System and method for interpretation and analysis of manufacturing activity
AssigneeAmper Technologies, Inc.
InventorsAkshat Thirani, Philip House
Life of this patent
Estimating the term…
Abstract
A system and method for monitoring manufacturing machines that includes a sensor system that monitors a plurality of machines, wherein the sensor system is comprised of a set of current transformers, and wherein the set of current transformers transform electric currents to the plurality of machines into electric signals; an activity processing engine, wherein the activity processing engine generates diagnostic metrics from the electric signals for each machine from the plurality of machines; and a management platform configured to provide access to diagnostic metrics of the plurality of machines.
Claims20
A system for monitoring manufacturing status for manufacturing machines comprising:
- a sensor system that monitors electrical signals from a plurality of manufacturing machines, wherein the sensor system comprises a set of current transformers, wherein each current transformer of the set of current transformers couples to a power line connected to a corresponding manufacturing machine, and wherein each current transformer transforms electric currents of the power line into an electrical signal of the corresponding manufacturing machine;
- an activity processing engine, wherein the activity processing engine generates manufacturing diagnostic metrics from the electrical signal for each manufacturing machine and assign a part identifier based on the classification by the trained machine learning model to a subset of the machine cycles, wherein the subset of the machine cycles assigned to the part identifier are detectable in the electrical signals of different manufacturing machines when manufacturing a similar part;
- and
- a management platform configured to provide access to the manufacturing diagnostic metrics of the plurality of manufacturing machines, the management platform comprising at least one user interface configured to express the manufacturing diagnostic metrics of the plurality of manufacturing machines organized by part identifiers.
A method for monitoring manufacturing status for manufacturing machines comprising:
- electrically coupling a set of electric current sensors to a set of power lines, wherein each power line is connected to one of a plurality of manufacturing machines;
- collecting, using the set of electric current sensors, electric signals for each manufacturing machine from the plurality of manufacturing machines;
- generating manufacturing diagnostic metrics from the electric signals for each manufacturing machine from the plurality of manufacturing machines, wherein the generating comprises, for each electric signal of the electric signals:
- classifying, using a machine learning model, a machine cycle through detection of machine cycle signature features in the electric signal, the machine cycle selected from a set of machine cycles that includes at least a subset of the machine cycles associated with part identifiers, wherein the subset of machine cycles associated with the part identifiers are detectable in the electric signals of a different manufacturing machine when manufacturing a similar part, and
- assigning a part identifier based on the classification by the machine learning model to the machine cycle;
- and
- outputting, at a user interface of a management platform, the manufacturing diagnostic metrics of the plurality of manufacturing machines organized by the part identifiers.
Dependencies are inferred from the claim text.