Quickly Enable Predictive Maintenance Using Amazon Monitron, an End-to-End System
Quickly Enable Predictive Maintenance Using Amazon Monitron, an End-to-End System
Successfully implementing predictive maintenance requires installing sensors and gateways in order to collect data from your industrial equipment, and building a machine learning (ML) model that analyzes your sensor data to give you valuable insights. However, companies have historically needed skilled technicians and data scientists to piece together these complex solutions from scratch. In this Tech Talk, we will introduce you to Amazon Monitron, an end-to-end system that uses machine learning (ML) to monitor the condition of your equipment and detect abnormal behavior, enabling you to implement predictive maintenance and reduce unplanned downtime. Amazon Monitron includes wireless sensors to capture vibration and temperature data from equipment, a gateway device to securely transfer data to AWS, the Amazon Monitron service that analyzes the data for abnormal equipment conditions using machine learning, and a companion mobile app to set up the devices and receive reports on operating behavior and alerts to potential failures in your machinery. You can start monitoring equipment condition in minutes without any development work or ML experience required.
Learning Objectives:
*Learn how you can install wireless Amazon Monitron Sensors and Gateways using the Amazon Monitron app and start monitoring your equipment in minutes
*Learn how Amazon Monitron secures the sensors, gateways, and the communication between them and the Amazon Monitron service, and encrypts your data at rest and in transit
*Learn how Amazon Monitron continually improves over time, by allowing your technicians to easily enter feedback on the alerts in the mobile app
***To learn more about the services featured in this talk, please visit: https://aws.amazon.com/monitron Subscribe to AWS Online Tech Talks On AWS:
https://www.youtube.com/@AWSOnlineTec…
Follow Amazon Web Services:
Official Website: https://aws.amazon.com/what-is-aws
Twitch: https://twitch.tv/aws
Twitter: https://twitter.com/awsdevelopers
Facebook: https://facebook.com/amazonwebservices
Instagram: https://instagram.com/amazonwebservices
☁️ AWS Online Tech Talks cover a wide range of topics and expertise levels through technical deep dives, demos, customer examples, and live Q\u0026A with AWS experts. Builders can choose from bite-sized 15-minute sessions, insightful fireside chats, immersive virtual workshops, interactive office hours, or watch on-demand tech talks at your own pace. Join us to fuel your learning journey with AWS.
#AWS
Content
0.16 -> hello everyone i'm karthik dharan
3.28 -> i'm a senior product manager on amazon
5.279 -> monitoring
6.64 -> today i'm here to share with you how you
9.2 -> can quickly enable
10.4 -> predictive maintenance for your
11.92 -> industrial equipment using amazon
13.84 -> monitoring
14.719 -> an end-to-end system
18.56 -> imagine an industrial site be it a
20.8 -> manufacturing plant
22.08 -> or a fulfillment warehouse typically
25.359 -> such a site has hundreds of equipment
27.92 -> that drive complex processes
30.8 -> when such an equipment fails
32.16 -> unexpectedly it is a nightmare
35.28 -> it will fire fight and fix it on
37.28 -> priority
38.879 -> at the same time you incur significant
41.12 -> cost from
42.239 -> lost productivity and this is a common
45.76 -> and costly affair in the industrial
47.6 -> world
49.2 -> 82 of the industrial companies
52 -> experience have experienced some sort of
53.76 -> sudden breakdowns
54.96 -> in the past few years experiencing
57.52 -> anywhere between five and twenty percent
59.199 -> reduction of productive capacity
61.76 -> and annual cost of 50 billion dollars
65.76 -> to counter this there's a growing shift
68 -> to predictive maintenance strategy
70.72 -> let us look at this chart that shows
73.76 -> the the equipment uptime with in
76.24 -> relation to
77.04 -> time to detecting failure as you see
80.72 -> the earlier you detect failures the
83.04 -> higher the equipment of time
85.6 -> now let us overlay the different
87.52 -> maintenance strategies on top of this
90 -> chart
93.36 -> quite a few customers adopt a reactive
95.68 -> maintenance strategy
97.2 -> in this strategy you fix an equipment
99.68 -> when it breaks down
102.24 -> because you have to stop your processes
104.32 -> in order to fix the equipment
106.24 -> you lose crucial operating hours during
108.799 -> the fix
109.36 -> fixing process this means a reactive
112.799 -> strategy is typically associated
114.88 -> with that with a low equipment of time
118.479 -> to offset this customers adopt
122.32 -> a planned or preventative maintenance
124.159 -> strategy
125.439 -> in this strategy customers schedule
128.239 -> maintenance activities
129.679 -> periodically for example every three
132 -> months or every six months
134.08 -> the benefit of scheduling periodic
136.319 -> activities is you can catch failures
138.48 -> during these maintenance activities
141.36 -> however
142.4 -> there are downsides failures can still
145.28 -> happen
145.76 -> outside of these scheduled planned
147.92 -> maintenance activities
149.68 -> which you may not be able to catch at
152.16 -> the same time
153.12 -> because you are performing maintenance
155.84 -> based on time and not based on the
157.76 -> actual condition of the equipment
160 -> you you may end up over maintaining your
162.16 -> machine which may also not be good
165.2 -> to address these challenges few
167.12 -> customers
168.4 -> have started adopting a predictive
170.08 -> maintenance strategy
171.76 -> in this strategy you only schedule
174.16 -> maintenance depending on the need
176.08 -> and this is done by tracking the
177.84 -> condition of the equipment
180.239 -> predictive maintenance has multiple
182 -> components
183.68 -> at the very first at the very foundation
185.84 -> of it is anomaly detection
188.159 -> here you track the condition of the
190.159 -> equipment you're monitoring
191.76 -> and detect anomalies in this condition
194 -> to take actions based on them
196.64 -> next is diagnostics here you detect the
199.68 -> root cause
200.72 -> that is a precursor indicator of a
202.8 -> failure
203.76 -> this could either be lack of lubrication
206.159 -> or misalignment
207.44 -> that could result in an impending
209.44 -> failure
211.92 -> third you can also do predictions you
214.159 -> can estimate
215.2 -> the remaining useful life of the
216.879 -> equipment
218.4 -> this has two benefits one it allows you
221.44 -> to stretch the operating life of that
223.519 -> specific equipment
224.959 -> till its very end second it allows you
228.08 -> to schedule maintenance
229.519 -> at a time that's suitable to you and
232.319 -> finally
232.959 -> closing it all out as part of the
235.04 -> predictive maintenance strategy
236.56 -> you schedule a maintenance at a specific
239.12 -> point in time
240.159 -> ahead of its impending failure
245.519 -> data indicates that 80 percent of
247.28 -> maintenance activities
248.56 -> are based on time today however only 11
252.08 -> percent
252.799 -> of the failures that happen with
254.56 -> industrial equipment are related to age
257.84 -> this clearly signifies a need for
260.56 -> adopting a predictive maintenance
262 -> strategy
263.52 -> the good news here is technology has
266.4 -> been evolving
267.36 -> we now have affordable iot and ai
269.759 -> technologies
270.56 -> that can detect changes in the condition
272.96 -> and enable predictive maintenance
276.16 -> that said predictive maintenance
279.04 -> solutions are not easy to implement
282.479 -> imagine you had to monitor an equipment
284.56 -> at your industrial site
287.04 -> you would have to start by retrofitting
289.12 -> sensors to this equipment
291.759 -> if you look at sensors out in the market
293.919 -> most of them are quite expensive and
296 -> they require you to invest in complex
297.84 -> cablings
299.52 -> after you retrofit these sensors you'll
301.52 -> have to have to you'll have to hire
303.44 -> a hardware engineer to write code on the
305.759 -> microcontroller
306.88 -> and connect that sensor to the iot core
309.759 -> you'll also have to do this
311.12 -> in a secure way in an age of increasing
313.6 -> connectivity
317.039 -> collecting and storing data is just the
319.28 -> beginning of the puzzle
321.44 -> the core piece is to interpret the data
324.32 -> so that you can take actions
325.759 -> based on it in order to interpret the
328.479 -> data
329.199 -> for anomalies and to do it accurately
331.68 -> you may have to apply sophisticated
333.52 -> techniques
334.4 -> including machine learning to do this
337.199 -> you'll have to invest
338.4 -> in machine learning or data science
340.4 -> science talent
343.6 -> finally it's not just you have to detect
346.56 -> these
346.96 -> issues but you have to inform your end
349.28 -> users of these
350.16 -> issues end users in this case being
352.4 -> technicians and reliability managers
354.24 -> that are going to
355.6 -> maintain these machines you'll have to
358.72 -> require that
359.52 -> this these technicians with capabilities
362.16 -> to track these issues
363.919 -> to look at historical data and to take
366.08 -> actions based on them
368.72 -> all these steps mean that it takes
371.44 -> months
371.919 -> if not year if not a year to
375.039 -> even do pilot in in predictive
377.44 -> maintenance
378.8 -> and it incurs hundreds of thousands of
380.639 -> dollars of cost
382.319 -> depending on the scale
385.919 -> this practically puts predictive
387.759 -> maintenance out of the reach
389.6 -> of most industrial customers
393.199 -> to address these challenges we built
395.12 -> amazon monetron
396.479 -> an end-to-end system for equipment
398.56 -> monitoring
400.08 -> by making it easy quick and cost
403.44 -> effective
404.479 -> we are rema re-imagining how we can do
407.12 -> how we can enable predictive maintenance
408.8 -> for industrial equipment
412.479 -> amazon monitoring is an out of the box
414.56 -> system to enable predictive maintenance
417.759 -> it includes everything you need to get
420.319 -> started with predictive maintenance
422.56 -> it comes with monotron sensors the
425.039 -> sensors they capture the vibration and
426.8 -> temperature data
428.16 -> of the equipment this data is then
430.72 -> transferred over
432.08 -> to a local gateway which forwards the
434.08 -> data to the monitoring service
437.12 -> in the service we run we analyze the
439.36 -> sensor data
440.319 -> using a combination of vibration iso
442.639 -> standards and machine learning
445.68 -> these insights are then delivered over
447.84 -> to the monitoring app
449.28 -> which is used for two purposes precisely
451.759 -> one
452.319 -> to first in the first place configure
454.319 -> and provision these
455.44 -> devices that is the sensors and the
457.36 -> gateways and second
459.12 -> to to consume the insights coming from
461.28 -> the monitor on service
462.639 -> and to take actions based on these
464.4 -> insights
468.72 -> monotron can be used to monitor a
470.879 -> variety of rotating equipment
472.96 -> these include motors gearboxes pumps
477.68 -> compressors and fans these rotating
481.039 -> equipment
482.4 -> they they are critical components of
484.319 -> most complex processes
486.96 -> vibration and temperature are key
489.12 -> leading indicators
490.319 -> of the health of these rotating
491.599 -> equipment by allowing you
494.08 -> to monitor which is a combination of to
497.199 -> measure
497.68 -> and analyze the vibration and
499.039 -> temperature data coming from this
500.8 -> equipment
501.599 -> we are equipping you to enable
503.68 -> predictive maintenance for this
505.28 -> rotating equipment
510.319 -> let us look at a few key features of
512.24 -> amazon monitoring
516 -> first it comes with fully managed low
518.479 -> cost wireless sensors and gateways
521.839 -> the sensors are wireless which means you
524.64 -> don't have to invest
525.68 -> in complex cabling to get them up and
527.92 -> running
529.12 -> both the sensors and the gateways are
531.44 -> fully managed
533.12 -> they are pre-configured to run with the
535.04 -> monitor on service
536.72 -> this means yeah you do not have to write
539.36 -> a single line of code
541.04 -> in order to get started with monotron
544.959 -> monotron comes pre-built with machine
547.279 -> learning and iso standards-based
549.36 -> analytics
551.36 -> we have taken years of experience of
553.6 -> maintaining machines
555.279 -> in our fulfillment centers and put that
557.68 -> knowledge
558.48 -> into amazon monitoring with
561.6 -> with monetron you get out of the box
563.36 -> experience for these analytics
567.519 -> monitoring delivers timely notifications
569.68 -> in the monitor on mobile application
572.32 -> using these notifications technicians
575.04 -> and reliability managers
576.56 -> can take actions and fix
580.08 -> those those issues that come from those
581.92 -> machines
585.36 -> finally monitoring has support for
587.92 -> adding machine learning feedback
589.519 -> within the app with just a few taps
593.2 -> in the app technicians can add feedback
596.48 -> on the alerts they received
598.32 -> they can give a thumbs up if the alert
600.08 -> was useful or a thumbs down if the
601.839 -> outlet wasn't
603.36 -> monotron uses this feedback to
605.2 -> continuously train
606.399 -> the underlying machine learning model to
608.48 -> improve the ongoing performance of these
610.48 -> alerts
613.2 -> let us look at some of the benefits of
614.88 -> amazon monitoring
617.12 -> first and foremost monitoring is built
619.92 -> to help you reduce
621.12 -> unplanned downtime for your industrial
623.04 -> equipment
624.56 -> by by by delivering early insights
627.839 -> into the condition of the equipment
629.6 -> monitoring equips you
631.04 -> to fix it before it fails and helps you
633.92 -> reduce
634.88 -> costly cost associated with unpanned
637.92 -> downtime
640.32 -> second monitoring is quick to install
642.64 -> and no special skills are needed
644.72 -> to use it whatsoever using the monitor
648.24 -> on app
648.72 -> you can you can provision the monitoring
650.72 -> sensors and gateways in a matter of
652.8 -> minutes
654.32 -> no knowledge of machine learning or
656.32 -> technical expertise
657.68 -> is required to set up and use monotron
662 -> third monotron has low setup costs
666.079 -> you have a very small upfront device
668.32 -> cost associated with monotron which is a
670.56 -> fraction of the cost of using
672.399 -> conventional expensive conventional
674.32 -> sensors
676 -> monitor on service fee is pay as you go
678.959 -> so you only pay
680.16 -> for the amount of time you use the
681.76 -> service
683.44 -> there are no licensing fees or long term
686.839 -> commitments
688.56 -> is also end-to-end secure
691.68 -> that monotone provides end-to-end
693.92 -> encryption of the data at rest
695.68 -> and in transit the device the service
699.04 -> and the communication between them is
701.2 -> fully secure
703.76 -> last but not the least monitor monetron
706.959 -> offers continuous improvements
709.519 -> this is on two fronts one on the device
712.48 -> front
713.68 -> and one on the machine learning side on
716.399 -> the device front
717.6 -> both the sensors and the gateways can be
720.48 -> remotely upgraded
721.76 -> with newer software using over the air
724.72 -> upgrades
726.24 -> what this means is if we come up with
728.639 -> improved
729.36 -> software for these devices we can push
731.92 -> this onto the devices in the field
734 -> and the devices can benefit from
735.76 -> improving capabilities
737.76 -> as an example we recently increased the
740.32 -> life the battery life of our sensors
742.16 -> from three to five
743.279 -> years by running optimizations on the
745.839 -> software
747.76 -> the devices in the field were also able
749.92 -> to benefit from this increase in battery
751.68 -> life
752.16 -> due to the over-the-air upgrades
755.36 -> similar to the devices on the machine
757.12 -> learning side as well
758.399 -> monitoring offers continuous
759.92 -> improvements
762.16 -> monitoring uses the user feedback that
764.72 -> is provided within the app
766.399 -> to continuously retrain the underlying
768.399 -> machine learning models
769.839 -> with this on delivers improving alert
772.959 -> notifications
774.16 -> so that end users can take
777.6 -> can can can use these alerts in an in an
780.079 -> improving fashion
783.76 -> now that we have seen an overview of
785.2 -> monitron let us look at a short demo
788.399 -> to see how you can set up and use
790.079 -> monotron
792.8 -> let's say you are a food and beverage
794.399 -> company and you decide to monitor a pump
796.72 -> at your dallas site to get started you
799.519 -> order the amazon monitoring sensors and
801.44 -> gateways on amazon.com
803.839 -> let's assume you received these devices
806.72 -> next
807.279 -> you want to set them up in order to
808.959 -> monitor your equipment
810.72 -> you go to the aws console and search for
813.12 -> amazon monitoring
818.16 -> and launch the service
823.04 -> you start by creating a project a
825.839 -> project is where you register sites
827.68 -> and assets that you want to monitor you
829.92 -> enter a name for your project
831.839 -> typically you want to enter your company
833.839 -> name here let's say
835.12 -> company xyz
840.24 -> next you want to add an administrator to
842.56 -> this project
843.6 -> this could be the reliability manager at
845.6 -> your dallas site
846.959 -> let's call her jane doe you can create
849.6 -> jane as a user
864.8 -> next you add chain to the project
870.24 -> as you saw the project was successfully
872.24 -> created and jane was added as a user
875.12 -> next you invite jane to access this
877.36 -> project by sending her email
878.8 -> instructions
880.56 -> let's say you already did that once you
883.04 -> do this jen can use the monitor on app
885.6 -> to start monitoring the pump let's go
888.16 -> into the app to see how she would do
889.92 -> that
892.399 -> let's assume that following the
893.76 -> instructions in the invitation email
895.6 -> jin installs the monitor and android app
897.76 -> and logs into it
899.279 -> once inside the app jain can set up the
901.519 -> monitoring sensors and gateways and
903.36 -> start monitoring the pump
905.199 -> she starts off by installing a gateway
908.16 -> to do so she navigates to the menu
911.839 -> to the gateway section and adds a
914.079 -> gateway
915.12 -> while doing so she she makes sure that
917.199 -> the bluetooth of her phone
918.399 -> is turned on and she presses a
920 -> commissioning button on the side of the
921.68 -> gateway
922.959 -> the app looks for available gateways in
925.199 -> the vicinity
929.12 -> jain selects the gateway now the app
931.44 -> looks for
932.32 -> available wi-fi networks at the site
936 -> jn selects a suitable and secure
937.759 -> wireless network at the dallas site
939.68 -> and onboards the gateway on that site
946.399 -> the installation process of the gateway
948.32 -> is very similar to how you would install
950.079 -> a smart home assistant such as an echo
952 -> device at home
953.68 -> the function of the gateway is to
954.959 -> automatically and securely transfer the
956.959 -> sensor data to the cloud
958.639 -> it does so by communicating with the
960.32 -> sensor over bluetooth low energy
962.72 -> and with the cloud using wi-fi internet
965.6 -> a few moments later
966.88 -> the gateway is connected to the
968.16 -> monitoring service in the cloud
971.04 -> now that the gateway is connected jane
972.959 -> can register the pump and pair sensors
975.04 -> to it
976 -> to do so she goes to the asset page
979.839 -> here she adds a new asset she calls this
983.199 -> asset
983.759 -> pump
987.199 -> and identifies the iso 26 28 16 class of
990.32 -> the asset
994.48 -> now that the asset has been added jane
996.72 -> can pair sensors to one or more
998.24 -> positions
998.88 -> on this asset to do so she clicks
1002.88 -> pair sensor and identifies a name for
1005.199 -> the position
1007.36 -> let's say she wants to pair the pair the
1009.68 -> sensor to the motor of the pump
1016.72 -> next jane brings the smartphone close to
1018.959 -> the sensor
1020.88 -> amazon monitor uses near-field
1022.959 -> communication technology to seamlessly
1024.64 -> register the sensor in the cloud
1026.959 -> in a few seconds the sensor is spread to
1028.959 -> the position and goes online
1031.439 -> gen can now physically mount the sensor
1033.839 -> to the motor of the pump
1035.28 -> using the suggested adhesive once done
1038.4 -> the sensor starts capturing vibration
1040.16 -> and temperature data
1041.439 -> and monitoring will start monitoring the
1043.199 -> pump
1049.2 -> the first set of data on temperature and
1051.12 -> vibration already trickled in
1053.76 -> as you saw all these steps were simple
1056.16 -> and quick
1056.96 -> and jane was able to start monitoring
1058.72 -> her equipment in minutes
1060.72 -> jane did not have to write even a single
1062.799 -> line of code to start monitoring the
1064.84 -> pump
1067.12 -> once the sensor is installed amazon
1068.799 -> monitoring chains a machine learning
1070.32 -> model by baselining the vibration and
1072.08 -> temperature patterns of the equipment
1074.48 -> it sends an alert when it detects an
1076.08 -> anomaly based on this model
1078 -> these alerts are sent as push
1079.44 -> notifications and can be viewed in the
1081.76 -> bell on the top right corner
1084.24 -> as you see three alerts were received in
1086.72 -> the month of november
1089.44 -> when an alert is received jain can
1091.52 -> conduct physical investigations of the
1093.2 -> equipment and fix the issue
1095.12 -> in doing so jane can either click on the
1097.6 -> alert or navigate to the position under
1099.6 -> consideration
1101.12 -> to look for historical sensor
1102.72 -> measurements and draw insights
1105.039 -> once the issue is fixed jane can resolve
1108.16 -> the issue
1108.88 -> in the monotron app by clicking on the
1110.72 -> resolve button
1112.72 -> jin can also provide feedback at the
1114.64 -> same time by providing the closure codes
1116.64 -> which are failure mode failure cause and
1118.4 -> the actions taken
1120.48 -> if no issue was found in the equipment
1123.2 -> jain can also indicate that in the app
1126.559 -> when jain does that monitor treats it as
1129.039 -> a false alarm label for the machine
1130.559 -> learning model
1131.36 -> and retrains the model based on that
1134.72 -> as you see jane was able to leverage the
1137.36 -> power of machine learning and monitor
1139.039 -> the pump
1140.16 -> she required no machine learning
1141.6 -> knowledge or expertise and did not have
1144 -> to do any development work in doing so
1147.36 -> through this demo we saw how you can set
1149.52 -> up and use amazon monitoring to monitor
1151.76 -> your equipment
1153.76 -> against this backdrop of demo let us
1156.96 -> look at how the
1158.16 -> end-to-end system works as you see
1161.28 -> monotron has three distinct components
1163.679 -> there are the devices
1165.84 -> the monitoring service and the
1168 -> monitoring app
1169.6 -> the devices are are essentially
1172.24 -> installed
1172.88 -> on the premise or the site where your
1174.88 -> machines are
1176.88 -> you install these devices on the on the
1179.44 -> equipment say for example a motor or a
1181.6 -> gearbox or a pump
1183.12 -> by gluing by using industrial adjectives
1188 -> the sensors once provisioned
1189.679 -> communicates with the gateway
1191.28 -> over bluetooth low energy and then the
1193.919 -> gateway communicates
1195.36 -> to the to the aws cloud using wireless
1198.559 -> network in the cloud as we discussed we
1202.08 -> run our
1202.799 -> analytics which is a combination of
1205.28 -> vibration iso standards
1206.88 -> and machine learning based analytics and
1209.76 -> finally the app is used
1211.44 -> for consuming the insights coming from
1213.12 -> this analytics
1214.64 -> the app is also used to provision the
1216.559 -> devices in the first place
1220.24 -> let us look a but bit more into the
1222.799 -> monitoring experience that monotron
1224.72 -> provides
1225.6 -> as you know the product of the core
1227.919 -> value prop of monotron
1229.28 -> is equipment monitoring we enable this
1232.24 -> monitoring using a combination of
1234 -> machine learning
1235.039 -> and vibration iso standard space
1236.64 -> analysis
1239.2 -> monotron essentially learns over a
1242 -> period of
1242.799 -> first two to seven days and baselines a
1245.039 -> model during this learning period
1249.039 -> once learned monitron starts inferencing
1252 -> the incoming streams of
1253.44 -> vibration and temperature data and looks
1255.84 -> for anomalies in them
1258.08 -> however it does not raise an alert for
1261.679 -> one of anomalies but instead it looks
1264.48 -> for persistent
1266.08 -> anomalies the reason here is the
1268.88 -> application that monitoring is helping
1271.12 -> address is this application of condition
1274.4 -> monitoring and predictive maintenance
1276.88 -> we see that anomalies that can persist
1279.84 -> indicate
1281.28 -> persisting conditions that are
1284 -> indicators of failure
1285.52 -> and so by only alerting on these
1287.76 -> persistent
1288.72 -> anomalies we are giving true indicators
1291.679 -> that that require investigation and
1293.84 -> fixing
1298.24 -> based based on the user feedback
1299.76 -> collected monitoring retrains the model
1302.4 -> that was built during the learning
1304 -> period this is an
1305.76 -> ongoing improvement that happens and
1307.76 -> every time user feedback is
1309.36 -> received the model is retrained
1313.679 -> monitoring also classifies the alerts
1316.48 -> into different severities
1318.799 -> specifically two a warning and an alarm
1323.2 -> end users i the reliability managers and
1325.76 -> technicians can use these
1327.84 -> severities to suitably prioritize the
1330.48 -> right
1331.2 -> alerts to investigate
1334.48 -> align alongside raising alerts monotron
1337.039 -> also explains the reasons behind
1338.799 -> alerting
1340.48 -> the way it does it is to through a
1342.24 -> commentary section that is in the app
1346.159 -> for example if an alert was raised
1348.559 -> because of vibration
1350.4 -> monotron would call it out that the
1352.08 -> alert was driven by vibration
1354.24 -> in another instance possibly the alert
1356.48 -> could be raised by temperature and
1357.84 -> monitoring would call out that this was
1359.679 -> because
1360 -> of the temperature also monotron
1364 -> explains whether the alerts were raised
1366.24 -> by the machine learning model
1368.159 -> or because of the iso standards based
1371.12 -> analysis
1372.799 -> by providing this visibility monotron is
1375.679 -> assisting end users technicians
1378 -> to get a better handle on the alerts and
1380.96 -> help them in investigating the issues
1382.72 -> and fix it
1386.64 -> now take let us look at how customers
1388.96 -> have been using monetron
1390.159 -> multiple customers have used it and
1392.24 -> found it quite useful
1395.28 -> for example amazon is using monotron to
1397.84 -> monitor conveyor systems in the
1399.36 -> fulfillment warehouses
1401.36 -> let us look at a short video of how
1403.44 -> amazon is using that
1406.66 -> [Music]
1410.72 -> the volume plan for this building is
1412.4 -> really close tied to the max machine
1414.4 -> capacity
1415.28 -> we can run through through the systems
1417.28 -> so meaning uh every downtime of
1419.44 -> at least half an hour or 50 minutes will
1421.84 -> cause a lot of packages which are
1424 -> somehow piling up somewhere in the
1426.32 -> building so amazon
1427.6 -> monotron will be used mainly on the on
1429.84 -> the conveyance we have 26 kilometers of
1431.84 -> conveyance in this building
1433.44 -> and that has got um to be running at all
1436.559 -> time in every corner of the
1438.159 -> building the moment around thousand of
1440.08 -> these centers are just located all over
1442.24 -> the the conveyance system in our
1444.159 -> building
1444.72 -> and the first couple of days it's
1446.24 -> learning so it's learning how the system
1448.08 -> is running
1448.72 -> what is the normal temperature of the of
1450.96 -> a motor what is the normal vibration of
1453.12 -> it
1453.52 -> and every deviation is directly alerted
1456 -> by our warehouse management system to
1457.76 -> our
1458 -> meet technicians and they can just go
1460.799 -> there and have a look
1461.76 -> we have an indicator and sensor give us
1463.84 -> the signal
1464.96 -> we have an um defect motor so we saw the
1468.32 -> temperature
1469.039 -> was very high the team make the decision
1471.919 -> to exchange the motor before we have an
1474 -> outage
1474.64 -> and we can work together with operations
1476.72 -> that we set up the maintenance window
1479.44 -> in the right time and to make sure that
1481.52 -> we have no impact on the customer
1486.56 -> [Music]
1491.12 -> over the next 12 months amazon
1493.679 -> fulfillment centers
1494.96 -> will deploy tens of thousands of
1496.96 -> monetron sensor sensors
1498.64 -> across dozens of sites worldwide this
1501.84 -> will help them reduce
1502.88 -> unplanned equipment downtime and improve
1505.36 -> the customer experience in terms of
1507.279 -> package delivery
1510.159 -> externally multiple customers have used
1512.96 -> monitoring
1514.559 -> for example ge gas gas power use
1517.2 -> monotron
1518 -> to monitor the health of tumblers at
1520.4 -> their turbine plant
1521.76 -> in maine tumblers are these
1525.039 -> equipment that are used to polish the
1527.52 -> airfoils
1528.799 -> that go into the turbines these tumblers
1533.279 -> in themselves have a variety of rotating
1536 -> equipment
1537.039 -> from bearings shaft to gearboxes
1541.039 -> previously gcash power
1544.08 -> had been meaning to monitor this
1545.52 -> equipment but because of the cost
1548.08 -> and the complexity of building a
1550.159 -> technology solution
1551.52 -> they had deep prioritized it with
1554.84 -> monetron g
1556.24 -> gas power was was able to connect these
1558.48 -> retrofit these tumblers with sensors
1560.559 -> and connect them to real-time analytics
1562.64 -> in the aws cloud in a matter of 10
1564.72 -> minutes
1566.559 -> it was extremely easy to use for the
1568.32 -> apparat operators
1569.679 -> they did not require any technical
1571.76 -> skills or
1572.96 -> knowledge of how to configure their itn
1575.12 -> ot networks
1579.44 -> fender a guitar manufacturer has used
1582.64 -> monotron to monitor
1584 -> the blowers that basically keep their
1586.4 -> paint shops dutch free
1590.96 -> fender found the monotron product
1592.64 -> extremely easy and cost effective to use
1595.919 -> based on their experience they feel that
1598.32 -> monotron
1599.12 -> is not just useful for the large
1600.88 -> industrial customers
1602.32 -> but also for small mom and pop shops
1605.44 -> because of the ease of use and the
1607.2 -> cost-effective nature
1610.32 -> rx components is a leading electrical
1613.36 -> and maintenance component distributor
1616.4 -> it distributes over 500 000 products
1619.76 -> from over 200 2500 suppliers
1624.32 -> rs is excited to partner with aws
1627.44 -> and monotron to offer first of its kind
1631.279 -> end-to-end wireless vibration and
1633.52 -> temperature condition monitoring
1634.88 -> solution in its portfolio as you've seen
1641.44 -> over the last few slides monitron is
1644.32 -> very easy to use
1645.84 -> and it can help you quickly get started
1648.159 -> and enable predictive maintenance for
1650 -> industrial equipment
1652 -> you can also get started with it in an
1654.159 -> easy way here is how
1656.799 -> you can go on to amazon.com or amazon
1659.44 -> business
1660 -> and purchase the amazon amazon
1662.24 -> monitoring sensors and gateways
1664.88 -> the sensors and the gateways are
1667.039 -> available in three
1668 -> different skus you can buy a pack of
1670.159 -> five sensors
1671.44 -> or you can buy a gateway or you can buy
1673.76 -> a starter kit which is a pack of five
1675.36 -> sensors and one gateway
1678.159 -> these sensors and gateways are available
1680.08 -> on amazon.com
1681.279 -> in the us and also on the european
1683.279 -> marketplaces
1684.32 -> namely france italy spain germany
1687.44 -> and uk once these sensors and gateways
1691.84 -> are received
1693.2 -> you can use the monitor on app to set up
1695.6 -> these devices
1696.88 -> and physically install them on your
1698.399 -> machine machines
1700.88 -> and once you do that you can monitor and
1703.36 -> start managing your equipment
1706.64 -> as you see as you've seen in in the
1708.799 -> previous examples it's extremely easy
1711.2 -> quick and cost effective and you can
1713.52 -> quickly enable
1714.48 -> predictive maintenance for industrial
1716.32 -> equipment
1718.72 -> thank you questions
Source: https://www.youtube.com/watch?v=2O-UvaVKVlQ