Numerical control machine tool energy consumption state monitoring method based on recurrent neural network

Abstract

The invention discloses a numerical control machine tool energy consumption state monitoring method based on a recurrent neural network, and relates to the technical field of numerical control machine tool health monitoring, and the method comprises the following steps: S1, collecting energy consumption data of a numerical control machine tool in a machining process; s2, preprocessing the collected data; s3, classifying and storing the preprocessed data; s4, predicting the future energy consumption condition according to the historical data of the energy consumption of the numerical control machine tool; s5, fault diagnosis is carried out according to the future energy consumption condition of the numerical control machine tool, the energy consumption condition of the numerical control machine tool in a certain time in the future can be predicted according to the historical data of the energy consumption condition of the numerical control machine tool, and the operation state of the numerical control machine tool is diagnosed according to the predicted energy consumption condition. And early warning is conducted on possible faults in advance, workpiece damage caused by the faults in the machining process can be avoided, the machining efficiency can be effectively improved, and the purposes of energy conservation and emission reduction are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of health monitoring of numerically controlled machine tools, and in particular to a method for monitoring the energy consumption status of numerically controlled machine tools based on a recurrent neural network. Background Art

[0002] Health monitoring of CNC machine tools refers to reducing downtime and improving machine utilization by monitoring various data during the processing of CNC machine tools, diagnosing possible machine failures and providing early warnings, which is of great significance for improving production efficiency and reducing production costs.

[0003] Current health monitoring methods for CNC machine tools generally only focus on the performance and remaining life of components such as the spindle and tool, but ignore the monitoring of energy data such as electricity, gas (compressed air), and liquid (cutting fluid) required in the machining process, which cannot improve the energy utilization rate of machine tools. Therefore, a CNC machine tool energy consumption status monitoring method based on recurrent neural network is proposed. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the prior art and to propose a method for monitoring the energy consumption status of a CNC machine tool based on a recurrent neural network.

[0005] A method for monitoring the energy consumption state of a CNC machine tool based on a recurrent neural network comprises the following steps:

[0006] S1. Collect energy consumption data of CNC machine tools during processing;

[0007] S2, preprocessing the collected data;

[0008] S3, classify and store the preprocessed data;

[0009] S4. Predict future energy consumption based on historical data of energy consumption of CNC machine tools;

[0010] S5. Perform fault diagnosis based on the future energy consumption of the CNC machine tool.

[0011] Preferably, in step S1, the data acquisition object is specified according to the goal of CNC machine tool status monitoring, which may be one or more of the three types of electrical, pneumatic and hydraulic data, and the sensor used for data acquisition is selected according to the data acquisition object.

[0012] Preferably, in step S2, data preprocessing includes the following steps:

[0013] a. Clean the original data to remove the error information in the data, including null values, illegal characters and abnormal values;

[0014] b. Format the data and unify the format of data representation;

[0015] c. Mark the data. According to the actual working status of the CNC machine tool, mark the energy consumption data as normal or various abnormal conditions to form marked data pairs.

[0016] Preferably, in step S3, the classification and storage of data includes the following steps:

[0017] a. Classify the collected energy consumption data according to the different materials, processing technology and data types of CNC machine tools and workpieces, and store each type of data separately;

[0018] b. Use the Storm stream processing framework to receive and process the data stream of energy consumption data;

[0019] c. Use the Spark engine to divide energy consumption data into small data units to achieve batch processing and offline data access;

[0020] d. Use Hive distributed storage system to store energy consumption data.

[0021] Preferably, in step S4, predicting future energy consumption includes the following steps:

[0022] a. Build an asynchronous sequence-to-sequence recurrent neural network;

[0023] b. Input the historical data of energy consumption of CNC machine tools, use the back propagation algorithm to train the above recurrent neural network, and obtain the energy consumption state prediction model;

[0024] c. Input the current energy consumption into the above energy consumption state prediction model to predict and output the energy consumption within a certain period of time in the future.

[0025] Preferably, in step S5, performing fault diagnosis according to the future energy consumption of the CNC machine tool includes the following steps:

[0026] a. Build a recurrent neural network for sequence-to-category model;

[0027] b. Input the historical data of energy consumption of CNC machine tools and their corresponding states, use the back propagation algorithm to train the above recurrent neural network, and obtain a CNC machine tool fault diagnosis model to determine whether the state of the CNC machine tool is normal;

[0028] c. Input the obtained energy consumption in a certain period of time in the future into the above-mentioned CNC machine tool fault diagnosis model to determine whether the CNC machine tool will fail in a certain period of time in the future;

[0029] d. Provide early warning of possible failures.

[0030] Compared with the prior art, the present invention has the following advantages:

[0031] The present invention can predict the energy consumption of a CNC machine tool within a certain period of time in the future based on the historical data of the energy consumption of the CNC machine tool, and diagnose the operating status of the CNC machine tool based on the predicted energy consumption, and give early warning of possible faults, thereby avoiding damage to workpieces caused by faults during the machining process, and effectively improving the machining efficiency. At the same time, it can avoid waste of resources, reduce energy consumption, and achieve the purpose of energy conservation and emission reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The figure is a flow chart of the method for monitoring the energy consumption status of a CNC machine tool according to the present invention.

[0033] Figure 2 It is a schematic diagram of the energy consumption state prediction model in the present invention.

[0034] Figure 3 It is a schematic diagram of the energy consumption fault diagnosis model in the present invention. DETAILED DESCRIPTION

[0035] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0036] Reference Figure 1 As shown, a method for monitoring the energy consumption status of a CNC machine tool based on a recurrent neural network comprises the following steps:

[0037] S1. Collect energy consumption data of CNC machine tools during processing;

[0038] S2, preprocessing the collected data;

[0039] S3, classify and store the preprocessed data;

[0040] S4. Predict future energy consumption based on historical data of energy consumption of CNC machine tools;

[0041] S5. Perform fault diagnosis based on the future energy consumption of the CNC machine tool.

[0042] In this embodiment, in the step S1, the object of data collection is specified according to the goal of CNC machine tool status monitoring, which can be one or more of the three types of data: electricity, gas, and liquid. The energy required by the CNC machine tool includes electricity, gas, and liquid. The energy-consuming components include hydraulic systems, transmission systems, peripheral equipment, lubrication systems, and servo transmission systems; and according to the data collection object, the sensor used for data collection is selected, and a smart meter is used for the collection of electricity data, a gas flow meter is used for the collection of gas data, and a liquid flow meter is used for the collection of liquid data.

[0043] In this embodiment, in step S2, data preprocessing includes the following steps:

[0044] a. Clean the original data to remove the error information in the data, including null values, illegal characters and abnormal values;

[0045] b. Format the data, unify the format of data representation, and improve data processing performance;

[0046] c. Mark the data and mark the energy consumption data as normal or various abnormal conditions according to the actual working status of the CNC machine tool to form a marked data pair (x i ,y i ), where x i Represents the i-th data, y i Represents x i The corresponding label.

[0047] In this embodiment, in step S3, the classification and storage of data includes the following steps:

[0048] a. Classify the collected energy consumption data according to the different materials, processing technology and data types of CNC machine tools and workpieces, and store each type of data separately;

[0049] b. Use the Storm stream processing framework to receive and process the data stream of energy consumption data;

[0050] c. Use the Spark engine to divide energy consumption data into small data units to achieve batch processing and offline data access;

[0051] d. Use Hive distributed storage system to store energy consumption data.

[0052] In this embodiment, in step S4, predicting future energy consumption includes the following steps:

[0053] a. Build an asynchronous sequence-to-sequence recurrent neural network, such as Figure 2 As shown;

[0054] b. Input the historical data of energy consumption of CNC machine tools, use the back propagation algorithm to train the above recurrent neural network, and obtain the energy consumption state prediction model;

[0055] c. The energy consumption data x at the current time t and the previous n times t ,x t-1 ,…,x t-n Input the above energy consumption state prediction model and output the energy consumption y at the next m moments t ,y t+1 ,…,y t-+m .

[0056] In this embodiment, in step S5, fault diagnosis according to the future energy consumption of the CNC machine tool includes the following steps:

[0057] a. Build a recurrent neural network from sequence to category mode, such as Figure 3 As shown;

[0058] b. Input the historical data of energy consumption of CNC machine tools and their corresponding states, use the back propagation algorithm to train the above recurrent neural network, and obtain a CNC machine tool fault diagnosis model to determine whether the state of the CNC machine tool is normal;

[0059] c. The energy consumption y in the future m moments t ,y t+1 ,…,y t-+m And the energy consumption data x at the current time t and the previous n times t ,x t-1 ,…,x t-n Input the above CNC machine tool fault diagnosis model to determine the operating status of the CNC machine tool in the next m moments;

[0060] d. Provide early warning of possible failures.

[0061] It is known from common technical knowledge that the present invention can be implemented by other embodiments that do not deviate from its spirit or essential features. Therefore, the above disclosed embodiments are only illustrative in all respects and are not exclusive. All changes within the scope of the present invention or within the scope equivalent to the present invention are included in the present invention.

Claims

1. A method for monitoring energy consumption status of CNC machine tools based on recurrent neural network, characterized in that: The following steps are involved: S1. Collect energy consumption data of CNC machine tools during processing; S2, preprocessing the collected data; S3, classify and store the preprocessed data; S4. Predict future energy consumption based on historical data of energy consumption of CNC machine tools; S5. Perform fault diagnosis based on the future energy consumption of the CNC machine tool.

2. The method for monitoring energy consumption status of a CNC machine tool based on a recurrent neural network according to claim 1, characterized in that: In the step S1, the data acquisition object is specified according to the goal of the CNC machine tool status monitoring, which can be one or more of the three types of electrical, pneumatic and hydraulic data, and the sensor used for data acquisition is selected according to the data acquisition object.

3. The method for monitoring energy consumption status of a CNC machine tool based on a recurrent neural network according to claim 1, characterized in that: In step S2, data preprocessing includes the following steps: a. Clean the original data to remove the error information in the data, including null values, illegal characters and abnormal values; b. Format the data and unify the format of data representation; c. Mark the data. According to the actual working status of the CNC machine tool, mark the energy consumption data as normal or various abnormal conditions to form marked data pairs.

4. The method for monitoring energy consumption status of a CNC machine tool based on a recurrent neural network according to claim 1, characterized in that: In step S3, the classification and storage of data includes the following steps: a. Classify the collected energy consumption data according to the different materials, processing technology and data types of CNC machine tools and workpieces, and store each type of data separately; b. Use the Storm stream processing framework to receive and process the data stream of energy consumption data; c. Use the Spark engine to divide energy consumption data into small data units to achieve batch processing and offline data access; d. Use Hive distributed storage system to store energy consumption data.

5. The method for monitoring energy consumption status of a CNC machine tool based on a recurrent neural network according to claim 1, characterized in that: In step S4, predicting future energy consumption includes the following steps: a. Build an asynchronous sequence-to-sequence recurrent neural network; b. Input the historical data of energy consumption of CNC machine tools, use the back propagation algorithm to train the above recurrent neural network, and obtain the energy consumption state prediction model; c. Input the current energy consumption into the above energy consumption state prediction model to predict and output the energy consumption within a certain period of time in the future.

6. The method for monitoring energy consumption status of a CNC machine tool based on a recurrent neural network according to claim 1, characterized in that: In step S5, fault diagnosis according to the future energy consumption of the CNC machine tool includes the following steps: a. Build a recurrent neural network for sequence-to-category model; b. Input the historical data of energy consumption of CNC machine tools and their corresponding states, use the back propagation algorithm to train the above recurrent neural network, and obtain a CNC machine tool fault diagnosis model to determine whether the state of the CNC machine tool is normal; c. Input the obtained energy consumption in a certain period of time in the future into the above-mentioned CNC machine tool fault diagnosis model to determine whether the CNC machine tool will fail in a certain period of time in the future; d. Provide early warning of possible failures.

Images