WO2019127502A1 - 工业物联网装置的监控预测装置、系统及方法 - Google Patents

工业物联网装置的监控预测装置、系统及方法 Download PDF

Info

Publication number
WO2019127502A1
WO2019127502A1 PCT/CN2017/120228 CN2017120228W WO2019127502A1 WO 2019127502 A1 WO2019127502 A1 WO 2019127502A1 CN 2017120228 W CN2017120228 W CN 2017120228W WO 2019127502 A1 WO2019127502 A1 WO 2019127502A1
Authority
WO
WIPO (PCT)
Prior art keywords
industrial internet
monitoring
variable
things device
rolling bearing
Prior art date
Application number
PCT/CN2017/120228
Other languages
English (en)
French (fr)
Inventor
朱加兴
博芬西彭·丹尼尔
吴腾飞
库明希·克里斯托夫
周文晶
Original Assignee
西门子公司
朱加兴
博芬西彭·丹尼尔
吴腾飞
库明希·克里斯托夫
周文晶
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 西门子公司, 朱加兴, 博芬西彭·丹尼尔, 吴腾飞, 库明希·克里斯托夫, 周文晶 filed Critical 西门子公司
Priority to PCT/CN2017/120228 priority Critical patent/WO2019127502A1/zh
Publication of WO2019127502A1 publication Critical patent/WO2019127502A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring

Definitions

  • the invention relates to the field of industrial internet of things, in particular to a monitoring and forecasting device, system and method for an industrial internet of things device.
  • IIoT Industrial Internet of Things
  • One way to optimize resource utilization is to apply data science at the top of existing processes, that is, to identify hidden patterns, which can limit optimal resource utilization.
  • Another way to optimize resource utilization is to apply the data-lake concept to each unit to provide data for big data storage. Data analysis logic is applied at the top level of such storage. In this way, monitoring and forecasting mechanisms borrowed from other Internet companies are not suitable for industrial environments, especially industrial IoT environments, due to:
  • Manufacturing or transmitting company process data is protected as a trade secret, and copying this data offline to the cloud is not acceptable.
  • a first aspect of the present invention provides a method for monitoring and predicting an industrial Internet of Things device, comprising the steps of: receiving a plurality of sample values describing a characteristic variable of an industrial Internet of Things device, and establishing a plurality of variable relationships for one or a plurality of feature variables; Separately analyzing a plurality of variable relationships, and selecting a first variable relationship for each variable relationship; calculating a multidimensional result corresponding to the industrial IoT device based on the first variable relationship and the sampled value of the characteristic variable; based on the industrial internet of things
  • the processing hierarchy requirements are sent to the processing hierarchy in the Industrial Internet of Things.
  • the monitoring and prediction method of the industrial internet of things device further comprises the step of selecting the first variable relationship by using a residual square sum method.
  • monitoring and prediction method of the industrial internet of things device further comprises the step of: updating the variable relationship.
  • the industrial internet of things device is a smart motor, wherein the characteristic variables include:
  • the monitoring and prediction method of the industrial internet of things device further comprises the steps of: receiving the characteristic variable cavity temperature, the outer temperature of the shell, the rotational speed, the vibration acceleration of the rolling bearing, and the sampling value of the working current, respectively, and establishing the smart motor respectively. N-dimensional variable relationship.
  • the one-dimensional variable relationship of the intelligent motor includes a variable relationship based on a rolling bearing vibration acceleration and a rolling bearing vibration acceleration threshold.
  • the first relationship of the variable relationship based on the rolling bearing vibration acceleration and the rolling bearing vibration acceleration threshold includes:
  • V4 i is the vibration acceleration of the rolling bearing
  • th(V4 i ) is the vibration acceleration threshold of the rolling bearing
  • i is a natural number.
  • the two-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature of the cavity and the operating current.
  • the three-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature of the cavity, the operating current, and the vibration acceleration of the rolling bearing.
  • the first relationship of the n-dimensional variable relationship of the smart motor includes:
  • V xntheory , V xn are respectively obtained.
  • the multi-dimensional result is that the smart motor 300 is faulty, wherein l 1 , l 2 ... l n are empirical values, and the multi-dimensional is greater than one dimension.
  • a second aspect of the present invention provides a monitoring and predicting apparatus for an industrial Internet of Things device, comprising: a variable relationship establishing module that receives a plurality of sample values describing characteristic variables of an industrial IoT device and establishes one or more characteristic variables a plurality of variable relationships; an analysis control module that receives a plurality of variable relationships from the variable relationship building module, selects a first variable relationship for each variable relationship, and then based on the first variable relationship and the sampled value of the characteristic variable Calculating the multi-dimensional results corresponding to the industrial IoT device, and transmitting the multi-dimensional results to the processing level in the industrial Internet of Things based on the requirements of the processing hierarchy in the industrial Internet of Things.
  • the analysis control module selects the first variable relationship by using a residual square sum method.
  • the monitoring and predicting device of the industrial internet of things device further comprises at least one variable relationship storage module disposed in the network card of the industrial internet of things device and/or in the industrial internet of things industrial cloud.
  • the industrial internet of things device is a smart motor, wherein the characteristic variables include:
  • variable relationship establishing module receives the characteristic values of the characteristic variable cavity temperature, the outer temperature of the shell, the rotational speed, the vibration acceleration of the rolling bearing, and the sampling value of the working current, and respectively establishes an n-dimensional variable relationship of the intelligent motor.
  • the one-dimensional variable relationship of the intelligent motor includes a variable relationship based on a rolling bearing vibration acceleration and a rolling bearing vibration acceleration threshold.
  • the first relationship of the variable relationship based on the rolling bearing vibration acceleration and the rolling bearing vibration acceleration threshold includes:
  • V4 i is the vibration acceleration of the rolling bearing
  • th(V4 i ) is the vibration acceleration threshold of the rolling bearing
  • i is a natural number.
  • the two-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature of the cavity and the operating current.
  • the three-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature of the cavity, the operating current, and the vibration acceleration of the rolling bearing.
  • the first relationship of the n-dimensional variable relationship of the smart motor includes:
  • V xntheory , V xn are respectively obtained.
  • the multi-dimensional result is that the smart motor 300 is faulty, wherein l 1 , l 2 ... l n are empirical values, and the multi-dimensional is greater than one dimension.
  • a third aspect of the present invention provides a monitoring and forecasting system for an industrial Internet of Things device, comprising: an intelligent motor coupled to the industrial Internet of Things; and a plurality of collecting devices connected to the smart motor, wherein
  • the monitoring and forecasting system of the industrial Internet of things device further comprises the monitoring and predicting device of the industrial internet of things device according to any one of claims 11 to 20.
  • the plurality of collection devices include an acquisition board, and the acquisition board is connected with a frequency converter, a first temperature sensor, a second temperature sensor and a vibration sensor, and the acquisition board is coupled to the intelligent motor via an industrial Ethernet network.
  • a network card wherein the frequency converter reads an operating current and/or a rotational speed of the intelligent motor;
  • the first temperature sensor is coupled to a lateral side of a bearing stator of the intelligent motor, and collects a cavity of the intelligent motor a temperature sensor;
  • the second temperature sensor is connected to the outer casing of the smart motor, and collects the temperature outside the casing of the smart motor;
  • the vibration sensor is connected to the rolling bearing of the intelligent motor, and collects the vibration acceleration of the rolling bearing of the intelligent motor .
  • the monitoring and prediction mechanism of the industrial internet of things device provided by the invention is an architectural model, which covers cloud computing and edge computing, improves calculation accuracy and reduces cost.
  • the present invention has a flexible deployment, especially when the device encounters an infinite amount of conditions in an industrial Internet of Things.
  • FIG. 1 is a schematic block diagram of a monitoring and predicting device of an industrial internet of things device according to an embodiment of the present invention
  • FIG. 2 is a graph of a two-dimensional variable relationship in accordance with an embodiment of the present invention.
  • FIG. 3 is a graph of a three-dimensional variable relationship in accordance with an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of a curve for selecting a first relationship in accordance with an embodiment of the present invention.
  • an industrial cloud 100 has an industrial cloud 100 and a plurality of gateways and devices coupled to the industrial cloud.
  • the smart motor 300 corresponding to the gateway 200 is shown.
  • a first temperature sensor 310, a second temperature sensor 320, a vibration sensor 330, and a frequency converter 340 are further connected to the smart motor 300, and the first temperature sensor 310, the second temperature sensor 320, the vibration sensor 330, and the frequency converter 340 are respectively connected. Both are connected to a collection board 350.
  • the acquisition board 350 is connected to the network card 200 of the smart motor 300 through industrial Ethernet (profinet).
  • the monitoring and forecasting device provided by the present invention includes a variable relationship establishing module 220 and an analysis control module 210, and the above modules may be disposed in the network card 200. Alternatively, the entire monitoring and prediction device 120 may in turn be placed in an industrial cloud.
  • a method for monitoring and predicting an industrial Internet of Things device includes the following steps:
  • step S1 the variable relationship establishing module 220 receives a plurality of sample values describing the feature variables of the smart motor 300, and establishes a plurality of variable relationships for one or a plurality of feature variables.
  • the characteristic variables of the smart motor 300 include the temperature inside the cavity, the temperature outside the casing, the rotational speed, the vibration acceleration of the rolling bearing, and the operating current.
  • the first temperature sensor 310 is connected to the outside of the bearing stator of the smart motor 300, and collects the temperature inside the cavity of the smart motor 300.
  • the second temperature sensor 320 is coupled to the outer casing of the smart motor 300 and collects the outside temperature of the smart motor 300.
  • the vibration sensor 330 is coupled to the rolling bearing of the smart motor 300 and collects the rolling bearing vibration acceleration of the smart motor 300.
  • the frequency converter 340 is used to read the operating current and/or the rotational speed of the smart motor 300.
  • the first temperature sensor 310, the second temperature sensor 320, and the vibration sensor 330 transmit the sample values of the cavity temperature, the outside temperature of the case, and the vibration acceleration of the rolling bearing to the acquisition board 350 through analog-to-digital conversion, and the inverter 340 operates the current through the modbus protocol. / or the speed is transmitted to the acquisition board 350.
  • the acquisition board 350 continuously collects the sample values of the feature data, the feature variables are transmitted to the variable relationship establishment module 220 of the network card 200 of the smart motor 300 through Industrial Ethernet (Profinet).
  • Profilet Industrial Ethernet
  • the collecting board 350 continuously collects the sampled values of the feature data and sends the sampled values to the network card 200, and then the network card 200 transmits the data to the monitoring and forecasting in the industrial cloud 100.
  • Device 120 performs the processing.
  • variable relationship establishing module 220 receives the sampled value of the feature data continuously collected by the acquisition board 350, and then is one or more of the characteristic variable cavity temperature, the outer temperature of the case, the rotation speed, the vibration acceleration of the rolling bearing, and the working current of the intelligent motor. , respectively, establish the n-dimensional variable relationship of the intelligent motor.
  • the one-dimensional variable relationship of the intelligent motor includes a variable relationship based on a rolling bearing vibration acceleration and a rolling bearing vibration acceleration threshold
  • the two-dimensional variable relationship of the intelligent motor includes a variable relationship based on a temperature of the cavity and an operating current
  • the intelligent motor The three-dimensional variable relationship includes a variable relationship based on the temperature of the cavity, the operating current, and the acceleration of the rolling bearing vibration.
  • step S2 is performed.
  • the analysis control module 210 in the network card 200 of the smart motor 300 receives a plurality of variable relationships from the variable relationship establishing module, and selects a first variable relationship for each variable relationship.
  • the present invention utilizes the residual square sum method to select the first variable relationship.
  • the first curve S 1 represents the first variable relationship
  • the second curve S 2 represents the second variable relationship
  • the third curve S 3 represents the third variable relationship.
  • the sample values of the characteristic variables are respectively brought into the first curve S 1 , the second curve S 2 and the third curve S 3 to obtain three sets of theoretical values.
  • the set of sampled value arguments of the feature variable y is ⁇ y 1sample , y 2sample , . . .
  • the set of sample values of the feature variable y is ⁇ y 1sample , y 2sample , . . . , y nsample ⁇
  • the relationship of the variables corresponding to a curve S 1 yields the theoretical value of y ⁇ y 11theory , y 12theory ,..., y 1ntheory ⁇
  • the set of sample values of the feature variable y is ⁇ y 1sample , y 2sample ,..., y nsample ⁇
  • the variable relationship corresponding to the curve S 2 yields the theoretical value of y ⁇ y 21theory , y 22theory ,..., y 2ntheory ⁇
  • the set of sample values of the feature variable y is ⁇ y 1sample , y 2sample ,..., y nsample ⁇ is brought into the third curve
  • the variable relationship corresponding to S 3 yields
  • the sum of the squares of the three residuals corresponds to the three curves, which are the sum of squared residuals of the first curve S 1 respectively.
  • Sum of squared residuals of the second curve S 2 Sum of squared residuals of the third curve S 3
  • the curve S is an actual value, and then the above three residual square sum values are compared, wherein the first curve is the smallest sum of squared residuals, which is closest to the true value.
  • the first curve is also different, for example, the regions Z1 and Z2 are included in FIG. 4, and the SSE (Relation1) ⁇ SSE (Relation2) ⁇ SSE (Relation 3) in Z1, in Z1.
  • the curve corresponding to the first relationship is SSE (Relation1).
  • in the Z2 area there may be different conclusions.
  • step S3 the analysis control module 210 in the network card 200 of the smart motor 300 calculates the multi-dimensional result corresponding to the industrial internet of things device based on the first variable relationship and the sampled value of the characteristic variable.
  • the one-dimensional variable relationship of the intelligent motor includes a variable relationship based on a vibration acceleration of a rolling bearing and a vibration acceleration threshold of a rolling bearing.
  • the first relationship of the variable relationship based on the rolling bearing vibration acceleration and the rolling bearing vibration acceleration threshold includes:
  • V4 i is the vibration acceleration of the rolling bearing
  • th(V4 i ) is the vibration acceleration threshold of the rolling bearing
  • i is a natural number.
  • the threshold of the feature variable is preset, and an alarm is triggered when the feature variable exceeds its threshold. For example, if the amplitude dynamics V4 i of the motor 300 is greater than th (V4 i ), an alarm is triggered.
  • the two-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature of the cavity and the operating current.
  • FIG 2 is a graph showing the relationship between a two-dimensional variables of a particular embodiment of the present invention, wherein the abscissa represents the chamber temperature V x, V y and the vertical axis indicates operating current.
  • the first relationship based on a two-dimensional variable relationship between the temperature of the cavity and the operating current includes:
  • l 1 and l 2 are empirical values.
  • the three-dimensional variable relationship of the intelligent motor includes a variable relationship based on the temperature inside the cavity, the working current, and the vibration acceleration of the rolling bearing.
  • 3 is a graph of a three-dimensional variable relationship in which the axis V x1 represents the temperature in the cavity, the axis V x3 represents the operating current, and the axis V x2 represents the vibration acceleration of the rolling bearing, in accordance with an embodiment of the present invention.
  • the theoretical values of the intracavity temperature V x1theory , V x12theory , V x11theory , the theoretical value of the operating current Vx 33sample , the vibration acceleration of the rolling bearing Vx 22sample , V x21sample , V x2sample are obtained .
  • the first relationship of the n-dimensional variable relationship of the intelligent motor includes:
  • the multi-dimensional result is that the smart motor 300 is faulty, wherein l 1 , l 2 ... l n are empirical values, and the multi-dimensional is greater than one dimension.
  • step S4 is performed to transmit the multi-dimensional result to the processing level 400 in the industrial Internet of Things based on the requirements of the processing layer, 400 in the industrial Internet of Things.
  • the monitoring and prediction method of the industrial internet of things device provided by the first aspect of the present invention further includes the step of updating the variable relationship.
  • the variable relationship storage module 140 can be set in the industrial cloud 100.
  • a third aspect of the present invention also provides a monitoring and forecasting system for an industrial Internet of Things device, comprising: a smart motor 300, a plurality of collecting devices, and a monitoring and predicting device provided by the second aspect of the present invention.
  • the smart motor 300 is coupled to the industrial internet of things, and a plurality of collection devices are connected to the smart motor 300.
  • the plurality of collection devices include an acquisition board 350.
  • the frequency converter 340, the first temperature sensor 310, the second temperature sensor 320, and the vibration sensor 330 are connected to the collection board.
  • the acquisition board 350 is coupled through an industrial Ethernet network.
  • the inverter 340 reads the operating current and/or the rotational speed of the smart motor 300; the first temperature sensor 310 is connected to the outside of the bearing stator of the smart motor 300, and the smart motor 300 is collected.
  • the second temperature sensor 320 is connected to the outer casing of the smart motor 300 to collect the outside temperature of the smart motor 300; the vibration sensor 330 is connected to the rolling bearing of the smart motor 300, and the The vibration acceleration of the rolling bearing of the intelligent motor.
  • the monitoring and prediction mechanism of the industrial internet of things device provided by the invention is an architectural model, which covers cloud computing and edge computing, improves calculation accuracy and reduces cost.
  • the present invention has a flexible deployment, especially when the device encounters an infinite amount of conditions in an industrial Internet of Things.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

一种工业物联网装置的监控预测装置、系统及方法，其中，包括如下步骤：变量关系建立模块（220）接收复数个描述工业物联网装置的特征变量的采样值，为一个或复数个特征变量建立复数个变量关系；分析控制模块（210）分别分析复数个变量关系，并为每个变量关系选择一个第一变量关系；基于所述第一变量关系和特征变量的采样值，计算出工业物联网装置所对应的多维结果；分析控制模块（210）基于工业物联网中的处理层级（400）的需求，将所述多维结果发送给工业物联网中的处理层级（400）。上述工业物联网装置的监控预测机制是一个体系结构模型其涵盖了云计算和边缘计算，提高了计算准确度，具有灵活的部署，尤其是在工业物联网中装置遇到不量状况时非常可靠。

Description

工业物联网装置的监控预测装置、系统及方法 技术领域
本发明涉及工业物联网领域，尤其涉及工业物联网装置的监控预测装置、系统及方法。
背景技术
随着工业4.0的到来，由于可利用信息的激增，工业物联网(IIoT，Industrial Internet of Things)需要优化资源的利用。优化资源利用的其中一个方式在现有流程的顶层应用数据科学，也就是确定隐含模式(identify hidden pattern)，这样可以限制最优的资源利用。而优化资源利用的另外方式是在每个单元应用数据池(data-lake)概念，对大数据存储提供数据。在这样的存储的顶层应用了数据分析逻辑。在这样的方式下，从别的互联网公司借用的监控预测机制并不适用于工业环境特别是工业物联网环境，这是由于：
基于具有低潜在因素需求(low latency requirements)的工业应用，应用数据分析于工业云中在很多情况下是不可行的。
制造或者传输公司的过程数据(process data)是作为商业秘密保护起来的，线下拷贝这些数据到云端是不被接受的。
虽然在线和线下处理这些资源可能是可行的(租或买)，但是并不能在这些资源的全容量(full capacity)情况下运行它们。
发明内容
本发明第一方面提供了工业物联网装置的监控预测方法，其中，包括如下步骤：接收复数个描述工业物联网装置的特征变量的采样值，为一个或复数个特征变量建立复数个变量关系；分别分析复数个变量关系，并为每个变量关系选择一个第一变量关系；基于所述第一变量关系和特征变量的采样值，计算出工业物联网装置所对应的多维结果；基于工业物联网中的处理层级的需求，将所述多维结果发送给工业物联网中的处理层级。
进一步地，所述工业物联网装置的监控预测方法还包括如下步骤：利用 残差平方和方法选择所述第一变量关系。
进一步地，工业物联网装置的监控预测方法还包括如下步骤：所述变量关系的更新步骤。
进一步地，所述工业物联网装置为智能电机，其中，所述特征变量包括：
-腔内温度；
-壳外温度；
-转速；
-滚动轴承振动加速度；
-工作电流。
进一步地，所述工业物联网装置的监控预测方法还包括如下步骤：接收智能电机描述的特征变量腔内温度、壳外温度、转速、滚动轴承振动加速度和工作电流的采样值，分别建立智能电机的n维变量关系。
进一步地，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系。
进一步地，所述基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系的所述第一关系包括：
A(V4 i)＝(V4 i＜th(V4 i))，
其中，V4 i为滚动轴承振动加速度，th(V4 i)为滚动轴承振动加速度阈值，i为自然数。
进一步地，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系。
进一步地，所述智能电机的三维变量关系包括基于腔内温度，工作电流，滚动轴承振动加速度的变量关系。
进一步地，所述智能电机的n维变量关系的所述第一关系包括：
f(V x1，V x2，…，V xn)＝0，
其中，首先将智能电机的特征变量采样值V x1sample，V x2sample......V xnsample分别带入f(V x1，V x2，…，V xn)＝0并分别得到V xntheory，V xn-1theory......V x1theory，如果满足以下任一项：
Figure PCTCN2017120228-appb-000001
则多维结果为提示智能电机300会发生故障，其中，l 1，l 2……l n为经验值，所述多维为大于一维。
本发明第二方面提供了工业物联网装置的监控预测装置，其中，包括：变量关系建立模块，其接收复数个描述工业物联网装置的特征变量的采样值，并为一个或复数个特征变量建立复数个变量关系；分析控制模块，其接收来自所述变量关系建立模块的复数个变量关系，为每个变量关系选择一个第一变量关系，然后基于所述第一变量关系和特征变量的采样值，计算出所所述工业物联网装置对应的多维结果，并基于工业物联网中的处理层级的需求，将所述多维结果发送给工业物联网中的处理层级。
进一步地，所述分析控制模块利用残差平方和方法选择所述第一变量关系。
进一步地，工业物联网装置的监控预测装置还包括至少一个设置在所述工业物联网装置的网卡中和/或所述工业物联网工业云中的所述变量关系存储模块。
进一步地，所述工业物联网装置为智能电机，其中，所述特征变量包括：
-腔内温度；
-壳外温度；
-转速；
-滚动轴承振动加速度；
-工作电流。
进一步地，变量关系建立模块接收智能电机描述的特征变量腔内温度、壳外温度、转速、滚动轴承振动加速度和工作电流的采样值，分别建立智能电机的n维变量关系。
进一步地，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系。
进一步地，所述基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系的所述第一关系包括：
A(V4 i)＝(V4 i＜th(V4 i))，
其中，V4 i为滚动轴承振动加速度，th(V4 i)为滚动轴承振动加速度阈值，i为自然数。
进一步地，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系。
进一步地，所述智能电机的三维变量关系包括基于腔内温度，工作电流， 滚动轴承振动加速度的变量关系。
进一步地，所述智能电机的n维变量关系的所述第一关系包括：
f(V x1，V x2，…，V xn)＝0，
其中，首先将智能电机的特征变量采样值V x1sample，V x2sample......V xnsample分别带入f(V x1，V x2，…，V xn)＝0并分别得到V xntheory，V xn-1theory......V x1theory，如果满足以下任一项：
Figure PCTCN2017120228-appb-000002
则多维结果为提示智能电机300会发生故障，其中，l 1，l 2……l n为经验值，所述多维为大于一维。
本发明第三方面提供了工业物联网装置的监控预测系统，其中，包括：智能电机，其耦合于所述工业物联网中；复述个采集装置，其连接于所述智能电机中，其中，所述工业物联网装置的监控预测系统还包括所述权利要求11至20任一项所述的工业物联网装置的监控预测装置。
进一步地，所述复数个采集装置包括采集板，所述采集板上连接有变频器、第一温度传感器、第二温度传感器和振动传感器，所述采集板通过工业以太网耦合于所述智能电机的网卡，其中，所述变频器读取所述所述智能电机的工作电流和/或转速；所述第一温度传感器连接至所述智能电机的轴承定子外侧，采集所述智能电机的腔内温度；所述第二温度传感器连接至所述智能电机的外壳，采集所述智能电机的壳外温度；所述振动传感器，连接至所述智能电机的滚动轴承，采集所述智能电机的滚动轴承振动加速度。
本发明提供的工业物联网装置的监控预测机制是一个体系结构模型(architectural model)，其涵盖了云计算(cloud computing)和边缘计算(edge computing)，提高了计算准确度，并降低了成本。本发明具有灵活的部署，尤其是在工业物联网中装置遇到不量状况时非常可靠。
附图说明
图1是根据本发明一个具体实施例的工业物联网装置的监控预测装置的框架示意图；
图2是根据本发明一个具体实施例的二维变量关系的曲线图；
图3是根据本发明一个具体实施例的三维变量关系的曲线图；
图4是根据本发明一个具体实施例的选择第一关系的曲线示意图。
具体实施方式
以下结合附图，对本发明的具体实施方式进行说明。
如图1所示，工业物联网中具有一个工业云100以及耦合到工业云的多个网关和设备，在本实施例中，仅示出一个网关200，以及该网关200所对应的智能电机300，在智能电机300上还分别连接有第一温度传感器310、第二温度传感器320、振动传感器330以及变频器340，上述第一温度传感器310、第二温度传感器320、振动传感器330和变频器340都连接至一个采集板350。其中，所述采集板350通过工业以太网(profinet)连接到智能电机300的网卡200。本发明提供的监控预测装置包括变量关系建立模块220和分析控制模块210，上述模块可以设置在网卡200中。可选地，整个监控预测装置120又可以设置在工业云中。
下面结合附图和优选实施例对本发明提供的工业物联网装置的监控预测方法及装置进行详细说明。
工业物联网装置的监控预测方法，包括如下步骤：
首先执行步骤S1，变量关系建立模块220接收复数个描述智能电机300的特征变量的采样值，为一个或复数个特征变量建立复数个变量关系。
具体地，智能电机300的特征变量包括腔内温度、壳外温度、转速、滚动轴承振动加速度、工作电流。其中，第一温度传感器310连接至智能电机300的轴承定子外侧，并采集所述智能电机300的腔内温度。第二温度传感器320连接至智能电机300的外壳，并采集智能电机300的壳外温度。振动传感器330连接至智能电机300的滚动轴承，并采集所述智能电机300的滚动轴承振动加速度。变频器340用于读取智能电机300的工作电流和/或转速。第一温度传感器310、第二温度传感器320和振动传感器330通过模拟数字转换将腔内温度、壳外温度和滚动轴承振动加速度的采样值传输至采集板350，变频器340通过modbus协议将工作电流和/或转速传输至采集板350。采集板350持续收集上述特征数据的采样值以后，通过工业以太网(Profinet)将上述特征变量发送给智能电机300的网卡200的变量关系建立模块220。
需要说明的是，如果监控预测装置120设置在工业云100中，则采集板350持续收集上述特征数据的采样值并发送给网卡200，然后网卡200再将数据传输至工业云100中的监控预测装置120进行处理。
接着，变量关系建立模块220收到采集板350持续收集的述特征数据的 采样值以后为智能电机的特征变量腔内温度、壳外温度、转速、滚动轴承振动加速度、工作电流中的一个或复数个，分别建立智能电机的n维变量关系。例如，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系，所述智能电机的三维变量关系包括基于腔内温度，工作电流，滚动轴承振动加速度的变量关系。
然后执行步骤S2，在本实施例中，智能电机300的网卡200中的分析控制模块210，接收来自所述变量关系建立模块的复数个变量关系，为每个变量关系选择一个第一变量关系。
图4是选择第一关系的曲线示意图。优选地，本发明利用残差平方和方法选择所述第一变量关系。如果有三种变量关系，分别对应图4中的三条曲线，其中，第一曲线S 1表示第一变量关系，第二曲线S 2表示第二变量关系，第三曲线S 3表示第三变量关系。具体地，将特征变量的采样值分别带入第一曲线S 1、第二曲线S 2和第三曲线S 3得到三组理论值。例如，特征变量y的采样值论值集合为{y 1sample，y 2sample，…，y nsample}，因此，特征变量y的采样值集合为{y 1sample，y 2sample，…，y nsample}带入第一曲线S 1对应的变量关系得到y的理论值{y 11theory，y 12theory，…，y 1ntheory}，特征变量y的采样值集合为{y 1sample，y 2sample，…，y nsample}带入第二曲线S 2对应的变量关系得到y的理论值{y 21theory，y 22theory，…，y 2ntheory}，特征变量y的采样值集合为{y 1sample，y 2sample，…，y nsample}带入第三曲线S 3对应的变量关系得到y的理论值{y 31theory，y 32theory，…，y 3ntheory}。利用求残差平方和的方法，得到三个残差平方和对应于三条曲线，分别为第一曲线S 1的残差平方和
Figure PCTCN2017120228-appb-000003
第二曲线S 2的残差平方和
Figure PCTCN2017120228-appb-000004
第三曲线S 3的残差平方和
Figure PCTCN2017120228-appb-000005
如图4所示，曲线S是实际值，然后比较上述三个残差平方和值，其中，第一曲线是残差平方和最小的，最接近真实值的。并且，在曲线每个区域中，第一曲线也是不同的，例如在图4中包括区域Z1和Z2，如果在Z1中SSE(Relation1)＜SSE(Relation2)＜SSE(Relation3)，则在Z1中的第一关系所对应的曲线是SSE(Relation1)。然而，在Z2区域，或许又有不同的结论。
然后执行步骤S3，智能电机300的网卡200中的分析控制模块210基于所述第一变量关系和特征变量的采样值，计算出工业物联网装置所对应的多 维结果。
下面结合具体实施例对选择出来的智能电机的一维变量关系，二维变量关系和三维变量关系的第一关系进行说明。
其中，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系。所述基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系的所述第一关系包括：
A(V4 i)＝(V4 i＜th(V4 i))，
其中，V4 i为滚动轴承振动加速度，th(V4 i)为滚动轴承振动加速度阈值，i为自然数。我们通过其中一个特征变量来判断工业装置的状态，并保证是实时判断。具体地，预先设定特征变量的阈值，当该特征变量超过了它的阈值则触发报警。例如，电机300的振幅动率V4 i如果大于th(V4 i)则触发报警。
其中，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系。图2是根据本发明一个具体实施例的二维变量关系的曲线图，其中横轴V x表示腔内温度，纵轴V y表示工作电流。基于腔内温度和工作电流的二维变量关系的所述第一关系包括：
f(x，y)＝0，
其中，首先将腔内温度的采样值V xsample带入f(x，y)＝0，可以得到工作电流的理论值V ytheory。同理，将工作电流的采样值V ysample带入f(x，y)＝0，可以得到腔内温度的理论值V xtheory。如果
Figure PCTCN2017120228-appb-000006
或者
Figure PCTCN2017120228-appb-000007
则二维结果为提示智能电机300会发生故障。其中，l 1和l 2为经验值。
其中，所述智能电机的三维变量关系包括基于腔内温度，工作电流，滚动轴承振动加速度的变量关系。图3是根据本发明一个具体实施例的三维变量关系的曲线图，其中轴V x1表示腔内温度，轴V x3表示工作电流，轴V x2表示滚动轴承振动加速度。同理，分别将腔内温度的采样值，工作电流的采样值V x31sample、Vx 32sample和滚动轴承振动加速度的采样值Vx 22sample、V x21sample两两带入f(x，y，z)＝0，可以分别得到腔内温度理论值V x1theory、V x12theory、V x11theory，工作电流理论值Vx 33sample，滚动轴承振动加速度Vx 22sample、V x21sample、V x2sample。
因此，所述智能电机的n维变量关系的所述第一关系包括：
f(V x1，V x2，…，V xn)＝0，
其中，首先将智能电机的特征变量采样值V x1sample，V x2sample......V xn-1sample分别带入f(V x1，V x2，…，V xn)＝0并得到V xntheory，当
Figure PCTCN2017120228-appb-000008
则提示智能电机会缓慢发生故障。然后将V x1sample，V x2sample......V xn-2ample，V xnsample分别带入f(V x1，V x2，...，V xn)＝0并得到V xn-1theory，如果
Figure PCTCN2017120228-appb-000009
则提示智能电机会缓慢发生故障。执行上述过程三次，最终，我们将V x2sample，V x3sample......V xnsample分别带入f(V x1，V x2，…，V xn)＝0则得到V x1theory，如果
Figure PCTCN2017120228-appb-000010
则多维结果为提示智能电机300会发生故障，其中，l 1，l 2……l n为经验值，所述多维为大于一维。
最后执行步骤S4，基于工业物联网中的处理层，400的需求，将所述多维结果发送给工业物联网中的处理层级400。
其中，本发明第一方面提供的工业物联网装置的监控预测方法还包括变量关系的更新步骤。例如，如图1所示，可以在工业云100中设置所述变量关系存储模块140。
本发明第三方面还提供了工业物联网装置的监控预测系统，其中，包括智能电机300、复数个采集装置以及本发明第二方面所提供的监控预测装置。其中，所述智能电机300耦合于所述工业物联网中，复述个采集装置连接于所述智能电机300中。
其中，所述复数个采集装置包括采集板350，所述采集板上连接有变频器340、第一温度传感器310、第二温度传感器320和振动传感器330，所述采集板350通过工业以太网耦合于所述智能电机300的网卡200。其中，所述变频器340读取所述所述智能电机300的工作电流和/或转速；所述第一温度传感器310连接至所述智能电机300的轴承定子外侧，采集所述智能电机300的腔内温度；所述第二温度传感器320连接至所述智能电机300的外壳，采集所述智能电机300的壳外温度；所述振动传感器330连接至所述智能电机300的滚动轴承，采集所述智能电机的滚动轴承振动加速度。
本发明提供的工业物联网装置的监控预测机制是一个体系结构模型(architectural model)，其涵盖了云计算(cloud computing)和边缘计算(edge computing)，提高了计算准确度，并降低了成本。本发明具有灵活的部署，尤 其是在工业物联网中装置遇到不量状况时非常可靠。
尽管本发明的内容已经通过上述优选实施例作了详细介绍，但应当认识到上述的描述不应被认为是对本发明的限制。在本领域技术人员阅读了上述内容后，对于本发明的多种修改和替代都将是显而易见的。因此，本发明的保护范围应由所附的权利要求来限定。此外，不应将权利要求中的任何附图标记视为限制所涉及的权利要求；“包括”一词不排除其它权利要求或说明书中未列出的装置或步骤；“第一”、“第二”等词语仅用来表示名称，而并不表示任何特定的顺序。

Claims (22)

  1. 工业物联网装置的监控预测方法，其中，包括如下步骤：
    接收复数个描述工业物联网装置的特征变量的采样值，为一个或复数个特征变量建立复数个变量关系；
    分别分析复数个变量关系，并为每个变量关系选择一个第一变量关系；
    基于所述第一变量关系和特征变量的采样值，计算出工业物联网装置所对应的多维结果；
    基于工业物联网中的处理层级的需求，将所述多维结果发送给工业物联网中的处理层级。
  2. 根据权利要求1所述的工业物联网装置的监控预测方法，其特征在于，所述工业物联网装置的监控预测方法还包括如下步骤：
    利用残差平方和方法选择所述第一变量关系。
  3. 根据权利要求1所述的工业物联网装置的监控预测方法，其特征在于，工业物联网装置的监控预测方法还包括如下步骤：
    所述变量关系的更新步骤。
  4. 根据根据权利要求1所述的工业物联网装置的监控预测方法，其特征在于，所述工业物联网装置为智能电机，其中，所述特征变量包括：
    -腔内温度；
    -壳外温度；
    -转速；
    -滚动轴承振动加速度；
    -工作电流。
  5. 根据根据权利要求4所述的工业物联网装置的监控预测方法，其特征在于，所述工业物联网装置的监控预测方法还包括如下步骤：
    接收智能电机描述的特征变量腔内温度、壳外温度、转速、滚动轴承振动加速度和工作电流的采样值，分别建立智能电机的n维变量关系。
  6. 根据权利要求5所述的工业物联网装置的监控预测方法，其特征在于，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系。
  7. 根据权利要求6所述的工业物联网装置的监控预测方法，其特征在于， 所述基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系的所述第一关系包括：
    A(V4 i)＝(V4 i＜th(V4 i))，
    其中，V4 i为滚动轴承振动加速度，th(V4 i)为滚动轴承振动加速度阈值，i为自然数。
  8. 根据权利要求5所述的工业物联网装置的监控预测方法，其特征在于，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系。
  9. 根据权利要求5所述的工业物联网装置的监控预测方法，其特征在于，所述智能电机的三维变量关系包括基于腔内温度，工作电流，滚动轴承振动加速度的变量关系。
  10. 根据权利要求5所述的工业物联网装置的监控预测方法，其特征在于，所述智能电机的n维变量关系的所述第一关系包括：
    f(V x1，V x2，...，V xn)＝0，
    其中，首先将智能电机的特征变量采样值V x1sample，V x2sample......V xnsample分别带入f(V x1，V x2，...，V xn)＝0并分别得到V xntheory，V xn-1theory......V x1theory，如果满足以下任一项：
    Figure PCTCN2017120228-appb-100001
    则多维结果为提示智能电机300会发生故障，其中，l 1，l 2……l n为经验值，所述多维为大于一维。
  11. 工业物联网装置的监控预测装置，其中，包括：
    变量关系建立模块，其接收复数个描述工业物联网装置的特征变量的采样值，并为一个或复数个特征变量建立复数个变量关系；
    分析控制模块，其接收来自所述变量关系建立模块的复数个变量关系，为每个变量关系选择一个第一变量关系，然后基于所述第一变量关系和特征变量的采样值，计算出所所述工业物联网装置对应的多维结果，并基于工业物联网中的处理层级的需求，将所述多维结果发送给工业物联网中的处理层级。
  12. 根据权利要求11所述的工业物联网装置的监控预测装置，其特征在于，所述分析控制模块利用残差平方和方法选择所述第一变量关系。
  13. 根据权利要求11所述的工业物联网装置的监控预测装置，其特征在于，工业物联网装置的监控预测装置还包括至少一个设置在所述工业物联网 装置的网卡中和/或所述工业物联网工业云中的所述变量关系存储模块。
  14. 根据根据权利要求11所述的工业物联网装置的监控预测装置，其特征在于，所述工业物联网装置为智能电机，其中，所述特征变量包括：
    -腔内温度；
    -壳外温度；
    -转速；
    -滚动轴承振动加速度；
    -工作电流。
  15. 根据根据权利要求14所述的工业物联网装置的监控预测装置，其特征在于，变量关系建立模块接收智能电机描述的特征变量腔内温度、壳外温度、转速、滚动轴承振动加速度和工作电流的采样值，分别建立智能电机的n维变量关系。
  16. 根据权利要求15所述的工业物联网装置的监控预测装置，其特征在于，所述智能电机的一维变量关系包括基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系。
  17. 根据权利要求16所述的工业物联网装置的监控预测装置，其特征在于，所述基于滚动轴承振动加速度以及滚动轴承振动加速度阈值的变量关系的所述第一关系包括：
    A(V4 i)＝(V4 i＜th(V4 i))，
    其中，V4 i为滚动轴承振动加速度，th(V4 i)为滚动轴承振动加速度阈值，i为自然数。
  18. 根据权利要求15所述的工业物联网装置的监控预测装置，其特征在于，所述智能电机的二维变量关系包括基于腔内温度和工作电流的变量关系。
  19. 根据权利要求15所述的工业物联网装置的监控预测装置，其特征在于，所述智能电机的三维变量关系包括基于腔内温度，工作电流，滚动轴承振动加速度的变量关系。
  20. 根据权利要求15所述的工业物联网装置的监控预测装置，其特征在于，所述智能电机的n维变量关系的所述第一关系包括：
    f(V x1，V x2，...，V xn)＝0，
    其中，首先将智能电机的特征变量采样值V x1sample，V x2sample......V xnsample分别 带入f(V x1，V x2，...，V xn)＝0并分别得到V xntheory，V xn-1theory......V x1theory，如果满足以下任一项：
    Figure PCTCN2017120228-appb-100002
    则多维结果为提示智能电机300会发生故障，其中，l 1，l 2……l n为经验值，所述多维为大于一维。
  21. 工业物联网装置的监控预测系统，其中，包括：
    智能电机，其耦合于所述工业物联网中；
    复述个采集装置，其连接于所述智能电机中，
    其中，所述工业物联网装置的监控预测系统还包括所述权利要求11至20任一项所述的工业物联网装置的监控预测装置。
  22. 根据权利要求21所述的工业物联网装置的监控预测系统，其中，所述复数个采集装置包括采集板，所述采集板上连接有变频器、第一温度传感器、第二温度传感器和振动传感器，所述采集板通过工业以太网耦合于所述智能电机的网卡，其中，
    所述变频器读取所述所述智能电机的工作电流和/或转速；
    所述第一温度传感器连接至所述智能电机的轴承定子外侧，采集所述智能电机的腔内温度；
    所述第二温度传感器连接至所述智能电机的外壳，采集所述智能电机的壳外温度；
    所述振动传感器，连接至所述智能电机的滚动轴承，采集所述智能电机的滚动轴承振动加速度。
PCT/CN2017/120228 2017-12-29 2017-12-29 工业物联网装置的监控预测装置、系统及方法 WO2019127502A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/120228 WO2019127502A1 (zh) 2017-12-29 2017-12-29 工业物联网装置的监控预测装置、系统及方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/120228 WO2019127502A1 (zh) 2017-12-29 2017-12-29 工业物联网装置的监控预测装置、系统及方法

Publications (1)

Publication Number Publication Date
WO2019127502A1 true WO2019127502A1 (zh) 2019-07-04

Family

ID=67062875

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/120228 WO2019127502A1 (zh) 2017-12-29 2017-12-29 工业物联网装置的监控预测装置、系统及方法

Country Status (1)

Country Link
WO (1) WO2019127502A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112073531A (zh) * 2020-09-15 2020-12-11 常熟理工学院 一种基于边缘计算的物联网实时监测系统的实现方法
WO2021114044A1 (en) * 2019-12-09 2021-06-17 Siemens Aktiengesellschaft Information acquiring method, apparatus, and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101158693A (zh) * 2007-09-26 2008-04-09 东北大学 基于多核独立元分析的批量生产过程故障检测方法
CN105022273A (zh) * 2015-07-25 2015-11-04 南通大学 一种基于物联网的多级带式输送机协调控制系统及方法
CN105607617A (zh) * 2015-12-18 2016-05-25 广州市澳视光电子技术有限公司 一种基于物联网技术的安防故障诊断系统及方法
CN106485589A (zh) * 2016-10-20 2017-03-08 河南省农业科学院 一种基于物联网的农业企业集团信息化管理系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101158693A (zh) * 2007-09-26 2008-04-09 东北大学 基于多核独立元分析的批量生产过程故障检测方法
CN105022273A (zh) * 2015-07-25 2015-11-04 南通大学 一种基于物联网的多级带式输送机协调控制系统及方法
CN105607617A (zh) * 2015-12-18 2016-05-25 广州市澳视光电子技术有限公司 一种基于物联网技术的安防故障诊断系统及方法
CN106485589A (zh) * 2016-10-20 2017-03-08 河南省农业科学院 一种基于物联网的农业企业集团信息化管理系统

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021114044A1 (en) * 2019-12-09 2021-06-17 Siemens Aktiengesellschaft Information acquiring method, apparatus, and system
US11838367B2 (en) 2019-12-09 2023-12-05 Siemens Aktiengesellschaft Information acquiring method, apparatus, and system
CN112073531A (zh) * 2020-09-15 2020-12-11 常熟理工学院 一种基于边缘计算的物联网实时监测系统的实现方法

Similar Documents

Publication Publication Date Title
CN112203282B (zh) 一种基于联邦迁移学习的5g物联网入侵检测方法及系统
Zhang et al. An IoT-based online monitoring system for continuous steel casting
Hafeez et al. Edge intelligence for data handling and predictive maintenance in IIOT
US11336543B2 (en) Edge HMI module server system and method
CN110929934A (zh) 设备故障预测方法、装置、计算机设备和存储介质
CN109743356B (zh) 工业互联网数据采集方法及装置、可读存储介质和终端
EP3376731A1 (en) Rule-based information exchange in internet of things
Mahajan et al. Prediction of network traffic in wireless mesh networks using hybrid deep learning model
WO2019127502A1 (zh) 工业物联网装置的监控预测装置、系统及方法
US10693841B2 (en) System and method for transmitting data relating to an object
Parto et al. A novel three-layer IoT architecture for shared, private, scalable, and real-time machine learning from ubiquitous cyber-physical systems
EP3475846B1 (en) Transactional-unstructured data driven sequential federated query method for distributed systems
Wang et al. Anomaly detection in wireless sensor networks based on KNN
CN116224791A (zh) 智能制造协作机器人边缘系统的协作训练控制方法
Zhang et al. Intelligent fault diagnosis system based on vibration signal edge computing
Chen et al. Heterogeneous multiview crowdsensing based on half quadratic optimization for the visual internet of things
CN115473688A (zh) 面向软件定义网络的异常检测方法、装置及设备
WO2016156656A2 (en) Arrangement for implementation of scalable the internet of things platform
Latif et al. Cloudlet Federation Based Context-Aware Federated Learning Approach
JP6530353B2 (ja) ライブデータ検索システムおよびライブデータ検索方法
Chen et al. NNFacet: Splitting Neural Network for Concurrent Smart Sensors
Sheeba et al. WFCM based big sensor data error detection and correction in wireless sensor network
US20230385603A1 (en) Neural architecture search system and search method
Yang et al. Semisupervised Graph Neural Networks for Traffic Classification in Edge Networks
CN112188455B (zh) 基于网关的定位方法、装置、设备及存储介质

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: 17936056

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: 17936056

Country of ref document: EP

Kind code of ref document: A1