US 12,359,669 B2 · Adaptive learning system for improving sump pump control

Abstract

A sump pump system may implement adaptive learning and machine learning techniques to facilitate improved control of sump pumps. A sump pump system may implement the described techniques to generate, train, and/or implement a machine learning model that is capable of predicting or estimating one or more conditions of the sump pump system (e.g., water level in the basin, motor malfunction, stuck impeller, geyser effect, blocked outlet pipe, faulty level sensor/switch, faulty bearing, failure to engage pump at high-water mark, etc.) based on one or more detected input variables (e.g., acceleration or vibration patterns detected in water, on a pump, or on a pipe; capacitance values of water; audio signatures; electrical signatures, such as power or current draw; pump motor rotation speed; water pressure signatures or values, such as those detected at the bottom of a sump basin; etc.).

Claims20

Dependent claims

A sump pump system configured to implement adaptive learning, the sump pump system comprising:

a sump pump;
and
a sump pump control system coupled to the sump pump comprising:
a first sensor configured to detect water levels;
a second sensor configured to detect acceleration;
and
one or more processors coupled to the first sensor and the second sensor, the one or more processors configured to:
(A) access a machine-learning (ML) model previously trained by performing one or more statistical analyses on one or more sets of data indicative of the water levels in a sump basin and one or more sets of data indicative of detected acceleration patterns, the ML model trained to discover one or more relationships between the one or more water levels and the one or more acceleration patterns;
and
(B) implement the ML model, wherein the sump pump control system is configured to activate or deactivate the sump pump based on the one or more predicted water levels that are predicted via the ML model based on (i) one or more acceleration data collected during operation of the sump pump and (ii) the one or more discovered relationships.

7 more elements

Dependencies are inferred from the claim text.

Adaptive learning system for improving sump pump control

Life of this patent

Adaptive learning system for improving sump pump control

Life of this patent

Abstract

Claims20