nawi-seedling-project-foundational-control-methods-for-water-treatment-systems-online-actuator-

The National Renewable Energy Laboratory in the US specializes in the research and development of renewable energy, energy efficiency, energy systems integration, and sustainable transportation.

national-renewable-energy-laboratory-nrel

National Renewable Energy Laboratory (NREL)

This is a data set corresponding to conventional operation of a closed-circuit reverse osmosis (CCRO) system for desalination with feedback controls enabled. To be used to describe reference CCRO pilot operation.

Dataset 1 Conventional Operation.zip

This folder contains data describing the closed-circuit reverse osmosis (CCRO) system for desalination (meta-data). These meta-data are considered globally valid, i.e. information that is accurate over the whole repository.


Global Meta Data.zip

The main purpose of these experiments is to collect data from a closed-circuit reverse osmosis (CCRO) system for desalination under open-loop conditions (no feedback control) in the expected range of operation. To this end, the flux through the system and the recirculation flow rates are brought to a nominal setpoint by means of manipulation of the power delivered to each pump through conventional feedback control (PI control). Once reached, the feedback controllers are switched off and a sinusoidal perturbation is applied to the power delivered to (a) one pump, (b)  both pumps, or(c) none of the pumps. The system continues in filtration mode until an upper control limit for conductivity is reached. When this happens, the system is switched to drainage mode. When the conductivity drops below a lower control limit, the system is switched back to filtration mode. Each file in this data set corresponds to a single filtration phase followed by a drainage phase.

Dataset 2 Open Loop Experiments.zip

This folder contains the data corresponding to experiments executed to test extremum seeking control (ESC). The ESC was used to provide a setpoint for the critical reject conductivity at the beginning of each sequence in order to minimize the following objective function.
Obj=SEC+b*SBC
where SEC is the specific energy consumption, SBC is the specific brine management cost, and b is a user-defined parameter determining the trade-off between SEC and SBC.

A total of three experiments with different conditions were carried out, which are as follows:
Exp1: Feed concentration of 1.8 mS/cm and the weighting parameter as b=1
Exp2: Feed concentration of 1.8 mS/cm and the weighting parameter as b=10
Exp3: An increase of 50% in the feed concentration from 1.8 mS/cm to 2.7 mS/cm and the weighting parameter as b=10

Dataset 3 ESC Experiments.zip

This folder contains experimental data for system identification of a closed-circuit reverse osmosis (CCRO) system for desalination. The data are obtained under open-loop conditions (no feedback control). 

Dataset 4 Open Loop Experiments.zip

This folder contains the data corresponding to experiments executed to test a model predictive controller (MPC) on a closed-circuit reverse osmosis (CCRO) system for desalination.

Dataset 5 Model Predictive Control.zip

This folder contains the data corresponding to experiments executed to test Bayesian optimization (BO) on a closed-circuit reverse osmosis (CCRO) system for desalination. The BO was used to provide a setpoint for the critical reject conductivity at the beginning of each sequence in order to minimize the following objective function.
Obj=SEC+b*SBC
where SEC is the specific energy consumption, SBC is the specific brine management cost, and b is a user-defined parameter determining the trade-off between SEC and SBC.

Filters