantaneous temperature residual has positive and negative values, the positive value represents that additional heat is generated in the actual process, and the negative value represents that additional heat dissipation is generated in the actual process.

[0036] Step S3: According to the instruction power, the planned injection energy in each sampling period is calculated, and the open-loop thermal instability degree is obtained according to the instantaneous temperature residual change characteristics under each sampling period; the instruction power, the planned injection energy and the open-loop thermal instability degree in the sampling period are taken as input data, and the machine tool background temperature of the next sampling period is taken as output data, and the machine tool background temperature prediction model is trained; The machine tool background temperature prediction model is used to predict the machine tool thermal deformation driven by the additive process and slowly accumulated. The machine tool thermal deformation is mainly obtained by comparing the predicted machine tool background temperature in the future sampling period with the actual machine tool background temperature. Because the prediction of the machine tool background temperature in the prior art is usually only for the change of the machine tool background temperature itself, and does not consider the fundamental heat source driving of the additive process, the driving features under the additive process are extracted in the embodiment of the present application, and in order to quantify the thermal load of the additive process on the machine tool, the data in each sampling period is calculated to obtain two features of the planned injection energy and the open-loop thermal instability degree as the heat source driving features.

[0037] The planned injection energy can be counted according to the instruction power in the sampling period, that is, the energy corresponding to the power signal is obtained by time integration of the instruction power. The planned injection energy quantifies the total heat injected by the additive process plan into the system in the sampling period. It should be noted that the numerical integration method for calculating energy is a well-known technique for those skilled in the art, and the specific content will not be repeated.

[0038] The open-loop thermal instability degree is obtained by counting the change characteristics of all instantaneous temperature residuals corresponding to all time points in the sampling period. The greater the change characteristics, the greater the degree of fluctuation of the molten pool temperature in the sampling period, which can reflect the impact of heat input in the current sampling period.

[0039] Preferably, in the embodiment of the present application, the open-loop thermal instability degree is the standard deviation of the instantaneous temperature residual of all time points in the sampling period.

[0040] In the embodiment of the present application, the input of the machine tool background temperature prediction model is the instruction power sequence, the planned injection energy and the open-loop thermal instability degree in a sampling period, and the output is the machine tool background temperature of the next sampling period. Similar to the instantaneous temperature residual prediction model, the machine tool background temperature prediction model in the embodiment of the present application also uses a long short-term memory model, and the loss function also uses a mean square error function. The training target is to minimize the mean square error between the machine tool background temperature of the future sampling period output by the model and the machine tool background temperature corresponding to the label.

[0041] Step S4: processing the data in the real-time sampling period using the instantaneous temperature residual prediction model to obtain the predicted instantaneous temperature residual; obtaining the predicted machine tool background temperature residual according to the difference between the machine tool background temperature in the real-time sampling period and the predicted machine tool background temperature output by the machine tool background temperature prediction model.

[0042] The training of the two prediction models is completed through the above steps. In actual use, the two trained models can be loaded, and the data collected in the real-time sampling period is integrated as input data and input into the two models to obtain the predicted instantaneous temperature residual and the predicted machine tool background temperature. The predicted machine tool background temperature also needs to be further compared with the machine tool background temperature in the real-time sampling period to obtain the predicted machine tool background temperature residual, which is used to represent the temperature change of the machine tool background temperature under the influence of thermal driving.

[0043] Preferably, in the embodiment of the present application, because the instantaneous temperature residual prediction model is a high-frequency model relative to the machine tool background temperature prediction model, and the output data of the instantaneous temperature residual prediction model is also a high-frequency data, that is, the high-frequency data of the predicted instantaneous temperature residual makes the power of the laser to be compensated and fine-tuned, in order to avoid the input data distribution mismatch caused by the compensation intervention of the high-frequency model, and the mismatch result causes the detected instantaneous temperature residual to approach 0 and loses the physical meaning. Therefore, in the embodiment of the present application, online input reconstruction is performed, the planned injection energy and the command power are normally detected under the real-time sampling period, and the open-loop thermal instability degree is obtained through the instantaneous temperature residual prediction model, that is, the predicted instantaneous temperature residual sequence of the real-time sampling period is obtained by using the instantaneous temperature residual prediction model, and the predicted open-loop thermal instability degree is obtained according to the predicted instantaneous temperature residual sequence; the predicted open-loop thermal instability degree, the planned injection energy under the real-time sampling period, and the command power under the real-time sampling period are taken as input data, and the predicted machine tool background temperature is output by the machine tool background temperature prediction model. The predicted open-loop thermal instability degree obtained by the method can simulate the change characteristics of the predicted instantaneous temperature residual sequence that will occur without compensation, and avoids the data distribution mismatch caused by the intervention of high-frequency compensation.

[0044] Preferably, in the embodiment of the present application, the predicted machine tool background temperature residual is the difference between the predicted machine tool background temperature and the machine tool background temperature under the real-time sampling period.

[0045] Step S5: adjusting the laser power according to the predicted instantaneous temperature residual, and adjusting the machine tool feed amount according to the predicted instantaneous temperature residual and the predicted machine tool background temperature residual.

[0046] Based on the analysis of the real-time data in step S4, the predicted instantaneous temperature residual at the future time point and the predicted machine tool background temperature residual at the future sampling time can be obtained. Based on the two kinds of residuals, the numerical control machine tool can be compensated and controlled, and the double-scale thermal error can be comprehensively suppressed.

[0047] The predicted instantaneous temperature residual is a high-frequency characteristic, which corresponds to each time point under the sampling period, that is, the adjustment frequency of the laser power is high, and the local overheating of the workpiece caused by the set and path factors is actively suppressed by real-time adjustment of the laser power. In the embodiment of the present application, the compensation channel adopts a model-based feedforward control. At each high-frequency time point of 10 milliseconds, the system first calculates a basic power command according to the preset target molten pool temperature and the background temperature determined by the current machine tool thermal state, and the power is the power required to maintain the target temperature in an ideal case without any overheating risk. Then, the adjusted laser power can be obtained by adjusting the laser basic power according to the predicted temperature residual.

[0048] Preferably, in the embodiment of the present application, the laser power is adjusted according to the predicted instantaneous temperature residual, comprising: The predicted instantaneous temperature residual is divided by a preset conversion coefficient to obtain a compensation power; and the laser base power is subtracted by the compensation power to obtain the adjusted laser power. The conversion coefficient is obtained by the method for obtaining the machining temperature as described above, and will not be repeated here. By this adjustment method, the power can be automatically reduced according to an overheating prediction taking into account the global thermal state before the laser spot reaches the risk area, thereby effectively preventing the melting and collapse of the fine structure of the part.

[0049] The second compensation in the embodiment of the present application is to adjust the machine tool feed amount with a sampling period as a time unit, i.e. at a low frequency, to avoid further amplification of mechanical errors in the case of accumulated heat.

[0050] Preferably, in the embodiment of the present application, the machine tool feed amount is adjusted according to the predicted instantaneous temperature residual and the predicted machine tool background temperature residual, comprising: The sum of the predicted instantaneous temperature residual and the predicted machine tool background temperature residual is taken as the machine tool thermal error factor cumulative value; if the machine tool thermal error factor cumulative value is greater than a preset first threshold value and less than a preset second threshold value, it indicates that the feed amount needs to be adjusted at this time, and a linearly decreasing method is used to lower the feed amount; if the machine tool thermal error factor cumulative value is greater than or equal to the second threshold value, it indicates that the thermal error risk is large, and in order to avoid deterioration, the feed amount is adjusted to a preset minimum value. In the embodiment of the present application, the first threshold value is set to 75% of the median value in the dimension of the machine tool thermal error factor cumulative value; and the second threshold value is set to a higher value in the first 90% of the dimension of the machine tool thermal error factor cumulative value.

[0051] It should be noted that the linearly decreasing is to set a small decreasing amount, and each adjustment is performed by lowering the step size according to the decreasing amount. The specific decreasing amount can be set according to the specific work task, and will not be repeated here.

[0052] In summary, the application collects the static heat dissipation characteristics before processing and multiple heat characteristics in the processing process, trains two prediction models according to the collected data, respectively, which are the instantaneous temperature residual prediction model and the machine tool background temperature prediction model. The loss temperature residual prediction model output prediction instantaneous temperature residual can ensure the reference while not causing control lag; the machine tool background temperature prediction model is a low-frequency model with a sampling period as a time unit, and the planned energy injection and the open-loop thermal instability degree are taken as the driving characteristics of the machine tool heat source physical properties in the sampling period, so that the model can accurately predict the predicted machine tool background temperature in the future period. Then, according to the prediction results of the two models, the laser power and the machine tool feed rate can be controlled in time and reasonably. Through the effective training and application of the two prediction models, the application avoids the technical problems that the transient overheating control is relatively lagging and the machine tool thermal drift compensation does not consider the strong heat source in the load increasing process.

[0053] It should be noted that the above-mentioned embodiment sequence of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0054] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing, characterized in that: The method comprises: The static heat dissipation characteristics of the CNC machine tool processing task are obtained based on the task information. During the execution of the processing task, the command power and instantaneous melt pool temperature at each time point in the preset sampling period are collected. Temperature readings at different locations on the characteristic structural parts of the CNC machine tool are collected to obtain the machine tool structure temperature characteristic vector and the machine tool background temperature. At each moment, the processing temperature is obtained based on the machine tool background temperature and the command power, and the instantaneous temperature residual at each moment is obtained based on the difference between the instantaneous molten pool temperature and the processing temperature. The machine tool structure temperature feature vector, command power, and static heat dissipation characteristics are used as input data, and the instantaneous temperature residual at the next moment is used as output data to train the instantaneous temperature residual prediction model. The planned injection energy in each sampling cycle is calculated based on the command power, and the degree of open-loop thermal instability is obtained based on the instantaneous temperature residual variation characteristics in each sampling cycle. The command power, planned injection energy, and degree of open-loop thermal instability in the sampling cycle are used as input data, and the machine tool background temperature in the next sampling cycle is used as output data to train the machine tool background temperature prediction model. The instantaneous temperature residual prediction model is used to process data in a real-time sampling period to obtain a predicted instantaneous temperature residual; the predicted machine tool background temperature residual is obtained based on the difference between the machine tool background temperature in the real-time sampling period and the predicted machine tool background temperature output by the machine tool background temperature prediction model; The laser power is adjusted according to the predicted instantaneous temperature residual, and the machine feed rate is adjusted according to the predicted instantaneous temperature residual and the predicted machine background temperature residual.

2. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The static heat dissipation characteristics include geometric heat dissipation conditions and path repeated heating risks; the volume ratio of the solidified entity to the powder to be melted in the local area of ​​each path point is obtained, and the average value of all volume ratios is used as the geometric heat dissipation conditions; the number of repeated heating times for each path point is counted, and the average value of the number of repeated heating times for all path points is used as the path repeated heating risk.

3. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The machine tool background temperature is the average value of all elements in the machine tool structure temperature characteristic vector.

4. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The method for obtaining the processing temperature includes: The command power is multiplied by a preset conversion coefficient to obtain an input temperature, and the sum of the input temperature and the background temperature of the machine tool is used as the processing temperature.

5. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The method for obtaining the instantaneous temperature residual includes: The difference between the instantaneous molten pool temperature and the processing temperature is taken as the instantaneous temperature residual.

6. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The degree of open-loop thermal instability is the standard deviation of the instantaneous temperature residuals at all time points within the sampling period.

7. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The method for obtaining the predicted machine tool background temperature includes: An instantaneous temperature residual prediction model is used to obtain a predicted instantaneous temperature residual sequence of a real-time sampling period, and a predicted open-loop thermal instability degree is obtained based on the predicted instantaneous temperature residual sequence. The predicted open-loop thermal instability degree, the planned injection energy under the real-time sampling period, and the command power under the real-time sampling period are used as input data, and the predicted machine tool background temperature is output through the machine tool background temperature prediction model.

8. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The predicted machine tool background temperature residual is the difference between the predicted machine tool background temperature and the machine tool background temperature in the real-time sampling period.

9. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The adjusting the laser power according to the predicted instantaneous temperature residual comprises: The predicted instantaneous temperature residual is divided by a preset conversion coefficient to obtain a compensation power; and the compensation power is subtracted from the laser base power to obtain an adjusted laser power.

10. The method for intelligent thermal error compensation of a numerical control system based on multi-source temperature sensing according to claim 1, characterized in that: The adjusting of the machine tool feed rate according to the predicted instantaneous temperature residual and the predicted machine tool background temperature residual comprises: The sum of the predicted instantaneous temperature residual and the predicted machine tool background temperature residual is used as the cumulative value of the machine tool thermal error factor; if the cumulative value of the machine tool thermal error factor is greater than the preset first threshold and less than the preset second threshold, the feed rate is reduced by a linear descent method; if the cumulative value of the machine tool thermal error factor is greater than or equal to the second threshold, the feed rate is adjusted to the preset minimum value.

Images