WO2021248769A1 - 一种机电设备运行状态的监测方法、装置及*** - Google Patents

一种机电设备运行状态的监测方法、装置及*** Download PDF

Info

Publication number
WO2021248769A1
WO2021248769A1 PCT/CN2020/123106 CN2020123106W WO2021248769A1 WO 2021248769 A1 WO2021248769 A1 WO 2021248769A1 CN 2020123106 W CN2020123106 W CN 2020123106W WO 2021248769 A1 WO2021248769 A1 WO 2021248769A1
Authority
WO
WIPO (PCT)
Prior art keywords
electromechanical equipment
data
neural network
training
safety
Prior art date
Application number
PCT/CN2020/123106
Other languages
English (en)
French (fr)
Inventor
程涛
温浩凯
Original Assignee
深圳技术大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳技术大学 filed Critical 深圳技术大学
Publication of WO2021248769A1 publication Critical patent/WO2021248769A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
    • G01D21/00—Measuring or testing not otherwise provided for
    • G01D21/02—Measuring two or more variables by means not covered by a single other subclass
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Definitions

  • the invention relates to the technical field of control engineering, in particular to a method, device and system for monitoring the operating state of electromechanical equipment.
  • Electromechanical equipment can be divided into three categories according to its purpose: industrial electromechanical equipment, information electromechanical equipment, and people's death electromechanical equipment.
  • Industrial electromechanical equipment refers to electromechanical equipment used in production enterprises, such as ordinary lathes, ordinary milling machines, CNC machine tools, etc.
  • information electromechanical equipment refers to electronic mechanical products used for information collection, transmission, and storage processing, such as computer terminals, Communication equipment, fax machines, printers, etc.
  • people's livelihood electromechanical equipment refers to electronic and mechanical products used in people's lives, such as VCD, air conditioners, refrigerators, etc.
  • CNC machine tools have been widely used in various production enterprises due to their high degree of automation, easy computer control, high continuity of CNC machining, good CNC machining consistency, suitable for processing complex parts, and easy to establish networked systems. middle.
  • the CNC machine tool is an automated machine tool equipped with a program control system, real-time monitoring of the operating status of electromechanical equipment such as CNC machine tools is a vital task.
  • problems such as low operating efficiency of electromechanical equipment, poor coordination, low accuracy of acquired data, heavy burden on the control center, and relatively backward control system architecture.
  • the technical problem to be solved by the present invention is to provide a method, device and system for monitoring the operating state of electromechanical equipment, aiming to solve the problem of low accuracy of acquired data in the existing method for monitoring the operating state of electromechanical equipment.
  • the first aspect of the embodiments of the present invention provides a method for monitoring the operating state of an electromechanical device.
  • the method includes the following steps:
  • each set of training data includes a variety of training operating parameters related to the operating state of the electromechanical equipment to be monitored
  • the optimized neural network and actual data are used to obtain accurate values of each actual operating parameter in the actual data.
  • the network structure of the preset neural network includes an input layer, a plurality of hidden layers, and an output layer
  • the data flow direction in the preset neural network includes a forward data flow direction and a reverse data flow direction
  • the forward data flow direction is an input layer, multiple hidden layers, and an output layer in sequence
  • the reverse data flow direction is an output layer, multiple hidden layers, and an input layer in sequence
  • the network structure and data of the optimized neural network are The flow direction is consistent with the preset neural network.
  • the training a preset neural network using multiple sets of the training data to obtain an optimized neural network specifically includes:
  • the error is substituted into the preset neural network in the reverse data flow direction, and the weight coefficient of each hidden layer neuron is modified based on the error to obtain the modified weight of each hidden layer neuron. Coefficients, and then turn to substituting multiple sets of the training data into the preset neural network in the forward data flow direction to obtain the output result of each training operating parameter in each set of the training data.
  • the use of the optimized neural network and actual data to obtain the accurate value of each actual operating parameter in the actual data specifically includes:
  • the actual data is substituted into the optimized neural network in the forward data flow direction, and the accurate value of each actual operating parameter in the actual data is obtained.
  • the method before the acquiring actual data related to the operating state of the electromechanical equipment to be monitored, the method further includes:
  • a safe value or safety range corresponding to each safe operation parameter is synthesized.
  • the method further includes:
  • the safety value or safety range of the actual operating parameter is the safety value or safety range corresponding to each safe operating parameter.
  • the method further includes:
  • the safety problem report is sent to the associated platform of the electromechanical equipment to be monitored, so that the associated platform can solve the safety problem existing in the safety problem report.
  • the solution of the security issue in the security issue report by the associated platform specifically includes:
  • the actual operating parameters that exceed the safety value or the safety range in the safety problem report are compared with the preset safety knowledge base, and the comparison result is obtained.
  • the safety knowledge base includes electromechanical equipment experts, books, and all on the Internet. Safety knowledge related to the operating status of the electromechanical equipment to be monitored;
  • a second aspect of the embodiments of the present invention provides a device for monitoring the operating state of electromechanical equipment, the device including:
  • the first acquisition module is used to acquire multiple sets of training data related to the operating state of the electromechanical equipment to be monitored, wherein each set of the training data includes various training operating parameters related to the operating state of the electromechanical equipment to be monitored;
  • the training module is used to train preset neural networks by using multiple sets of the training data to obtain an optimized neural network, where the optimized neural network is used to perform various training operating parameters in each set of training data Perform intra-species fusion respectively to obtain the accurate value of each of the operating parameters for training;
  • the second acquisition module is configured to acquire actual data related to the operating state of the electromechanical equipment to be monitored, where the actual data includes various actual operating parameters related to the operating state of the electromechanical equipment to be monitored;
  • the monitoring module is used to obtain the accurate value of each actual operating parameter in the actual data by using the optimized neural network and actual data.
  • the third aspect of the embodiments of the present invention provides a monitoring system for the operation status of electromechanical equipment, the system includes a cloud platform and a monitoring device for the operation status of the electromechanical equipment as described in the second aspect of the embodiment of the present invention, which is set on the electromechanical equipment to be monitored ,
  • the cloud platform is the associated platform described in the first aspect of the embodiments of the present invention, and the cloud platform and the monitoring device for the operation status of electromechanical equipment are connected through at least one of GPRS, Wifi, 3G, 4G, and 5G. Make a communication connection.
  • the multiple actual operating parameters in the actual data are fused within the species, and the value of each actual operating parameter can be obtained more accurately. This process improves the accuracy of data acquisition and makes electromechanical equipment more efficient.
  • FIG. 1 is a flowchart of a method for monitoring the operating state of electromechanical equipment according to the first embodiment of the present invention
  • FIG. 2 is a flowchart of a method for monitoring the operating state of electromechanical equipment according to a second embodiment of the present invention
  • Fig. 3 is a flow chart of the method for monitoring the operating state of electromechanical equipment provided by the second embodiment of the present invention connected to Fig. 2;
  • FIG. 4 is a schematic diagram of training a preset neural network provided by the second embodiment of the present invention.
  • FIG. 5 is a diagram of the training result of the preset neural network provided by the second embodiment of the present invention.
  • FIG. 6 is a flowchart of a method for an associated platform to solve a security problem in a security problem report provided by the second embodiment of the present invention
  • FIG. 7 is a block diagram of a module of a monitoring device for the operating state of an electromechanical equipment according to a third embodiment of the present invention.
  • FIG. 8 is a block diagram of modules of a monitoring system for the operating state of electromechanical equipment according to a fourth embodiment of the present invention.
  • FIG. 9 is an architecture diagram of the edge-cloud collaboration mode provided by the fourth embodiment of the present invention.
  • FIG. 10 is an architecture diagram of a cloud platform based on an improved Hadoop distributed system infrastructure provided by the fourth embodiment of the present invention.
  • Fig. 11 is a schematic diagram of a fault knowledge base of a CNC machine tool provided by a fourth embodiment of the present invention.
  • the "training data, training operating parameters”, "actual data, actual operating parameters” and “safety data, safe operating parameters” described below can be the same parameter, such as the same temperature, but in the training data,
  • the temperature parameter is used to train the preset neural network; in the actual data and actual operating parameters, the temperature parameter is used in the actual monitoring of the operating status of the electromechanical equipment to be monitored; in the safety data and safe operating parameters , The temperature parameter is used to obtain the temperature safety value or safety range of the electromechanical equipment to be monitored.
  • FIG. 1 is a flowchart of a method for monitoring the operating state of an electromechanical device according to a first embodiment of the present invention.
  • the method for monitoring the operating status of electromechanical equipment includes the following steps:
  • acquiring multiple sets of training data/actual data related to the operating state of the electromechanical equipment to be monitored is implemented based on multiple sensors set on the electromechanical equipment to be monitored, and The multiple sensors on electromechanical equipment are divided into 3 types of combinations:
  • the same detection parameters such as temperature, light intensity, etc.
  • the same sensor structure and a combination of multiple sensors located at different positions of the electromechanical equipment to be monitored.
  • three photosensitive sensors are set at different positions of the electromechanical equipment to be monitored.
  • a combination of multiple sensors with the same detection parameters, different sensor structures, and the same operating parameters (such as position changes, operating safety factors, etc.) related to the operating state of the electromechanical equipment to be monitored for example, in the electromechanical equipment to be monitored
  • Displacement sensors and ultrasonic sensors are set up in some locations to detect position changes
  • a combination of multiple sensors with different detection parameters, different sensor structures, and the same operating parameters that are related to the operating state of the electromechanical equipment to be monitored are used to detect the operating safety factor, or temperature sensors and thermal sensors are set up at certain locations of the electromechanical equipment to be monitored to detect temperature.
  • a preset neural network is trained using multiple sets of training data related to the operating state of the electromechanical equipment to be monitored to obtain an optimized Neural network; Secondly, the optimized neural network is used to perform intra-species fusion of various actual operating parameters in the actual data related to the operating state of the electromechanical equipment to be monitored, so as to obtain the value of each actual operating parameter more accurately. This process improves the accuracy of data acquisition and makes electromechanical equipment more efficient.
  • the combination of multiple sensors with the same detection parameter, the same sensor structure, and different positions of the electromechanical equipment to be monitored avoids the result deviation caused by the single detection area of the sensor; it has the same detection parameter, different sensor structures,
  • the combination of multiple sensors that detect the same operating parameter related to the operating state of the electromechanical equipment to be monitored avoids the result deviation caused by the large detection error of a single sensor; it has different detection parameters, different sensor structures, and detection and electromechanical equipment to be monitored.
  • the combination of multiple sensors with the same operating parameter related to the operating state of the equipment cuts into the problem from different angles, ensuring the accuracy of the results.
  • the third aspect when acquiring multiple sets of training data/actual data related to the operating state of the electromechanical equipment to be monitored, three types of sensor combinations are set on the electromechanical equipment to be monitored, which in a true sense realizes the multiplicity of electromechanical equipment to be monitored. Source information collection further improves the accuracy of data acquisition.
  • Figure 2 is a flowchart of a method for monitoring the operating state of electromechanical equipment according to a second embodiment of the present invention
  • Figure 3 is a second embodiment of the present invention connected to Figure 2
  • the embodiment provides a flowchart of a method for monitoring the operating status of an electromechanical device.
  • FIG. 4 is a schematic diagram of the training of a preset neural network provided by the second embodiment of the present invention
  • FIG. 5 is a preset neural network provided by the second embodiment of the present invention.
  • FIG. 6 is a flowchart of the method for solving the security problem in the security problem report provided by the associated platform according to the second embodiment of the present invention.
  • the network structure of the preset neural network includes an input layer, a plurality of hidden layers, and an output layer
  • the data flow direction in the preset neural network includes a forward data flow direction and a reverse data flow direction.
  • the forward data flow direction is Input layer, multiple hidden layers and output layer
  • the reverse data flow direction is output layer, multiple hidden layers and input layer in turn.
  • the network structure and data flow direction of the optimized neural network are consistent with the preset neural network.
  • the state of neurons in any hidden layer will only affect the state of neurons in the next hidden layer.
  • step S12 includes:
  • S122 Determine whether the output result of each training operating parameter in each group of training data is consistent with the expected result of each training operating parameter, and if so, obtain the optimized neural network;
  • the “box” in Figure 4 represents the input layer or the output layer, and the “circle” represents the hidden layer.
  • the data of the collision sensor, sound sensor, vibration sensor and flame sensor located on the electromechanical equipment to be monitored As the input value of the preset neural network, the first hidden layer in the middle accepts data from 4 kinds of sensors, and the input is passed to the second hidden layer after weight calculation and activation function processing.
  • the second hidden layer is 3 Input and 1 output, the second hidden layer inputs the data of the first hidden layer and after the weight calculation and activation function processing of the second hidden layer, the output layer outputs the operating safety factor. At this time, the output Compare the operating safety factor of the output with the expected coefficient.
  • the preset neural network After updating the weight coefficients of the middle two hidden layers at the same time, the data of the 4 kinds of sensors are propagated in the preset neural network in the forward direction of the data flow. This is a loop, and finally the value of each hidden layer The weight coefficient is optimized to obtain an optimized neural network.
  • the preset neural network After training the preset neural network, after 469 iterations, the accuracy of the results output by the preset neural network has reached 98.85%. At this time, the preset neural network It is an optimized neural network.
  • step S14 includes:
  • the weight coefficients of each hidden layer in the optimized neural network are already optimal values, so at this time, the actual data that is substituted into the optimized neural network can be obtained by the optimized neural network. The exact value of each actual operating parameter in.
  • the method includes:
  • the optimized neural network since the weight coefficients of each hidden layer in the optimized neural network are already optimal values, the optimized neural network is used to obtain the safety value or safety value corresponding to each safe operating parameter.
  • the range has a high degree of accuracy.
  • step S14 the method includes:
  • the safety value or safety range of each actual operating parameter is the safety value or safety range corresponding to each safe operating parameter in step S23.
  • the associated platform to resolve the security issues in the security issue report includes the following steps:
  • the safety knowledge base includes electromechanical equipment experts, books, all and waiting on the Internet Safety knowledge related to monitoring the operating status of electromechanical equipment;
  • the associated platform is the control center.
  • multiple sets of training data are propagated in a forward data flow direction in a preset neural network, and each set of training data is propagated in a reverse data flow direction.
  • the error between the output result of each training operating parameter and the expected result of each training operating parameter is continuously looped to update the weight coefficients of multiple hidden layers in the preset neural network, so that multiple hidden layers
  • the weighting coefficient of is optimized, and an optimized neural network is generated, so that when the optimized neural network is used to process data such as actual data and safety data, the output results are more accurate.
  • the electromechanical equipment to be monitored can solve the safety problem in the safety problem report by itself
  • the electromechanical equipment to be monitored can solve the safety problem in the safety problem report by itself
  • the security issue report is sent to the associated platform (control center) of the electromechanical equipment to be monitored, so that the associated platform can solve the security issue in the security issue report.
  • the third aspect is to compare the actual operating parameters that exceed the safety value or the safety range in the safety problem report with the preset safety knowledge base to obtain the comparison result, and obtain a solution based on the comparison result, so that the maintenance personnel can improve Quickly repair the failure of the electromechanical equipment to be monitored.
  • FIG. 7 is a block diagram of a module of an apparatus for monitoring the operating state of an electromechanical equipment according to a third embodiment of the present invention.
  • the apparatus 100 for monitoring the operating state of electromechanical equipment provided by the third embodiment of the present invention includes:
  • the first acquisition module 101 is configured to acquire multiple sets of training data related to the operating state of the electromechanical equipment to be monitored, wherein each set of training data includes various training operating parameters related to the operating state of the electromechanical equipment to be monitored;
  • the training module 102 is used to train a preset neural network using multiple sets of training data to obtain an optimized neural network, where the optimized neural network is used to separately seed a variety of training operating parameters in each set of training data Internal fusion to obtain the accurate value of each training operating parameter;
  • the second acquisition module 103 is configured to acquire actual data related to the operating state of the electromechanical device to be monitored, where the actual data includes various actual operating parameters related to the operating state of the electromechanical device to be monitored;
  • the monitoring module 104 is used to obtain the accurate value of each actual operating parameter in the actual data by using the optimized neural network and actual data.
  • FIG. 8 is a module block diagram of a system for monitoring the operation status of electromechanical equipment according to a fourth embodiment of the present invention
  • FIG. 9 is a side-cloud provided by the fourth embodiment of the present invention.
  • FIG. 10 is an architecture diagram of a cloud platform based on an improved Hadoop distributed system infrastructure provided by the fourth embodiment of the present invention.
  • FIG. 11 is a schematic diagram of a CNC machine tool fault knowledge base provided by the fourth embodiment of the present invention.
  • the monitoring system 200 for the operation status of electromechanical equipment provided by the fourth embodiment of the present invention includes a cloud platform 201 and the monitoring of the operation status of the electromechanical equipment provided on the electromechanical equipment to be monitored as provided by the third embodiment of the present invention.
  • the monitoring device 100 for the operating state of the electromechanical device to be monitored is equivalent to an edge computing node.
  • the cloud platform 201 is composed of SaaS, PaaS, and IaaS.
  • SaaS is composed of Flex, CSS, HTML, etc.
  • PaaS is composed of parallel computing, distributed cache, test environment, resource pool, etc.
  • IaaS is composed of storage equipment, virtualization, network equipment, and servers.
  • the cloud platform 201 also has the functions of resource integration and scheduling, data processing and analysis, facility management and monitoring, and data security storage.
  • the collaborative work between multiple edge computing nodes and the cloud platform 201 constitutes an edge-cloud collaborative mode.
  • the cloud platform 201 and the device 100 for monitoring the operation status of the electromechanical equipment are connected wirelessly through at least one of GPRS, Wifi, 3G, 4G, and 5G, and the data information is transmitted through wireless transmission. Pass it to the cloud platform 201 for big data integration in the cloud.
  • the software can be classified according to module functions for flexible combination; on the other hand, it also has real-time network data processing and man-machine interface interactive control functions.
  • private network communication, wireless image transmission, computer network and multimedia technology can be used to establish a monitoring command platform; on the other hand, private network communication, wireless image transmission and other methods can be used to establish mobile command.
  • the communication platform provides a multi-functional command and dispatch platform and a remote decision-making platform.
  • the software system can adopt C/S architecture and modular design, and is composed of functional modules such as scheduling, GPS, short message, reporting, and track playback.
  • the cloud platform 201 is provided with a security knowledge base subsystem, which includes a knowledge base (for example, the CNC machine tool fault knowledge base in FIG. 11), a database, and an inference engine.
  • the knowledge base includes security knowledge proposed by machine tool experts, books, networks and other things with professional decision-making and judgment. It often uses multiple rules to standardize operations and adds confidence factors to improve accuracy;
  • the database can store factual data , Consists of dynamic and static databases.
  • the static database stores parameters that are not greatly transformed, such as the size of the machine tool, the range of motion of the robotic arm, etc.
  • the dynamic database stores various parameters in the operation of the machine tool, such as the current running speed of the motor and the current robotic arm.
  • the position and the current temperature of the machine tool are all important components in decision-making; the inference engine is responsible for inferring certain conclusions based on the input data parameters using the relevant knowledge content of the knowledge base, which includes forward, reverse and mixed reasoning, and the inference engine
  • the performance and structure of is generally related to the representation method of knowledge, and has nothing to do with the content of knowledge.
  • the Hadoop architecture has the characteristics of high scalability, high efficiency, and high fault tolerance, it is very suitable for building an edge-cloud collaborative CNC machine tool system, so the cloud platform 201 can It is a cloud platform 201 based on an improved Hadoop distributed system infrastructure.
  • the cloud platform 201 uses Linux as the operating system, HDFS and MapReduce as the main cores, database query and programming languages SQL and EIL, and is equipped with data interfaces and wireless communications to realize machine tool facility management, edge device management, data management, system management and User management functions.
  • the facility management function of the machine tool is mainly to monitor the running status of the related software and hardware of the machine tool, and make relevant decisions by calling the safety knowledge base when a fault occurs to ensure the normal operation of the machine tool.
  • Edge equipment management mainly obtains the operating status of the machine tool supporting computer and each sensor on the machine tool through real-time monitoring of each edge node to ensure the normal operation of the edge node.
  • Data management is mainly responsible for the functions of machine tool and edge equipment data acquisition, neural network data processing, and knowledge base data analysis to ensure that instructions are issued correctly and quickly, and it also has the function of data storage and recording.
  • System management monitors the operating status of each subsystem of the entire cloud platform to ensure the safe and efficient operation of the cloud platform.
  • the user module is responsible for giving administrators and operators different permissions, monitoring and saving operation records, and ensuring the security of the operating system.
  • the collaborative work between multiple edge computing nodes and the cloud platform constitutes an edge-cloud collaboration mode.
  • Edge computing nodes can be spread over most areas of electromechanical equipment), quickly obtain the overall data of electromechanical equipment, taking into account data interaction and data set processing functions, making the entire electromechanical equipment system operate flexibly, and its own computing power can achieve nerves
  • the network data fusion function reduces the computing burden of the cloud platform and improves the work efficiency of the entire system; second, it utilizes the powerful computing power, big data processing and storage functions of the cloud platform, and is equipped with a deep learning system based on the services of hardware resources and software resources.
  • the cloud platform and the monitoring device for the operation status of the electromechanical equipment are connected wirelessly through at least one of GPRS, Wifi, 3G, 4G, and 5G, which can solve the problem of the mechanical arms, cameras, and motors in the electromechanical equipment. Data transmission problem.
  • the software is classified according to module functions and combined flexibly, which solves various interconnection problems in the integration of multi-standard emergency communication information systems and realizes integrated communication.
  • the software has real-time network data processing and man-machine interface interactive control functions, which can realize real-time monitoring and management of network information, and provide users with a reliable man-machine interface to meet the requirements of information processing for data exchange control, computing performance, and Graphic processing and display, as well as corresponding command and control requirements, provide support for the realization of intelligent human-computer interaction for the emergency command system.
  • a monitoring command platform can be established by various means such as private network communication, wireless image transmission, computer network, and multimedia technology, so that on-site information can be transmitted back to the cloud platform in the form of images and sounds.
  • the software system adopts C/S architecture and modular design, and is composed of functional modules such as scheduling, GPS, short message, reporting, and trajectory playback, which can meet various deployment requirements.
  • the cloud platform is designed as a cloud platform based on the improved Hadoop distributed system infrastructure, equipped with machine tool facility management, edge device management, data management, system management, and user management functions to ensure the normal operation of electromechanical equipment and cloud platforms, The edge nodes work normally, the instructions are correct and quickly issued, and the operating system is safe. At the same time, it also has the function of data storage and recording.
  • the knowledge base often uses multiple rules to standardize operations, and a confidence factor is added to improve the accuracy, which improves the reasoning ability of the secure knowledge base subsystem.
  • the performance of the inference engine is related to the structure and the representation method of knowledge, and has nothing to do with the content of the knowledge, which helps to ensure the independence of the inference engine and the knowledge base, and improves the flexibility of reasoning.
  • the method, device and system for monitoring the operating status of electromechanical equipment provided by the present invention have the following beneficial effects:
  • the multiple actual operating parameters in the actual data are fused within the species, and the value of each actual operating parameter can be obtained more accurately. This process improves the accuracy of data acquisition and makes electromechanical equipment more efficient.
  • the steps of the method or algorithm described in combination with the embodiments disclosed herein can be directly implemented by hardware, a software module executed by a processor, or a combination of the two.
  • the software module can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or all areas in the technical field. Any other known storage media.
  • the computer can be implemented in whole or in part by software, hardware, firmware, or any combination thereof.
  • software it can be implemented in the form of a computer program product in whole or in part.
  • the computer program product includes one or more computer instructions.
  • the computer program instructions When the computer program instructions are loaded and executed on the computer, the procedures or functions according to the present invention are generated in whole or in part.
  • the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • computer instructions can be transmitted from a website, computer, server, or data center through a cable (such as Coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) to transmit to another website site, computer, server or data center.
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk).

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

一种机电设备运行状态的监测方法、装置（100）及***（200），其中，该方法包括获取多组与待监测机电设备的运行状态相关的训练数据（S11）；利用多组训练数据对预置的神经网络进行训练，得到优化的神经网络（S12）；获取与待监测机电设备的运行状态相关的实际数据（S13）；利用优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值（S14）。该方法利用优化的神经网络对与待监测机电设备的运行状态相关的实际数据中的多种实际运行参数进行种内融合，更加精确地得出每种实际运行参数的值，不仅提升了获取数据的精确度，而且使得机电设备具有较高的工作效率。

Description

一种机电设备运行状态的监测方法、装置及*** 技术领域
本发明涉及控制工程技术领域，尤其是指一种机电设备运行状态的监测方法、装置及***。
 背景技术
随着时代的不断发展，工业化水平的不断提高，机电设备的种类也越来越多。机电设备按照用途可分为三大类：产业类机电设备、信息类机电设备、民生类机电设备。产业类机电设备是指用于生产企业的机电设备，例如普通车床、普通铣床、数控机床等；信息类机电设备是指用于信息的采集、传输和存储处理的电子机械产品，例如计算机终端、通讯设备、传真机、打印机等；民生类机电设备是指用于人民生活领域的电子机械产品，例如VCD、空调、电冰箱等。
近些年来，数控机床凭借其自动化程度高、易实现计算机控制、数控加工连续性高、数控加工一致性好、适合加工复杂零件、便于建立网络化***等特性，被广泛应用于各种生产企业中。而由于数控机床是一种装有程序控制***的自动化机床，故对像数控机床这样的机电设备的运行状态进行实时监测是一项至关重要的工作。目前，在对机电设备的运行状态进行监测时，存在机电设备的运行效率较低、协同性较差、获取的数据精确度不高、控制中心负担较大、控制***架构相对落后等问题。
因此，有必要对上述机电设备运行状态的监测方法进行改进。
 技术问题
本发明所要解决的技术问题是：提供一种机电设备运行状态的监测方法、装置及***，旨在解决在现有的机电设备运行状态的监测方法中，获取的数据精确度不高的问题。
 技术解决方案
为了解决上述技术问题，本发明采用的技术方案为：
本发明实施例第一方面提供了一种机电设备运行状态的监测方法，该方法包括如下步骤：
获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组所述训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，所述优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种所述训练用运行参数的准确值；
获取与所述待监测机电设备的运行状态相关的实际数据，其中，所述实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
在一些实施方案中，所述预置的神经网络的网络结构包括输入层、多个隐藏层及输出层，所述预置的神经网络中的数据流向包括正向数据流向及反向数据流向，所述正向数据流向依次为输入层、多个隐藏层及输出层，所述反向数据流向依次为输出层、多个隐藏层及输入层，且所述优化的神经网络的网络结构、数据流向均与预置的神经网络一致。
在一些实施方案中，所述利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，具体包括：
将多组所述训练数据分别以正向数据流向代入至预置的神经网络，得到每组所述训练数据中每种训练用运行参数的输出结果；
分别判断每组所述训练数据中每种训练用运行参数的输出结果是否与每种训练用运行参数的期望结果一致，若是，则得到所述优化的神经网络；
若否，则获取每组所述训练数据中每种训练用运行参数的输出结果与每种训练用运行参数的期望结果之间的误差；
将所述误差以反向数据流向代入至预置的神经网络，并基于所述误差分别对每一隐藏层的神经元的权系数进行修改，得到修改后的每一隐藏层的神经元的权系数，而后转到将多组所述训练数据分别以正向数据流向代入至预置的神经网络，得到每组所述训练数据中每种训练用运行参数的输出结果。
在一些实施方案中，所述利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值，具体包括：
将所述实际数据以正向数据流向代入至优化的神经网络，得到所述实际数据中每种实际运行参数的准确值。
在一些实施方案中，所述获取与所述待监测机电设备的运行状态相关的实际数据之前，还包括：
获取多组与待监测机电设备的运行状态相关的安全数据，其中，每组所述安全数据均包括与待监测机电设备的运行状态相关的多种安全运行参数；
将多组所述安全数据分别以正向数据流向代入至优化的神经网络，得到每组所述安全数据中每种安全运行参数的输出结果；
根据每组所述安全数据中每种安全运行参数的输出结果，综合得到与每种所述安全运行参数相对应的安全值或安全范围。
在一些实施方案中，所述利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值之后，还包括：
分别判断每种所述实际运行参数的准确值是否对应超过每种实际运行参数的安全值或安全范围，并根据超过所述安全值或安全范围的实际运行参数生成安全问题报告，其中，每种所述实际运行参数的安全值或安全范围即为与每种安全运行参数相对应的安全值或安全范围。
在一些实施方案中，所述根据超过所述安全值或安全范围的实际运行参数生成安全问题报告之后，还包括：
根据所述安全问题报告判断待监测机电设备是否能够自行解决安全问题报告中存在的安全问题；
若否，则将所述安全问题报告发送至待监测机电设备的关联平台，以由所述关联平台解决安全问题报告中存在的安全问题。
在一些实施方案中，所述由所述关联平台解决安全问题报告中存在的安全问题，具体包括：
将所述安全问题报告中超过安全值或安全范围的实际运行参数与预置的安全知识库进行比对，得到比对结果，其中，所述安全知识库包括机电设备专家、书籍、网络上所有与待监测机电设备的运行状态相关的安全知识；
对所述比对结果进行分析，得出解决方案。
本发明实施例第二方面提供了一种机电设备运行状态的监测装置，该装置包括：
第一获取模块，用于获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组所述训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
训练模块，用于利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，所述优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种所述训练用运行参数的准确值；
第二获取模块，用于获取与所述待监测机电设备的运行状态相关的实际数据，其中，所述实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
监测模块，用于利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
本发明实施例第三方面提供了一种机电设备运行状态的监测***，该***包括云平台及设置在待监测机电设备上如本发明实施例第二方面所述的机电设备运行状态的监测装置，所述云平台即为本发明实施例第一方面所述的关联平台，所述云平台与机电设备运行状态的监测装置之间通过GPRS、Wifi、3G、4G、5G中的至少一种方式进行通信连接。
 有益效果
从上述描述可知，与现有技术相比，本发明的有益效果在于：
首先，利用多组与待监测机电设备的运行状态相关的训练数据对预置的神经网络进行训练，以得到优化的神经网络；其次，利用优化的神经网络对与待监测机电设备的运行状态相关的实际数据中的多种实际运行参数进行种内融合，更加精确地得出每种实际运行参数的值。此过程提升了获取数据的精确度，使得机电设备具有较高的工作效率。
 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1为本发明第一实施例提供的机电设备运行状态的监测方法的流程图；
图2为本发明第二实施例提供的机电设备运行状态的监测方法的流程图；
图3为连接图2的本发明第二实施例提供的机电设备运行状态的监测方法的流程图；
图4为本发明第二实施例提供的预置的神经网络的训练示意图；
图5为本发明第二实施例提供的预置的神经网络的训练结果图；
图6为本发明第二实施例提供的关联平台解决安全问题报告中存在的安全问题的方法的流程图；
图7为本发明第三实施例提供的机电设备运行状态的监测装置的模块方框图；
图8为本发明第四实施例提供的机电设备运行状态的监测***的模块方框图；
图9为本发明第四实施例提供的边-云协同模式架构图；
图10为本发明第四实施例提供的基于改进Hadoop分布式***基础架构的云平台架构图；
图11为本发明第四实施例提供的数控机床故障知识库的示意图。
 本发明的实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。此外，下面所描述的本发明的各个实施方式中所涉及到的技术特征只要彼此之间未构成冲突就可以相互组合。
下文所描述的“训练数据、训练用运行参数”、“实际数据、实际运行参数”以及“安全数据、安全运行参数”之间可以为同一种参数，比如同为温度，只不过在训练数据、训练用运行参数中，该温度参数用于训练预置的神经网络；在实际数据、实际运行参数中，该温度参数用于待监测机电设备的运行状态的实际监测；在安全数据、安全运行参数中，该温度参数用于求取待监测机电设备的温度安全值或安全范围。
请参阅图1，图1为本发明第一实施例提供的机电设备运行状态的监测方法的流程图。
如图1所示，本发明第一实施例提供的机电设备运行状态的监测方法包括如下步骤：
S11、获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
S12、利用多组训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种训练用运行参数的准确值；
S13、获取与待监测机电设备的运行状态相关的实际数据，其中，实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
S14、利用优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
具体的，于本实施例中，获取多组与待监测机电设备的运行状态相关的训练数据/实际数据是基于设置在待监测机电设备上的多种多个传感器实现的，且设置在待监测机电设备上的多种多个传感器分为3种组合类型：
其一，具有同一种检测参数（如温度、光强等）、同种传感器结构、位于待监测机电设备不同位置的多个传感器组合，例如，在待监测机电设备的不同位置设置3个光敏传感器以检测光强；
其二，具有同一种检测参数、不同传感器结构、检测与待监测机电设备的运行状态相关的同一运行参数（如位置变化、运行安全系数等）的多个传感器组合，例如，在待监测机电设备的某些位置设置位移传感器及超声波传感器以检测位置变化；
其三，具有不同检测参数、不同传感器结构、检测与待监测机电设备的运行状态相关的同一运行参数的多个传感器组合，例如，在待监测机电设备的某些位置设置碰撞传感器、声音传感器、振动传感器及火焰传感器以检测运行安全系数，或者在待监测机电设备的某些位置设置温度传感器及热敏传感器以检测温度。
本发明第一实施例提供的机电设备运行状态的监测方法，第一方面，首先，利用多组与待监测机电设备的运行状态相关的训练数据对预置的神经网络进行训练，以得到优化的神经网络；其次，利用优化的神经网络对与待监测机电设备的运行状态相关的实际数据中的多种实际运行参数进行种内融合，更加精确地得出每种实际运行参数的值。此过程提升了获取数据的精确度，使得机电设备具有较高的工作效率。第二方面，具有同一种检测参数、同种传感器结构、位于待监测机电设备不同位置的多个传感器组合，避免了传感器检测区域单一所导致的结果偏差；具有同一种检测参数、不同传感器结构、检测与待监测机电设备的运行状态相关的同一运行参数的多个传感器组合，避免了由于单种传感器检测误差较大所导致的结果偏差；具有不同检测参数、不同传感器结构、检测与待监测机电设备的运行状态相关的同一运行参数的多个传感器组合，从不同的角度切入问题，保证了结果的准确性。第三方面，获取多组与待监测机电设备的运行状态相关的训练数据/实际数据时，在待监测机电设备上设置了3种类型的传感器组合，真正意义上实现了对待监测机电设备的多源信息采集，进一步提升了获取数据的精确度。
请参阅图2、图3、图4、图5以及图6，图2为本发明第二实施例提供的机电设备运行状态的监测方法的流程图，图3为连接图2的本发明第二实施例提供的机电设备运行状态的监测方法的流程图，图4为本发明第二实施例提供的预置的神经网络的训练示意图，图5为本发明第二实施例提供的预置的神经网络的训练结果图，图6为本发明第二实施例提供的关联平台解决安全问题报告中存在的安全问题的方法的流程图。
以本发明第一实施例提供的机电设备运行状态的监测方法为基础，在本发明第二实施例中：
进一步地，预置的神经网络的网络结构包括输入层、多个隐藏层及输出层，且预置的神经网络中的数据流向包括正向数据流向及反向数据流向，正向数据流向依次为输入层、多个隐藏层及输出层，反向数据流向依次为输出层、多个隐藏层及输入层。
需要说明的是，于本实施例中，优化的神经网络的网络结构、数据流向均与预置的神经网络一致。另外，当数据流经预置的神经网络中的多个隐藏层时，任一隐藏层的神经元的状态只会对下一隐藏层的神经元的状态产生影响。
进一步地，如图2所示，步骤S12包括：
S121、将多组训练数据分别以正向数据流向代入至预置的神经网络，得到每组训练数据中每种训练用运行参数的输出结果；
S122、分别判断每组训练数据中每种训练用运行参数的输出结果是否与每种训练用运行参数的期望结果一致，若是，则得到所述优化的神经网络；
S123、若否，则获取每组训练数据中每种所述训练用运行参数的输出结果与每种训练用运行参数的期望结果之间的误差；
S124、将误差以反向数据流向代入至预置的神经网络，并基于误差分别对每一隐藏层的神经元的权系数进行修改，得到修改后的每一隐藏层的神经元的权系数，而后转到S121。
为清楚地理解本实施例提供的步骤S121-S124，下面对其进行举例说明：
如图4所示，图4中的“方框”代表输入层或输出层，“圆形”代表隐藏层，将位于待监测机电设备上的碰撞传感器、声音传感器、振动传感器及火焰传感器的数据分别作为预置的神经网络的输入值，中间第一层隐藏层接受4种传感器的数据，经权值计算和激活函数处理后将输入传至第二层隐藏层，第二层隐藏层为3输入及1输出，第二层隐藏层将第一层隐藏层的数据输入并经第二层隐藏层的权值计算和激活函数处理后，由输出层输出运行安全系数，此时，需要对输出的运行安全系数与期望系数进行比对，若比对结果不一致，即未达到迭代次数，则获取输出的运行安全系数与期望系数之间的误差，并将误差在预置的神经网络中进行反向数据流向传播，同时更新中间两层隐藏层的权系数后，再对4种传感器的数据在预置的神经网络中进行正向数据流向传播，以此为循环，最终使得每一隐藏层的权系数达到最优，以得到优化的神经网络。另外，如图5所示，在对预置的神经网络进行训练后，经过469次迭代运算，预置的神经网络所输出的结果的准确率已达到98.85%，此时，预置的神经网络即为优化的神经网络。
进一步地，如图2所示，步骤S14包括：
S141、将实际数据以正向数据流向代入至优化的神经网络，得到实际数据中每种实际运行参数的准确值。
需要说明的是，于本实施例中，优化的神经网络中各隐藏层的权系数已是最优值，故此时被代入至优化的神经网络的实际数据，可经优化的神经网络得到实际数据中每种实际运行参数的准确值。
进一步地，如图2所示，步骤S13之前包括：
S21、获取多组与待监测机电设备的运行状态相关的安全数据，其中，每组安全数据均包括与待监测机电设备的运行状态相关的多种安全运行参数；
S22、将多组安全数据分别以正向数据流向代入至优化的神经网络，得到每组安全数据中每种安全运行参数的输出结果；
S23、根据每组安全数据中每种安全运行参数的输出结果，综合得到与每种安全运行参数相对应的安全值或安全范围。
需要说明的是，于本实施例中，由于优化的神经网络中各隐藏层的权系数已是最优值，所以利用优化的神经网络得到的与每种安全运行参数相对应的安全值或安全范围具有较高的精确度。
进一步地，如图3所示，步骤S14之后包括：
S31、分别判断每种实际运行参数的准确值是否对应超过每种实际运行参数的安全值或安全范围，并根据超过安全值或安全范围的实际运行参数生成安全问题报告。
S32、根据安全问题报告判断待监测机电设备是否能够自行解决安全问题报告中存在的安全问题；
S33、若否，则将安全问题报告发送至待监测机电设备的关联平台，以由关联平台解决安全问题报告中存在的安全问题。
其中，每种实际运行参数的安全值或安全范围即为步骤S23中的与每种安全运行参数相对应的安全值或安全范围。
进一步地，如图6所示，由关联平台解决安全问题报告中存在的安全问题包括如下步骤：
S41、将安全问题报告中超过安全值或安全范围的实际运行参数与预置的安全知识库进行比对，得到比对结果，其中，安全知识库包括机电设备专家、书籍、网络上所有与待监测机电设备的运行状态相关的安全知识；
S42、对比对结果进行分析，得出解决方案。
需要说明的是，于本实施例中，关联平台即为控制中心。
为清楚地理解本发明第二实施例提供的机电设备运行状态的监测方法，下面对该方法的步骤进行完整的说明：
S101、获取多组与待监测机电设备的运行状态相关的训练数据；
S102、将多组训练数据分别以正向数据流向代入至预置的神经网络，得到每组训练数据中每种训练用运行参数的输出结果；
S103、判断每组训练数据中每种训练用运行参数的输出结果是否与每种训练用运行参数的期望结果一致，若是，则转到S105；
S104、若否，则获取每组训练数据中每种训练用运行参数的输出结果与每种训练用运行参数的期望结果之间的误差，将误差以反向数据流向代入至预置的神经网络，并基于误差分别对每一隐藏层的神经元的权系数进行修改，得到修改后的每一隐藏层的神经元的权系数，而后转到S103；
S105、得到优化的神经网络；
S106、获取多组与待监测机电设备的运行状态相关的安全数据；
S107、将多组安全数据分别以正向数据流向代入至优化的神经网络，得到每组安全数据中每种安全运行参数的输出结果；
S108、根据每组安全数据中每种安全运行参数的输出结果，综合得到与每种安全运行参数相对应的安全值或安全范围；
S109、获取与待监测机电设备的运行状态相关的实际数据；
S110、将实际数据以正向数据流向代入至优化的神经网络，得到实际数据中每种实际运行参数的准确值；
S111、判断每种实际运行参数的准确值是否对应超过每种实际运行参数的安全值或安全范围；
S112、根据超过安全值或安全范围的实际运行参数生成安全问题报告；
S113、判断待监测机电设备是否能够自行解决安全问题报告中存在的安全问题；
S114、若否，则将安全问题报告发送至待监测机电设备的关联平台。
本发明第二实施例提供的机电设备运行状态的监测方法，第一方面，通过在预置的神经网络中以正向数据流向传播多组训练数据，以反向数据流向传播每组训练数据中每种训练用运行参数的输出结果与每种训练用运行参数的期望结果之间的误差的方式，不断地循环以更新预置的神经网络中多个隐藏层的权系数，使多个隐藏层的权系数达到最优，生成优化的神经网络，使得在利用优化的神经网络处理诸如实际数据、安全数据等数据时，输出的结果更加精准。第二方面，若待监测机电设备能够自行解决安全问题报告中存在的安全问题，则待监测机电设备自行解决安全问题报告中存在的安全问题；若待监测机电设备不能自行解决安全问题报告中存在的安全问题，则将安全问题报告发送至待监测机电设备的关联平台（控制中心），以由关联平台解决安全问题报告中存在的安全问题，此时，不仅通过待监测机电设备与关联平台的协同工作方式，使得整个***具有较好的协同性，而且关联平台也不必承担所有的处理工作，避免了控制中心负担较大的问题。第三方面，将安全问题报告中超过安全值或安全范围的实际运行参数与预置的安全知识库进行比对，得到比对结果，并根据比对结果得出解决方案，使得维修人员可以更快地修复待监测机电设备的故障。
请参阅图7，图7为本发明第三实施例提供的机电设备运行状态的监测装置的模块方框图。
如图7所示，与本发明第一实施例提供的机电设备运行状态的监测方法相对应，本发明第三实施例提供的机电设备运行状态的监测装置100包括：
第一获取模块101，用于获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
训练模块102，用于利用多组训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种训练用运行参数的准确值；
第二获取模块103，用于获取与所述待监测机电设备的运行状态相关的实际数据，其中，实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
监测模块104，用于利用优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
请参阅图8、图9、图10以及图11，图8为本发明第四实施例提供的机电设备运行状态的监测***的模块方框图，图9为本发明第四实施例提供的边-云协同模式架构图，图10为本发明第四实施例提供的基于改进Hadoop分布式***基础架构的云平台架构图，图11为本发明第四实施例提供的数控机床故障知识库的示意图。
如图8所示，本发明第四实施例提供的机电设备运行状态的监测***200包括云平台201及设置在待监测机电设备上如本发明第三实施例所提供的机电设备运行状态的监测装置100，其中，云平台201即为本发明第二实施例所提供的关联平台，且云平台201与机电设备运行状态的监测装置100之间通过GPRS、Wifi、3G、4G、5G中的至少一种方式进行通信连接。
具体的，于本实施例中，设置在待监测机电设备上的机电设备运行状态的监测装置100相当于边缘计算节点。云平台201由SaaS、PaaS和IaaS构成，SaaS由Flex、CSS、HTML等组成，PaaS由并行计算、分布式缓存、测试环境、资源池等组成，IaaS由存储设备、虚拟化、网络设备、服务器设置等组成，且云平台201还具有资源整合调度、数据处理分析、设施管理监控和数据安全存储的功能。于其他实施例中，如图9所示，多个边缘计算节点与云平台201之间的协同工作组成了边-云协同模式。
具体的，于本实施例中，云平台201与机电设备运行状态的监测装置100之间通过GPRS、Wifi、3G、4G、5G中的至少一种方式进行无线通信连接，利用无线传输将数据信息传至云平台201，在云端进行大数据融合。而软件一方面可按模块功能分类，进行灵活组合；另一方面还具有实时网络数据处理、人机界面交互控制功能。除此之外，一方面可利用专网通信、无线图像传输、计算机网络以及多媒体技术等多种手段建立监控指挥平台；另一方面可利用专网通信、无线图像传输等多种手段建立移动指挥通信平台，提供多功能的指挥调度平台和远程决策平台。软件***可采用C/S架构与模块化设计，由调度、GPS、短消息、报表和轨迹回放等功能模块组成。
需要说明的是，于本实施例中，云平台201内设有安全知识库子***，该安全知识库子***包括知识库（例如图11中的数控机床故障知识库）、数据库及推理机，其中，知识库包括机床专家、书籍、网络等各种具备专业决策评判的事物提出的安全知识，其常以多种规则来规范操作，并加以置信度因子来提高准确度；数据库能够存储事实数据，由动态与静态数据库组成，静态数据库存储变换不大的参数，例如机床的大小、机械臂的运动范围等等，动态数据库存储机床运行中的各项参数，例如电机当前运行速度、机械臂当前位置、机床当前温度，这些参数均是决策中的重要构成；推理机负责根据输入的数据参数运用知识库的相关知识内容推理出一定的结论，其包含正向、反向和混合推理，推理机的性能与构造一般与知识的表示方法有关，与知识的内容无关。
还需要说明的是，于其他实施例中，如图10所示，由于Hadoop架构拥有高扩展、高效和高容错性的特点，非常适合构建边-云协同的数控机床***，故云平台201可以为基于改进Hadoop分布式***基础架构的云平台201。该云平台201以Linux为操作***，HDFS和MapReduce为主要核心，数据库查询和程序设计语言SQL与EIL，并配备数据接口和无线通信，实现机床设施管理、边缘设备管理、数据管理、***管理及用户管理的功能。机床设施管理功能主要是对机床相关软硬件运行状态进行监控，出现故障时通过调用安全知识库给出相关决策，保证机床正常运行。边缘设备管理主要通过对各个边缘节点的实时监控获取机床配套电脑和机床上各个传感器的运行状态，保证边缘节点正常工作。数据管理主要负责机床、边缘设备数据的获取、神经网络数据处理、知识库数据分析的功能，保证指令正确迅速下达，兼具数据存储记录功能。***管理则监控整个云平台的各个子***运行状况，保证云平台安全高效运行。用户模块负责给予管理员和操作人员不同权限，监控保存操作记录，保证操作***安全。
本发明第四实施例提供的机电设备运行状态的监测***，第一方面，多个边缘计算节点与云平台之间的协同工作组成边-云协同模式，其一，利用边缘计算的灵活性（边缘计算节点可遍布机电设备的大部分区域），快速地获取机电设备的整体数据，兼顾数据交互和数据集处理功能，可使整个机电设备***运作灵活，且其自带的计算力可实现神经网络数据融合功能，减轻云平台运算负担，提高整个***的工作效率；其二，利用云平台的强大算力、大数据处理与存储功能，同时基于硬件资源和软件资源的服务，配备深度学***台与机电设备运行状态的监测装置之间通过GPRS、Wifi、3G、4G、5G中的至少一种方式进行无线通信连接，可解决机电设备中机械臂、相机、电机之间的数据传输问题。第三方面，软件按模块功能分类，进行灵活组合，解决了多制式应急通信信息***集成中的各项互联问题，实现了融合通信。第四方面，软件具有实时网络数据处理、人机界面交互控制功能，可实现对网络信息的实时监控与管理，为用户提供可靠的人机交互界面，满足信息处理对数据交换控制、计算性能、图形处理和显示以及相应的指挥控制要求，为应急指挥***实现智能化人机交互提供支持，同时可通过各车型、各任务规划设计，满足具体指挥和云平台的电源***、音视频控制可视化操作需求。第五方面，利用专网通信、无线图像传输、计算机网络以及多媒体技术等多种手段建立监控指挥平台，可实现现场信息以图像、声音等形式回传到云平台。第六方面，软件***采用C/S架构与模块化设计，由调度、GPS、短消息、报表和轨迹回放等功能模块组成，可以满足各种部署需求。第七方面，云平台设计成基于改进Hadoop分布式***基础架构的云平台，配备机床设施管理、边缘设备管理、数据管理、***管理及用户管理的功能，保证了机电设备及云平台正常运行、边缘节点正常工作、指令正确且迅速下达及操作***的安全，同时还兼具数据存储记录功能。第八方面，知识库常以多种规则来规范操作，并加以置信度因子来提高准确度，提高了安全知识库子***的推理能力。第九方面，推理机的性能与构造与知识的表示方法有关，与知识的内容无关，有利于保证推理机与知识库的独立性，提高了推理的灵活性。
综上所述，本发明提供的机电设备运行状态的监测方法、装置及***，其有益效果在于：
首先，利用多组与待监测机电设备的运行状态相关的训练数据对预置的神经网络进行训练，以得到优化的神经网络；其次，利用优化的神经网络对与待监测机电设备的运行状态相关的实际数据中的多种实际运行参数进行种内融合，更加精确地得出每种实际运行参数的值。此过程提升了获取数据的精确度，使得机电设备具有较高的工作效率。
结合本文中所公开的实施例描述的方法或算法的步骤可以直接用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本发明所述的流程或功能。计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线)或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘Solid  State Disk)等。
需要说明的是，本发明内容中的各个实施例均采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似的部分互相参见即可。对于产品类实施例而言，由于其与方法类实施例相似，所以描述的比较简单，相关之处参见方法类实施例的部分说明即可。
还需要说明的是，在本发明内容中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本发明内容。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本发明内容中所定义的一般原理可以在不脱离本发明内容的精神或范围的情况下，在其它实施例中实现。因此，本发明内容将不会被限制于本发明内容所示的这些实施例，而是要符合与本发明内容所公开的原理和新颖特点相一致的最宽的范围。

Claims (10)

  1. 一种机电设备运行状态的监测方法，其特征在于，包括如下步骤：
    获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组所述训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
    利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，所述优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种所述训练用运行参数的准确值；
    获取与所述待监测机电设备的运行状态相关的实际数据，其中，所述实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
    利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
  2. 如权利要求1所述的机电设备运行状态的监测方法，其特征在于，所述预置的神经网络的网络结构，包括：输入层、多个隐藏层及输出层，所述预置的神经网络中的数据流向包括正向数据流向及反向数据流向，所述正向数据流向依次为输入层、多个隐藏层及输出层，所述反向数据流向依次为输出层、多个隐藏层及输入层，且所述优化的神经网络的网络结构、数据流向均与预置的神经网络一致。
  3. 如权利要求2所述的机电设备运行状态的监测方法，其特征在于，所述利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，具体包括：
    将多组所述训练数据分别以正向数据流向代入至预置的神经网络，得到每组所述训练数据中每种训练用运行参数的输出结果；
    分别判断每组所述训练数据中每种训练用运行参数的输出结果是否与每种训练用运行参数的期望结果一致，若是，则得到所述优化的神经网络；
    若否，则获取每组所述训练数据中每种训练用运行参数的输出结果与每种训练用运行参数的期望结果之间的误差；
    将所述误差以反向数据流向代入至预置的神经网络，并基于所述误差分别对每一隐藏层的神经元的权系数进行修改，得到修改后的每一隐藏层的神经元的权系数，而后转到将多组所述训练数据分别以正向数据流向代入至预置的神经网络，得到每组所述训练数据中每种训练用运行参数的输出结果。
  4. 如权利要求2所述的机电设备运行状态的监测方法，其特征在于，所述利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值，具体包括：将所述实际数据以正向数据流向代入至优化的神经网络，得到所述实际数据中每种实际运行参数的准确值。
  5. 如权利要求2所述的机电设备运行状态的监测方法，其特征在于，所述获取与所述待监测机电设备的运行状态相关的实际数据之前，还包括：
    获取多组与待监测机电设备的运行状态相关的安全数据，其中，每组所述安全数据均包括与待监测机电设备的运行状态相关的多种安全运行参数；
    将多组所述安全数据分别以正向数据流向代入至优化的神经网络，得到每组所述安全数据中每种安全运行参数的输出结果；
    根据每组所述安全数据中每种安全运行参数的输出结果，综合得到与每种所述安全运行参数相对应的安全值或安全范围。
  6. 如权利要求5所述的机电设备运行状态的监测方法，其特征在于，所述利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值之后，还包括：分别判断每种所述实际运行参数的准确值是否对应超过每种实际运行参数的安全值或安全范围，并根据超过所述安全值或安全范围的实际运行参数生成安全问题报告，其中，每种所述实际运行参数的安全值或安全范围即为与每种安全运行参数相对应的安全值或安全范围。
  7. 如权利要求6所述的机电设备运行状态的监测方法，其特征在于，所述根据超过所述安全值或安全范围的实际运行参数生成安全问题报告之后，还包括：
    根据所述安全问题报告判断待监测机电设备是否能够自行解决安全问题报告中存在的安全问题；
    若否，则将所述安全问题报告发送至待监测机电设备的关联平台，以由所述关联平台解决安全问题报告中存在的安全问题。
  8. 如权利要求7所述的机电设备运行状态的监测方法，其特征在于，所述由所述关联平台解决安全问题报告中存在的安全问题，具体包括：
    将所述安全问题报告中超过安全值或安全范围的实际运行参数与预置的安全知识库进行比对，得到比对结果，其中，所述安全知识库包括机电设备专家、书籍、网络上所有与待监测机电设备的运行状态相关的安全知识；
    对所述比对结果进行分析，得出解决方案。
  9. 一种机电设备运行状态的监测装置，其特征在于，包括：
    第一获取模块，用于获取多组与待监测机电设备的运行状态相关的训练数据，其中，每组所述训练数据均包括与待监测机电设备的运行状态相关的多种训练用运行参数；
    训练模块，用于利用多组所述训练数据对预置的神经网络进行训练，得到优化的神经网络，其中，所述优化的神经网络用于对每组训练数据中的多种训练用运行参数分别进行种内融合，得到每种所述训练用运行参数的准确值；
    第二获取模块，用于获取与所述待监测机电设备的运行状态相关的实际数据，其中，所述实际数据包括与待监测机电设备的运行状态相关的多种实际运行参数；
    监测模块，用于利用所述优化的神经网络及实际数据获取实际数据中每种实际运行参数的准确值。
  10. 一种机电设备运行状态的监测***，其特征在于，包括：云平台及设置在待监测机电设备上如权利要求9所述的机电设备运行状态的监测装置，所述云平台即为权利要求7-8任一项所述的关联平台，所述云平台与机电设备运行状态的监测装置之间通过GPRS、Wifi、3G、4G、5G中的至少一种方式进行通信连接。
PCT/CN2020/123106 2020-06-11 2020-10-23 一种机电设备运行状态的监测方法、装置及*** WO2021248769A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010530898.9 2020-06-11
CN202010530898.9A CN111623830A (zh) 2020-06-11 2020-06-11 一种机电设备运行状态的监测方法、装置及***

Publications (1)

Publication Number Publication Date
WO2021248769A1 true WO2021248769A1 (zh) 2021-12-16

Family

ID=72258349

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/123106 WO2021248769A1 (zh) 2020-06-11 2020-10-23 一种机电设备运行状态的监测方法、装置及***

Country Status (2)

Country Link
CN (1) CN111623830A (zh)
WO (1) WO2021248769A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116893614A (zh) * 2023-06-06 2023-10-17 苏州优世达智能科技有限公司 一种基于多传感器融合的水陆两栖无人艇的控制方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111623830A (zh) * 2020-06-11 2020-09-04 深圳技术大学 一种机电设备运行状态的监测方法、装置及***
CN113393143B (zh) * 2021-06-24 2022-06-17 重庆大学 基于信息融合的翅片机加工状态监测方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104085789A (zh) * 2014-05-06 2014-10-08 新乡市起重机厂有限公司 一种起重机运行状态的智能监测方法
CN109357749A (zh) * 2018-09-04 2019-02-19 南京理工大学 一种基于dnn算法的电力设备音频信号分析方法
CN109376620A (zh) * 2018-09-30 2019-02-22 华北电力大学 一种风电机组齿轮箱故障的迁移诊断方法
CN109543814A (zh) * 2018-08-31 2019-03-29 南京理工大学 一种地铁信号***各设备故障预测方法
US20190244504A1 (en) * 2016-10-24 2019-08-08 Hochiki Corporation Fire monitoring system
CN110554657A (zh) * 2019-10-16 2019-12-10 河北工业大学 一种数控机床运行状态健康诊断***及诊断方法
CN111623830A (zh) * 2020-06-11 2020-09-04 深圳技术大学 一种机电设备运行状态的监测方法、装置及***

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0704775A1 (de) * 1994-08-22 1996-04-03 Zellweger Luwa Ag Verfahren und Vorrichtung zur Bestimmung von relevanten Grössen bei einer Verarbeitung von textilen Gebilden
CN101799367B (zh) * 2010-01-27 2011-08-10 北京信息科技大学 一种机电设备神经网络故障趋势预测方法
CN102880115B (zh) * 2012-09-29 2015-08-26 陈广义 一种基于物联网的数控机床远程协作诊断***
CN102937534B (zh) * 2012-11-26 2014-08-27 北京信息科技大学 一种基于组合预测模型对机电设备的故障预测方法
CN107040555A (zh) * 2016-02-03 2017-08-11 上海迪麦科能源科技有限公司 一种基于云端服务器的机电设备远程管理方法
CN106203627A (zh) * 2016-07-08 2016-12-07 中国电子科技集团公司电子科学研究院 一种对网络靶场进行评价的方法
CN106323633A (zh) * 2016-10-28 2017-01-11 华中科技大学 一种基于指令域分析的进给轴装配故障的诊断方法
CN106874581B (zh) * 2016-12-30 2021-03-30 浙江大学 一种基于bp神经网络模型的建筑空调能耗预测方法
CN107168244A (zh) * 2017-07-10 2017-09-15 吉林工程技术师范学院 一种数控机床运行状态在线监测及报警***
US10997429B2 (en) * 2018-04-11 2021-05-04 Micron Technology, Inc. Determining autonomous vehicle status based on mapping of crowdsourced object data
CN109632355B (zh) * 2018-12-20 2021-03-16 广州航天海特***工程有限公司 基于机电设备状态数据漂移的故障预测方法及***
CN109931678B (zh) * 2019-03-13 2020-09-25 中国计量大学 基于深度学习lstm的空调故障诊断方法
CN110119851B (zh) * 2019-05-23 2021-11-09 上海建工四建集团有限公司 一种建筑机电***故障智能预测方法和***
CN110362037A (zh) * 2019-07-05 2019-10-22 南京简睿捷软件开发有限公司 一种用于数控机床的综合维护***平台
CN110361987A (zh) * 2019-08-08 2019-10-22 上海博坤信息技术有限公司 一种基于建筑信息模型的机电设备运行状态模拟仿真方法
CN111007799B (zh) * 2019-12-18 2021-04-27 宁波财经学院 一种基于神经网络的数控装备远程诊断***

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104085789A (zh) * 2014-05-06 2014-10-08 新乡市起重机厂有限公司 一种起重机运行状态的智能监测方法
US20190244504A1 (en) * 2016-10-24 2019-08-08 Hochiki Corporation Fire monitoring system
CN109543814A (zh) * 2018-08-31 2019-03-29 南京理工大学 一种地铁信号***各设备故障预测方法
CN109357749A (zh) * 2018-09-04 2019-02-19 南京理工大学 一种基于dnn算法的电力设备音频信号分析方法
CN109376620A (zh) * 2018-09-30 2019-02-22 华北电力大学 一种风电机组齿轮箱故障的迁移诊断方法
CN110554657A (zh) * 2019-10-16 2019-12-10 河北工业大学 一种数控机床运行状态健康诊断***及诊断方法
CN111623830A (zh) * 2020-06-11 2020-09-04 深圳技术大学 一种机电设备运行状态的监测方法、装置及***

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116893614A (zh) * 2023-06-06 2023-10-17 苏州优世达智能科技有限公司 一种基于多传感器融合的水陆两栖无人艇的控制方法
CN116893614B (zh) * 2023-06-06 2023-12-15 苏州优世达智能科技有限公司 一种基于多传感器融合的水陆两栖无人艇的控制方法

Also Published As

Publication number Publication date
CN111623830A (zh) 2020-09-04

Similar Documents

Publication Publication Date Title
WO2021248769A1 (zh) 一种机电设备运行状态的监测方法、装置及***
US12013842B2 (en) Web services platform with integration and interface of smart entities with enterprise applications
WO2023246568A1 (zh) 一种数字孪生架构、模型评价体系以及模型运行方法
CN102882969B (zh) 一种工矿企业的安全生产云服务平台
CN102932419B (zh) 一种用于面向工矿企业的安全生产云服务平台的数据存储***
CN112600891B (zh) 一种基于信息物理融合的边云协同***及工作方法
CN102917032B (zh) 一种工矿企业的安全生产云服务平台
CN107194565B (zh) 一种基于云决策的电网调度优化方法和***
CN102739420A (zh) 基于数字化的排水管网运营管理与维护的分布式复杂巨***
CN114169579B (zh) 核电工业互联网综合智能平台***
US11595271B2 (en) Digital twin architecture for multi-access edge computing environment
TW201525710A (zh) 適用於電信應用之雲端化服務彈性組裝系統及方法
WO2017173972A1 (zh) 流程跟踪方法及装置、存储介质
KR20220097299A (ko) 인공 지능 기반 실시간 음성인식 원격 협업 방법 및 그 시스템
Wu et al. Enterprise Digital Intelligent Remote Control System Based on Industrial Internet of Things
WO2023103248A1 (zh) 一种设备自动调试的方法、装置、设备、***及存储介质
TW202013304A (zh) 整合ｉｃｔ服務供裝與虛實資源監控之系統
CN116523187A (zh) 一种基于bim的工程进度监控方法及其***
CN114170041A (zh) 应用建筑主题数据建立智慧建筑运维管理***的方法
CN116861708B (zh) 生产设备多维模型构建方法及装置
CN115439015B (zh) 基于数据中台的局域电网数据管理方法、装置及设备
CN114745616B (zh) 一种地下热信息远程监控预警***和方法
TWI764470B (zh) 智慧建築整合管理系統及其管理方法
Mendes et al. The Digitalization of Manufacturing: A Case Study
Cui et al. Large scale MTConnect data collection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20940170

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20940170

Country of ref document: EP

Kind code of ref document: A1