| Title |
Optimal Control Simulation of a Ventilation System using Occupant's Prediction Model and Ping-Pong Approach |
| Authors |
Kim Young-Jin ; Park Cheol-Soo |
| Keywords |
Ventilation ; Optimal Control ; DCV-CO2 ; Ping-Pong ; Markov Chain |
| Abstract |
In general, On-off and Multi step controls are widely applied to DCV-CO2 ventilation systems. The problem of the on-off and Multi step control is that their controls are not based on an optimal algorithm. Therefore, this study suggests a simulation assisted optimal control. The simulation assisted optimal control uses optimization algorithm to solve for optimal control variables to minimize a cost function over the time horizon. For this study, CONTAMW 2.4 simulation tool and EnergyPlus are coupled in MATLAB platform to simulate thermal and air-flow phenomena using Ping-Pong method. And the optimal control of ERV (Energy Recovery Ventilator) system is performed by a gradient-based search that uses the derivative of the cost function. The cost elements are energy flow and CO2 concentration (bedroom1, living room). A prominent characteristic presents that occupant’s schedule applies to a stochastic model based prediction of occupants' presence using the Markov Chain method. To perform Markov Chain method, the number of occupants every hour in each of the rooms (20 households) was examined and then the transition probability matrix was generated. By comparing the optimal control with existing controls (On-off and Multi step controls), it is shown that the proposed optimal control can lead to significant improvements for ventilation system performance. |