Neural-Network Power Regulator for Self-Tuning Converter Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of power conversion circuits, including switched-mode DC-DC converters, requires careful tuning and is sensitive to design changes and aging components, making it challenging to achieve optimal performance across varying operating conditions, especially requiring expertise for peak performance and long-term stability.
Innovation Solution
Incorporating a machine-learning-based artificial neural network into the power conversion regulator circuit to predict output parameters and generate control signals, combining this with conventional error-based control using a weighted approach to adapt to changing conditions and improve regulation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional PID regulator with parameter tuning is used, then the power converter can be controlled, but the tuning process is time-consuming and requires highly skilled engineers
Solution Approach 1:
The system employs self-tuning capability where the regulator automatically adjusts its parameters based on system responses without requiring manual intervention by skilled engineers. The processor circuit implements algorithms that learn optimal control parameters through observation of system behavior, enabling the system to tune itself autonomously.
Solution Approach 2:
The system performs preliminary characterization of the power converter circuit during manufacturing or initial operation. The processor circuit captures system responses to test inputs and pre-computes optimal control parameters before actual operation begins, eliminating the need for time-consuming on-site tuning by experts.
2Reliability
If advanced non-linear control algorithms are used, then system behavior is improved, but the algorithms need careful tuning for specific applications and may have unwanted side effects
Solution Approach 1:
The system dynamically adjusts control parameters based on operating conditions rather than using fixed complex algorithms. The processor circuit modifies control characteristics in real-time according to the actual state of the power converter, simplifying the control approach while maintaining reliability across varying conditions.
Solution Approach 2:
The control system transitions from static pre-tuned algorithms to dynamic adaptive control. The processor circuit continuously monitors system responses and adjusts control parameters on-the-fly, enabling the system to handle non-linear behavior without requiring complex predetermined algorithms for each specific application.
3Reliability
If high-quality components with small aging effects are chosen, then longtime-stability and high performance are ensured, but the circuit costs are dramatically impacted
Solution Approach 1:
The system implements continuous feedback monitoring of power converter performance parameters. The processor circuit measures actual system responses and compares them against expected behavior, detecting drift caused by component aging in real-time and triggering recalibration or compensation actions to maintain stability without requiring premium aging-resistant components.
Solution Approach 2:
The regulator performs self-diagnosis and self-calibration to compensate for component aging effects. Through automated monitoring and adjustment of control parameters, the system maintains optimal performance despite the natural degradation of standard components over time, eliminating the need for expensive aging-resistant parts.
4Manufacturing precision
If application-specific test plans are created, then the objectives for conversion circuit performance are met, but the complexity increases requiring only experts to design and tune
Solution Approach 1:
The system automatically generates and executes its own test sequences during manufacturing and operation. The processor circuit performs self-characterization by applying test inputs and measuring responses, then uses this data to optimize control parameters without requiring external expert intervention or application-specific test plan design.
Solution Approach 2:
The control system implements a universal tuning approach that works across different power converter applications. The processor circuit uses the same fundamental algorithms and characterization methods regardless of the specific application, eliminating the need for experts to design custom test plans for each case while maintaining optimized performance.
Data Source
AI summary
A power conversion regulator circuit includes: a regulator input configured to be dynamically supplied with a feedback signal representative of an output parameter of a power converter circuit; a regulator output configured to dynamically provide a control signal to the power converter circuit, for making adjustments to the output of the power converter circuit; a processing circuit configured to (a) implement an artificial neural network having a plurality of artificial neurons, wherein the artificial neural network is configured to compute a machine-learning-based (ML-based) error signal, based on at least the feedback signal and a target level for the output parameter, and (b) output a correction signal, based at least in part on the ML-based error signal; and regulator circuitry configured to generate the control signal for outputting via the regulator output, based at least in part on the correction signal.