A²/O 공정의 유출수 NH₄-N에 대한 모델기반 예측 제어 알고리즘 개발 및 평가

우대준(Dae Joon Woo),김효수(Hyosoo Kim),김예진(Ye Jin Kim),차재환(Jae Hwan Cha),최수정(Soo Jung Choi),김민수(Min Soo Kim),김창원(Chang Won Kim) 大韓環境工學會 2011 대한환경공학회지 Vol.33 No.1

본 연구에서는 하수처리공정의 유출 수질을 안정적인 유지하기 위하여, 유입수 패턴을 이용한 모델 기반 NH₄-N 예측제어 알고리즘을 개발하고 A₂/O 공정을 대상으로 적용 및 평가하였다. 평가에 사용된 자료는 B 시 S 하수처리장에 위치한 pilot 규모의 A₂/O 공정 운전자료를 사용하였다. 생물학적 반응조를 모사하기 위해 수정된 ASM³⁺bio-P 모델(Lee, 2003)을 사용하였고, 침전조 농도 거동 모사를 위해 일차원 이중 지수 함수 모델(Takacs et al., 1991)을 사용하였다. 유입수 패턴을 이용하여 하루 뒤 유출수 NH₄-N 농도를 예측하고, 사전 작성된 NH₄-N 제어 schedule을 사용하여 pilot plant의 호기조 DO를 조절하는 제어 로직을 적용하였다. 제어 적용성을 평가하기 위해 제어를 적용하지 않은 경우와 적용한 경우를 비교하였고, 계절적 영향을 알아보기 위해 여름철과 겨울철에 NH₄-N 제어 schedule을 적용한 실험을 하였다. 여름철 및 겨울철 모두 제어를 적용하지 않은 경우 수질기준을 초과하는 사례가 발생하였지만, 제어를 적용한 경우 목표수질 이내의 안정적인 유출수질이 방출됨을 확인하였다. 제어를 적용하지 않은 경우에 비교해서, 예측 제어를 적용한 경우에는 송풍기의 RPM이 약 9.1% 증가하였고, 유출수의 NH₄-N 농도는 약 45.2% 감소하였다. 이를 통해 본 연구에서 개발된 유출수 NH₄-N 예측 제어 알고리즘 적용으로 인한 운영비용 증가 대비 수질 개선 효과가 크게 나타났기 때문에, 안정적인 유출수 확보를 위한 측면에서 효율적인 제어기법으로 판단된다. In this study, model-based NH₄-N predictive control algorithm by using influent pattern was developed and evaluated for effective control application in A₂/O process. A pilot-scale A₂/O process at S wastewater treatment plant in B city was selected. The behaviors of organic, nitrogen and phosphorous in the biological reactors were described by using the modified ASM³⁺Bio-P model.11) A one-dimensional double exponential function model12) was selected for modeling of the secondary settlers. The effluent NH₄-N concentration on the next day was predicted according to model-based simulation by using influent pattern. After the objective effluent quality and simulation result were compared, the optimal operational condition which able to meet the objective effluent quality was deduced through repetitive simulation. Next the effluent NH₄-N control schedule was generated by using the optimal operational condition and this control schedule on the next day was applied in pilot-scale A₂/O process. DO concentration in aerobic reactor in predictive control algorithm was selected as the manipulated variable. Without control case and with control case were compared to confirm the control applicability and the study of the applied NH₄-N control schedule in summer and winter was performed to confirm the seasonal effect. In this result, the effluent NH₄-N concentration without control case was exceeded the objective effluent quality. However the effluent NH₄-N concentration with control case was not exceeded the objective effluent quality both summer and winter season. As compared in case of without predictive control algorithm, in case of application of predictive control algorithm, the RPM of air blower was increased about 9.1%, however the effluent NH₄-N concentration was decreased about 45.2%. Therefore it was concluded that the developed predictive control algorithm to the effluent NH₄-N in this study was properly applied in a full-scale wastewater treatment process and was more efficient in aspect to stable effluent.

Model Predictive Current Control Strategy for Improved Dynamic Response in Cascaded H-Bridge Multilevel Inverters

Baek Ju-Yoen,Lee Kyo-Beum 대한전기학회 2024 Journal of Electrical Engineering & Technology Vol.19 No.3

This paper proposes a model predictive current control technique for fast dynamic response in cascaded H-bridge multilevel inverters. Typically, cascaded H-bridge multilevel inverters are controlled using the PI controller with a pulse-width modulation or the model predictive control method through system modeling. However, PI controller has a limitation in terms of dynamic responses. In contrast, the model predictive control method does not require tuning of proportional and integral gain values and does not use pulse-width modulation. Thus, the model predictive control has a fast response and a fexible control technique. Applying the conventional model predictive control to multilevel inverters results in a large amount of computation because of a large number of voltage vectors. The proposed method suggests a model predictive control strategy to determine output voltage vectors and switching states based on the position of a reference voltage. Furthermore, the proposed method reduces the computation compared to conventional model predictive control techniques, which calculate all voltage vectors, and achieves faster responses than conventional method that uses adjacent vectors. The efectiveness of the proposed model predictive current control method was validated through simulation and experimental results.

A Three-Vector-Based Model Predictive Flux Control for PMSM Drives

Xu Yanping,Hu Miaomiao,Yan Zhongqiao,Zhang Yanping,Ma Hao 대한전기학회 2021 Journal of Electrical Engineering & Technology Vol.16 No.5

The model predictive fl ux control (MPFC) strategy is an improved method for the weighting factor that is diffi cult to adjust in the traditional model predictive torque control (MPTC). By analyzing the relationship among torque, stator fl ux and load angle, the simultaneous control of torque and stator fl ux amplitude is converted into the control of equivalent references stator fl ux vector, which eliminates directly the weighting factor. However, MPFC also suff ers from high torque and fl ux ripples if only one voltage vector is applied during each control period. In order to solve this problem, a three-vector-based model predictive fl ux control strategy is proposed for permanent magnet synchronous motor in this paper. Three voltage vectors are applied in one control cycle to achieve good control performance of torque and fl ux, and the durations of three vectors are determined based on the principle of stator fl ux deadbeat control. The experimental results show that, compared with the model predictive fl ux control strategy and the optimal duty model predictive fl ux control strategy, the three-vector-based model predictive fl ux control strategy can eff ectively reduce the torque and fl ux ripples, improve the system’s steady-state performance.

A novel multi-objective tuning strategy for model predictive control in trajectory tracking

Jianqiao Chen,Guofu Tian,Yanbo Fu 대한기계학회 2023 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.37 No.12

Accuracy and efficiency are two important performances of model predictive control in trajectory tracking and they are seriously affected by the control parameters of model predictive control. To make the model predictive control with high accuracy and efficiency simultaneous, the paper proposed a strategy to tune the control parameters for model predictive control. The proposed strategy converts the tuning problem to a multi-objective optimization problem and employs non-dominated sorting genetic algorithm (NSGA-II) to solve it. The proposed strategy is employed to tune the control parameters for a classical model predictive control in a typical trajectory tracking condition. The simulation results show that the comprehensive performances of model predictive controller tuned by the proposed method are better than other tuning methods. The proposed tuning strategy is validated and it can be applied to tune the control parameters for model predictive control in trajectory tracking.

단상 부스트 PFC 컨버터의 제어 기법 리뷰: 평균 전류 모드 제어, 예측 전류 모드 제어 및 모델 기반 예측 전류 제어

고현준(Hyeon-Joon Ko),최영준(Yeong-Jun Choi) 한국컴퓨터정보학회 2023 韓國컴퓨터情報學會論文誌 Vol.28 No.12

부스트 PFC (Power Factor Correction)컨버터는 AC 입력 전류의 단일 역률과 낮은 THD (Total Harmonic Distortion)를 달성하기 위해 다양한 제어기법들이 연구되고 있다. 그중 인덕터 전류의 평균값을 전류지령에 추종하도록 제어하는 평균전류 모드 제어가 있으며 가장 널리 사용되고 있다. 하지만, 오늘날 디지털 프로세서의 발달로 고도화된 디지털 제어가 가능해지면서 부스트 PFC 컨버터의 예측제어가 관심을 받고 있다. 예측제어에는 예측 알고리즘으로 듀티를 미리 생성하는 예측전류 모드 제어 및 모델을 기반으로 한 비용함수를 선정하여 스위칭 동작을 하는 모델예측제어로 분류된다. 따라서 본 논문에서는 부스트 PFC 컨버터의 평균전류 모드 제어, 예측전류 모드 제어, 모델예측 전류 제어를 간단히 설명한다. 또한, 시뮬레이션을 통해 전체 부하 및 다양한 외란 조건에서의 전류 제어를 비교 분석한다. For boost PFC (Power Factor Correction) converters, various control methods are being studied to achieve unity power factor and low THD (Total Harmonic Distortion) of AC input current. Among them, average current mode control, which controls the average value of the inductor current to follow the current reference, is the most widely used. However, nowadays, as advanced digital control becomes possible with the development of digital processors, predictive control of boost PFC converters is receiving attention. Predictive control is classified into predictive current mode control, which generates duty in advance using a predictive algorithm, and model predictive current control, which performs switching operations by selecting a cost function based on a model. Therefore, this paper simply explains the average current mode control, predictive current mode control, and model predictive current control of the boost PFC converter. In addition, current control under entire load and disturbance conditions is compared and analyzed through simulation.

Development of an energy cost prediction model for a VRF heating system

Park, Bo Rang,Choi, Eun Ji,Hong, Jongin,Lee, Je Hyeon,Moon, Jin Woo Elsevier 2018 Applied thermal engineering Vol.140 No.-

<P><B>Abstract</B></P> <P>This study developed a predictive model using artificial neural network (ANN) to forecast the energy cost for a variable refrigerant flow (VRF) heating system. The energy cost is predicted with the ANN model by considering the set-points for the refrigerant condensation temperature, condenser fluid temperature, condenser fluid pressure, and air handling unit supply air temperature together with past operational data and other climatic data. The predicted energy cost was used as a determinant for the control algorithm to optimize the heating system operation in terms of cost.</P> <P>The study consisted of three steps: initial model development, model optimization, and performance evaluation. The neural network toolbox in the Matrix laboratory was used to develop the model and conduct the performance tests. For the model training and performance evaluation, data sets were collected in the winter from a test building.</P> <P>Initial model consisted of a structure that included ten input neurons and a learning method. Then, the optimization process was used to find the optimal structure of the ANN model, which was 1 hidden layer with 15 hidden neurons, while the optimal learning method had a 0.5 learning rate and 0.4 momentum. In the performance evaluation, the optimized model demonstrated its prediction accuracy to be within the recommended level, with 0.8417 r<SUP>2</SUP> and 4.87% coefficient of variation root mean squared error between the measured and the predicted costs, thus proving its applicability in the control algorithm to supply a comfortable indoor thermal environment in a cost-efficient manner.</P> <P><B>Highlights</B></P> <P> <UL> <LI> A predictive and adaptive ANN model was developed for controlling heating system. </LI> <LI> The model predicted heating energy cost for the different variable settings. </LI> <LI> Model optimization was conducted for the accurate and stable prediction. </LI> <LI> The optimized model demonstrated its prediction accuracy within the recommended level. </LI> </UL> </P>

A Sliding Mode Based Model Predictive Control Structure for Permanent Magnet Synchronous Motor

Ilro Lee,Youngwoo Lee,Donghoon Shin,Chung Choo Chung 제어로봇시스템학회 2015 제어로봇시스템학회 국제학술대회 논문집 Vol.2015 No.10

This paper presents a sliding mode based model predictive controller for permanent magnet synchronous motor(PMSM). The proposed controller consists of model predictive controller(MPC) for outer velocity control loop and sliding mode controller(SMC) for inner current control loop. The SMC is proposed to track the desired currents in finite time. The MPC is developed to obtain optimal velocity control input for minimized cost function with constraint conditions. The closed-loop stability of the proposed controller is proven by Lyapunov theorem. Finally, simulation results are attached to validate the performance of the proposed controller.

모델오차 제어기 통합기법을 이용한 무인 헬리콥터 강건제어기 설계

박세욱,김유단 한국항공우주학회 2011 한국항공우주학회 학술발표회 논문집 Vol.2011 No.11

본 논문에서는 무인 헬리콥터의 모델 불확실성을 고려한 모델오차 통합제어기를 설계 하였다. 모델오차 통합제어기의 구성은 기준제어기와 모델오차 예측필터로 구성하였으며, 기준제어기는 출력되먹임 선형최적 제어기법을 이용하였다. 제어입력은 모델 불확실성을 상쇄시키기 위해 예측필터를 통해 추정된 모델오차를 이용하여 실시간으로 갱신된다. 제안된 제어기법은 무인 헬리콥터의 선형모델을 이용하여 모델오차 제어기 통합기법의 사용 유무를 통해서 검증하였으며, 시뮬레이션 결과 모델오차 제어기 통합기법이 모델 불확실성을 상쇄시킬 뿐만 아니라 기준명령 추종 성능을 향상시킴을 확인할 수 있었다. In this paper, the Model-Error Control Synthesis (MECS) for a small unmanned helicopter dynamics is presented, considering the model uncertainties. The Model-Error Control Synthesis consists of a nominal controller using Linear Quadratic Regulator (LQR) with output feedback and a model-error predictive filter. The control input is updated realtime using the estimated model-error from a predictive filter to cancel the model uncertainties or disturbance inputs. The performance the proposed controller is verified using a linear model of a small unmanned helicopter. Using numerical simulation, MECS is shown to have better performance than the stand-alone nominal controller in the sense of the model uncertainties cancellation and reference tracking.

제어 입력 기반 예측된 종방향 속도를 이용하는 자율주행 자동차의 차선 변경을 위한 모델 예측 조향 제어 알고리즘

오광석(Kwangseok Oh),오세찬(Sechan Oh) 한국자동차공학회 2021 한국 자동차공학회논문집 Vol.29 No.7

Steering control when changing lanes involving autonomous vehicles is the most important task of autonomous driving for driving strategy and safety. This paper presents a model predictive steering control algorithm for changing lanes among autonomous vehicles by using a control input that is based on predicted velocity. Two model predictive control algorithms have been designed for longitudinal and lateral autonomous driving with physical constraints. The model predictive longitudinal controller computes optimal longitudinal accelerations that use relative information between subject and preceding vehicles. Based on the optimal accelerations from the longitudinal controller, longitudinal velocities have been predicted, and the predicted velocities have been used to compute the optimal steering angle for changing lanes through a model predictive controller. The proposed model predictive control algorithms for lane change behavior have been constructed in a Matlab/Simulink environment. A performance evaluation has been conducted by using a commercial software(CarMaker) to ensure reasonable evaluation under variable lane change conditions of the subject vehicle.

A Model Predictive Controller for Autonomous Vehicle Path Tracking by Considering Dynamics of Automatic Steering System

Euiyun Kim,Junsoo Kim,Minkwang Lee,Sunwoo Myoungho 한국자동차공학회 2012 한국자동차공학회 학술대회 및 전시회 Vol.2012 No.11

Path tracking control is an essential function for autonomous driving systems. These autonomous driving systems require high tracking accuracy, ride comfort, and smoothness of steering action in the controller’s design. In order to satisfy various requirements, model predictive control (MPC) approaches, which derive an optimal steering trajectory with respect to a pre-defined path, have been widely used. The conventional predictive controller is based on a simple bicycle model for estimation of vehicle motion. However, since there is the dissimilarity between the simplified model and the real autonomous vehicle, the conventional controller is limited in the improvement of vehicle maneuvers. To overcome the limitation, dynamics and constraints of an automatic steering system are taken into account as practical control challenges. In this paper, we propose a model predictive control technique by considering dynamics of the automatic steering system. The optimal trajectory of steering angle is obtained by using the quadratic programming (QP) optimization method. The proposed model predictive controller was verified by simulation using a commercial vehicle model. The simulation results show that an improved control performance can be achieved by considering dynamics of automatic steering system.