Data-driven control policy for power generator stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing power generation systems face challenges in synchronizing multiple generators without accurate knowledge of system models or parameters, leading to instability and increased operational costs due to the need for extensive sensor measurements.
Innovation Solution
A data-driven control system that learns a control policy from input-and-output sequences of rotor angles and excitation voltages, allowing for optimal and stable control of power generators without requiring full state knowledge or accurate system models, using a processor to iteratively update the control policy based on real-time measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If model-based optimal control design is used, then control performance is improved, but accurate knowledge of system model and parameters is required which is difficult or impractical to determine
Solution Approach 1:
The patent replaces model-based control mechanisms with a data-driven neural network approach. Instead of relying on mechanical/mathematical system models and parameters (damping constant, inertia constant, governor time constant), the invention uses a neural network controller trained on operational data to directly map system states to control actions, eliminating the need for accurate parameter knowledge.
Solution Approach 2:
The invention changes the fundamental parameter representation from physical system parameters (inertia constant, damping constant) to data-driven parameters (neural network weights and biases) that are learned from operational data rather than theoretically determined.
2Ease of operation
If model-free controllers are designed without accurate knowledge of the model, then ease of operation is improved, but dependence on real-time measurement of all states requires multitude of expensive sensors
Solution Approach 1:
The patent extracts and utilizes only the most critical measurable parameter (rotor angle) from the full system state, discarding the requirement for complete state measurement. The neural network is designed to achieve effective control using this single extracted parameter, eliminating the need for expensive sensors to measure all states.
Solution Approach 2:
The neural network controller serves itself by learning from historical operational data what combinations of readily available measurements ( rotor angle) are sufficient for effective control, without requiring external provision of complete state information through expensive sensor networks.
3Stability of the object's composition
If synchronization parameters are matched within desired tolerance, then system stability is improved, but any mismatch during connection results in undesired transients and disruption
Solution Approach 1:
The neural network controller performs preliminary learning and adaptation during normal operation, building up knowledge of system dynamics and optimal control strategies. This preliminary action prepares the controller to handle synchronization events more effectively, reducing transients and disruptions when generators are connected.
Solution Approach 2:
The invention implements feedback through the neural network's continuous learning from operational data and its ability to adjust control actions based on observed system responses. This feedback mechanism enables the controller to maintain synchronization stability and minimize disruptive transients by adapting to actual system behavior rather than relying on theoretical models.
Data Source
AI summary
A control system for controlling a power generator of a power generation system executes a control policy to map an input-and-output sequence to a current value of the excitation voltage, submits the current value of the excitation voltage to the power generator, accepts a current value of the rotor angle caused by actuating the power generator according to the current value of the excitation voltage, and updates the input-and-output sequence with the corresponding current values of the rotor angle and the excitation voltage. The input-and-output sequence of values of the operation of the power generator includes a sequence of multiple values of the rotor angle of the power generator and a corresponding sequence of multiple values of excitation voltage to the power generator causing the values of the rotor angle. The control policy maps the input-and-output sequence to a current control input defining the current value of the excitation voltage.