An integrated management system for CNC tool production

Abstract

The present invention relates to the technical field of CNC tool production management, and specifically to a CNC tool production integrated management system, comprising a data collection module for real-time collection of CNC tool status data, an intelligent analysis module for performing status analysis on the CNC tool according to the collected data, and a dynamic management module for dynamically managing a production plan according to the status of the CNC tool. The present invention improves the accuracy and response speed of data collection by constructing a multi-sensor network and using multimodal data fusion technology. It combines a convolutional neural network and a dynamic Bayesian network to construct an intelligent analysis model, which can accurately predict the tool status, and adaptively adjust the tool use and maintenance strategy through a reinforcement learning algorithm, thereby improving the automation level of tool management. At the same time, dynamic management is performed according to the tool status, and production plans and task scheduling are optimized in real time, thereby reducing tool wear and downtime, and reducing production costs.

Description

Technical Field

[0001] The invention relates to the technical field of CNC tool production management, and in particular to a CNC tool production integrated management system. Background Art

[0002] Integrated management of CNC tool production can not only improve production efficiency and reduce costs, but also promote intelligent manufacturing and sustainable development. However, the management process faces the following challenges:

[0003] Although sensor technology has been used to monitor the status of tools (such as vibration, temperature, wear, etc.), the existing monitoring technology and algorithms still cannot accurately predict the wear status of tools in some complex machining tasks. In particular, in complex and harsh machining environments, the prediction model of tool life may be biased. Existing systems are limited to providing basic tool inventory and life management functions, and lack automated and intelligent analysis and decision support. Currently, most models used to analyze tool status are based on specific assumptions or experience, but in actual machining, tool status may be affected by multiple factors, and a single model cannot accurately describe tool status changes. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a CNC tool production integrated management system, comprising:

[0006] The data collection module is used to collect CNC tool status data in real time, specifically:

[0007] By building a multi-sensor network, the frequency and range of data collection are automatically adjusted according to the real-time parameter changes of the tool under different production conditions, and multi-modal data fusion is performed based on the weight of each sensor data using a dynamic weighting method;

[0008] And, the intelligent analysis module is used to analyze the status of the CNC tool based on the collected data, specifically:

[0009] The convolutional neural network algorithm is combined with the dynamic Bayesian network algorithm to build an intelligent analysis model, and the fused data is used as the input of the model. The state analysis of the CNC tool is performed based on the output results of the model, and finally the reinforcement learning algorithm is used to manage the state of the CNC tool.

[0010] Also includes:

[0011] The dynamic management module is used to dynamically manage the production plan according to the status of the CNC tool, specifically:

[0012] The production task scheduling is adaptively adjusted according to the predicted status of the CNC tool, and a comprehensive analysis is performed based on the adjusted comprehensive cost of the production task and the status of the CNC tool to complete the secondary analysis of the production plan.

[0013] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the frequency and range of automatically adjusting data collection are as follows:

[0014] Set the initial data collection of the vibration sensor to , the initial data collected by the temperature sensor is ;

[0015] Simultaneously set the reference parameter values ​​of each sensor as well as ;

[0016] Dynamically adjust the acquisition frequency according to the changes in each sensor data, then, in, Indicates the reference acquisition frequency, Indicates the adjusted acquisition frequency, Represents the adjustment coefficient, which is the sensitivity of frequency adjustment. Indicates the collected data of the vibration sensor at the current moment, Indicates the data collected by the temperature sensor at the current moment. Indicates the data change of the sensor, which is used to measure whether to adjust the acquisition frequency.

[0017] When the sensor data changes to meet the formula When , the acquisition frequency is dynamically reduced by adjusting the coefficient;

[0018] When the sensor data changes to meet the formula When , the current acquisition frequency is maintained by adjusting the coefficient;

[0019] When the sensor data changes to meet the formula When the acquisition frequency is increased, the coefficient can be adjusted.

[0020] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the multimodal data fusion is specifically as follows:

[0021] Set the initial weight for each sensor data, then we have, in, Indicates the number of sensors, Indicates The initial weights corresponding to the sensors;

[0022] According to the real-time feedback data from the sensor, the weight is dynamically adjusted, then: in, Indicates The weight of each sensor is dynamically adjusted. Indicates The weights before sensor adjustment, Indicates Real-time feedback values ​​of sensors;

[0023] According to the adjusted weights, the data is fused, and then, in, Indicates The data collected by the sensors, represents the fused data, Indicates The initial weights corresponding to the sensors.

[0024] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the convolutional neural network algorithm completes the feature extraction of the fused data set through convolution operation, as follows:

[0025] The fused data is constructed as a multidimensional time series, then we have: in, represents the number of time steps, Represents the data dimension of each time step, which is the number of dimensions corresponding to the data of different sensors. Indicates the fusion data The data corresponding to the time step, Indicated in The data of the sensor at all times, express dimensional real number space;

[0026] Convolutional neural network is used to extract local features from time series data, which is used to extract the spatial features of CNC tool state from each time step, specifically:

[0027] Performing convolution operation on multidimensional time series data, we have: in, Indicates Moment, The output features of the convolution kernels are the extracted local features. represents the weight of the convolution kernel, represents the bias term, represents the activation function, Represents the data dimension of each time step, which is the number of dimensions corresponding to the data of different sensors. Indicates the time series data The data corresponding to the sequence;

[0028] The extracted local features are subjected to dimensionality reduction, and we have: in, Represents the local features after dimensionality reduction, which is Moment The sensor data features are the factors affecting the status of CNC tools. Indicates the size of the pooling window in the dimensionality reduction process.

[0029] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the construction of the intelligent analysis model is as follows:

[0030] The vibration data and temperature data in the sensor data features are used as the influencing factors of the CNC tool state. At the moment, the state of the CNC tool is related in time. Status By the moment Status The impact of

[0031] According to the established dependencies, an intelligent analysis model is constructed, then: in, Represents the state transition probability, which is the probability of switching from the state of the CNC tool at the previous moment to the state of the CNC tool at the current moment. represents the state prediction of CNC tool under vibration and temperature data, Indicates the prediction results of the intelligent analysis model for CNC tools, including normal state and abnormal state. represents the joint probability of vibration and temperature data under a given tool state, represents the prior probability of the CNC tool state, is the marginal probability of vibration and temperature data, , Indicates that the CNC tool is in status The probability distribution of the vibration sensor data and the probability distribution of the temperature sensor under , Respectively represent the standard deviation of vibration data and temperature data, , Respectively indicate the CNC tool in state Below, the mean of vibration data and temperature data.

[0032] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the CNC tool state management using the reinforcement learning algorithm is specifically as follows:

[0033] Set current state Select the action set include,

[0034] Continue the current processing task without any maintenance; interrupt the processing task and perform tool maintenance; or directly replace the tool and continue the production task;

[0035] At the same time, according to the state change of the selected action at the next moment, the reward strategy for subsequent value evaluation is set, then,

[0036] When the selected action causes the state to change in the opposite direction at the next moment, the currently selected action is set as a negative reward;

[0037] When the selected action does not cause the state at the next moment to change in the opposite direction, the currently selected action is set as a positive reward;

[0038] According to the selected action, the state of the next moment is predicted based on the reinforcement learning algorithm. To conduct a value assessment, we have: in, Action indicating selection The next moment state The value assessment of Represents the learning efficiency, which is the sensitivity to new actions in each action update. represents the reward for the selected action, Represents the discount factor, with a value between 0 and 1. Indicates the optimal action at the next moment value assessment.

[0039] As a preferred solution of the integrated management system for production of numerical control tools described in the present invention, the production task scheduling is adaptively adjusted according to the predicted state of the numerical control tools as follows:

[0040] For the optimal scheduling of tasks under normal conditions of CNC tools, it is based on the comprehensive evaluation of the processing cost and importance of the production tasks. in, Indicates production tasks The weight is the importance of the production task, which is set by the implementer according to the actual application scenario. Indicates production tasks The processing cost includes the implementation cost and resource cost of completing the task. Indicates production tasks The comprehensive priority of the tool is determined by the user, and the CNC tool in normal state is adaptively selected according to the priority level.

[0041] For the secondary matching of CNC tools and production tasks under abnormal conditions, a comprehensive assessment is made based on the task suspension cost and loss. in, It indicates the probability that the tool status is abnormal after the tool is replaced to perform the production task. Indicates that the task is paused The loss cost, It represents the expected loss of failed generation task. According to the level of loss, the tool and production task are adaptively matched for the second time.

[0042] As a preferred solution of the integrated management system for CNC tool production described in the present invention, the comprehensive analysis based on the adjusted comprehensive cost of production tasks and the status of CNC tools is as follows:

[0043] Record the relevant data of the generated tasks after adjustment, and feed the recorded data back to the intelligent analysis model to conduct a secondary analysis of the CNC tool. If the analysis result is normal, it means that the adjustment of the current production plan is normal; if the analysis result is abnormal, it means that the adjustment of the current production plan is abnormal, and readjustment is performed until the secondary analysis result is normal;

[0044] Also record the processing costs of the adjusted production plan , if the adjusted processing cost meets the formula , indicating that the adjusted production plan exceeds the preset cost and needs to be readjusted until the preset cost is met.

[0045] Beneficial effects of the present invention: The present invention constructs a multi-sensor network to collect data such as vibration and temperature of CNC tools in real time, and uses multimodal data fusion technology to dynamically weight and process data from different sensors, thereby improving the accuracy and response speed of data collection; the convolutional neural network and the dynamic Bayesian network are combined to construct an intelligent analysis model, which can accurately predict the tool status (such as normal, abnormal, etc.), and adaptively adjust the tool use and maintenance strategy through a reinforcement learning algorithm to achieve intelligent decision support and improve the degree of automation of tool management; at the same time, dynamic management is performed according to the tool status, and production plans and task scheduling are optimized in real time, which reduces tool wear and downtime and reduces production costs; at the same time, based on the prediction of abnormal conditions, the system can arrange tool maintenance in advance, avoid production interruptions caused by faults, and improve production efficiency; through an adaptive task scheduling mechanism and feedback-based optimization and adjustment, the present invention can flexibly adjust tool use and maintenance plans under different production environments and task requirements, ensuring efficient completion of production tasks while extending tool life. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0047] Figure 1 The present invention is a schematic diagram of the overall structure of a CNC tool production integrated management system.

[0048] Figure 2 The present invention is a schematic diagram of the logical structure of data collection for a CNC tool production integrated management system. DETAILED DESCRIPTION

[0049] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0050] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0052] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0053] Meanwhile, in the description of the present invention, it should be noted that the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0054] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] Example 1: Reference Figure 1 , is an embodiment of the present invention, which provides a CNC tool production integrated management system, including a data collection module, an intelligent analysis module and a dynamic management module. The data collection module is used to collect CNC tool status data in real time; the intelligent analysis module is used to perform status analysis on the CNC tool based on the collected data; and the dynamic management module is used to perform dynamic management based on the status of the CNC tool.

[0056] Reference Figure 2 , an implementation logic of data collection is proposed. Specifically, the data collection module uses multiple sensors to collect CNC tool status data, and combines the data fusion algorithm to dynamically weight different sensors, and adjust the weights based on real-time feedback to improve the accuracy of CNC tool status data. The module includes a data collection unit and a data fusion unit. The data collection unit uses dynamic adaptive sensor data collection technology to collect data on CNC tools, and the data fusion unit uses a dynamic weighting method to perform multimodal data fusion according to the weights of each sensor data.

[0057] Specifically, the dynamic adaptive sensing data acquisition technology achieves accurate and flexible data acquisition and monitoring by constructing a multi-sensor network to automatically adjust the frequency and range of data acquisition according to the real-time parameter changes of the tool under different production conditions. The multi-sensor network is a data acquisition unit composed of multiple sensors, including a vibration sensor and a temperature sensor. The vibration sensor is used to monitor the vibration state of the CNC tool, and then capture the dynamic characteristics of the CNC tool during the cutting process. The temperature sensor is used to measure the temperature change between the CNC tool and the workpiece, and then evaluate the thermal impact caused by friction and cutting.

[0058] Furthermore, the frequency and range of data collection are automatically adjusted according to the parameter changes of the sensor, and the frequency of data collection is automatically adjusted as follows:

[0059] Set the initial data collection of the vibration sensor to , the initial data collected by the temperature sensor is ;

[0060] Simultaneously set the reference parameter values ​​of each sensor as well as ;

[0061] Dynamically adjust the acquisition frequency according to the changes in each sensor data, then, in, Indicates the reference acquisition frequency, Indicates the adjusted acquisition frequency, Indicates the adjustment coefficient, which is set by the implementer according to the actual application scenario and is the sensitivity of the frequency adjustment. Indicates the data change of the sensor, which is used to measure whether to adjust the acquisition frequency.

[0062] When the sensor data changes to meet the formula When , it means that the sensor data exceeds the benchmark data, and the acquisition frequency is dynamically reduced by adjusting the coefficient;

[0063] When the sensor data changes to meet the formula When , it indicates that the sensor data is normal, and the current acquisition frequency is maintained by adjusting the coefficient;

[0064] When the sensor data changes to meet the formula , it means that the sensor data is lower than the benchmark data, and the acquisition frequency can be increased by adjusting the coefficient.

[0065] It should be noted that for the adjusted acquisition frequency, the collection of sensor data is reversely controlled through the feedback regulation mechanism, then,

[0066] Taking the adjusted acquisition frequency as the current acquisition frequency, we have: According to the current sensor data change rate, the acquisition frequency at the next moment is adjusted, then: in, Indicates the current collection frequency. Indicates the collection frequency at the next moment, Represents the data in the current sensor, including vibration sensor and temperature sensor, Indicates the data change rate of the sensor, It represents the feedback control coefficient, which is set by the implementer according to the actual application scenario.

[0067] The multimodal data fusion is to dynamically adjust the weight of each sensor data according to real-time feedback, thereby realizing multimodal fusion of data, as follows:

[0068] Set the initial weight for each sensor data, then we have, in, Indicates the number of sensors, Indicates The initial weights corresponding to the sensors;

[0069] According to the real-time feedback data from the sensor, the weight is dynamically adjusted, then: in, Indicates The weight of each sensor is dynamically adjusted. Indicates The weights before sensor adjustment, Indicates Real-time feedback values ​​of sensors;

[0070] According to the adjusted weights, the data is fused, and then, in, Indicates The data collected by the sensors, represents the fused data, Indicates The initial weights corresponding to the sensors.

[0071] Furthermore, the intelligent analysis module is used to perform status analysis on the CNC tool based on the collected data by constructing an intelligent analysis model, using the fused data as the input of the model, and performing status analysis on the CNC tool based on the output results of the model, including an intelligent analysis model construction unit and an intelligent analysis unit.

[0072] Specifically, the intelligent analysis model construction unit is based on a machine learning algorithm, and constructs an intelligent analysis model by combining a convolutional neural network algorithm and a dynamic Bayesian network algorithm. The neural network algorithm extracts features of the fused data set through a convolution operation, as follows:

[0073] The fused data is constructed as a multidimensional time series, then we have: in, represents the number of time steps, Represents the data dimension of each time step, which is the value of different sensors, Indicated in The data of the sensors at all times;

[0074] Convolutional neural network is used to extract local features from time series data, which is used to extract the spatial features of CNC tool state from each time step, specifically:

[0075] Performing convolution operation on multidimensional time series data, we have: in, Indicates Moment, The output features of the convolution kernels are the extracted local features. represents the weight of the convolution kernel, represents the bias term, Represents the activation function, which is selected by the implementer according to the actual application scenario;

[0076] The extracted local features are subjected to dimensionality reduction, and we have: in, Represents the local features after dimensionality reduction, which is Moment The sensor data features are the factors affecting the status of CNC tools. Indicates the size of the pooling window in the dimensionality reduction process.

[0077] The dynamic Bayesian network algorithm predicts the state of the CNC tool in time series based on the extracted feature data. The specific prediction process is as follows:

[0078] Based on the influencing factors and state correlation, the dependency relationship between the three is constructed, specifically: in, express The status of CNC tools at all times, express The status of CNC tools at all times, express The node of the factors affecting the vibration data at the moment, express The temperature data at the moment affects the node of the factor, Represents the time dependency of the CNC tool status nodes, Indicates the dependency between the CNC tool status and influencing factors at the same time. Indicates the mutual independence between the influencing factor nodes;

[0079] The vibration data and temperature data in the sensor data features are used as the influencing factors of the CNC tool state. At the moment, the state of the CNC tool is related in time. Status By the moment Status The impact of

[0080] According to the established dependencies, an intelligent analysis model is constructed, then: in, Represents the state transition probability, which is the probability of switching from the state of the CNC tool at the previous moment to the state of the CNC tool at the current moment. represents the state prediction of CNC tool under vibration and temperature data, Indicates the prediction results of the intelligent analysis model for CNC tools, including normal state and abnormal state. represents the joint probability of vibration and temperature data under a given tool state, represents the prior probability of the CNC tool state, is the marginal probability of vibration and temperature data, , Indicates that the CNC tool is in status The probability distribution of the vibration sensor data and the probability distribution of the temperature sensor under , Respectively represent the standard deviation of vibration data and temperature data, , Respectively indicate the CNC tool in state Below, the mean of vibration data and temperature data.

[0081] It should be noted that in order to improve the accurate analysis of the status of CNC tools, the reinforcement learning algorithm is used to formulate the CNC tool status management strategy, as follows:

[0082] The CNC tool management strategy based on the reinforcement learning algorithm is formulated by Select the action set Conduct a comprehensive value assessment to evaluate the impact of the selected action on the next state, and then complete the precise analysis of the CNC tool, as follows:

[0083] Set current state Select the action set include,

[0084] Continue the current processing task without any maintenance; interrupt the processing task and perform tool maintenance; or directly replace the tool and continue the production task;

[0085] At the same time, according to the state change of the selected action at the next moment, the reward strategy for subsequent value evaluation is set, then,

[0086] When the selected action causes the state to change in the opposite direction at the next moment, the currently selected action is set as a negative reward;

[0087] When the selected action does not cause the state at the next moment to change in the opposite direction, the currently selected action is set as a positive reward;

[0088] According to the selected action, the state of the next moment is predicted based on the reinforcement learning algorithm. To conduct a value assessment, we have: in, Action indicating selection The next moment state The value assessment of Indicates learning efficiency, which is the sensitivity to new actions in each action update. It is set by the implementer according to the actual application scenario. Represents the reward for the selected action, which is determined by the set reward strategy. Represents the discount factor, which is between 0 and 1 and is set by the implementer according to the actual application scenario. Indicates the optimal action at the next moment value assessment;

[0089] The intelligent analysis model is used to predict the current state of the tool, and the reinforcement learning algorithm is used to select the optimal action to update the next state of the CNC tool until the updated CNC tool state meets production requirements.

[0090] Furthermore, the dynamic management module is used to dynamically adjust the CNC tool production plan according to the status of the CNC tool. It dynamically adjusts the CNC tool production plan according to the output results of the intelligent analysis model, including a dynamic adjustment of the production plan unit and a feedback optimization adjustment unit. The dynamic adjustment of the production plan unit adaptively adjusts the production task scheduling according to the predicted status of the CNC tool, and the feedback optimization adjustment unit optimizes the production plan based on the secondary analysis of the adjusted production plan.

[0091] Specifically, the adaptive adjustment of production task scheduling according to the predicted state of the CNC tool is to schedule the production task according to the predicted result of the intelligent analysis simulation, as follows:

[0092] When the predicted status of the CNC tool is normal, the urgent production task is adjusted to the time period when the tool status is normal to ensure the quality and efficiency of the production task. At the same time, according to different production tasks, the CNC tool for processing is dynamically selected to improve the processing accuracy and reduce tool wear;

[0093] When the predicted state of the CNC tool is abnormal, the current processing task is suspended and the matching between the production task and the CNC tool is re-performed.

[0094] Furthermore, for the optimal scheduling of tasks under normal conditions of CNC tools, it is based on the comprehensive evaluation of the processing cost and importance of production tasks, then, in, Indicates production tasks The weight is the importance of the production task, which is set by the implementer according to the actual application scenario. Indicates production tasks The processing cost includes the implementation cost and resource cost of completing the task. Indicates production tasks The comprehensive priority of the tool is determined by the user, and the CNC tool in normal state is adaptively selected according to the priority level.

[0095] For the secondary matching of CNC tools and production tasks under abnormal conditions, a comprehensive assessment is made based on the task suspension cost and loss. in, It indicates the probability that the tool status is abnormal after the tool is replaced to perform the production task. Indicates that the task is paused The loss cost, It represents the expected loss of failed generation task. According to the level of loss, the tool and production task are adaptively matched for the second time.

[0096] Furthermore, the secondary analysis based on the adjusted production plan is a comprehensive analysis based on the adjusted comprehensive cost of the production task and the state of the CNC tool, then,

[0097] Record the relevant data of the generated tasks after adjustment, and feed the recorded data back to the intelligent analysis model to conduct a secondary analysis of the CNC tool. If the analysis result is normal, it means that the adjustment of the current production plan is normal; if the analysis result is abnormal, it means that the adjustment of the current production plan is abnormal, and readjustment is performed until the secondary analysis result is normal;

[0098] Also record the processing costs of the adjusted production plan , if the adjusted processing cost meets the formula , indicating that the adjusted production plan exceeds the preset cost and needs to be readjusted until the preset cost is met.

[0099] It should be noted that the real-time data provided by the data collection module is the input of the subsequent intelligent analysis module. Through multimodal data fusion, the weight of the sensor data is adjusted in real time according to the feedback, which improves the accuracy of the CNC tool status data; the intelligent analysis module analyzes the fused data provided by the data collection module and passes the prediction results (tool status) to the dynamic management module; the dynamic management module relies on the prediction results of the intelligent analysis module. When a tool abnormality is detected, the current task will be suspended and resources will be reallocated, thereby maximizing production efficiency and reducing tool wear;

[0100] Through the collaborative work of data collection, analysis and dynamic management, the system can understand the status of the tool in real time, and dynamically optimize the production plan based on the prediction results to achieve intelligent manufacturing; the combination of the three enables the system to make timely adjustments to avoid downtime or task interruption caused by tool abnormalities, reduce tool waste, and improve production efficiency; the system can adaptively adjust production tasks according to the real-time status of the tool to ensure that the task is executed in the optimal state. Through the reinforcement learning algorithm, the system can also continuously optimize the maintenance strategy to increase the service life of the tool.

[0101] Furthermore, if the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0102] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0103] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0104] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment may not be described (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those features that are not relevant to implementing the invention).

[0105] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A CNC tool production integrated management system, characterized by: include, The data collection module is used to collect CNC tool status data in real time, specifically: By building a multi-sensor network, the frequency and range of data collection are automatically adjusted according to the real-time parameter changes of the tool under different production conditions, and multi-modal data fusion is performed based on the weight of each sensor data using a dynamic weighting method; And, the intelligent analysis module is used to analyze the status of the CNC tool based on the collected data, specifically: The convolutional neural network algorithm is combined with the dynamic Bayesian network algorithm to build an intelligent analysis model, and the fused data is used as the input of the model. The state analysis of the CNC tool is performed based on the output results of the model, and finally the reinforcement learning algorithm is used to manage the state of the CNC tool. The construction of the intelligent analysis model is as follows: The vibration data and temperature data in the sensor data features are used as the influencing factors of the CNC tool state. At the same time, at time t, the state of the CNC tool is set to be related in time. The state S of the CNC tool at time t t The state S at time t-1 is t-1 The impact of According to the established dependencies, an intelligent analysis model is constructed, then: P(S t |h t ′ ,V ,h t ′ ,T ,S t-1 )=P(S t |S t-1 )·P(S t |h t ′ ,V ,h t ′ ,T ) P(h t ′ ,V ,h t ′ ,T |S t )=P(h t ′ ,V |S t )·P(h t ′ ,T |S t ) Among them, P(S t |S t-1 ) represents the state transition probability, which is the probability of switching from the state of the CNC tool at the previous moment to the state of the CNC tool at the current moment. P(S t |h t ' ,V ,h t ' ,T ) represents the state prediction of the CNC tool under vibration and temperature data, P(S t |h t ' ,V ,h t ' ,T ,S t-1 ) represents the prediction result of the intelligent analysis model for the CNC tool, including normal state and abnormal state, P(h′ t,V ,h′ t,T |S t ) represents the joint probability of vibration and temperature data under a given tool state, P(S t ) represents the prior probability of the CNC tool state, P(h′ t,V ,h′ t,T ) is the marginal probability of vibration and temperature data, P(h′ t,V |S t ), P(h′ t,T |S t ) indicates that the CNC tool is in state S t The probability distribution of the vibration sensor data under and the probability distribution of the temperature sensor, σ V(t) , σ T(t) Represent the standard deviation of vibration data and temperature data respectively, Respectively represent the CNC tool in state S t Below, the mean of vibration data and temperature data; The specific implementation of CNC tool state management using reinforcement learning algorithm is as follows: Set the current state S t The action set a selected below includes: Continue the current processing task without any maintenance; interrupt the processing task and perform tool maintenance; or directly replace the tool and continue the production task; At the same time, according to the state change of the selected action at the next moment, the reward strategy for subsequent value evaluation is set, then, When the selected action causes the state to change in the opposite direction at the next moment, the currently selected action is set as a negative reward; When the selected action does not cause the state at the next moment to change in the opposite direction, the currently selected action is set as a positive reward; According to the selected action, the state S at the next moment is predicted based on the reinforcement learning algorithm. t+1 To conduct a value assessment, we have: Among them, Q(S t+1 ,a) represents the effect of the selected action a on the next state S t+1 The value assessment of , α represents the learning efficiency, which is the sensitivity to new actions in each action update, r represents the reward for the selected action, β represents the discount factor, which is between 0 and 1, Represents the value evaluation of the optimal action a′ in the next state; Also includes: The dynamic management module is used to dynamically manage the production plan according to the status of the CNC tool, specifically: Adaptively adjust the production task scheduling according to the predicted status of the CNC tool, conduct a comprehensive analysis based on the adjusted comprehensive cost of the production task and the status of the CNC tool, and complete the secondary analysis of the production plan; The adaptive adjustment of production task scheduling according to the predicted state of the CNC tool is specifically as follows: For the optimal scheduling of tasks under normal conditions of CNC tools, it is based on the comprehensive evaluation of the processing cost and importance of the production tasks. Among them, w i Represents the production task T i The weight, C i Represents the production task T i The processing cost includes the implementation cost and resource cost of completing the task, P(T i ) represents the production task T i The comprehensive priority of the tool is determined by the user, and the CNC tool in normal state is adaptively selected according to the priority level. For the secondary matching of CNC tools and production tasks under abnormal conditions, a comprehensive assessment is made based on the task suspension cost and loss. Cost(T i ,Pause)=P(S t = Abnormal)·C lost (T i ) Among them, P(S t = abnormal) indicates the probability that the tool status is abnormal after the tool is replaced to perform the production task, Cost(T i ,Pause) means pausing task T i The loss cost, C lost (T i ) represents the expected loss of failed generation task. According to the level of loss, the tool and production task are adaptively matched for the second time.

2. The integrated management system for CNC tool production according to claim 1, characterized in that: The frequency and range of automatic adjustment data collection are as follows: Set the initial data collected by the vibration sensor to V0(t) and the initial data collected by the temperature sensor to T0(t); At the same time, set the reference parameter value V of each sensor base (t) and T base (t); Dynamically adjust the acquisition frequency according to the changes in each sensor data, then, f new =f base ·(1+α·ΔP) Among them, f base Indicates the reference acquisition frequency, f new represents the adjusted acquisition frequency, α represents the adjustment coefficient, which is the sensitivity of the frequency adjustment, V(t) represents the data collected by the vibration sensor at the current moment, T(t) represents the data collected by the temperature sensor at the current moment, ΔP represents the data change of the sensor, which is used to measure whether to adjust the acquisition frequency, then, When the sensor data changes satisfy the formula ΔP ≥ 1, the acquisition frequency is dynamically reduced by adjusting the coefficient; When the sensor data changes satisfy the formula 0≤ΔP<1, the current acquisition frequency is maintained by adjusting the coefficient; When the sensor data changes satisfy the formula ΔP<0, the acquisition frequency is increased by adjusting the coefficient.

3. A CNC tool production integrated management system as claimed in claim 2, characterized in that: The multimodal data fusion is specifically as follows: Set the initial weight for each sensor data, then we have, Where n represents the number of sensors, w i represents the initial weight corresponding to the i-th sensor; According to the real-time feedback data from the sensor, the weight is dynamically adjusted, then: in, represents the dynamically adjusted weight of the i-th sensor, represents the weight of the i-th sensor before adjustment, F i Represents the real-time feedback value of the i-th sensor; According to the adjusted weights, the data is fused, and then, Among them, X i represents the data collected by the i-th sensor, X represents the fused data, and w i Represents the initial weight corresponding to the i-th sensor.

4. A CNC tool production integrated management system as claimed in claim 3, characterized in that: The convolutional neural network algorithm completes the feature extraction of the fused data set through convolution operation, as follows: The fused data is constructed as a multidimensional time series, then we have: X={x1,x2,x3,...,x T } x t ∈R n Among them, T represents the number of time steps, n represents the data dimension of each time step, and x represents the number of dimensions corresponding to the data of different sensors. T represents the data corresponding to the Tth time step in the fused data, x t represents the data of the sensor at time t, R n represents n-dimensional real number space; Convolutional neural network is used to extract local features from time series data, which is used to extract the spatial features of CNC tool state from each time step, specifically: Performing convolution operation on multidimensional time series data, we have: Among them, h i,j represents the output feature of the jth convolution kernel at the i-th moment, which is the extracted local feature, W j,k represents the weight of the convolution kernel, b j represents the bias term, σ represents the activation function, n represents the data dimension of each time step, and is the number of dimensions corresponding to the data of different sensors, x i+k-1 Represents the data corresponding to the i+k-1th sequence in the time series data; The extracted local features are subjected to dimensionality reduction processing, and then we have: h i ′ ,j =max{h i,j ,h i+1,j ,...,h i+p,j } Among them, h i ' ,j It represents the local feature after dimensionality reduction, is the j-th sensor data feature at the i-th moment, is the influencing factor of the CNC tool state, and p represents the size of the pooling window in the dimensionality reduction process.

5. A CNC tool production integrated management system as claimed in claim 4, characterized in that: The comprehensive analysis based on the adjusted comprehensive cost of production tasks and the status of CNC tools is as follows: Record the relevant data of the generated tasks after adjustment, and feed the recorded data back to the intelligent analysis model to conduct secondary analysis on the CNC tool. If the analysis result is normal, it means that the adjustment of the current production plan is normal; If the analysis result is abnormal, it means that the adjustment of the current production plan is abnormal, and it should be readjusted until the secondary analysis result is normal; At the same time, the processing cost C of the adjusted production plan is recorded i ′, if the adjusted processing cost is compared with the processing cost before adjustment, it satisfies the formula C i ′≥2C i , indicating that the adjusted production plan exceeds the preset cost and needs to be readjusted until the preset cost is met.

Images