Engine and Pump Controller Threshold Learning Across RPM Ranges

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing control systems for engines and motors struggle to accurately detect parameter deviations due to varying operating conditions, leading to insufficient fault detection at different RPMs, as conventional warning/fault levels are often set too broadly or narrowly, missing potential issues.

Innovation Solution

Implementing machine learning-based controllers that learn normal parameter levels across a full operational range, allowing for real-time comparison and detection of deviations within configurable error thresholds, issuing warnings or shutdowns as necessary.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional control panels use single fault levels for monitoring, then the device complexity is low, but the measurement precision of fault detection deteriorates due to inability to account for RPM variations

Engineering Contradiction:

Improvefault detection precisionVSAvoidcontroller complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The controller dynamically adjusts fault threshold levels based on the operational RPM of the engine or motor. Instead of using fixed single fault levels, the system continuously adapts the threshold values to match the current operating conditions, enabling accurate fault detection across varying speeds while maintaining reasonable controller complexity through algorithmic adaptation

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the monitoring parameters (fault threshold levels) according to the RPM state. By establishing state-specific threshold values for different operational ranges, the controller achieves high measurement precision in fault detection without requiring overly complex hardware, as the adaptation is achieved through parameter adjustment rather than structural complexity

Inventive Principle:

Principle #35Parameter changes

2Reliability

If conventional control panels use fixed fault levels, then the ease of operation is high, but the reliability of fault detection deteriorates due to false alarms or insufficient warnings

Engineering Contradiction:

Improvefault detection reliabilityVSAvoidparameter configuration complexity

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The controller performs preliminary learning during an initial operational phase to establish baseline parameter levels for different RPM states. This preliminary action enables the system to automatically configure appropriate fault thresholds for each operational state without requiring manual intervention, thereby maintaining ease of operation while achieving high reliability in fault detection

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback mechanisms where the controller continuously monitors operational parameters, compares them against learned baselines, and adjusts fault detection decisions accordingly. This feedback loop enables reliable fault detection across varying conditions while the system automatically manages the complexity of parameter configuration through adaptive learning rather than manual setup

Inventive Principle:

Principle #23Feedback

3Measurement precision

If machine learning is implemented to learn normal levels across full operational range, then the measurement precision of fault detection improves, but the use of energy increases due to continuous monitoring and processing

Engineering Contradiction:

Improvefault detection precisionVSAvoidcontroller energy consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The operational range is segmented into discrete RPM states or ranges, with specific baseline levels learned for each segment. This segmentation allows the controller to use simpler comparison logic rather than continuous complex machine learning inference, reducing energy consumption while maintaining high measurement precision for fault detection at each operational segment

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs partial machine learning actions by establishing baseline levels for key RPM states rather than continuously adapting to every possible operating condition. This approach achieves sufficient measurement precision for practical fault detection while significantly reducing the computational energy burden compared to full continuous adaptation

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12607971B2Machine learning-based systems for parameter deviation monitoring and alerting in engines and pump controls

Publication Date: 2026.04.21 CATTRON NORTH AMERICA INC

AI summary

Exemplary embodiments are disclosed of controllers with machine learning and systems including the same. In exemplary embodiments, a controller is configured to include or be operable for executing a machine learning process during which the controller learns normal level(s) of parameter(s) (e.g., system component(s) and/or sensor(s) parameters, etc.) to be monitored across a full operational range of a system component, which, in turn, enables the controller to compare the monitored parameter(s) to the learned normal level(s) for detection of issue(s) or problem(s) associated with the monitored parameter(s), e.g., regardless of the RPM at which the engine, machine, motor, or other system component is operating, etc.