CN114330740A - A federated learning-based training system for manufacturing equipment fault monitoring model - Google Patents
A federated learning-based training system for manufacturing equipment fault monitoring modelDownload PDFInfo
- Publication number
- CN114330740A CN114330740ACN202111549090.6ACN202111549090ACN114330740ACN 114330740 ACN114330740 ACN 114330740ACN 202111549090 ACN202111549090 ACN 202111549090ACN 114330740 ACN114330740 ACN 114330740A
- Authority
- CN
- China
- Prior art keywords
- equipment
- fault monitoring
- monitoring model
- data
- local
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Images
Classifications
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
Description
Translated fromChinese技术领域technical field
本发明属于计算机应用技术领域，具体地说涉及一种基于联邦学习的制造业装备故障监测模型训练系统。The invention belongs to the technical field of computer applications, in particular to a manufacturing equipment fault monitoring model training system based on federated learning.
背景技术Background technique
随着计算机技术的发展，越来越多的技术应用在制造业领域，通过现场采集制造业装备的历史运行状态数据，并筛选出正常运行的历史数据得到正常状态数据集，进一步构建制造业装备的故障监测模型并利用正常状态数据集进行模型训练，最后使得训练后的故障监测模型可以实时从数据库中获取制造业装备的数据并进行故障甄别。在现有技术中，故障监测模型主要利用装备的历史数据进行模型训练，并通过增加训练次数、修改模型参数以提高监测模型的精确率。但是，当装备的数据量较少时，会因缺少数据而导致模型训练的精确率不高，另外，对于具有隐私性的装备数据，无法进行现场集中采集，从而难以获取更多的装备数据。With the development of computer technology, more and more technologies are applied in the manufacturing field. By collecting the historical operating status data of manufacturing equipment on site, and filtering out the normal operating historical data to obtain a normal status data set, the manufacturing equipment can be further constructed. The fault monitoring model is based on the normal state data set for model training, and finally the trained fault monitoring model can obtain the data of manufacturing equipment from the database in real time and perform fault identification. In the prior art, the fault monitoring model mainly uses the historical data of the equipment for model training, and increases the training times and modifies the model parameters to improve the accuracy of the monitoring model. However, when the amount of equipment data is small, the accuracy of model training will be low due to lack of data. In addition, for equipment data with privacy, it is impossible to collect on-site centralized data, so it is difficult to obtain more equipment data.
在保护装备数据隐私的条件下，为了提高故障监测模型的准确率，需要使用大量的数据进行模型训练，但是，当装备的数据涉及用户隐私时，无法获取更多的设备数据进行集中训练模型。目前，在保障制造业装备的数据隐私性和安全性的条件下，对制造业装备数据的获取变得越来越困难，导致现有故障监测模型存在模型精确度较低的问题，这使得上述模型训练方式面临着极大的挑战。Under the condition of protecting the privacy of equipment data, in order to improve the accuracy of the fault monitoring model, a large amount of data needs to be used for model training. However, when the equipment data involves user privacy, more equipment data cannot be obtained to train the model centrally. At present, under the condition of ensuring the data privacy and security of manufacturing equipment, the acquisition of manufacturing equipment data becomes more and more difficult, resulting in the problem of low model accuracy in the existing fault monitoring model, which makes the above Model training methods face great challenges.
因此，现有技术还有待于进一步发展和改进。Therefore, the existing technology still needs to be further developed and improved.
发明内容SUMMARY OF THE INVENTION
针对现有技术的种种不足，为了解决上述问题，现提出一种基于联邦学习的制造业装备故障监测模型训练系统。本发明提供如下技术方案：In view of various deficiencies of the existing technology, in order to solve the above problems, a training system for manufacturing equipment fault monitoring model based on federated learning is proposed. The present invention provides the following technical solutions:
一种基于联邦学习的制造业装备故障监测模型训练系统，包括：A federated learning-based training system for manufacturing equipment fault monitoring model, including:
装备，设置有至少一个，用于收集自身运行状态数据，并处理后形成本地故障监测模型；Equipment, at least one is provided for collecting its own operating status data, and processing it to form a local fault monitoring model;
聚合服务器，设置有一个，用于聚合各装备本体上传的本地故障监测模型参数，生成全局故障监测模型参数并发送给各装备；An aggregation server is provided, which is used to aggregate the local fault monitoring model parameters uploaded by each equipment body, generate global fault monitoring model parameters and send them to each equipment;
所述装备与聚合服务器双向信号连接。The equipment is in a bidirectional signal connection with the aggregation server.
进一步的，装备获取自身的运行状态数据，并对运行状态数据进行预分析；Further, the equipment obtains its own operating status data, and pre-analyzes the operating status data;
各装备基于预分析后的运行状态数据进行各自本地故障检测模型训练，将训练后的本地故障监测模型参数发送给聚合服务器，聚合服务器通过加权平均聚合各本地故障监测模型参数，生成全局故障监测模型参数并发送给各装备，各装备接收后更新本地故障监测模型，继续利用本地运行数据重复训练模型，直至全局故障监测模型收敛，模型训练结束；Each equipment trains its own local fault detection model based on the pre-analyzed operating status data, and sends the trained local fault monitoring model parameters to the aggregation server. The aggregation server aggregates the parameters of each local fault monitoring model through a weighted average to generate a global fault monitoring model. The parameters are sent to each equipment, and each equipment updates the local fault monitoring model after receiving it, and continues to use the local operating data to repeatedly train the model until the global fault monitoring model converges and the model training ends;
将训练好的故障监测模型应用在各装备，对装备的运行状态进行实时监测，并将故障诊断结果实时通知到现场检修人员的手持终端或自动控制设备，并将维修信息及时上传至数据云端，形成智能服务闭环。Apply the trained fault monitoring model to each equipment, monitor the running status of the equipment in real time, notify the fault diagnosis results to the handheld terminal or automatic control equipment of the on-site maintenance personnel in real time, and upload the maintenance information to the data cloud in time. Form a closed loop of intelligent services.
进一步的，将训练后的本地故障监测模型参数通过同态加密后发送给聚合服务器，聚合服务器分别解密各本地故障监测模型参数；Further, the trained local fault monitoring model parameters are sent to the aggregation server through homomorphic encryption, and the aggregation server decrypts the parameters of each local fault monitoring model respectively;
所述同态加密算法包括：每个装备和聚合服务器都具有相同的公私钥，则有F(E(x),E(y))＝E(xΘy)，其中，F为同态加密算法，x和y是明文空间M中的元素，Θ为M上的运算，E(·)是M上密钥空间为k的加密函数，即对数据x和y加密后运算的结果与x和y运算后再加密的结果是相同的。The homomorphic encryption algorithm includes: each equipment and aggregation server have the same public and private keys, then there is F(E(x), E(y))=E(xΘy), where F is a homomorphic encryption algorithm, x and y are the elements in the plaintext space M, Θ is the operation on M, E( ) is the encryption function with the key space k on M, that is, the result of the operation after encrypting the data x and y and the operation on x and y The result of re-encryption is the same.
进一步的，对运行状态数据进行预分析的方法包括：Further, the method for pre-analyzing the operating state data includes:
采集制造业装备的历史运行数据，对所述制造业装备历史运行数据进行预处理，其中，预处理包括对所述装备历史数据进行清洗，剔除异常、缺失的样本数据；Collect historical operation data of manufacturing equipment, and preprocess the historical operation data of manufacturing equipment, wherein the preprocessing includes cleaning the historical data of the equipment, and removing abnormal and missing sample data;
对预处理后的历史运行数据构建装备运行时间序列数据；Construct equipment operation time series data from the preprocessed historical operation data;
对所述装备运行时间序列数据进行归一化处理。Normalize the equipment operation time series data.
进一步的，对预处理后的历史运行数据构建装备运行时间序列数据方法包括：对经过预处理后的历史运行数据按每时段统计，并按时间顺序构建装备运行时间序列数据。Further, the method for constructing equipment operation time series data on the preprocessed historical operation data includes: collecting statistics on the preprocessed historical operation data by time period, and constructing the equipment operation time series data in chronological order.
进一步的，对所述装备运行时间序列数据进行归一化处理方法包括：Further, the normalization processing method for the equipment operation time series data includes:
获取装备运行时间序列数据中的最大值和最小值；Obtain the maximum and minimum values in the equipment operating time series data;
对每个时段的装备运行时间序列数据归一化在0-1之间，归一化公式为
进一步的，聚合服务器通过加权平均聚合各本地故障监测模型参数的方法包括：聚合服务器在接收到所有装备的本地故障监测模型参数后，解密所有本地故障监测模型参数，基于加权平均法聚合所有本地故障监测模型参数，得到全局故障监测模型参数并建立全局故障监测模型；Further, the method for the aggregation server to aggregate the parameters of each local fault monitoring model by weighted average includes: after the aggregation server receives the local fault monitoring model parameters of all equipment, decrypts all the local fault monitoring model parameters, and aggregates all the local faults based on the weighted average method. Monitor model parameters to obtain global fault monitoring model parameters and establish a global fault monitoring model;
其中，聚合本地故障监测模型参数的加权平均法公式为
进一步的，所述全局故障监测模型收敛的判断方法包括：对建立的全局故障监测模型进行检测，通过测试数据计算预测误差值判断全局故障模型是否收敛，若未收敛，聚合服务器将全局故障监测模型参数加密后发送给各装备，各装备解密全局故障监测模型参数并更新本地故障监测模型参数，继续迭代训练，直到全局故障监测模型收敛。Further, the method for judging the convergence of the global fault monitoring model includes: detecting the established global fault monitoring model, and calculating a prediction error value through the test data to determine whether the global fault monitoring model has converged; The parameters are encrypted and sent to each equipment. Each equipment decrypts the parameters of the global fault monitoring model and updates the parameters of the local fault monitoring model, and continues iterative training until the global fault monitoring model converges.
进一步的，将收敛的全局故障监测模型作为最终的制造业装备故障监测模型。Further, the converged global fault monitoring model is used as the final manufacturing equipment fault monitoring model.
进一步的，每个装备利用本地装备运行时间序列数据每次进行200轮本地故障监测模型训练后再发送至聚合服务器。Further, each equipment uses the local equipment running time series data to perform 200 rounds of local fault monitoring model training each time before sending it to the aggregation server.
有益效果：Beneficial effects:
1、将联邦学习用于制造业装备故障监测模型训练方法上，能够在保障装备的数据隐私和安全的条件下训练故障监测模型，提高故障监测模型的精确度；1. Using federated learning in the training method of manufacturing equipment fault monitoring model can train the fault monitoring model under the condition of ensuring the data privacy and security of the equipment, and improve the accuracy of the fault monitoring model;
2、联邦学习可在不获取装备数据的条件下，训练并优化故障监测模型，从而提高模型的精确率，避免调取用户装备数据，进而避免被他人将数据出于商业目的而利用，甚至滥用；2. Federated learning can train and optimize the fault monitoring model without acquiring equipment data, thereby improving the accuracy of the model and avoiding the retrieval of user equipment data, thereby avoiding the data being used or even abused by others for commercial purposes ;
3、将模型训练任务分散到各个设备中进行循环，提高训练效率，降低单设备训练样本数量及复杂度；3. Disperse model training tasks to various devices for circulation, improve training efficiency, and reduce the number and complexity of training samples per device;
4、按照200批次为一组循环训练，提高训练的可信度以及构建速度；4. According to 200 batches as a group of circular training, improve the credibility of training and construction speed;
5、通过同态加密算法对收集和下发的训练模型数据进行加密处理，避免原始数据的泄露，保证数据的安全性。5. Encrypt the collected and distributed training model data through the homomorphic encryption algorithm to avoid the leakage of the original data and ensure the security of the data.
附图说明Description of drawings
图1是本发明具体实施例中一种基于联邦学习的制造业装备故障监测模型训练系统结构示意图；1 is a schematic structural diagram of a federated learning-based manufacturing equipment fault monitoring model training system in a specific embodiment of the present invention;
图2是本发明具体实施例中一种基于联邦学习的制造业装备故障监测模型训练系统流程示意图；2 is a schematic flowchart of a training system for a federated learning-based manufacturing equipment fault monitoring model training system according to a specific embodiment of the present invention;
图3是本发明具体实施例中制造业装备故障监测模型训练流程示意图。3 is a schematic diagram of a training flow of a manufacturing equipment fault monitoring model in a specific embodiment of the present invention.
具体实施方式Detailed ways
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
在本申请实施例中使用的术语是仅仅出于描述特定实施例的目的，而非旨在限制本发明。在本申请实施例中所使用的单数形式的“一种”和“该”也旨在包括多数形式，除非上下文清楚地表示其他含义。The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. As used in the embodiments of this application, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise.
应当理解，本文中使用的术语“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A 和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。It should be understood that the term "and/or" used in this document is only an association relationship to describe associated objects, indicating that there can be three kinds of relationships, for example, A and/or B, which can mean that A exists alone, and A and B exist simultaneously. B, there are three cases of B alone. In addition, the character "/" in this document generally indicates that the related objects are an "or" relationship.
取决于语境，如在此所使用的词语“如果”、“若”可以被解释成为“在……时”或“当……时”或“响应于确定”或“响应于检测”。类似地，取决于语境，短语“如果确定”或“如果检测(陈述的条件或事件)”可以被解释成为“当确定时”或“响应于确定”或“当检测(陈述的条件或事件)时”或“响应于检测(陈述的条件或事件)”。Depending on the context, the words "if", "if" as used herein may be interpreted as "at" or "when" or "in response to determining" or "in response to detecting". Similarly, the phrases "if determined" or "if detected (the stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detected (the stated condition or event)," depending on the context )" or "in response to detection (a stated condition or event)".
还需要说明的是，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商品或者系统不仅包括那些要素，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商品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括要素的商品或者系统中还存在另外的相同要素。It should also be noted that the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a commodity or system comprising a list of elements includes not only those elements, but also includes not explicitly listed other elements, or elements inherent to the commodity or system. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the commodity or system that includes the element.
本发明选取某N个装备的历史运行数据作为实施例，其中每个装备作为一个本地方，如图1所示，共有N个装备和一个聚合服务器，具体流程如下：The present invention selects the historical operation data of a certain N equipment as an embodiment, wherein each equipment is used as a local place, as shown in FIG.
首先，每个装备利用本地装备运行时间序列数据进行200轮本地故障监测模型训练，然后取出本地监测模型参数并经过同态加密发送给聚合服务器，聚合服务器对收到的所有装备的本地故障监测模性参数进行解密，并根据每个装备的本地装备运行时间序列数据量通过加权平均聚合本地故障监测模型参数，获得每轮联邦学习的全局故障监测模型参数，并建立全局故障监测模型并通过测试数据集检测全局故障监测模型的收敛性，若全局故障监测模型未收敛，则将全局故障监测模型参数加密发送给各装备，各装备解密全局故障监测模型参数并更新本地故障监测模型，继续利用本地装备运行时间序列数据进行200轮本地故障监测模型训练，重复以上过程，直至全局故障监测模型收敛。最后，将收敛的全局故障监测模型作为最终的制造业故障监测模型。First, each equipment uses the local equipment running time series data for 200 rounds of local fault monitoring model training, and then takes out the local monitoring model parameters and sends them to the aggregation server through homomorphic encryption. The aggregation server receives the local fault monitoring model of all equipment. According to the local equipment running time series data volume of each equipment, the local fault monitoring model parameters are aggregated by weighted average to obtain the global fault monitoring model parameters of each round of federated learning, and the global fault monitoring model is established to pass the test data. Set to detect the convergence of the global fault monitoring model. If the global fault monitoring model fails to converge, the global fault monitoring model parameters are encrypted and sent to each equipment. Each equipment decrypts the global fault monitoring model parameters and updates the local fault monitoring model, and continues to use the local equipment. Run the time series data for 200 rounds of local fault monitoring model training, and repeat the above process until the global fault monitoring model converges. Finally, the converged global fault monitoring model is used as the final manufacturing fault monitoring model.
如图2所示，本实施例具体方法如下：As shown in Figure 2, the specific method of this embodiment is as follows:
步骤S101，采集制造业装备的历史运行数据，对所述制造业装备历史运行数据进行预处理；Step S101, collect historical operation data of manufacturing equipment, and preprocess the historical operation data of manufacturing equipment;
步骤S102，对预处理后的历史运行数据构建装备运行时间序列数据；Step S102, constructing equipment operation time series data on the preprocessed historical operation data;
步骤S103，对所述装备运行时间序列数据进行归一化处理；Step S103, normalizing the equipment operation time series data;
步骤S104，根据归一化后的装备运行时间序列数据进行联邦学习，由每个装备利用本地数据训练本地装备故障监测模型，并将训练后的本地装备故障监测模型参数加密后上传至聚合服务器；In step S104, federated learning is performed according to the normalized equipment operation time series data, and each equipment uses local data to train a local equipment fault monitoring model, and the trained local equipment fault monitoring model parameters are encrypted and uploaded to the aggregation server;
步骤S105，由聚合服务器收集并聚合所有本地装备故障监测模型参数，以获得全局装备故障监测模型参数并建立全局装备故障监测模型；Step S105, the aggregation server collects and aggregates all local equipment fault monitoring model parameters to obtain global equipment fault monitoring model parameters and establish a global equipment fault monitoring model;
步骤S106，判断全局装备故障监测模型是否收敛，若未收敛，则将全局装备故障监测模型参数发送给各装备继续迭代训练，直到全局装备故障监测模型收敛；Step S106, judging whether the global equipment fault monitoring model converges, if not, sending the parameters of the global equipment fault monitoring model to each equipment to continue iterative training until the global equipment fault monitoring model converges;
步骤S107，将收敛的全局装备故障监测模型作为制造业装备故障监测模型。Step S107, taking the converged global equipment fault monitoring model as the manufacturing equipment fault monitoring model.
进一步的，采集装备的历史运行数据，对所述装备历史运行数据进行预处理，包括：Further, the historical operation data of the equipment is collected, and the historical operation data of the equipment is preprocessed, including:
采集装备历史运行数据；Collect equipment historical operation data;
对所述装备历史数据进行清洗，剔除异常、缺失的样本数据。The historical data of the equipment is cleaned to remove abnormal and missing sample data.
进一步的，对所述历史运行数据进行统计，构建装备运行时间序列数据，包括：Further, perform statistics on the historical operation data to construct equipment operation time series data, including:
对所述经过预处理后的历史运行数据按每时段统计，并按时间顺序构建装备运行时间序列数据。The preprocessed historical operation data is counted at each time period, and the equipment operation time series data is constructed in chronological order.
进一步的，对所述装备运行时间序列数据进行归一化处理，包括：Further, normalizing the equipment running time series data, including:
获取装备运行时间序列数据中的最大值和最小值；Obtain the maximum and minimum values in the equipment operating time series data;
对每个时段的装备运行时间序列数据归一化在0-1之间，归一化公式为
进一步的，根据所述归一化后的装备运行数据时间序列数据进行联邦学习，由每个装备分别在本地训练本地模型，并将训练后的本地模型参数上传至聚合服务器，包括：Further, the federated learning is performed according to the normalized time series data of the equipment operation data, the local model is trained locally by each equipment, and the trained local model parameters are uploaded to the aggregation server, including:
每个装备在本地利用归一化后的装备运行数据时间序列数据训练本地故障监测模型，模型训练结束后，由每个装备将模型参数加密后上传至聚合服务器，其中，加密方式选用同态加密算法，每个装备和聚合服务器都具有相同的公私钥，同态加密算法的相关定义为：定义x和y是明文空间M中的元素，Θ为M上的运算，E(·)是M上密钥空间为k的加密函数，设同态加密算法为F，则有 F(E(x),E(y))＝E(xΘy)，即对数据x和y加密后运算的结果与x和y运算后再加密的结果是相同的。通过应用同态加密算法，即使其他人获取了已加密的模型参数，由于缺少密钥信息，无法得知任何原始数据信息，进一步保证了数据的安全。Each equipment uses the normalized equipment operation data time series data to train the local fault monitoring model locally. After the model training is completed, each equipment encrypts the model parameters and uploads them to the aggregation server. The encryption method is homomorphic encryption. Algorithm, each equipment and aggregation server has the same public and private key, the related definition of homomorphic encryption algorithm is: define x and y as elements in plaintext space M, Θ is the operation on M, E( ) is on M The encryption function whose key space is k, let the homomorphic encryption algorithm be F, then there is F(E(x), E(y))=E(xΘy), that is, the result of encrypting the data x and y is the same as x The result of encrypting after the y operation is the same. By applying the homomorphic encryption algorithm, even if others obtain the encrypted model parameters, due to the lack of key information, they cannot know any original data information, which further ensures the security of the data.
进一步的，由聚合服务器聚合所有本地模型参数，包括：Further, all local model parameters are aggregated by the aggregation server, including:
聚合服务器在接收到所有装备的本地故障监测模型参数后，解密所有本地故障监测模型参数，基于加权平均法聚合所有本地故障监测模型参数，得到全局故障监测模型参数并建立全局故障监测模型。其中，聚合本地故障监测模型参数的加权平均法公式为
进一步的，判断全局故障监测模型是否收敛，包括：Further, judging whether the global fault monitoring model has converged, including:
对建立的全局故障监测模型进行检测，通过测试数据计算预测误差值判断全局故障模型是否收敛，若未收敛，聚合服务器将全局故障监测模型参数加密后发送给各装备，各装备解密全局故障监测模型参数并更新本地故障监测模型参数，继续迭代训练，直到全局故障监测模型收敛。The established global fault monitoring model is tested, and the prediction error value is calculated by the test data to judge whether the global fault model is converged. If not, the aggregation server encrypts the parameters of the global fault monitoring model and sends it to each equipment, and each equipment decrypts the global fault monitoring model. parameters and update the parameters of the local fault monitoring model, and continue iterative training until the global fault monitoring model converges.
进一步的，获得制造业装备故障监测模型，包括：Further, obtain the manufacturing equipment failure monitoring model, including:
将收敛的全局故障监测模型作为最终的制造业装备故障监测模型。The converged global fault monitoring model is used as the final manufacturing equipment fault monitoring model.
如图3所示，制造业装备故障监测模型训练及应用流程如下：As shown in Figure 3, the training and application process of the manufacturing equipment fault monitoring model is as follows:
S1、各装备对本地历史运行数据进行预处理，包括清洗、标准化等；S1. Each equipment preprocesses local historical operation data, including cleaning, standardization, etc.;
S2、应用联邦学习，进行分布式训练故障监测模型，获取制造业装备故障监测模型；S2. Apply federated learning to perform distributed training of fault monitoring models, and obtain manufacturing equipment fault monitoring models;
S3、将训练好的制造业装备故障监测模型应用在各装备上；S3. Apply the trained manufacturing equipment fault monitoring model to each equipment;
S4、各装备故障监测模型实时监测运行状态数据；S4. The fault monitoring model of each equipment monitors the running status data in real time;
S5、将发生故障的装备信息及故障诊断发送给检修人员的手持终端或自动控制设备；S5. Send the faulty equipment information and fault diagnosis to the hand-held terminal or automatic control equipment of the maintenance personnel;
S6、检修人员现场维修指导，并将维修信息及时上传至数据云端。S6. The maintenance personnel provide on-site maintenance guidance, and upload the maintenance information to the data cloud in time.
对于本领域技术人员而言，显然本发明不限于上述示范性实施例的细节，而且在不背离本发明的精神或基本特征的情况下，能够以其他的具体形式实现本发明。因此，无论从哪一点来看，均应将实施例看作是示范性的，而且是非限制性的，本发明的范围由所附权利要求而不是上述说明限定，因此旨在将落在权利要求的等同要件的含义和范围内的所有变化囊括在本发明内。It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the invention is to be defined by the appended claims rather than the foregoing description, which are therefore intended to fall within the scope of the claims. All changes within the meaning and scope of the equivalents of , are included in the present invention.
以上已将本发明做一详细说明，以上所述，仅为本发明之较佳实施例而已，当不能限定本发明实施范围，即凡依本申请范围所作均等变化与修饰，皆应仍属本发明涵盖范围内。The present invention has been described in detail above. The above descriptions are only preferred embodiments of the present invention, and should not limit the scope of implementation of the present invention. inventions are covered.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549090.6ACN114330740A (en) | 2021-12-17 | 2021-12-17 | A federated learning-based training system for manufacturing equipment fault monitoring model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549090.6ACN114330740A (en) | 2021-12-17 | 2021-12-17 | A federated learning-based training system for manufacturing equipment fault monitoring model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114330740Atrue CN114330740A (en) | 2022-04-12 |
Family
ID=81053413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111549090.6APendingCN114330740A (en) | 2021-12-17 | 2021-12-17 | A federated learning-based training system for manufacturing equipment fault monitoring model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114330740A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114580087A (en)* | 2022-05-06 | 2022-06-03 | 山东大学 | A method, device and system for predicting federal remaining useful life of shipborne equipment |
| CN114662618A (en)* | 2022-05-23 | 2022-06-24 | 湖南工商大学 | A fault diagnosis method, device and related equipment based on federated learning |
| WO2023197453A1 (en)* | 2022-04-13 | 2023-10-19 | 苏州浪潮智能科技有限公司 | Fault diagnosis method and apparatus, device, and storage medium |
| CN117762464A (en)* | 2023-12-29 | 2024-03-26 | 中睿信数字技术有限公司 | Cloud computing-based software operation and maintenance system and method |
| CN118427562A (en)* | 2024-07-04 | 2024-08-02 | 国网山东省电力公司信息通信公司 | Multivariate time series prediction method and system for multi-device federated dynamic graph |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110161953A (en)* | 2019-05-30 | 2019-08-23 | 淄博鲁工造粒装备科技有限公司 | A kind of Granulation Equipments intelligence control novel maintenance platform |
| CN110572253A (en)* | 2019-09-16 | 2019-12-13 | 济南大学 | A method and system for enhancing the privacy of federated learning training data |
| AU2020101613A4 (en)* | 2020-08-02 | 2020-09-10 | B.P.R.V.S, Priyatham MR | Collaborative federative learning approach for internet of vehicles traffic prediction |
| CN111678696A (en)* | 2020-06-17 | 2020-09-18 | 南昌航空大学 | A Machine Intelligence Fault Diagnosis Method Based on Federated Learning |
| CN112101489A (en)* | 2020-11-18 | 2020-12-18 | 天津开发区精诺瀚海数据科技有限公司 | Equipment fault diagnosis method driven by united learning and deep learning fusion |
| CN112906859A (en)* | 2021-01-27 | 2021-06-04 | 重庆邮电大学 | Federal learning algorithm for bearing fault diagnosis |
| CN113469370A (en)* | 2021-06-22 | 2021-10-01 | 河北工业大学 | Industrial Internet of things data sharing method based on federal incremental learning |
| CN113592102A (en)* | 2021-07-23 | 2021-11-02 | 青岛亿联信息科技股份有限公司 | Parking prediction model training method and system based on federal learning and block chain |
- 2021
- 2021-12-17CNCN202111549090.6Apatent/CN114330740A/enactivePending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110161953A (en)* | 2019-05-30 | 2019-08-23 | 淄博鲁工造粒装备科技有限公司 | A kind of Granulation Equipments intelligence control novel maintenance platform |
| CN110572253A (en)* | 2019-09-16 | 2019-12-13 | 济南大学 | A method and system for enhancing the privacy of federated learning training data |
| CN111678696A (en)* | 2020-06-17 | 2020-09-18 | 南昌航空大学 | A Machine Intelligence Fault Diagnosis Method Based on Federated Learning |
| AU2020101613A4 (en)* | 2020-08-02 | 2020-09-10 | B.P.R.V.S, Priyatham MR | Collaborative federative learning approach for internet of vehicles traffic prediction |
| CN112101489A (en)* | 2020-11-18 | 2020-12-18 | 天津开发区精诺瀚海数据科技有限公司 | Equipment fault diagnosis method driven by united learning and deep learning fusion |
| CN112906859A (en)* | 2021-01-27 | 2021-06-04 | 重庆邮电大学 | Federal learning algorithm for bearing fault diagnosis |
| CN113469370A (en)* | 2021-06-22 | 2021-10-01 | 河北工业大学 | Industrial Internet of things data sharing method based on federal incremental learning |
| CN113592102A (en)* | 2021-07-23 | 2021-11-02 | 青岛亿联信息科技股份有限公司 | Parking prediction model training method and system based on federal learning and block chain |
Non-Patent Citations (2)
| Title |
|---|
| 姚剑波: "《大数据安全与隐私》", 31 July 2017, 电子科技大学出版社, pages: 279* |
| 王健宗: "《深入浅出联邦学习 原理与实践》", 31 May 2021, 机械工业出版社, pages: 54* |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023197453A1 (en)* | 2022-04-13 | 2023-10-19 | 苏州浪潮智能科技有限公司 | Fault diagnosis method and apparatus, device, and storage medium |
| CN114580087A (en)* | 2022-05-06 | 2022-06-03 | 山东大学 | A method, device and system for predicting federal remaining useful life of shipborne equipment |
| CN114662618A (en)* | 2022-05-23 | 2022-06-24 | 湖南工商大学 | A fault diagnosis method, device and related equipment based on federated learning |
| CN117762464A (en)* | 2023-12-29 | 2024-03-26 | 中睿信数字技术有限公司 | Cloud computing-based software operation and maintenance system and method |
| CN117762464B (en)* | 2023-12-29 | 2024-07-09 | 中睿信数字技术有限公司 | Cloud computing-based software operation and maintenance system and method |
| CN118427562A (en)* | 2024-07-04 | 2024-08-02 | 国网山东省电力公司信息通信公司 | Multivariate time series prediction method and system for multi-device federated dynamic graph |
| CN118427562B (en)* | 2024-07-04 | 2024-11-08 | 国网山东省电力公司信息通信公司 | Multi-device federal dynamic graph-oriented multi-element time sequence prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114330740A (en) | A federated learning-based training system for manufacturing equipment fault monitoring model | |
| CN114330741A (en) | A training method for fault monitoring model of manufacturing equipment based on federated learning | |
| CN113674509B (en) | Edge node disaster monitoring and early warning system and method based on block chain | |
| CN111928888A (en) | Intelligent monitoring and analyzing method and system for water pollution | |
| CN117833464A (en) | Online operation state safety monitoring method for electricity consumption information acquisition terminal | |
| CN111552933A (en) | Method and device for identifying abnormal login of account | |
| CN108494802A (en) | Key message infrastructure security based on artificial intelligence threatens Active Defending System Against | |
| CN117596119A (en) | A device data collection and monitoring method and system based on SNMP protocol | |
| CN116679643B (en) | Industrial equipment control system based on block chain technology | |
| CN119651923B (en) | Substation control system based on cloud control and local area network | |
| CN110795260B (en) | Smart customer care system | |
| CN109861825B (en) | Internal attack detection method based on weighted rules and consistency in CPS system | |
| CN113112038B (en) | Intelligent monitoring and diagnostic analysis system, device, electronic equipment and storage medium | |
| CN119885184A (en) | Power grid enterprise power data storage security risk assessment system | |
| CN113506096B (en) | An inter-system interface method based on industrial Internet identification resolution system | |
| CN118262889B (en) | A ventilator fault prediction method and system based on medical Internet of Things | |
| CN119697002A (en) | Power communication network fault location analysis method and system based on deep learning | |
| CN118631594A (en) | A two-way authenticated electronic detonator system and control method thereof | |
| CN108833414B (en) | A kind of online service abnormality monitoring method | |
| CN114785617B (en) | 5G network application layer anomaly detection method and system | |
| Manicavasagam | Identifying and Detecting Network Indicators of Compromise (IOCs) for Demand Response Programs | |
| CN116820896B (en) | A non-intrusive industrial control terminal anomaly detection method based on physical signals | |
| CN119834974B (en) | Safe collection method and system for CRRT equipment data | |
| TWI810853B (en) | Cloud platform with an industrial iot model forecast function and a proactive audit function | |
| CN120011832B (en) | An intelligent equipment operation and maintenance method and system based on railway operation and maintenance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |