New Method Based on Model-Free Adaptive Control Theory and Kalman Filter for Multi-Product Unsteady Flow State Estimation

2021 ◽  
Vol 143 (7) ◽  
Author(s):  
Lei He ◽  
Jing Gong ◽  
Kai Wen ◽  
Changchun Wu ◽  
Yuan Min
Keyword(s):  
Adaptive Control ◽  
Unsteady Flow ◽  
State Estimation ◽  
Flow Model ◽  
Transient Flow ◽  
Pipeline System ◽  
Flow State ◽  
Flow Estimation ◽  
Model Based ◽  
Model Free

Abstract In this paper, a new methodology is proposed to realize real-time unsteady flow estimation for a multi-product pipeline system. Integrating transient flow model, adaptive control theory, and adaptive filter, this method is developed to solve the contradiction between the efficiency and accuracy in traditional model-based methods. In terms of improving computational efficiency, the linear flow model based on frequency response and difference transforming is established to replace the traditional nonlinear flow model for transient flow state estimation. To reduce the deviation between actual observations and linear model estimates, we first introduce a model-free adaptive control method as linear compensation of the reduced order unsteady flow state model. To overcome the interference of observation noise, the Kalman filter method is applied to the modified state space model to obtain the one-step-ahead transient flow estimation. The proposed method is applied to the transient flow state estimation of a multi-product pipeline system and compared with the model-based method and two data-driven methods. The proposed method can reduce the deviation of transient flow estimation between the reduced order linear model and the traditional nonlinear model to less than 0.5% under unforeseen conditions and shows strong robustness to noise interference and parameter drift.

Kalman Filter and Model-Free Adaptive Control Theory Applied to the Unsteady Flow State Estimation of Product Pipelines

2020 ◽  
Author(s):  
Lei He ◽  
Kai Wen ◽  
Jing Gong
Keyword(s):  
Unsteady Flow ◽  
Linear Model ◽  
Flow Model ◽  
Transient Flow ◽  
Linear Method ◽  
Model Parameters ◽  
Flow State ◽  
Transient Pressure ◽  
Linear Flow ◽  
Model Based

Abstract The accurate online estimation of unsteady flow state provides important operation information for product pipelines real-time scheduling. In practice, affected by the parameter drift and observation noises, traditional estimation methods based on the first principle can hardly provide accurate results within acceptable time. The nonlinear and fast transient characteristics of pipeline flow make it difficult to realize on-line adaptive modification of model parameters. In order to meet the requirements of computational efficiency and accuracy simultaneously, this paper proposes a methodology with two-level adaptive adjustment to realize the digital twin of pipeline nonlinear transient flow process by using simplified linear flow model. In terms of improving computing efficiency, the linear flow model based on frequency response and difference transforming is established to process the on-line state estimation of transient flow. To reduce the deviation between the actual observed value and the linear model estimation, we first introduce mode-free adaptive control method as linear compensation of the reduced order unsteady flow model. The compact form dynamic linearization method has been adopted to design the virtual input of the linear flow model. To further improve the adaptability of the linear model, the model parameters are online adjusted by using the recursive least squares with forgetting factor method. The uncertainty of the model and the interference of observation noise is eliminated by adopting Kalman filter to the state space model based on modified linear model. The effectiveness of the proposed methodology is evaluated by applying to the digital twin process of a product pipeline transient pressure in a multistation pipeline. The results show that the proposed method can make transient pressure estimation of second-order linear model agree well with the value of nonlinear flow model even under unforeseen conditions and noise interference. The performance of the proposed method is better than model-based linear method, data-driven linear method and nonlinear method.

scholarly journals Prefrontal solution to the bias-variance tradeoff during reinforcement learning

2020 ◽  
Author(s):  
Dongjae Kim ◽  
Jaeseung Jeong ◽  
Sang Wan Lee
Keyword(s):  
Adaptive Control ◽  
Reinforcement Learning ◽  
Prediction Error ◽  
Brain Regions ◽  
Decision Task ◽  
Prediction Errors ◽  
Model Based ◽  
Model Free ◽  
Bias Variance ◽  
The Brain

AbstractThe goal of learning is to maximize future rewards by minimizing prediction errors. Evidence have shown that the brain achieves this by combining model-based and model-free learning. However, the prediction error minimization is challenged by a bias-variance tradeoff, which imposes constraints on each strategy’s performance. We provide new theoretical insight into how this tradeoff can be resolved through the adaptive control of model-based and model-free learning. The theory predicts the baseline correction for prediction error reduces the lower bound of the bias–variance error by factoring out irreducible noise. Using a Markov decision task with context changes, we showed behavioral evidence of adaptive control. Model-based behavioral analyses show that the prediction error baseline signals context changes to improve adaptability. Critically, the neural results support this view, demonstrating multiplexed representations of prediction error baseline within the ventrolateral and ventromedial prefrontal cortex, key brain regions known to guide model-based and model-free learning.One sentence summaryA theoretical, behavioral, computational, and neural account of how the brain resolves the bias-variance tradeoff during reinforcement learning is described.

State estimation in wall-bounded flow systems. Part 3. The ensemble Kalman filter

2011 ◽  
Vol 682 ◽  
pp. 289-303 ◽  
Author(s):  
C. H. COLBURN ◽  
J. B. CESSNA ◽  
T. R. BEWLEY
Keyword(s):  
Kalman Filter ◽  
State Estimation ◽  
Channel Flow ◽  
Ensemble Kalman Filter ◽  
Accurate Estimation ◽  
Wall Region ◽  
Flow State ◽  
Model Based ◽  
Wall Flow ◽  
Near Wall

State estimation of turbulent near-wall flows based on wall measurements is one of the key pacing items in model-based flow control, with low-Re channel flow providing the canonical testbed. Model-based control formulations in such settings are often separated into two subproblems: estimation of the near-wall flow state via skin friction and pressure measurements at the wall, and (based on this estimate) control of the near-wall flow field fluctuations via actuation of the fluid velocity at the wall. In our experience, the turbulent state estimation sub-problem has consistently proven to be the more difficult of the two. Though many estimation strategies have been tested on this problem (by our group and others), none have accurately captured the turbulent flow state at the outer boundary of the buffer layer (5 ≤ y+ ≤ 30), which is deemed to be an important milestone, as this is the approximate range of the characteristic near-wall turbulent structures, the accurate estimation of which is important for the control problem. Leveraging the ensemble Kalman filter (an effective variant of the Kalman filter which scales well to high-dimensional systems), the present paper achieves at least an order of magnitude improvement (in the near-wall region) over the best results available in the published literature on the estimation of low-Reynolds number turbulent channel flow based on wall information alone.

scholarly journals Research on viscoelastic fluid unsteady flow model based on torque loss correction

2021 ◽  
Vol 93 ◽  
pp. 106927
Author(s):  
Bingfan Li ◽  
Gang Liu ◽  
Shiyuan Liu ◽  
Lei Chen
Keyword(s):  
Unsteady Flow ◽  
Viscoelastic Fluid ◽  
Flow Model ◽  
Model Based ◽  
Torque Loss

Representation, abstraction, and simple-minded sophisticates

2020 ◽  
Vol 43 ◽  
Author(s):  
Peter Dayan
Keyword(s):  
Decision Theory ◽  
Predictive Coding ◽  
Bayesian Decision Theory ◽  
Bayesian Decision ◽  
Model Based ◽  
Model Free

Abstract Bayesian decision theory provides a simple formal elucidation of some of the ways that representation and representational abstraction are involved with, and exploit, both prediction and its rather distant cousin, predictive coding. Both model-free and model-based methods are involved.

Sign in / Sign up

Export Citation Format

Share Document