Edge cloud-based resin material drying system and method

Information

  • Patent Grant
  • 10921792
  • Patent Number
    10,921,792
  • Date Filed
    Friday, December 21, 2018
    6 years ago
  • Date Issued
    Tuesday, February 16, 2021
    3 years ago
Abstract
A method of evaluating factory production machinery up time and down time performance provides a collection of sensors in individual communication with factory production machinery, with each sensor collecting high frequency vector data as respecting a physical parameter associated with the machinery, extracts the data from the sensors to produce a sensor data set, transforms the data set into the frequency domain, extracts statistical and mathematical information from the data set, transfers the data set, to an associated edge cloud, and within the associated edge cloud processes the data set to provide a repair, maintenance and operation board for the machinery to evaluate up time and down time performance for the factory production machinery.
Description
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

Not applicable—this invention was conceived and developed entirely using private source funding; this patent application is being filed and paid for entirely by private source funding.


INCORPORATION BY REFERENCE

Applicant incorporates by reference the disclosures of the following United States patent Publications:


US 2018/0077522 A1, published 15 Mar. 2018, entitled “Predicting Dryer Maintenance Using Machine Wearables”;


US 2017/0178030 A1, published 22 Jun. 2017, entitled “Method, System And Apparatus Using Field Learning To Upgrade Trending Sensor Curves Into Fuel Gauge Based Visualization Of Predictive Maintenance By User Driven Feedback Mechanism;”


US 2016/0245765 A1, published 21 Feb. 2016, entitled “Method For Accurately Measuring Real-Time Dew-Point Value And Total Moisture Content Of A Material”;


US 2016/0245279 A1, published 23 Feb. 2015, entitled “Real Time Machine Learning Based Predictive and Preventive Maintenance of a Vacuum Pump”;


US 2017/0051978 A1, published 23 Aug. 2015, entitled “Machine Learning Based Predictive Maintenance of a Dryer”;


US 2017/0011298 A1, published 2 Jul. 2015, entitled “Internet of Things Based Determination of Machine Reliability and Automated Maintenance, Repair, and Operation (MRO) Logs”;


US 2016/0313216 A1, published 6 Nov. 2015, entitled “Fuel Gauge Visualization of IOT Based Predictive Maintenance System Using Multi-Classification Based Machine Learning”;


US 2016/0291552 A1, published 2 Dec. 2015, entitled “System For Rule Management, Predictive Maintenance And Quality Assurance Of A Process And Machine Using Reconfigurable Sensor Networks And Big Data Machine Learning”;


US 2017/0160328 A1, published 2 Dec. 2015, entitled “Distributed IOT Based Sensor Analytics for Power Line Diagnosis”;


US 2016/0245686 A1, published 22 Dec. 2015, entitled “Fault Detection In Rotor Driven Equipment Using Rotational Invariant Transform Of Sub-Sampled 3-Axis Vibrational Data”;


US 2017/0201585 A1, published 13 Jul. 2017, entitled “Distributed Edge Processing of Internet of Things Device Data in Co-Location Facilities”;


US 2017/0060574, published 2 Mar. 2017, entitled “Edge Intelligence Platform and Internet of Things Censor Stream System”;


US 2017/0163444, published 8 Jun. 2017, entitled “Apparatus and Method for Using and Internet of Things Edge Secure Gateway”


US 2017/123425, published 20 Jul. 2017, entitled “Distributed Edge Processing of Internet of Things Device Data in Co-Location Facilities”.


DESCRIPTION OF THE PRIOR ART

The term “cloud” has become familiar not only to data processing professionals but also to anyone familiar with computer technology on anything more than a casual basis. The so-called “cloud” is made up of more than 10,000 data centers scattered over the world. In the next five years, driven by the need to get data, solutions and applications closer to the user, hundreds of thousands of scaled-down data centers are expected to sprout up at the edge of the Internet network, to form what is sometimes collectively called the “edge cloud”. Accordingly, edge cloud computing means computing that makes use of or is performed wholly or partially in such scaled-down data centers defining the “edge cloud”.


In todays networked world, data centers run by large entities such as Amazon, Google, Equinox, DRT, and the like are interconnected and stream data, information, and applications, namely software, over the Internet to end users. Large providers of content lease capacity in data centers or operate their own data centers. Global operations such as Amazon have one, two, or more data centers per continent. These large data centers may be considered to sit at the virtual center of the cloud, meaning they are all well removed, distance wise, from the end users.


There is burgeoning activity involving applications and content, which activities include streaming video to mobile devices, “wait-and-see” sensitive Internet computing, data, software applications, and security filters, all that need to be pushed into the Internet network, namely closer to the edge of the Internet network and therefore closer to the user, in order that these applications, data, software, etc., may be accessed quickly and may provide their data, information, results, etc., faster to the Internet users. This shift to “edge cloud computing” is necessitated by a number of factors, with latency, namely the delay before a transfer of data begins following an instruction for its transfer, being the most often cited reason. Cost is also always a concern.


Accordingly, growth of the edge cloud is driven by the need to get data and applications closer to users. Hence, the edge cloud, in the broadest sense, is composed of orders of magnitude more data centers, each desirably being scaled down and each desirably being a shorter distance from the relevant end users. The rationale for the edge cloud is that if an application runs on a processor at a location closer to the user, latency is reduced as the data traverses a shorter segment of the Internet network, thereby ultimately reducing the resources required. By contrast, when compared to the conventional centralized data center world, in a major metropolitan area when information, data, etc., travels from a data center to the end user, the information, data, etc., might have to go through many routers and run over thousands of miles of fiber optics on its journey to the end user. As a result the transport costs and potentials for delay are considerable. The edge cloud approach minimizes if not eliminates these delays and associated risks and reduces costs.


Computing using edge cloud is currently used in factories and in public clouds such as those operated by Amazon, AWS, and Rack Space, for low latency scalable computations in Industrial Internet of Things applications. Such computing strategy, namely using the edge cloud, has gained momentum recently in both academic and industrial applications. It is acknowledged in current computational strategy that edge cloud computing could bring several advantages to Internet of Things applications, such as permitting low latency real time analytics that are needed for industrial control, providing lower cost of data processing, and providing more efficient computation, all as compared to use of the public cloud.


Contrasting, the visualization available through the public cloud facilitates low latency downloads for mobile and web application users who may be accessing information and formatting and computing data anywhere in the world.


There are already millions of “things” in the world of the Internet of Things, connected together using and through the public cloud, which “things” do not use any edge cloud computational capabilities.


SUMMARY OF THE INVENTION

In one of its aspects, this invention provides a “hybrid cloud” inclusive architecture for use in edge cloud computing, which addresses both brown field scenarios and green field scenarios.


In another of its aspects, this invention provides an integrated hybrid cloud architecture allowing a system of sensors to work seamlessly both with and without edge cloud computing capability. This aspect of the invention is particularly useful for new Industrial Internet of Things entrants experimenting with a few Internet of Things sensors. The overhead cost of edge cloud computing may be too high for those entrants seeking to provide an introductory proof of some concept.


The following detailed description is merely exemplary in nature and is not intended to limit the described embodiments of the invention or uses of the described embodiments. As used herein, the words “exemplary” and “illustrative” mean “serving as an example, instance, or for illustration.” Any implementation or embodiment or abstract disclosed herein as being “exemplary” or “illustrative” is not necessarily to be construed as preferred or advantageous over other implementations, aspects, or embodiments. All of the implementations or embodiments described in the detailed description are exemplary implementations and embodiments provided to enable persons of skill in the art to make and to use the implementations and embodiments as disclosed below, to otherwise practice the invention, and are not intended to limit the scope of the invention, which is defined by the claims.


Furthermore, by this disclosure, there is no intention on the part of the Applicant to be bound by any express or implied theory presented in the preceding materials, including but not limited to the summary of the invention or the description of the prior art, or in the following detailed description of the invention. It is to be understood that the specific implementations, devices, processes, aspects, and the like illustrated in the attached drawings and described in the following portion of the application, usually referred to as the “specification,” are simply exemplary embodiments of the inventive concepts defined in the claims. Accordingly, specific dimensions and other physical characteristics relating to the embodiments disclosed herein are not to be considered as limiting as respecting the invention unless the claims or the specification expressly state otherwise.





BRIEF DESCRIPTION OF THE DRAWINGS


FIG. 1 is a schematic diagram of a system, in accordance with the invention, illustrating system interaction among a sets of sensors, the sensors being both with and without computational capabilities and electronics, and being located in different factory facilities, providing data to associated edge clouds and to the public cloud.



FIG. 2 is a schematic flow chart depicting connection of legacy sensors and new sensors to a virtual edge cloud in accordance with aspects of the invention,



FIG. 3 is a schematic block diagram illustrating a preferred system architecture in accordance with the invention.



FIG. 4 is a schematic block diagram illustrating a preferred system architecture, the public cloud, and a virtual edge cloud in accordance with the invention.





DESCRIPTION OF THE INVENTION

As used herein, the term “Fog” means computation in the sensor electronics itself, while “Core” means computation in a central server. In accordance with the invention, for complex Industrial Internet of Things analytics, there are at least five layers of computation that are important. The five layers, or protocols of computation are as follows:

    • Fog 1: Where high frequency vector data is extracted from the sensors and transformed, either in the time domain or in the frequency domain, to produce a sensor data set.
    • Fog 2: Where useful, statistical, and mathematical information is extracted from a transformed sensor data set.
    • Core 1: Where data and features for one or multiple sensors are used in computations, using either a rule based algorithm or a machine learning engine, to predict reliability and performance and to provide further analytics for a machine or a process.
    • Core 2: Where the analytics obtained for a machine or process, preferably from Core 1, are used for one or more secondary layers or analytics such as a repair, maintenance, and operation board for the machine or process, or a rule set for the machine or process.
    • Core 3: Where data obtained preferably using the Core 2 protocol, is further used to process more advanced analytics such as the up time or down time performance of a factory in which machines having the sensors associated therewith are located.


In the traditional edge of “Fog” computing paradigm, the “Fog” computation, either Fog 1 or Fog 2 as identified above, increases the cost of sensor electronics as they necessarily use at least one gigabit or more of random access memory and require at least one gigahertz or higher processor speed. Performing core computing in the public cloud as per computations Core 2, Core 2, and Core 3 above, increases the cloud computing costs. This invention addresses this problem by providing a edge cloud architecture as a system for merging layers of computation one through five, namely Fog 1 and 2, and Core 1, 2, and 3 as identified above, into the edge cloud by performing them in the edge cloud and thus reducing the cost of hardware and the cost of a cloud subscription simultaneously, through the single architecture in accordance with the invention.


In addition to the computations described in the five protocols identified above, analytics obtained from the third and fourth protocols, Core 1 and Core 2, are used in accordance with the invention to provide real time feed data to control systems. The edge cloud computation approaches in accordance with the invention use soft integration of the layer consisting of either the third or fourth protocol, namely Core 1 and Core 2 identified above, with the system of the factory. The invention provides such edge cloud protocol for integration of the control plan with the edge cloud computations.


Specific to the problem of predictive maintenance where feedback is required from the users of the adaptive predictive analytics respecting maintenance issues in a course of computing within protocols three and four above, the invention in one of its aspects runs local and global adaptive predictive analytics. The hybrid cloud architecture of the invention caters to both of those needs, namely the local protocol is optimally only a single edge cloud computation, whereas the predictive analytics global uses multiple feedback from plural edge cloud computational protocols.


Fog level computation, in sensor electronics within or attached to the sensors, is vulnerable to cyber-attack, since typically there are many such devices in a single factory. In factories that are resource limited, advanced security measures are difficult to implement. This makes the entire factory network vulnerable to cyber attack and is one of the weakest points of the Industrial Internet of Things. With edge cloud computing in accordance with the invention, and now in the course of using proprietary protocols in accordance with the invention, preferably sensor devices used in the practice of the invention are ones that talk only to the associated edge cloud and to nothing else. In this way a factory Wi-Fi or Ethernet network in accordance with the invention remains much safer and essentially immune from compromise of the sensor devices. The invention accomplishes this with the edge cloud architecture addressing the critical issue of network security by use of proprietary protocol layers, all as disclosed and claimed herein.


As described above, FIG. 1 depicts a system in accordance with the invention, in which a set of sensors, with or without integrated computational capabilities and electronics, are sending data to an edge cloud, and to a low cloud in the case of real time and non-real time users, and to the public cloud. FIG. 1 further depicts legacy sensors sending data to the public cloud directly, since such legacy sensors are normally without edge cloud support. New sensors can communicate with end users via the edge cloud or, as often is the case of smaller scale operations, can connect to the public cloud directly, all as illustrated in FIG. 1.


Further regarding FIG. 1, two kinds of visualization are illustrated schematically where the public cloud provides visualization, which is non-real time but is available anywhere, while edge clouds serve the need for real-time visualization within the associated factory premises, with a user's viewing device and an associated edge cloud all connected to the same subnetwork. This is illustrated in FIG. 1. There could be hundreds of edge clouds in the system as illustrated in FIG. 1.


As described above FIG. 2 is a schematic flow chart depicting different ways of connecting legacy sensors, which have already been deployed in the field and work without any edge cloud, and new sensors; the connection is performed using virtual edge cloud, which is sometime abbreviated herein as “VEC”. In the schematic diagram of FIG. 2, the public cloud should be understood as being composed of two major functional blocks, namely a virtual edge cloud and an Internet of Things visualization micro service, both as shown. Still referring to FIG. 2, the virtual edge cloud is similar to an ordinary edge cloud, but the virtual edge cloud is scalable for many more sensors and does not facilitate real time visualization. With the virtual edge cloud, in accordance with the invention sensor data, whether processed using an edge cloud approach or in the public cloud, creates the same level of non-real time visualization. However, sensor data processed by the virtual edge cloud and without use of edge cloud per se, does not provide for real time visualization in the factory, all as illustrated schematically in FIG. 2.


As noted above, FIG. 3 is a block diagram illustrating the architecture of the instant invention. In FIG. 3, it is understood that in edge cloud, as known in distributed computing, it is important to distribute to and from all of the Industrial Internet of Things sensors information into five layers, with each layer being defined by a modular computational block as illustrated in FIG. 3. These computational layers are respectively in series in the illustrated hierarchy as follows:

    • Fog 1: This layer of the protocol of the invention includes high frequency vector data such as vibration information, current information, magnetic field information and the like, which is extracted from associated sensors connected to one or more machines in a factory with which the edge cloud of interest is associated. The vector data is transformed into either the time domain or the frequency domain.
    • Fog 2: This layer of the protocol of the invention involves useful statistical and mathematical data, including moments, peak to root mean square ratio and other parameters of physical data that are extracted from a transformed data set created in the Fog 1 layer.
    • Core 1: This layer of the protocol of the invention involves data for all the features for sensors using either a rule based or a machine learning engine approach and processing the same with a support vector machine or a random forest software, to predict reliability and provide analytics on the machine(s) on which the sensors reside.
    • Core 2: This layer of the protocol of the invention involves the analytics obtained for one or more machine from the Core 1 layer to form secondary layers of analytics such as machine repair and operation data, which may be provided on a dashboard or some other display device.
    • Core 3: This layer of the protocol of the invention involves data obtained from the layer identified as Core 2 and is further used to process more advanced enterprise resource planning software in order to evaluate parameters such as overall uptime or downtime performance of a factory or other facility.


In the practice of the invention each of the distributed computational layers described above requires three different data types. One of these data types is machine information or sensor information regarding which sensors are mounted on what kind of machines, the make or model of the machine, and the analytics required. This asset database includes unstructured text, image, and sound data captured from a machine for adaptive boosting of the analytics.


A second data type needed by each of the distributive computational layers is time series meta data processed from an earlier block in real time. So, as an example referring to FIG. 3, layer C receives metadata from computation in layer B., computation layer D receives metadata from computational layer C, and so on.


The third data type needed by each of the distributive computation layers is time series metadata stored from each block in a sensor time series database, with the data being from the relatively recent past. In the industrial and commercial contexts typically this will be data from the last two hours or two days of operation of the facility.


Each of the computational layers receives data via a broker service.


In the course of practice of the invention, data input to the edge cloud of interest can be raw sensor data, without layer A or layer B processing, or can be metadata generated by a computation in layer B. If a sensor with a Fog device is connected to the edge cloud, the sensor will send metadata directly to computing layer C for use thereby. Otherwise, raw sensor data is processed by and within layer A.


In the course of practice of the invention, metadata output from processing layers C, D, and E are preferably sent to the public cloud, and to a programmable logic controller/supervisory control and data acquisition system.


Raw sensor data is preferably input directly to edge cloud in one embodiment of the invention without processing by computation layers A and B. Alternatively, raw sensor data is metadata generated by computational layer B and then supplied to the edge cloud. In the embodiment of the invention where a sensor with a Fog device is connected to the edge cloud of interest, the invention sends metadata directly to layer C for processing. Otherwise in the preferred practice of the invention, raw data is processed in the edge cloud as it is received from computational layer A.


In the course of practice of the invention, metadata output from processing layers C, D, and/or E is sent to the public cloud, or to a programmable logic controller/supervisory control and data acquisition system, or to a hybrid programmable logic controller/supervisory control and data acquisition system, and/or to a real time listening service. Time series metadata is sent to be stored at a time series data base locally in the associated edge cloud. This time series database is synchronized and backed up with the time series database of the public cloud so that in the event of damage to the particular localized edge cloud of interest, no data is lost.


In the course of practice of the invention, visualization data, which can be JSON formatted data as required for analytic visualization, is sent to a visualization database in the public cloud.


The visualization data in another format is preferably sent to a mobile or other visualization device within the particular factory; these devices are preferably connected to the same subnetwork within the particular factory.


Yet another format of the visualization data, which will be formatted for an industrial bus, is preferably sent to the hybrid programmable logic controller/supervisory controller and data acquisition system.



FIG. 4 illustrates some architecture of the public cloud including a block identified as “virtual edge cloud,” which is similar to edge cloud except that with this block data does not need to be fed to a connected mobile device or to a hybrid programmable logic controller/supervisory control and data acquisition system, due to lower latency. Additionally, the sensor time series data base and the asset data base that are in or constitute the virtual edge cloud, in the preferred practice of the invention store such time series data, metadata and asset data of the sensors that are connected and report to the virtual edge cloud, which may be generated from the computations performed in block C and/or block D and/or block E of FIG. 3.


The sensors comprise both “brown” sensors and “green” sensors, where “brown” denotes sensors lacking computational capability and associated electronics and “green” denotes sensors having computational capability with associated electronics being either built into the sensor or located immediately adjacent thereto as respecting the machine from which the sensor is harvests data.


The architecture illustrated in FIG. 4 also performs back-up for all of the sensor time series database and asset database data generated in the edge cloud, for use in the event the edge cloud is damaged by hardware being damaged, or stolen, or misplaced, or taken away from the network by disconnection. Visualization data generated by and within computational layer C, and/or D, and/or E is fed directly to a visual database of the public cloud and then to a non-real time visualization device as requested by end users.


Further in the practice of the invention, the asset data base not only stores all of the information about the machines required to build the analytic model provided in blocks C, D and E of FIG. 3; the asset data base also stores all of the feedback information and data given by the users respecting the machines, in the form of texts, images and sound captured by a mobile phone or tablet. Such mobile phone or tablet optionally uses either the “on-premise” mode connected to the edge cloud or operates in a non-real time mode connected to the public cloud; either mode of operation is within the scope of the invention. Feedback by the users is used for adaptive modeling of predictive maintenance in computational layers C and D in FIG. 3. The feedback information is further processed by a feedback processing microservice in the public cloud and is desirably synchronized with the C and D layers of the edge cloud by the asset data base abbreviated “Asset DB” in FIG. 4.


Although schematic implementations of present invention and at least some of its advantages are described in detail hereinabove, it should be understood that various changes, substitutions and alterations may be made to the apparatus and methods disclosed herein without departing from the spirit and scope of the invention as defined by the appended claims. The disclosed embodiments are therefore to be considered in all respects as being illustrative and not restrictive with the scope of the invention being indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Moreover, the scope of this patent application is not intended to be limited to the particular implementations of apparatus and methods described in the specification, nor to any methods that may be described or inferentially understood by those skilled in the art to be present as described in this specification.


As disclosed above and from the foregoing description of exemplary embodiments of the invention, it will be readily apparent to those skilled in the art to which the invention pertains that the principles and particularly the compositions and methods disclosed herein can be used for applications other than those specifically mentioned. Further, as one of skill in the art will readily appreciate from the disclosure of the invention as set forth hereinabove, apparatus, methods, and steps presently existing or later developed, which perform substantially the same function or achieve substantially the same result as the corresponding embodiments described and disclosed hereinabove, may be utilized according to the description of the invention and the claims appended hereto. Accordingly, the appended claims are intended to include within their scope such apparatus, methods, and processes that provide the same result or which are, as a matter of law, embraced by the doctrine of the equivalents respecting the claims of this application.


As respecting the claims appended hereto, the term “comprising” means “including but not limited to”, whereas the term “consisting of” means “having only and no more”, and the term “consisting essentially of” means “having only and no more except for minor additions which would be known to one of skill in the art as possibly needed for operation of the invention.” The scope of the invention is, therefore, indicated by the appended claims rather than by the foregoing description and all changes which come within the range of equivalency of the claims are to be considered to be embraced within the scope of the claims. Additional objects, other advantages, and further novel features of the invention will become apparent from study of the appended claims as well as from study of the foregoing detailed discussion and description of the preferred embodiments of the invention, as that study proceeds.

Claims
  • 1. A method of evaluating factory production machinery up time and down time performance, comprising: a) providing a collection of sensors in individual communication with factory production machinery, each sensor collecting high frequency vector data as respecting a physical parameter associated with the machinery;b) extracting the data from the sensors to produce a sensor data set;c) transforming the data set into the frequency domain;d) extracting statistical and mathematical information from the data set;e) transferring data set, and optionally the extracted statistical and mathematical information, to an associated virtual edge cloud within a public cloud;within the associated virtual edge cloud, processing the data set using a rule based algorithm to predict machine reliability and future performance;g) within a second section of the associated virtual edge cloud, performing further analytical computations on the data set to provide a repair, maintenance and operation board for the machinery;h) within a third section of the associated virtual edge cloud, performing yet further analytical computations on the data set to evaluate up time and down time performance for the factory production machinery;i) storing time series data, metadata, and/or asset data from the sensors in a time series data base and an asset data base within the virtual edge cloud;j) feeding visualization data generated within the first, second, and third sections of the virtual edge cloud to a visualization data base within the public cloud; andk) storing user feedback information and data regarding the factory production machinery in the asset data base.
  • 2. A method of evaluating factory production machinery performance, comprising: a) providing a collection of sensors in communication with factory production machinery, each sensor collecting high frequency vector data as respecting physical parameters of the machinery;b) extracting the data from the sensors to produce a sensor data set;c) transforming the data set into the time domain;d) extracting statistical and mathematical information from the data set;e) transferring data set, and optionally the extracted statistical and mathematical information, to an associated virtual edge cloud within a public cloud;within the associated virtual edge cloud, processing the data set for one machine using a rule based algorithm to predict machine reliability and future performance;g) within a second section of the associated virtual edge cloud, performing further analytical computations on the data set to provide a repair, maintenance and operation board for the machine;h) within a third section of the associated virtual edge cloud, performing analytical computations on the data set to evaluate performance for the machine;i) storing time series data, metadata, and/or asset data from the sensors in a time series data base and an asset data base within the virtual edge cloud;j) feeding visualization data generated within the first, second, and third sections of the virtual edge cloud to a visualization data base within the public cloud; andk) storing user feedback information and data regarding the factory production machinery in the asset data base.
CROSS REFERENCE TO RELATED PATENT APPLICATIONS

This patent application claims the priority under 35 USC 120 of U.S. Provisional Application Ser. No. 62/608,705 filed 21 Dec. 2017 in the name of Biplab Pal and entitled “Dryer Maintenance Prediction Method and Apparatus”.

US Referenced Citations (122)
Number Name Date Kind
4023940 Shultz May 1977 A
4131011 Ling Dec 1978 A
5150289 Badavas Sep 1992 A
5487225 Downie Jan 1996 A
5610339 Haseley et al. Mar 1997 A
5825338 Salmon et al. Oct 1998 A
5995561 Yamasaki et al. Nov 1999 A
6104987 Farnsworth Aug 2000 A
6289606 Gillette et al. Sep 2001 B2
6330525 Hays Dec 2001 B1
6405108 Patel et al. Jun 2002 B1
7406399 Furem et al. Jul 2008 B2
7882394 Hosek et al. Feb 2011 B2
7938935 MacHattie et al. May 2011 B2
8021462 Moretto Sep 2011 B2
8094034 Patel et al. Jan 2012 B2
8112381 Yuan et al. Feb 2012 B2
8126574 Discenzo et al. Feb 2012 B2
8150340 Albsmeier et al. Apr 2012 B2
8334784 Patel et al. Dec 2012 B2
8390299 Laepple et al. Mar 2013 B2
8405940 Schweitzer, III et al. Mar 2013 B2
8421475 Thiim Apr 2013 B2
8433443 Hagerty et al. Apr 2013 B2
8560368 Maity et al. Oct 2013 B1
8571904 Guru et al. Oct 2013 B2
8726535 Garrido et al. May 2014 B2
8868242 Loutfi Oct 2014 B2
8920078 Woolever Dec 2014 B2
9052216 Kamel et al. Jun 2015 B2
9062536 Fischer Jun 2015 B2
9250275 Patel et al. Feb 2016 B2
9781243 Huang Oct 2017 B1
10041844 Brady Aug 2018 B1
20010038345 Satoh et al. Nov 2001 A1
20020143421 Wetzer Oct 2002 A1
20040102924 Jarrell May 2004 A1
20040176926 Edie Sep 2004 A1
20040199573 Schwarz et al. Oct 2004 A1
20050049834 Bottomfield Mar 2005 A1
20050222794 Baird et al. Oct 2005 A1
20060137105 Hong et al. Jun 2006 A1
20060168195 Maturana et al. Jul 2006 A1
20060208169 Breed et al. Sep 2006 A1
20060276949 Beck et al. Dec 2006 A1
20070100518 Cooper May 2007 A1
20070185685 Lannes et al. Aug 2007 A1
20070193056 Switalski Aug 2007 A1
20080103732 Stoupis May 2008 A1
20080109185 Cheung et al. May 2008 A1
20080289045 Fryer Nov 2008 A1
20080294382 Lim Nov 2008 A1
20090024359 Bibelhausen et al. Jan 2009 A1
20090043518 Roh et al. Feb 2009 A1
20090119243 Yuan et al. May 2009 A1
20100023307 Lee Jan 2010 A1
20100169030 Parlos Jul 2010 A1
20100199352 Hill et al. Aug 2010 A1
20100295692 Bjorn Nov 2010 A1
20110016199 De Carlo et al. Jan 2011 A1
20110131398 Chaturvedi et al. Jun 2011 A1
20110137697 Yedatore et al. Jun 2011 A1
20110216805 Fernando et al. Sep 2011 A1
20110307220 Lacaille Dec 2011 A1
20120045068 Kim et al. Feb 2012 A1
20120166142 Maeda et al. Jun 2012 A1
20120209569 Becourt et al. Aug 2012 A1
20120213098 Sun Aug 2012 A1
20120271576 Kamel Oct 2012 A1
20120290104 Holt et al. Nov 2012 A1
20120330499 Scheid et al. Dec 2012 A1
20120330614 Kar Dec 2012 A1
20130102284 Storozuk Apr 2013 A1
20130119047 Driussi May 2013 A1
20130170417 Thomas et al. Jul 2013 A1
20130173178 Poczka et al. Jul 2013 A1
20130201316 Binder et al. Aug 2013 A1
20130268469 Sharma et al. Oct 2013 A1
20130287060 Langdoc et al. Oct 2013 A1
20130304677 Gupta et al. Nov 2013 A1
20130318022 Yadav et al. Nov 2013 A1
20140129164 Gorbold May 2014 A1
20140132418 Lill May 2014 A1
20140163416 Shuck Jun 2014 A1
20140186215 Shinta et al. Jul 2014 A1
20140207394 Madden Jul 2014 A1
20140223767 Arno Aug 2014 A1
20140244836 Goel et al. Aug 2014 A1
20140262130 Yenni Sep 2014 A1
20140309805 Ricci Oct 2014 A1
20140314284 Movellan et al. Oct 2014 A1
20140335480 Asenjo et al. Nov 2014 A1
20140336791 Asenjo et al. Nov 2014 A1
20140337429 Asenjo et al. Nov 2014 A1
20150026044 Refaeli Jan 2015 A1
20150039250 Rank Feb 2015 A1
20150094914 Abreu Apr 2015 A1
20150139817 Kowalski May 2015 A1
20150181313 Murphy Jun 2015 A1
20150185251 Heydron et al. Jul 2015 A1
20150233792 Gao Aug 2015 A1
20150233856 Samuilov et al. Aug 2015 A1
20150247670 Robertson et al. Sep 2015 A1
20150261215 Blevins Sep 2015 A1
20160086285 Jordan Peters et al. Mar 2016 A1
20160147205 Kaufman May 2016 A1
20160189440 Cattone Jun 2016 A1
20160209831 Pal Jul 2016 A1
20160245279 Pal et al. Aug 2016 A1
20160245686 Pal et al. Aug 2016 A1
20160245765 Pal Aug 2016 A1
20160291552 Pal et al. Oct 2016 A1
20160299183 Lee Oct 2016 A1
20160313216 Pal et al. Oct 2016 A1
20160349305 Pal Dec 2016 A1
20170006135 Siebel Jan 2017 A1
20170032281 Hsu Feb 2017 A1
20170060574 Malladi et al. Mar 2017 A1
20170061608 Kim et al. Mar 2017 A1
20170163444 McLaughlin et al. Jun 2017 A1
20170201585 Doraiswamy et al. Jul 2017 A1
20200067789 Khuti Feb 2020 A1
Foreign Referenced Citations (21)
Number Date Country
201672991 Dec 2010 CN
102539911 Jul 2012 CN
103399486 Nov 2013 CN
203362223 Dec 2013 CN
203588054 May 2014 CN
104036614 Sep 2014 CN
1836576 Feb 2012 EP
2186613 May 2013 EP
2648393 Oct 2013 EP
WO 2005086760 Sep 2005 WO
WO 2010104735 Sep 2010 WO
WO 2013040855 Mar 2013 WO
WO 2013-041440 Mar 2013 WO
WO 2013093942 Jun 2013 WO
WO 2014044906 Mar 2014 WO
WO 2014085648 Jun 2014 WO
WO 2014089567 Jun 2014 WO
WO 2014117245 Aug 2014 WO
WO 2015022036 Feb 2015 WO
WO 2016137848 Sep 2016 WO
WO 2017-1234525 Jul 2017 WO
Non-Patent Literature Citations (28)
Entry
International Search Report and Written Opinion for PCT Application No. PCT/US16/18820; dated Aug. 4, 2016.
International Search Report and Written Opinion for PCT Application No. PCT/US15/066547; dated Mar. 17, 2016.
Sensors Drive Mobile IoT; Wong, William; Jan. 26, 2015; Electronic Design.
International Search Report and Written Opinion for PCT Application No. PCT/US16/028724; dated Aug. 22, 2016.
International Search Report and Written Opinion for PCT Application No. PCT/US16/18831; dated Aug. 12, 2016.
Fault Detection in Kerman Combined Cycle Power Plant Boilers by Means of Support Vector Machine Classifier Algorithms and PCA by M. Berahman, et al., 3rd International Conference on Control, Instrumentation, and Automation (ICCIA 2013), Dec. 28-30, 2013, Tehran, Iran.
Fault Monitoring and Diagnosis of Induction Machines Based on Harmonic Wavelet Transform and Wavelet neural Network by Qianjin Guo, et al., dated Sep. 2008, published at the Fourth International Conference on Natural Computation.
Fault Diagnosis Method Study in Roller Bearing Based on Wavelet Transform and Stacked Auto-encoder, by Junbo Tan, et al., dated Feb. 2015, published by IEEE.
A Diagnostic Expert System Embedded in a Portable Vibration Analysis Instrument by Dr. Robert Milne, et al., dated May 13, 1991, published at IEE Colloquium on Intelligent Instrumentation.
Detection of Precursor Wear Debris in Lubrication Systems by Jack Edmonds, et al., dated May 2000, published by IEEE.
Fault Diagnosis of Bearing Based on Fuzzy Support Vector Machine, by Haodong Ma, et al., dated Jan. 2015, published by IEEE.
Investigation of the Mechanical Faults Classification using Support Vector Machine Approach by Zhiqiang Jiang, et al., dated Aug. 2010, 2010 Second International Conference on Intelligent Human-Machine Systems and Cybernetics.
Impact Characterization of Multiple-Points-Defect on Machine Fault Diagnosis by Muhammad F. Yaqub, et al., 8th IEEE International Conference on Automation Science and Engineering, Aug. 20-24, 2012, Seoul, Korea.
Detection of Generalized-Roughness Bearing Fault by Spectral-Kurtosis Energy of Vibration or Current Signals by Fabio Immovilli, et al., IEEE Transations on Industrial Electronics, vol. 56, No. 11, Nov. 2009.
Intrinsic Mode Function Determination of Faulty Rolling Element Bearing Based on Kurtosis by Wei Kang, et al., Proceeding of the 2015 IEEE International Conference on Information and Automation, Lijiang, China, Aug. 2015.
Condition Monitoring and Fault Diagnosis of Rolling Element Bearings Based on Wavelet Energy Entropy and SOM by Shuai Shi, et al., dated Aug. 2012, published by IEEE.
Continuous Hidden Markov Model Based Gear Fault Diagnosis and Incipient Fault Detection by Jian-She Kang, et al., dated Jun. 2011, published by Institute of Electrical and Electronics Engineers (IEEE).
Study on Fault Diagnosis of Gear with Spall using Ferrography and Vibration Analysis by Wei Feng, et al., published in Aug. 2009 at the International Conference on Measuring Technology and Mechatronics Automation.
International Search Report and Written Opinion for PCT Application No. PCT/US2016/067814; dated Apr. 6, 2017.
International Search Report and Written Opinion for PCT Application No. PCT/US2016/067546; dated Apr. 11, 2017.
Krishnamurthy, S. et al. (2008) Automation of Facility Management Processes Using Machine-to-Machine Technologies. In: Floerkemeier C., Langheinrich M., Fleisch E., Mattern F., Sarma S.E. (eds) The Internet of Things. Lecture Notes in Computer Science, vol. 4952. DOI:10.1007/978-3-540-78731-0_5 (Year: 2008).
Holler, J. et al. (2014). “From Machine-to-machine to the Internet of Things: Introduction to a New Age of Intelligence.” Chapters 2, 4, 5, 7, 10, 12. Academic Press. DOI:10.1016/B978-0-12-407684-6.00002-4 (Year: 2014).
Azure IoT Edge open for developers to build for the intelligent edge, George, Sam; Azure Internet of Things; Nov. 15, 2017.
Predix Edge Technology Product Brief, General Electric, 2017.
http://ieeexplore.ieee.org/document/8089336/ Future Edge Cloud and Edge Computing for Internet of Things Applications—Janali Pan et al.
Challenges and Solutions of Protecting Variable Speed Drive Motors; Aversa, et al.; Feb. 11, 2013; Presented at the 2013 Texas A&M Conference for Protective Relay Engineers.
Dec. 21, 2018 Non-Final Office Action against Applicant's co-pending U.S. Appl. No. 14/833,111.
Vasudevan, Shobha, “Still a Fight to Get It Right: Verification in the Era of Machine Learning”, 2017 IEEE International Conference on Rebooting Computing (ICRC), Washington, DC, 2017, pp. 1-8. (Year: 2017).
Related Publications (1)
Number Date Country
20190265687 A1 Aug 2019 US
Provisional Applications (1)
Number Date Country
62608705 Dec 2017 US