MACHINE LEARNING BASED PREDICTIVE MAINTENANCE OF A DRYER

• US 20170051978A1
• Filed: 08/23/2015
• Published: 02/23/2017
• Est. Priority Date: 08/23/2015
• Status: Active Grant

• Current Assignee: Machinesense LLC
• Original Assignee: Prophecy Sensorlytics Llc
• Inventors: Pal, Biplab, Gillmeister, Steve
• Granted Patent: US 10,648,735 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: F26B 25/009 Alarm systems; Safety sytems, e.g. preventing fire and explosions using inert gases F26B21/14
• CPC Class Codes: G05B 23/0272 Presentation of monitored results, e.g. selection of status reports to be displayed; Filtering information to the user
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

A machine learning method and system for predictive maintenance of a dryer. The method includes obtaining over a communication network, an information associated with the dryer and receiving measurements of a vibration level of one of a process blower, a cassette motor and a regeneration blower associated with the dryer. Further, an anomaly is determined based on at least one of a back pressure and a fault and balance of at least one of the process blower and the regeneration blower is tracked. An alarm for maintenance is raised when one of an anomaly and an off-balance is detected.

Claims

1. A machine learning method for predictive maintenance of a dryer comprising:

i) obtaining over a communication network, an information associated with the dryer, wherein the information comprises measurements of a current associated with at least one heater of a heater bank, wherein the heater bank is associated with the dryer;

determining a failure data associated with the at least one heater, wherein the failure data indicates one of a poorly functioning and a failed heater, wherein the failure data is determined through a comparison of ratio of three phase currents with a healthy heater using one of a split-core transformer type machine wearable current sensor or a Hall effect based current sensor associated with the at least one heater;

ii) receiving measurements of a vibration level of at least a process blower, a cassette motor and a regeneration blower associated with at least one dryer;

iii) determining an anomaly through at least one of a back pressure and a fault associated with at least one of the cassette motor and the regeneration blower, wherein the anomaly is determined based on at least one of a vibration and magnetic field through an IoT based method;

iv) tracking at least one of the vibration and the magnetic field of at least one of the process blower and the regeneration blower; and

v) raising an alarm for maintenance when an anomaly is at least one of a detected in a real-time and predicted for a future time, wherein the prediction is based on at least one of a machine learning algorithm or a look up table.

2. The method of claim 1, further comprising:

receiving sensor data from at least one machine wearable sensor placed on at least one of the heater, the process blower, the cassette motor and a regeneration blower associated with the dryer

3. The method of claim 1,

wherein the measurements and the information are received at a data collection unit and wherein heaters in the heater bank are connected in phase.

4. The method of claim 1, wherein the communication network is one of WiFi, 2G, 3G, 4G, GPRS, EDGE, Bluetooth, ZigBee, Piconet of BLE, Zwave, or a combination thereof.

5. The method of claim 3, wherein the data collection unit is one of a mobile device and a wireless enabled device.

6. The method of claim 1, further comprising:

associating a machine learning architecture with the machine learning algorithm where normal states of the dryer are learned with a baseline reading and anomalous readings from the back pressure and the misalignment of the blowers, wherein the baseline reading and the anomalous readings are used as a training database.

7. The method of claim 1, further comprising:

acquiring data from multiple dryers associated with machine wearable sensors, wherein at least one of a mobile, a web and a desktop application acts as a mobile middleware to scale the machine learning architecture to a single data collection unit, and wherein the single data collection unit is at least one of a mobile device and a wireless device.

8. The method of claim 1, wherein the alarm is raised over the communication network through one of a notification on the mobile application, Short message service (SMS), email, or a combination thereof.

9. A machine learning architecture associated with a dryer comprising:

one or more heaters connected in phase;

one or more machine wearable sensors;

a process blower, a cassette motor and a regeneration blower associated with the one or more machine wearable sensors; and

a processor configured to execute instructions, which, when executed by the processor, causes the processor to;

receive a sensor data over a communication network, wherein the sensor data is at least one of a vibration, a magnetic field and a current measurement;

determine a failure data associated with the at least one heater, wherein the failure data indicates a time of heater failure through a calculation based on a reading of current at a machine wearable sensor associated with the at least one heater;

determine an anomaly based on at least one of a back pressure and a fault;

track preassigned balance of at least one of the process blower and the regeneration blower; and

raise an alarm for maintenance, when at least one of an anomaly and an off-balance is detected, wherein the anomaly and the misalignment is at least one of detected in real-time and predicted for a future time, wherein the prediction is based on a machine learning algorithm.

10. The architecture of claim 9, wherein the sensor data is received from at least one machine wearable sensor placed on at least one of the heater, the process blower, the cassette motor and a regeneration blower associated with the dryer.

11. The architecture of claim 9, wherein the measurements and the information are received at a data collection unit.

12. The architecture of claim 9, wherein the communication network is one of WiFi, 2G, 3G, 4G, GPRS, EDGE, Bluetooth, ZigBee, Piconet of BLE, Zwave, or a combination thereof,

13. The architecture of claim 11, wherein the data collection unit is one of a mobile device and a wireless enabled device.

14. The architecture of claim 9, wherein the machine learning architecture is associated with the machine learning algorithm where normal states of the dryer are learned with a baseline reading and anomalous readings from the back pressure and the misalignment, wherein the baseline reading and the anomalous readings are used as a training database, and wherein a machine learning module issues commands based on a learning outcome from the training database.

15. The architecture of claim 9,

wherein data is acquired from multiple dryers associated with machine wearable sensors,

wherein at least one of a mobile, a web and a desktop application acts as a mobile middleware to scale the machine learning architecture to a single data collection unit, and

wherein the single data collection unit is at least one of a mobile device and a wireless device.

16. The architecture of claim 9, wherein the alarm is raised over the communication network through one of a notification on the mobile application, Short message service (SMS), email, or a combination thereof.

17. The architecture of claim 9, wherein the machine learning architecture is associated with an IoT based system.

18. The architecture of claim 17, wherein the learning outcome is dependent on recognition of at least one of a pattern and deviation by the machine learning module.