Block interval demand calculations are completed at the end of each subsequent demand interval, and the result is used as the new present value for the peak demand evaluation. This method is similar to the way mechanical demand meters operate. The billing demand period for the block interval demand is defined by the DEMAND PERIOD parameter.
The billing demand period for the sliding window demand comprises a number of blocks and is defined a product of DEMAND PERIOD NUMBER OF DEMAND PERIODS. It represents a sliding window that moves one step forward after each subsequent block demand measurement. At the end of the present block demand interval, the information on the oldest block in the demand window is discarded, and the sliding window demand is calculated as average on the newest set of demand intervals.
The sliding window demand is updated as a new block demand is calculated at the end of a demand interval.
This method is similar to the way electro-mechanical thermal demand meters operate. The three-phase thermal maximum demand measuring element comprises a thermally lagged power demand pointer with an exponential response to a step change in instantaneous power load. The meter uses numerical techniques to approximate the thermal response characteristics of the thermal measuring element. A changing load is approximated by a series of step changes in load by calculating instantaneous power load at regular 1 second intervals. The thermal demand algorithm used allows calculating the thermal response with high accuracy, even for fast changing loads.
For thermal demand calculations, the billing demand period is defined as for the sliding window demand: DEMAND PERIOD NUMBER OF DEMAND PERIODS. The additional parameter needed to be defined by the user is a THERMAL TIME CONSTANT of the simulated thermal element. A thermal time constant depends on billing demand interval and on the electricity tariff rules applied by the power utility company.
In countries where thermal demand metering is still used, thermal demand meters may have a thermal time constant such that, for a step change in instantaneous load from 0 to 100%, the thermal demand pointer will reach 99%, or 63% of its final value after billing demand interval expired. The meter allows the user to adjust a thermal time constant in intervals of 1 to 3600 seconds with 0.1 second steps to match the power utility requirements. The following formula is used to define a thermal time constant for your application:
- thermal time constant, sec
t - billing demand interval, sec (demand interval number demand intervals)
S%(t) - the level that the thermal demand pointer will attain at the end of billing demand interval, expressed in percentage of the steady-state value
To approximate meters with S%(t) = 63%, the thermal time constant is considered to be equal to the billing demand interval, that is, = t, taken in seconds. For example, using a 15-minute billing demand interval, the thermal time constant is 900 seconds, and using a 30-minute demand interval - 1800 seconds.
For thermal demand meters with S%(t) = 99%, the thermal time constant will be 195.4 seconds using a 15-minute billing demand interval, and 390.9 seconds using a 30-minute billing demand interval.
The meters offer two evaluated parameters for power load control using predicted changes in power demand: accumulated demand and predicted sliding window demand.
Accumulated demand is the average power consumption over the fixed demand period, which is continually updated as power is being consumed during the demand period. Actually, accumulated demand is calculated from the consumed energy being integrated in the block interval demand storage register that is divided by the constant demand period. Thus, at any point during the demand period, accumulated demand represents the value proportional to the consumed energy converted to kW, kvar and kVA units. At the beginning of each demand period, accumulated demand is reset to zero. When power is being consumed, accumulated demand grows up to the maximum value, which is evaluated to block interval demand at the end of the demand period.
Accumulated demand can be checked for maximum demand allowed to trigger a setpoint at the moment when the actual demand has exceeded the predefined threshold and prior to the end of the demand interval, when only the new maximum demand value will be calculated.
Predicted sliding window demand is a predicted value that a sliding window demand will reach at the end of the present demand interval, assuming the instantaneous power load will not change. Predicted demand is updated as new instantaneous power is measured, and reflects changes in power load as they occur. It is calculated as average on the prior calculated set of partial block interval demands (N-1, assuming N = number of demand intervals in the sliding window), and on the predicted value for the present demand interval that is extrapolated to its end, considering power integrated from the beginning of the interval and instantaneous power being measured.
Due to extrapolation, predicted sliding window demand is very sensitive to the duration of the interval over which extrapolation is made. The predicted demand deviation might be slightly increased near the beginning of the demand interval where a base for extrapolation is too small.
In the event that sliding demand window comprises a solitary demand period (N = 1), the prediction will be made at any moment considering only present accumulated demand and measured instantaneous power.
Power Demand Interval
Demand interval measurements can use internal real-time clock (RTC), or an external source as a time reference.
When using the internal RTC, the demand period time is defined by the user from 1 and up to 60 minutes in preset intervals. For the external source, the external pulse sensed via a meter digital input denotes the start of the new demand interval. The number of demand periods for the sliding window technique can be defined from 1 to 15, for either time reference.
Using internal time base, the start of each demand period is always synchronized with the beginning of the nearest round interval divisible by the demand period. In the event of loss of power, or changing any demand parameter, the meter immediately begins a new shorter demand interval until the first synchronization.
The demand interval duration may by slightly shortened or prolonged by RTC time update. In all cases, the demand interval will be terminated at the nearest round boundary.
The demand minimum/maximum keeping registers can be cleared simultaneously via the front panel, communications, or by using a programmable setpoint. The demand accumulating registers and demand interval synchronization are never affected by reset, so the user can reset the extreme demands at any point within the demand interval without the risk of tampering with demand measurements.
The meters can provide a timing pulse indicating the beginning of new demand interval for consumer use. It can be used as synchronous time reference for other meters that have no internal demand interval synchronization.
The beginning of the demand interval is also asserted in the meter as an internal event that can be used to trigger a setpoint (e.g., to synchronize self-readings with demand interval, or to operate relay output in a customized order).