Diversity and Part Load
Diversity (simultaneity) and part load: time shift of demand profiles, target diversity curve, part load ratio and their effect on sizing and simulation.
VICUS Districts Page 1 of 3 in this chapter
Overview
Not all consumers of a heat network reach their peak load at the same time. VICUS Districts represents this through two separate quantities: diversity (also called diversity) reduces the consumer loads aggregated along shared routes, and the part load ratio reduces the load actually drawn by an individual consumer. Diversity is only taken into account if it is activated for the network; it then acts on pipe sizing, steady-state calculation and dynamic simulation. The part load ratio is applied independently of this - even when diversity is deactivated - and acts on the steady-state calculation and dynamic simulation; it enters pipe sizing only if Actual maximum power is selected as the dimensioning power.
This page explains the calculation model. Its operation is described on two follow-up pages:
- the network properties tab Diversity and part load
- the dialog Diversity calculation
What does the diversity factor mean?
The diversity factor of a route is the ratio of the peak load actually occurring simultaneously to the sum of the individual peak loads of all consumers supplied via that route:
A value of 1.0 means no diversity - all consumers peak together, there is no relief effect. Values smaller than 1.0 indicate that the peaks occur at different times and the route only has to carry a fraction of the summed individual peaks. The physical floor of the default curve is at about 0.45; as the optimization target, however, the curve is limited to 0.65 from below by default (see below).
Target diversity curve
For each route the calculation targets a diversity factor that depends on the number of downstream consumers. The default curve follows Winter et al. (Euroheat & Power, 2001):
with , , and . The curve is stored as a lookup table up to a consumer count of about 300 (support points , linearly interpolated in between). It decreases monotonically with rising consumer count from 1.0 (single consumer) towards the floor .
The target curve applies network-wide. For individual pipes it can be overridden by a custom diversity, and its support points can be edited in the dialog Diversity calculation.
Lower limit of the target curve
The optimization does not aim at the curve itself but at the curve limited from below by the Minimum diversity (default 0.65): wherever the curve drops below this value, the limit becomes the target. This keeps the target realistic for routes with very many downstream consumers, where the curve would otherwise approach its floor of about 0.45. With a value of 0 the curve is used unchanged. A custom diversity per pipe is not limited but used as entered. The setting is found in the dialog Diversity calculation under Detailed settings.
Time shift of the demand profiles
The achieved diversity factor does not result from a mere multiplication but from an optimization: each consumer profile is shifted forward or backward within a maximum permissible time shift, so that the peak loads on shared routes are spread out. The optimizer searches for the shifts that bring the calculated diversity of each route as close as possible to the target curve, and then verifies the result by an exact evaluation over the entire year.
As a result the calculation provides:
- per consumer an optimized time shift in hours (positive = delay, negative = advance of the profile),
- per route the achieved diversity factor.
Only routes with at least three downstream consumers enter the optimization; routes with fewer consumers keep their computed value. Only consumers that participate in diversity are counted (see below). Consumers that do not influence any such route keep a time shift of 0, unless they are shifted afterwards for the minimum load at the network peak.
Minimum load at the network peak
An optimization aimed purely at the target curve may shift individual consumers so far that they draw hardly any power at the time of the network peak. The Minimum load at peak (default 30 %) prevents this: after the optimization it is checked whether every consumer draws at least this share of its own maximum power at the time of the realized network peak load. Consumers below it are shifted again within the maximum time shift - to the shift closest to the optimized one that satisfies the condition. This ensures that all consumers are supplied in the steady-state calculation.
If the minimum load cannot be reached for a consumer within the maximum time shift, the closest possible shift is used and a Minimum load at network peak warning is shown. A value of 0 % disables the condition. The setting is found in the dialog Diversity calculation.
Excluding consumers from diversity
Individual consumers can be taken out of the optimization with Exclude from diversity in the Diversity and part load tab. An excluded consumer is not shifted to spread the peaks and does not count for the target diversity of the routes. Instead, its demand profile is placed within the maximum time shift so that its peak coincides with the peak load of the most heavily loaded route supplying it - it always contributes its full peak (diversity factor 1).
This is useful for large consumers such as commercial or industrial buildings whose peak should not be used to reduce the network peak.
Part load ratio
The part load ratio is an independent quantity per consumer in the range 0.0 - 1.0. It defines the fraction of the heating power that the consumer actually uses: 1.0 corresponds to the full design load, smaller values reduce the applied load accordingly. It is suitable, for example, for existing buildings whose actually drawn load is permanently below the installed connection power.
The part load ratio scales the consumer load in the steady-state calculation and in the dynamic simulation. In pipe sizing its effect depends on the selected dimensioning power: with Connection power it is not applied - every consumer is taken with its full connection power. With Actual maximum power the power of each consumer is multiplied by its part load ratio, see Running pipe sizing. In the diversity calculation the part load ratio weights the consumers when the network peak for the minimum load is determined.
Important in practice
Do not confuse diversity and the part load ratio - they act differently on purpose: diversity reduces pipe sizing (via the time shift of the peaks), whereas the part load ratio acts on the steady-state calculation and dynamic simulation, and in pipe sizing only with the dimensioning power Actual maximum power. If you size with the connection power, the pipes stay sized for the full design load, while permanently lower drawn loads (e.g. of existing buildings) are still represented realistically in calculation and simulation.
Effective diversity
For display purposes only, the achieved diversity of a route can be multiplied by the load-weighted mean part load ratio of its consumers. What is then displayed is the effective diversity, which combines time shift and reduced consumer load:
This is purely a display quantity in the tab Diversity and part load; it is not stored and does not enter the sizing.
Effect on the calculations
The diversity factor and the time shifts take effect only if Consider diversity is activated; otherwise the diversity factor is 1.0 and no time shifts are applied. The part load ratio is applied independently of this (even when diversity is deactivated).
| Quantity | Pipe sizing | Steady-state calculation | Dynamic simulation |
|---|---|---|---|
| Diversity factor per route | × nominal power | – (realized via time shift) | – (realized via time shift) |
| Time shift per consumer | – | applied to the demand profile | applied to the demand profile |
| Part load ratio per consumer | only with Actual maximum power: × consumer power | × applied load | × applied demand profile |
- Pipe sizing: the nominal power of each route is multiplied by its achieved diversity factor, see Calculation method. If diversity is activated and no up-to-date results exist, pipe sizing runs the calculation automatically beforehand.
- Steady-state calculation and dynamic simulation: instead of a blanket factor, the optimized time shifts are applied to the demand profiles - diversity thereby arises directly from the offset loads. In addition, the part load ratio scales the respective load.
- Time-series export: when exporting the building demands, the shifted profiles can be output (option Consider node time shift), see Editing building demands.
Results can become outdated
The calculated time shifts and diversity factors belong to exactly the network state for which they were calculated. If the network topology, a building demand, the part load ratio or exclusion of a consumer, a custom diversity, the target curve, the calculation settings or the dimensioning power of pipe sizing change, the stored results are considered outdated and must be recalculated before being applied. The tab points this out with a warning message.