NEOBIOME

CR_057: NZ Dairy-Shed Electricity Benchmarks — milking, milk cooling & micro-dairy (AI-compiled synthesis, anchored…

Properties4
typesource
source_typecompiled_research
authorAI deep research
topicsenergy_generation, food_systems_agriculture, nz_specific

13 min read

CR_057: NZ Dairy-Shed Electricity Benchmarks — milking, milk cooling & micro-dairy (AI-compiled synthesis, anchored…

NZ Dairy-Shed Electricity Benchmarks — Milking, Milk Cooling & Micro-Dairy Use

AI-compiled synthesis, anchored on four verified NZ primaries — model figures are cited from the primaries, not from this page

This is a compiled-research (cr_) synthesis of NZ dairy-shed electricity use, assembled for the NI energy-demand slice (community dairy-processing electricity load). It is data_quality: medium, but its four load-bearing primaries — Marshall-Tate 2017 (Massey), CAENZ 2007, MPI NZCP1 2024, and the EECA Technology Demonstration lists — were all retrieved and independently verified in this batch (listed in underlying_sources:). The corpus cites the numbers from those primaries. ⚠ Three restatement discrepancies in the synthesis were corrected at ingest (verified-source-only): CAENZ load split is vacuum 17% / other 34%, not the synthesis’s 16% / 35%; CAENZ vacuum daily use is 72→45 kWh/day, not 71→44; the EECA Inniskillen saving is 24,090 kWh/yr (a 24,089 variant appears in the fund-list PDF). ⚠ The micro-dairy figures (0.0326 kWh/L; 131-399 kWh/1,000 L; the 1-5 cow table) are the report’s own engineering DERIVATIONS, not measured NZ data.

Summary

An AI-compiled research synthesis of New Zealand dairy-shed electricity benchmarks, spanning three scales: (1) commercial whole-shed per-cow demand and its subsystem breakdown; (2) milk-cooling and refrigeration load in detail; and (3) a derived small-scale / house-cow / micro-dairy estimate for a 1-5 cow eco-village dairy. It is written explicitly for the Neobiome context, arguing that the practically relevant metric for a very small seasonal milk flow is kWh per litre, not the commercial kWh/cow/year, because refrigeration standby, cleaning hot-water volume, and vacuum setup do not scale down linearly.

It is ingested as the compiled index to the dairy-shed energy evidence retrieved in the 2026-07-25 batch. Its role is orientation and the derived micro-dairy scenario; every commercial and regulatory figure it quotes is verified and carried on its primary’s own page — Marshall-Tate 2017 for the per-cow benchmarks (LIT_117), CAENZ 2007 for the Southland load split and per-process intensities (OT_203), MPI NZCP1 2024 for the regulatory cooling floor (REG_033), and the EECA lists for real NZ milk-chilling demonstrations (OT_204).

Key claims

- claim: "FRAMING — this is an AI-compiled synthesis. Its model-relevant figures are eligible only because the four primaries behind them were retrieved and independently verified in this batch (underlying_sources: dairy_massey, dairy_caenz, dairy_nzcp1, dairy_eeca); the model cites those primaries, not this page. Three synthesis restatement errors were corrected at ingest (CAENZ split 17%/34% not 16%/35%; CAENZ vacuum 72->45 not 71->44 kWh/day; EECA Inniskillen 24,090 with a 24,089 fund-list variant). The micro-dairy figures below are the report's OWN engineering derivations, not measured NZ audit data."
  source_location: "This synthesis, whole document; corrections per RETRIEVAL_REPORT.md (2026-07-25) items 1-3."
- claim: "NZ COMMERCIAL WHOLE-SHED benchmark. Non-irrigated NZ dairy-shed electricity ranges 150-173 kWh/cow/year in the survey literature (Sims 2004; Barber & Pellow 2005; Miller & Glenn 2011; the 150 low from new Southland farms); the mechanistic base-farm model gives 172-177 kWh/cow/year (Manawatu 177 / Taranaki 174 / Bay of Plenty 172). Recommended top-level benchmark range 150-177 kWh/cow/year; broad observed middle-77% band 130-209 kWh/cow/year. Whole dairy shed, IRRIGATION EXCLUDED. [verified against the primary Massey MTech]"
  source_location: "This synthesis, 'Commercial NZ dairy-shed electricity benchmarks' + 'Practical benchmark ranges'. Verified on [[lit_117_marshall-tate-2017-dairy-sheds|LIT_117]] (Marshall-Tate 2017): literature range Fig 2 discussion (PDF p.15); base-farm Table 9 (PDF p.54); distribution Fig 5 (PDF p.18)."
- claim: "NZ COMMERCIAL MILK-COOLING per-cow. Massey's model separates milk cooling 25-26 kWh/cow/year from vat heat loss 7-8 kWh/cow/year; refrigeration validation 33 kWh/cow/year; actual uninsulated indoor vats averaged 31.7 kWh/cow (SD 3.12), insulated-vat average 22 kWh/cow (SD 0.9). Recommended commercial milk-cooling range 22-33 kWh/cow/year (insulated-vat low end to DX-refrigeration-validation high end). Vacuum/milking: ~20 kWh/cow/year for VSD machines, ~30-31 for non-VSD. [verified against the primary Massey MTech]"
  source_location: "This synthesis, 'Milk cooling and refrigeration' + 'Practical benchmark ranges'. Verified on [[lit_117_marshall-tate-2017-dairy-sheds|LIT_117]] (Table 9 PDF p.54; refrigeration/insulation discussion PDF pp.56-57)."
- claim: "NZ SOUTHLAND load split (CAENZ typical shed, ~100,000 kWh/year, ~15,000 L/day at peak). CORRECTED to water heating 29% / milk cooling (chillers) 20% / vacuum pumping 17% / pumping+lighting+other 34% (Fig 2.1 pie chart). ** The synthesis's 'vacuum 16% / other 35%' is wrong: 16% is a separate in-text worked example (a 10 kW pump x ~1,600 h/yr = 16,000 kWh), while the headline chart rounds vacuum to 17,000 kWh (17%).** [verified + CORRECTED against the primary]"
  source_location: "This synthesis, 'Commercial NZ dairy-shed electricity benchmarks' (Southland split). Verified/corrected on [[ot_203_caenz_2007_dairy_shed_energy|OT_203]] (CAENZ 2007 Fig 2.1, PDF p.17)."
- claim: "NZ CAENZ per-process intensities & vacuum trial. Water heating ~0.1 kWh/L (10->85 C); ~860 kWh/year saved per 1 C off the pre-cooler; vat refrigeration COP 2.7-2.9 when healthy; refrigeration ~10 kW running ~10 h/day = ~100 kWh/day = ~20,000 kWh/season (~20% of shed electricity) at 15,000 L/day; vat insulation ~80% heat-gain cut ~= 2,850 kWh/year electricity. Vacuum-pump trial: rotary-vane replacing water-ring cut daily use 72->45 kWh/day (Table 3.5), a 38% reduction / 6,750 kWh/year saving. ** The synthesis's '71->44 kWh/day' is wrong; the primary Table 3.5 reads 72->45 (the 38% and 6,750 kWh/yr are exact in both).** [verified + CORRECTED against the primary]"
  source_location: "This synthesis, 'Milk cooling' + 'Milking plant / vacuum pump'. Verified/corrected on [[ot_203_caenz_2007_dairy_shed_energy|OT_203]] (CAENZ 2007: intensities pp.41/79/80-82, refrigeration p.73, vacuum Table 3.5 p.29 / p.28)."
- claim: "NZ EECA real milk-chilling demonstrations (model-relevant as SAVINGS/example figures, not per-cow averages). Rylib Dairies Vari-Cool 22 = 31,424 kWh/year saving, -28% kWh/kg MS vs prior year. Inniskillen Dairy Vari-Cool 22 = 24,090 kWh/year saving, milk leaving site at an average 5 C (** 24,089 variant in the fund-list PDF — cite 24,090, note the 1-unit discrepancy**). Sullivan Farm Snapchill ice-bank with full heat recovery = 12,000 kWh/year saving, milk ~18 C -> 4-5 C, can use off-peak electricity / avoid a line upgrade. [verified against the primary EECA lists]"
  source_location: "This synthesis, 'Milk cooling and refrigeration' (EECA examples). Verified on [[ot_204_eeca-tech-demo-milk-chilling|OT_204]] (EECA Technology Demonstration Projects Jan 2024 + Fund project list; Rylib/Inniskillen/Sullivan rows)."
- claim: "NZ REGULATORY cooling floor that drives cooling design (NZCP1 2024). Cool to 10 C or below within 4 h of the commencement of milking; to 6 C or below within the SOONER of 6 h from commencement or 2 h from completion; hold at or below 6 C without freezing; subsequent milkings must not exceed 10 C. Continuous/robotic milking: cool to 6 C or below immediately after harvesting and hold. Primary cooling capable of reaching within 3 C of coolant/cooling-water temperature; coolant flow guideline >= 2.5x the maximum milk-pumping rate. Raw milk sold to consumers (RCS): <=6 C at all times initial-cooling-to-delivery, <=30 h harvest-to-delivery, label 'store at or below 4 C'. [regulatory rule, not an energy benchmark; verified against the primary code]"
  source_location: "This synthesis, 'NZ regulatory requirements that drive cooling design'. Verified on [[reg_033_mpi-nzcp1-farm-dairies|REG_033]] (MPI NZCP1, 31 May 2024, cooling clauses). Raw-milk RCS detail cross-refs [[reg_031_raw-milk-consumers-regs-2015|REG_031]] (2015 raw-milk regs already in corpus)."
- claim: "DERIVED MICRO-DAIRY ESTIMATE (the report's OWN engineering derivation; NOT measured NZ audit data). Cooling thermal duty of fresh milk 35->6 C = 1.03 kg/L x 3.93 kJ/kg.K x 29 K / 3600 = 0.0326 kWh of heat removed per litre (compressor electricity ~0.0163 kWh/L at COP 2.0; ~0.0045-0.0056 kWh/L if a plate cooler pre-drops 35->16 C first). Central 1-5 cow scenario (10 L/cow/day x 270 days, no plate cooler, COP 2.0, 0.8 kWh/day dedicated-fridge standby, MPI-style hot-water cleaning) spans ~399 kWh/1,000 L at 1 cow (2,700 L/yr, ~1,078 kWh/yr total) down to ~131 kWh/1,000 L at 5 cows (13,500 L/yr, ~1,762 kWh/yr total). Standby refrigeration (~292 kWh/yr) + fixed hot-water cleaning dominate at 1-2 cows, so per-litre intensity is several times a commercial shed even though the milk's own cooling duty is tiny. [derived in the report from IASRI courseware cp=3.93 kJ/kgK + NZCP1 6 C rule; NOT primary/measured]"
  source_location: "This synthesis, 'Small-scale / house-cow / micro-dairy estimates' (cooling physics + central estimate table). Physics inputs: IASRI dairy-engineering courseware (cp=3.93 kJ/kg.K); target from [[reg_033_mpi-nzcp1-farm-dairies|REG_033]] (6 C rule)."
- claim: "RECOMMENDED BENCHMARK RANGES for NI energy-demand modelling of a community dairy-processing load. Commercial NZ shed: 150-177 kWh/cow/year (top-level), 130-209 (observed middle-77%). Commercial milk cooling alone: 22-33 kWh/cow/year. Commercial vacuum/milking: ~20 (VSD) to 30-31 (non-VSD) kWh/cow/year. For a 1-5 cow micro-dairy: use kWh/1,000 L scenario modelling (~131-399), NOT commercial kWh/cow averages. Biggest low-energy levers: right-size refrigeration, ice-water / plate pre-cooling, shut off dedicated refrigeration outside lactation, minimise compliant hot-water volume, avoid an oversized fixed-speed vacuum. [synthesis recommendation; commercial ranges verified on the primaries above]"
  source_location: "This synthesis, 'Practical benchmark ranges to use in proposal modelling' + 'Design implications'. Commercial ranges verified on [[lit_117_marshall-tate-2017-dairy-sheds|LIT_117]]; micro-dairy scenario is the report's own derivation."

Underlying sources

The four primaries that make this synthesis model-eligible (cr_ needs-a-primary satisfied), each with its own verified page in this corpus:

IDPrimaryRole in this synthesis
[[lit_117_marshall-tate-2017-dairy-shedsLIT_117]]Marshall-Tate (2017) Power Systems for Dairy Sheds, Massey MTech, 142 pp
[[ot_203_caenz_2007_dairy_shed_energyOT_203]]CAENZ (2007) Improving Dairy Shed Energy Efficiency, 5 Southland sheds, 122 pp
[[reg_033_mpi-nzcp1-farm-dairiesREG_033]]MPI NZCP1: Design and Operation of Farm Dairies (31 May 2024)
[[ot_204_eeca-tech-demo-milk-chillingOT_204]]EECA Technology Demonstration project lists (2024)

Unretrieved secondary citations in the synthesis (leads only, NOT verified, NOT model-eligible): the DairyNZ milk-cooling webpage (milk cooling ~30% of dairy energy cost); the 2010-11 MAF/EECA/Fonterra 150-farm pilot reported second-hand by the Otago Daily Times / Rural News (average 112,100 kWh/farm/yr incl. irrigation; audit split water heating 24% / pumping 22% / refrigeration 17% / vacuum 15%) — journalism, not a retrieved primary, so its figures must not enter the model without retrieving the pilot report; the MPI Raw Milk RCS Notice 2022 (the 2015 raw-milk regs are held as REG_031); the IASRI dairy-engineering courseware (physics constants); and the Ireland/Australia/UK proxy fact-sheets (Teagasc; Upton et al. 2013; Dairy Australia; AHDB), which the synthesis itself flags as proxies.

Neobiome Intelligence relevance

Feeds the community energy-demand model as the dairy-shed electrical-load index. It is filed with the other cross-cutting NZ energy-demand primaries (RD_007, OT_028, RD_008) rather than under one EDT domain, because it is a kWh-load synthesis that serves both the energy model and the food/dairy pathway.

What it contributes that the primaries alone do not:

  1. A scale bridge. It states, and the derivation supports, that a commercial 150-177 kWh/cow/year figure LIT_117 is not a safe design basis for a 1-5 cow eco-village dairy, because standby refrigeration, cleaning hot-water volume, and vacuum setup do not scale down linearly. For a micro-dairy the model should use kWh per 1,000 L scenario modelling (~131-399 under MPI-style cleaning, the report’s own derivation), not kWh/cow.
  2. The cooling-duty floor. The milk’s own sensible-heat cooling duty is physically modest (~0.0326 kWh/L, 35->6 C — the report’s derivation from cp = 3.93 kJ/kg.K and the REG_033 6 C target), so at 1-2 cows the dedicated fridge/chiller standby can exceed the actual milk-cooling load — the design lever is right-sizing and switching the chiller off outside lactation, not a bigger vat.
  3. A regulatory hard constraint. Any legal raw-milk pathway must meet NZCP1’s <=6 C cooling floor REG_033 (and, for RCS raw-milk sale, the <=6 C chain + 30 h + <=4 C label, REG_031), which sets the cooling target the energy model must hit.

No engine cell is auto-changed by this ingest. The verified figures live on the four primary pages; wiring a sized community-dairy load from them (and from the derived micro-dairy scenario) is a modelling step, not a data entry. The derived micro-dairy numbers are carried as the report’s own estimates, to be treated as scenario inputs pending any measured NZ micro-dairy audit data.

Retrieval provenance

  • Upstream sources: the synthesis carries inline links to its own primaries. The four model-relevant ones were retrieved and verified in this batch and are listed in underlying_sources:; the remainder are secondary citations recorded above as leads.
  • Prepared by: an AI research tool. ⚠ The specific tool and the exact prompt were NOT recorded in the file and have not been inferred. The document’s internal figure-table structure indicates it was produced from a multi-figure research brief, but that brief is not in corpus.
  • Retrieved date: 2026-07-25 is the date the file entered the corpus, not necessarily the date the synthesis was generated. It cites the EECA Technology Demonstration Projects January 2024 release and the 31 May 2024 NZCP1, so it postdates those.
  • Verification: UNLIKE the usual cr_ posture, this synthesis’s model-relevant figures were independently retrieved and verified against their primaries on 2026-07-25 (Massey 6/6, CAENZ 16/18, NZCP1 7/7, EECA 6/6). The two CAENZ discrepancies and the one EECA inter-document discrepancy were corrected at ingest. data_quality: medium reflects that it is a compiled synthesis; the numbers the corpus uses are carried on the verified primary pages.

Notes

  • Why medium, not low. The usual cr_ risk is mis-attribution — right organisation, wrong document. Here the batch retrieved and read every primary behind the model-relevant figures, so the numbers are anchored, and the only defects found were three trivial restatement drifts (16 vs 17%, 35 vs 34%, 71 vs 72 -> corrected to 72->45, 24,090 vs 24,089), all corrected. It stays medium because it is a synthesis, not a primary, and because its unretrieved secondary citations (notably the 150-farm pilot journalism) are not verified.
  • Corrections applied, restated for the record: (1) CAENZ load split = vacuum 17% / other 34% (Fig 2.1), not 16% / 35%; (2) CAENZ vacuum daily = 72->45 kWh/day (Table 3.5), not 71->44; (3) EECA Inniskillen = 24,090 kWh/yr (projects PDF), with a 24,089 variant in the fund-list PDF.
  • Derived vs measured — the load-bearing distinction. The micro-dairy cooling duty (0.0326 kWh/L), the kWh/1,000 L band (131-399), and the 1-5 cow table are the report’s engineering derivations from milk specific heat (cp = 3.93 kJ/kg.K, IASRI courseware) and the NZCP1 6 C target, using stated assumptions (10 L/cow/day, 270 days, COP 2.0, 0.8 kWh/day standby, MPI-style cleaning). They are not measured NZ micro-dairy audit data and must be labelled as scenario estimates wherever used.
  • Scope: this is the electrical-load synthesis, distinct from the production/effluent dairy pages (OT_172, OT_171, LIT_086).

Research targets

Documents to retrieve

  • None. The dairy-shed electrical load is well-sourced by the four verified primaries in this batch. (The 2010-11 MAF/EECA/Fonterra 150-farm pilot report would upgrade the pilot figures from journalism to primary, but it is a low-priority enrichment, recorded here rather than opened as an RT.)

Research gaps

  • Not raised as an RT (recorded here): a measured NZ community-/house-cow-scale dairy electrical load (versus these commercial ~15,000 L/day sheds). The micro-dairy figures on this page are engineering derivations, not audit data; a measured small-scale figure would replace them. This is a future validation task, not a retrieval target.

Connections

Links to

Sources (11): LIT_086 · LIT_117 · OT_028 · OT_171 · OT_172 · OT_203 · OT_204 · RD_007 · RD_008 · REG_031 · REG_033

Referenced by

EDT domains (1): D01: Renewable Energy & Storage Systems

Graph View

Backlinks