Adaptive Plasma Power Control with Estimation Law Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive controllers struggle with nonlinear plasma processing systems, failing to ensure stability, scalability, and adaptability due to unbounded control values, modeling uncertainties, and asynchronous actuator responses, leading to inefficiencies and premature system failures.
Innovation Solution
An adaptive engine with estimation law modules that estimate model parameter tensors, apply control laws, and select optimal control signals based on estimated system outputs to manage plasma processing power systems, utilizing a nonlinear model with a bifurcated time-varying linear system for improved convergence and adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing adaptive controllers are used to control plasma processing systems, then the system can operate with multiple actuators, but the controllers fail to ensure stability and produce unbounded control values due to modeling uncertainties and asynchronous actuator responses
Solution Approach 1:
The controller segments the control task by creating separate control loops for each actuator, with individual models and control laws for each actuator-plasma parameter pair. This segmentation allows each actuator to be controlled independently according to its specific characteristics and response time, preventing instability caused by asynchronous responses while maintaining overall system reliability
Solution Approach 2:
The controller dynamically adjusts control parameters including adaptation gains and control law selection based on real-time system state, plasma conditions, and actuator performance. This parameter adaptation enables the system to handle modeling uncertainties and maintain stability across varying operating conditions while accommodating different actuator response characteristics
2Speed
If the rail voltage is held at a high level for much of a pulse cycle to accommodate fast power amplifier responses, then the power amplifier can respond quickly to control signals, but the DC section or rail overheats and causes premature system failure
Solution Approach 1:
The controller performs preliminary action by pre-charging the rail voltage to the required level before the power amplifier needs to deliver high power, rather than continuously maintaining high voltage. This allows the rail to be prepared in advance and then reduced to lower levels during periods when full power is not needed, preventing overheating while ensuring rapid response capability when required
Solution Approach 2:
The controller implements periodic action by pulsing the rail voltage at controlled intervals and duty cycles that match the plasma processing requirements. Instead of continuous high voltage, the system applies voltage in periodic bursts synchronized with the plasma cycle, allowing the rail to cool between pulses while maintaining the ability to deliver peak power when needed
3Adaptability or versatility
If existing adaptive controllers use transfer functions for control, then the control law can be applied, but the controllers are difficult to scale to arbitrary waveforms and coupled inputs and outputs
Solution Approach 1:
The controller implements a universal control architecture using state-space models and matrix-based control laws that can handle arbitrary waveforms, coupled inputs and outputs, and multiple actuators through a unified mathematical framework. This universal approach eliminates the need for separate transfer function derivations for each control scenario, making the system scalable while maintaining adaptability to different plasma processing recipes and actuator configurations
Data Source
AI summary
This disclosure describes systems, methods, and apparatus for estimation law modules of an adaptive engine. The estimation law modules estimate estimated model parameter tensors for each control sample within a frame. These estimated model parameter tensors are passed to control law modules and used to determine possible control signals as well as being passed to an estimation portion of a nonlinear model of the system, along with the possible control signals, to estimate estimated system outputs corresponding to each of the possible control signals. A selector module can then select a control as one of the possible control signals or a combination of two or more of the possible control signals, based on comparing the corresponding estimated system outputs to a reference signal, and/or measured system outputs.