A monitoring system for heterogeneous equipment in distributed manufacturing systems

Abstract

The present invention discloses a monitoring system for heterogeneous equipment in a distributed manufacturing system, comprising: a distributed manufacturing workshop and a cloud service unit, wherein one cloud service unit corresponds to a plurality of distributed manufacturing workshops; the distributed manufacturing workshop comprises: a heterogeneous data acquisition module and a heterogeneous data processing module; the cloud service unit comprises: a central database, a data monitoring and analysis module, and an interactive module; the present invention realizes effective access to the bottom layer heterogeneous equipment of the workshop, performs corresponding adaptation according to the type of equipment and the type of communication protocol, has a fast configuration speed, can adopt a one-to-one or one-to-many form, and flexibly realizes access to heterogeneous equipment.

Description

Technical Field

[0001] The present invention belongs to the technical field of data collection and communication of production equipment, and in particular relates to a monitoring system for heterogeneous equipment in a distributed manufacturing system. Background Art

[0002] With the continuous development of the manufacturing industry, the division of labor in the manufacturing process has become increasingly refined. Many manufacturing companies have adopted a distributed construction method for their factories and workshops, completing product production tasks through efficient and close cooperation, greatly improving product production efficiency.

[0003] Distributed manufacturing systems lead to the problem of distance that information transmission between factory workshops needs to overcome. At the same time, the equipment in the workshop is diverse and large in number. The connected equipment and the data generated are extremely heterogeneous, and the data communication specifications and construction standards between different workshops are different. These problems restrict the interconnection and information integration between factory workshops, which in turn makes it difficult to control the entire production process and to conduct effective and unified monitoring of key equipment in factory workshops.

[0004] Therefore, it is necessary to design a standardized equipment monitoring system that can adapt to information exchange between distributed manufacturing systems, effectively access heterogeneous equipment in factories, and effectively process heterogeneous data. Summary of the invention

[0005] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a monitoring system for heterogeneous equipment in a distributed manufacturing system, so as to solve the problem in the prior art that the distributed manufacturing system restricts the interconnection and information fusion between factory workshops, thereby causing difficulties in the control of the entire production process and it is difficult to effectively and uniformly monitor the key equipment in the factory workshops.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A monitoring system for heterogeneous equipment in a distributed manufacturing system of the present invention comprises: a distributed manufacturing workshop and a cloud service unit, wherein one cloud service unit corresponds to multiple distributed manufacturing workshops;

[0008] The distributed manufacturing workshop includes: heterogeneous data acquisition module and heterogeneous data processing module;

[0009] The heterogeneous data acquisition module is used to collect heterogeneous equipment data and sensor data in a distributed manufacturing system;

[0010] The heterogeneous data processing module is used to process the data collected by the heterogeneous data collection module; it includes: a workshop time series database, a data processing module, and an equipment model building module;

[0011] The workshop time series database is used to store the real-time data collected by the heterogeneous data collection modules;

[0012] The data processing module is used to process the data collected by the heterogeneous data collection module, including data noise reduction, data dimension reduction, abnormal data removal, and missing data filling;

[0013] Equipment model building module, used to model heterogeneous equipment in the workshop according to its basic information, status information and equipment function information;

[0014] The cloud service unit includes: central database, data monitoring and analysis module, and interaction module;

[0015] The central database receives data from different distributed manufacturing workshops; it is divided into a central time series database and a central relational database. The central time series database is used to store historical operation data of heterogeneous equipment in the workshop; the central relational database is used to store modeled data and associate the processed data and sensor data in the workshop with heterogeneous equipment;

[0016] The data monitoring and analysis module is used for real-time data calculation and offline data analysis;

[0017] The real-time data calculation is specifically as follows: the equipment model established by the equipment model building module is used to calculate the environment in the workshop and the status between heterogeneous equipment;

[0018] Offline data analysis specifically includes: analyzing the historical data of heterogeneous devices, using convolutional neural networks to predict the status of heterogeneous devices, and generating corresponding alarm notifications;

[0019] The interaction module is used to provide a human-computer interaction interface.

[0020] Furthermore, the heterogeneous data acquisition module collects data specifically including:

[0021] Use TCP / IP protocol to connect with the control system of heterogeneous equipment, so as to obtain the data released by the heterogeneous equipment, the data released by the heterogeneous equipment includes: current equipment operation status, workpiece processing status;

[0022] The sensor is connected via the ModBus bus or / and the USB bus to collect sensor data, which includes: vibration of the machine tool spindle, temperature signals of the tool and the oil tank, and current and voltage signals of the machine tool.

[0023] Furthermore, the data processing module processes the data specifically including: using the mean interpolation method to fill in the missing values, and using the min-max standardization method to eliminate the influence of different attributes of the samples having different magnitudes; (since noisy data is very likely to exist when the workshop data is collected, in order to avoid the influence of noise) using the wavelet threshold algorithm, selecting the wavelet basis to perform N-layer wavelet decomposition on the noisy signal, selecting different threshold quantization processing on coefficients of different layers, filtering out the coefficients of the noise signal, and then reconstructing the quantized wavelet coefficients to obtain an effective signal after noise reduction.

[0024] Furthermore, the basic information of the heterogeneous device includes the name, model, parameters and component information of the device; the device status information refers to the status parameters of the device itself during operation and the operating parameters of the device collected by sensors around the device, including operating status, perception status and spatial status; the device function information refers to the communication capability between devices and the communication capability between devices and the environment, including control function, measurement function, input function and output function.

[0025] Furthermore, the real-time data calculation is specifically: calculating the environment in the workshop and the status between heterogeneous equipment is specifically: using the fault diagnosis model based on deep transfer learning trained with offline data to monitor the real-time data, and the main inputs include the vibration signal of the lathe, current and voltage signals, etc.

[0026] Furthermore, the offline data analysis specifically includes:

[0027] The vibration and temperature data of heterogeneous equipment in offline data and the fault data of heterogeneous equipment are combined into a training set of a neural network. The vibration and temperature data of heterogeneous equipment are used as input, and the fault results of heterogeneous equipment are used as output. A convolutional neural network including an input layer, a convolutional layer, a pooling layer and an output layer is constructed, wherein the convolutional layer and the pooling layer are alternately arranged, and a convolutional layer is connected to a pooling layer, and then a convolutional layer is connected to set several layers. The structural features hidden in the data are extracted through the convolutional layer and the pooling layer, so as to obtain the relationship between the input vibration and temperature data of heterogeneous equipment and the fault of heterogeneous equipment, and realize the diagnosis of the equipment.

[0028] Then use the fault diagnosis model trained with offline data to analyze the real-time data, and use the real-time vibration and temperature data as the input of the fault diagnosis model to determine whether there is a fault in the current equipment.

[0029] Furthermore, the interactive module is used for system settings, equipment ledger, equipment real-time data viewing, equipment operation record viewing, alarm record query, workshop data viewing, and manufacturing system data viewing (workshop data and manufacturing system data viewing facilitates management personnel to control the production situation of the entire industrial chain).

[0030] Beneficial effects of the present invention:

[0031] 1. The present invention realizes the effective access of heterogeneous equipment at the bottom of the workshop, performs corresponding adaptation according to the type of equipment and the type of communication protocol, has a fast configuration speed, and can adopt a one-to-one or one-to-many form to flexibly realize the access of heterogeneous equipment.

[0032] 2. When expanding equipment, there is no need to change the heterogeneous data processing layer, which can achieve plug-and-play of data and improve the access speed of the workshop.

[0033] 3. The present invention adopts a combination of a time series database of a distributed manufacturing workshop and a cloud platform database, which not only ensures low latency and reliability of data transmission within the workshop, but also realizes data backup on the cloud platform and locally, ensuring the security of workshop data. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a structural diagram of the system of the present invention. DETAILED DESCRIPTION

[0035] In order to facilitate the understanding of those skilled in the art, the present invention is further described below in conjunction with embodiments and drawings. The contents mentioned in the implementation modes are not intended to limit the present invention.

[0036] Reference Figure 1 As shown, a monitoring system for heterogeneous equipment in a distributed manufacturing system of the present invention includes: a distributed manufacturing workshop and a cloud service unit, and one cloud service unit corresponds to multiple distributed manufacturing workshops;

[0037] The distributed manufacturing workshop includes: heterogeneous data acquisition module and heterogeneous data processing module;

[0038] The heterogeneous data acquisition module is used to collect heterogeneous equipment data and sensor data in a distributed manufacturing system; specifically includes:

[0039] Use TCP / IP protocol to connect with the control system of heterogeneous equipment to obtain the data released by the heterogeneous equipment, including: current equipment operation status, workpiece processing status (such as lathe spindle speed and various alarm information);

[0040] The sensor is connected via the ModBus bus or / and the USB bus to collect sensor data, which includes: vibration of the machine tool spindle, temperature signals of the tool and the oil tank, and current and voltage signals of the machine tool.

[0041] The collected equipment includes: lathes, milling machines, manipulators and AGVs; the collected data is the operation data of the equipment. For lathes and milling machines, the spindle speed, tool movement position, feed speed, tool number, whether there is a fault, etc. For AGV, the main information is the current position information and movement speed, etc. For manipulators, the main information is the current movement path, the real-time position and posture of each axis, etc.

[0042] The heterogeneous data processing module is used to process the data collected by the heterogeneous data collection module; it includes: a workshop time series database, a data processing module, and an equipment model building module;

[0043] The workshop time series database is used to store the real-time data collected by the heterogeneous data collection modules;

[0044] The data processing module is used to process the data collected by the heterogeneous data collection module, including data noise reduction, data dimension reduction, abnormal data removal, and missing data filling;

[0045] The data processing module processes the data specifically including: using the mean interpolation method to fill in the missing values, and using the min-max standardization method to eliminate the influence of different attributes of the samples having different magnitudes; (since noisy data is very likely to exist when the workshop data is collected, in order to avoid the influence of noise) using the wavelet threshold algorithm, selecting the wavelet basis to perform N-layer wavelet decomposition on the noisy signal, selecting different threshold quantization processing on the coefficients of different layers, filtering out the coefficients of the noise signal, and then reconstructing the quantized wavelet coefficients to obtain an effective signal with noise reduction.

[0046] Equipment model building module, used to model heterogeneous equipment in the workshop according to its basic information, status information and equipment function information;

[0047] The basic information of the heterogeneous device includes the name, model, parameters and component information of the device; the device status information refers to the status parameters of the device itself during operation and the operating parameters of the device collected by sensors around the device, including operating status, perception status and spatial status; the device function information refers to the communication capability between devices and the communication capability between devices and the environment, including control function, measurement function, input function and output function.

[0048] The cloud service unit includes: central database, data monitoring and analysis module, and interaction module;

[0049] The central database receives data from different distributed manufacturing workshops; it is divided into a central time series database and a central relational database. The central time series database is used to store historical operation data of heterogeneous equipment in the workshop (it realizes data backup of workshop operation, ensures the security of workshop data, and can be used as training data for fault diagnosis modules in the application service layer); the central relational database is used to store modeled data, and associates processed data and sensor data in the workshop with heterogeneous equipment;

[0050] The data monitoring and analysis module is used for real-time data calculation and offline data analysis;

[0051] The real-time data calculation is specifically as follows: the equipment model established by the equipment model building module is used to calculate the environment in the workshop and the status between heterogeneous equipment;

[0052] Offline data analysis specifically includes: analyzing the historical data of heterogeneous devices, using convolutional neural networks to predict the status of heterogeneous devices, and generating corresponding alarm notifications;

[0053] The interactive module is used to provide a human-computer interaction interface, which is divided into the Web side, PC side and mobile side.

[0054] Among them, the real-time data calculation is specifically: calculating the environment in the workshop and the status between heterogeneous equipment is specifically: using the fault diagnosis model based on deep transfer learning trained with offline data to monitor the real-time data, and the main inputs include the vibration signal of the lathe, current and voltage signals, etc.

[0055] Wherein, the offline data analysis specifically includes:

[0056] The vibration and temperature data of heterogeneous equipment in offline data and the fault data of heterogeneous equipment are combined into a training set of a neural network. The vibration and temperature data of heterogeneous equipment are used as input, and the fault results of heterogeneous equipment are used as output. A convolutional neural network including an input layer, a convolutional layer, a pooling layer and an output layer is constructed, wherein the convolutional layer and the pooling layer are alternately arranged, and a convolutional layer is connected to a pooling layer, and then a convolutional layer is connected to set several layers. The structural features hidden in the data are extracted through the convolutional layer and the pooling layer, so as to obtain the relationship between the input vibration and temperature data of heterogeneous equipment and the fault of heterogeneous equipment, and realize the diagnosis of the equipment.

[0057] Then use the fault diagnosis model trained with offline data to analyze the real-time data, and use the real-time vibration and temperature data as the input of the fault diagnosis model to determine whether there is a fault in the current equipment.

[0058] In this example, a deep convolutional neural network is used to build a fault diagnosis model; the model includes 3 convolutional layers and 3 fully connected layers;

[0059] The first convolution layer uses a 3*3 convolution kernel, the moving step of the convolution kernel is 1, the shape of the input data is 64*64, the padding method is the same, and the maximum pooling is performed;

[0060] The second convolution layer uses a 3*3 convolution kernel, the moving step of the convolution kernel is 1, the shape of the input data is 32*32, the padding is the same, and the maximum pooling is performed;

[0061] The third convolution layer uses a 3*3 convolution kernel, the moving step of the convolution kernel is 1, the shape of the input data is 16*16, the Padding is the same, and the maximum pooling is performed;

[0062] The fully connected layer is a three-layer neural network with 485, 300, and 100 neurons in each layer, and the activation function is Relu;

[0063] Finally, the fault diagnosis result is output through the Softmax layer.

[0064] The model training parameters include: learning rate is 0.001, batch size is 64, and loss function is root mean square error.

[0065] First, the source domain data samples of working condition 0 and working condition 1 are used for training; then a small amount of target domain unlabeled data samples of working condition 2 are used for training;

[0066] The sample-based migration strategy weighs the samples in the source domain to migrate the samples in the source domain and the target domain (or assigns weights to the samples); samples in the source domain that are similar to those in the target domain are given high weights, and samples that are dissimilar are given low weights.

[0067] Finally, the deep convolution-based fault diagnosis model is optimized through a sample-based transfer learning strategy, so that the fault diagnosis model can quickly acquire the generalization ability of the third working condition fault data.

[0068] The interactive module is used for system settings, equipment ledger, equipment real-time data viewing, equipment operation record viewing, alarm record query, workshop data viewing, manufacturing system data viewing. The workshop data and manufacturing system data viewing facilitate managers to control the production situation of the entire industrial chain.

[0069] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principle of the present invention. These improvements should also be regarded as the protection scope of the present invention.

Claims

1. A monitoring system for heterogeneous equipment in a distributed manufacturing system, characterized in that: include: Distributed manufacturing workshops and cloud service units, one cloud service unit corresponds to multiple distributed manufacturing workshops; The distributed manufacturing workshop includes: heterogeneous data acquisition module and heterogeneous data processing module; The heterogeneous data acquisition module is used to collect heterogeneous equipment data and sensor data in a distributed manufacturing system; The heterogeneous data processing module is used to process the data collected by the heterogeneous data collection module; it includes: a workshop time series database, a data processing module, and an equipment model building module; The workshop time series database is used to store the real-time data collected by the heterogeneous data collection modules; The data processing module is used to process the data collected by the heterogeneous data collection module, including data noise reduction, data dimension reduction, abnormal data removal, and missing data filling; Equipment model building module, used to model heterogeneous equipment in the workshop according to its basic information, status information and equipment function information; The cloud service unit includes: central database, data monitoring and analysis module, and interaction module; The central database receives data from different distributed manufacturing workshops; it is divided into a central time series database and a central relational database. The central time series database is used to store historical operation data of heterogeneous equipment in the workshop; the central relational database is used to store modeled data and associate the processed data and sensor data in the workshop with heterogeneous equipment; The data monitoring and analysis module is used for real-time data calculation and offline data analysis; The real-time data calculation is specifically as follows: the equipment model established by the equipment model building module is used to calculate the environment in the workshop and the status between heterogeneous equipment; Offline data analysis specifically includes: analyzing the historical data of heterogeneous devices, using convolutional neural networks to predict the status of heterogeneous devices, and generating corresponding alarm notifications; The interaction module is used to provide a human-computer interaction interface.

2. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The heterogeneous data acquisition module collects data specifically including: Use TCP / IP protocol to connect with the control system of heterogeneous equipment, so as to obtain the data released by the heterogeneous equipment, the data released by the heterogeneous equipment includes: current equipment operation status, workpiece processing status; The sensor is connected via the ModBus bus or / and the USB bus to collect sensor data, which includes: vibration of the machine tool spindle, temperature signals of the tool and the oil tank, and current and voltage signals of the machine tool.

3. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The data processing module processes the data specifically including: using the mean interpolation method to fill in the missing values, using the min-max standardization method to eliminate the influence of different attributes of the samples having different magnitudes; using the wavelet threshold algorithm, selecting the wavelet basis to perform N-layer wavelet decomposition on the noisy signal, selecting different threshold quantization processing on the coefficients of different layers, filtering out the coefficients of the noise signal, and then reconstructing the quantized wavelet coefficients to obtain an effective signal with noise reduction.

4. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The basic information of the heterogeneous device includes the name, model, parameters and component information of the device; the device status information refers to the status parameters of the device itself during operation and the operating parameters of the device collected by sensors around the device, including operating status, perception status and spatial status; the device function information refers to the communication capability between devices and the communication capability between devices and the environment, including control function, measurement function, input function and output function.

5. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The real-time data calculation is specifically: calculating the environment in the workshop and the status between heterogeneous equipment is specifically: using the fault diagnosis model based on deep transfer learning trained with offline data to monitor the real-time data.

6. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The offline data analysis specifically includes: The vibration and temperature data of heterogeneous equipment in offline data and the fault data of heterogeneous equipment are combined into a training set of a neural network. The vibration and temperature data of heterogeneous equipment are used as input, and the fault results of heterogeneous equipment are used as output. A convolutional neural network including an input layer, a convolutional layer, a pooling layer and an output layer is constructed, wherein the convolutional layer and the pooling layer are alternately arranged, and a convolutional layer is connected to a pooling layer, and then a convolutional layer is connected to set several layers. The structural features hidden in the data are extracted through the convolutional layer and the pooling layer, so as to obtain the relationship between the input vibration and temperature data of heterogeneous equipment and the fault of heterogeneous equipment, and realize the diagnosis of the equipment. Then use the fault diagnosis model trained with offline data to analyze the real-time data, and use the real-time vibration and temperature data as the input of the fault diagnosis model to determine whether there is a fault in the current equipment.

7. The monitoring system for heterogeneous equipment in a distributed manufacturing system according to claim 1, characterized in that: The interactive module is used for system settings, equipment ledger, equipment real-time data viewing, equipment operation record viewing, alarm record query, workshop data viewing, and manufacturing system data viewing.

Images