How to get meaningful energy efficiency results when there is a spaghetti of boards
Look at that mess! We focus on the result, not just ‘clip the shit out of it’. We need to be careful about which CTs go on which units, as you can’t (easily) calculate between units. Begin!
- Main useful breakdown is MAIN OVAL , SMALL OVAL, NEWBUILD, CAFE, KITCHEN , CANTEEN, DIETICIAN, PUMP, COMP, AIRCON, COKE, COLDROOMS and ‘U’ (Unmonitored: on MDB but no CT)
- ‘U’ can be calculated as ‘net MDB’ less ‘stuff we have measured’ on the MDB. However, with so many CTs, we need multiple units and can’t do inter-unit arithmetic. We would need to have MDB Total on the same unit as all its loads (too many for one unit). So we give up on U.
- By putting a unit capturing Net we can get Import and Export. The same unit can capture ALL the second level (MDB, OVAL, OVAL) which is useful for error checking. But that breakdown is poor: MDB needs to be broken down.
- For consumption results to be readable (think ‘pie chart’ (single level) not ‘starburst’ (multi level)) we need to our loads to add to 100% (i.e. Net). This means in the example:
- OVALs: Field Lights (x2)
- All sub boards and minor loads of MDB. (As noted, we can’t derive UM so we aim to capture everything on MDB.)
- This leaves NEWBUILD which have solar on it. We would ideally break this into COMM/TENANT1/etc but it’s an impossible board to get all tenants.
- Instead, we monitor at NEWBUILD and can get some breakout loads (e.g. AIRCON).
- Then we need to get “everything else on NEWBUILD”. Normally this is easy (UM= NEWBUILD – AIRCON). But with Production, we say UM = NET + PRODUCTION – AIRCON. See reference. We’ll do an example below. But what about the battery? Um… treat it like production? No, cos production is positive. Treat it like a load? Yes: when charging it’s a positive load, when discharging it’s negative load. A short-range pie chart will break with negative loads, but a battery over any day+ range will sum to zero and hence disappear from the pie chart. This is right: it’s just a time-shifter, it doesn’t use or produce any energy (overall). See below for more on this.
- We monitor NEWBUILD COMM BATTERY SOLAR. This allows us to use the equations to get “U” (Unmonitored) which is T1+T2+T3+T4.
- Hence on our single level pie chart we can remove “NEWBUILD” and replace with NB-COMM, NB-BATTERY NB-SOLAR and NB-T1T2T3T4. We expect the battery usually be zero (net).
Pies
That’s pretty confusing. Let’s summarise with three pie charts:
Notice in the final one we don’t have BATTERY. We didn’t need to clip battery! But we like to see it working, but just know it’s not going to show in energy results.
Notice too that we managed to extract PRODUCTION from NEWBUILD. Normally this would a problem, as we just see NEWBUILD which has self-consumption (i.e. some of the solar is consumed via NEWBUILD board). Instead, we got PRODUCTION directly and COMM directly, and we known NEWBUILD total. This allows us to get TENANTS (i.e. Unmonitored): PRODUCTION is ‘removed’. See below for details on that.
Deriving ‘Unmonitored’
How did we get ‘tenants’ (T1T2T3T4)? We didn’t clip them! We derived them:
| ct’d | ct’d | ct’d | ct’d | ct’d | derive* | result |
| NEWBUILD | SOLAR | NEWAIRCON | BATTERY | COMM | T1+T2+T3+T4 | . |
| 5 | 2 | 1 | 1 | 1 | 4 | Importing 5, producing 2, consuming 7: OK |
| 0 | 0 | 1 | 1 | 1 | -3 | OK: impossible as T1-4 must be positive |
| 10 | 2 | 1 | 1 | 1 | 9 | More import, more unknown load: OK |
| 5 | 0 | 1 | 1 | 1 | 2 | No solar, consumption=import |
| 5 | 5 | 1 | 1 | 1 | 7 | Solar but same import: must be consumed |
| 5 | 5 | 1 | -1 | 1 | 9 | Battery discharging, same import, more consumption |
| 3 | 0 | 0 | -3 | 0 | 6 | Battery dischanging + import = 6 consumption |
* derived uses equation: U = N+P-L or NEWBUILD+SOLAR-NEWAIRCON-BATTERY-COMM