Optimal control of chilled water system with ensemble learning and cloud-edge-terminal implementation
Abstract : In modern large buildings, the chilled water system regulates the indoor temperature through a series of heat exchanges. This paper studies the optimal control problem of the chilled water system and, proposes an ensemble learning method for the cooling load prediction under different operation conditions with imbalance sample distribution. A new control strategy is afterwards developed for optimal selection of the process control inputs that guarantees the demand for cooling load with a lower energy consumption. The optimal control strategy is also learned in real-time using a cloud-edge-terminal form, which can be used for big data modeling and increase the effectiveness of the system response. The proposed method is applied to a real high-rise building and, the results show a significant improvement in the proposed prediction model and the optimal control strategy, compared to the state-of-the-art methods. Regarding manual operation, the control strategy decreased the energy consumption by 5.59%, and on average 35645 KWh of electric energy per month was saved.
? They are presented here to supplement existing O&M procedures, or to merely serve as reminders of activities that should be taking place.
? In many installations, energy saving measures have reduced demand to the point that existing chillers are tremendously oversized, forcing the chiller to operate at greatly reduced loads even during peak demand times.
? By simply moving sensors to locations that are more representative of the air stream of interest, the existing control algorithms work as they were originally intended.
? The existing pumps are still used for the infrequent peak flows that the new smaller pump cannot handle.
? Aircooled condensers have the ability to operate in below-freezing weather, and can do so without the problems associated with operating the cooling tower in these conditions.
? Data points that may cause operational problems if corrective action is not taken.
? In addition to monitoring data, it is vital that the chiller controls alert operators to possible problems.
? While measuring the performance of the entire chiller plant is more difficult, it can help identify operating problems or evaluate the effectiveness of system control methods and setpoints.
? A proper energy management system can help trend and diagnose problems or changes over time.
• The MARL methods show that a range of flexibility requests can be met by providing an optimal energy portfolio of buildings and the proposed extended JAL performs best, considering responsibility and the commitment of allocation values.
• The tool allows the user to propose changes to existing equipment including chillers, pumps, and towers, and calculates how much energy and money the plant can save by implementing these changes.
• Operations and maintenance activities and equipment represent real costs to a facility and must be evaluated like any other proposed action.
• Boiler operators should take daily logs of stack temperature for trending purposes as this is a highly diagnostic indication of boiler heat-transfer-surface condition.
? Insufficient flow reduces heat-transfer efficiency and causes poor chiller performance, which might cause the chiller controls to invoke safeties.
? Excessive flow may result in high water velocity, erosion, vibration, or noise, while insufficient flow reduces heat transfer efficiency and causes poor chiller performance.
? In packaged air-cooled chillers, the manufacturers improve performance by staging fans in response to chiller load and ambient, dry-bulb temperature.
? To ensure optimum heat transfer performance, the condenser-heat transfer surfaces must be kept free of scale and sludge.
? Control capabilities run the gamut from slow-acting pneumatic controls, to electromechanical controls, to sophisticated digital controls that use "feedforward" algorithms tuned to give superior performance.
|