As defined by ASHRAE, thermal energy storage (TES) systems are systems that remove heat from a storage medium (like water or ice) for use at another time.  The primary objective of a TES system is to reduce the on-peak electrical demand charges and therefore the cost of system operation.  A TES system can significantly reduce the cost of energy by allowing power intensive, electric driven cooling equipment to be predominantly operated during utility off-peak hours when utility electrical demand charges are reduced.  A TES system can theoretically shift all or part of the demand charges for the generation of cooling energy to utility non-peak hours, dependent only on the arrangement and physical size (capacity) of the storage vessel.  Thus these systems are principally designed to reduce energy cost.
Over the years, TES systems have been credited with saving energy as well.  Energy is equal to power (kW) multiplied by time (hours) or for electricity, kilowatt-hours (kWh).  For cooling energy, the equation is still power (tons) multiplied by time (hours) to yield ton-hours of cooling energy.  The TES system has the capability of shifting the generation of these ton-hours from peak to non-peak utility times thus saving cost, but not energy.  Any slight energy savings that may be available from operating a chiller at full load for more hours or operating at night with lower condensing temperatures will generally be offset by tank temperature losses and inefficiency in operation of the control system storing and releasing chilled water to the loads.
In addition, TES systems have also been considered as a substitute for redundant chiller capacity.  Whether or not this is true depends on one’s definition of a redundant chiller. All TES systems are designed based on a time component, back to cooling energy in ton-hours.  Consider a three chiller plant with no redundant chiller capacity.  Adding a storage tank that is used principally for the peak shaving of electrical demands to provide redundant capacity only makes sense if excess chiller capacity was available the day before.  If the true design load was three chillers, by the evening of the first day, the capacity of the tank has certainly been depleted and if the design load continues with two chillers, no other mechanism is available to make the excess chilled water for the TES system.  Even if the tank is just kept cold and kept in reserve (not used for peak shaving) until one of the three chillers fail, it could only be considered a redundant chiller for one day.
The concept of storing chilled water in the volumes required for a campus TES system has a disadvantage that is concerning from a redundancy and safety standpoint.  The typical campus cooling system is a closed, pressurized chilled water system with very little leakage over time and therefore very little make-up water introduced into the system.  Since the storage tank sizes required for a TES system cannot be constructed as pressure vessels, the connection of the TES storage tank converts the closed chilled water system into an open system.  The obvious issue here is that controls and control valves must be relied upon to maintain the level of water in the tank even when the supply and return line connections both have the capability of quickly overflowing the tank and disabling the chilled water system.  A less obvious issue is that the tank must be vented, putting air above a very large surface of chilled water.  Such water readily absorbs oxygen, making the chilled water system and its piping significantly more prone to corrosion with the oxygenated water.  Neither of these are insurmountable issues but they are further considerations in the desirability of a TES system.
The chilled water system could remain closed if a heat exchanger were inserted at the union of the storage system and the closed chilled water system.  However even with extremely effective plate frame heat exchangers, there must be some temperature difference between the stored chilled water and the chilled water in the closed system, thus further complicating the operation of the storage system.
Lastly, since the system saves only energy cost, it is completely dependent on the local electrical rate for on-peak vs. off-peak power.  Some Owners may find that relying on a 10 year payback based on the current electrical rate schedule more risky that competing energy cost reduction techniques.

Author:  Jerry Williams