Chilled Water Systems 2026 | Complete Technical Guide Chillers IoT | Wattnow
⚠️ Copying content is not permitted
Industry Guide – 2026 Edition

Master
Chilled Water
Systems

Refrigeration cycle, components, 2026 refrigerants, sizing, data-driven IoT monitoring, F-Gas framework and energy regulations from sensor to KPI.

60%
of electricity bill in commercial buildings
-40%
savings via monitoring
18 months
average IoT ROI
2027
GWP < 150 mandatory

Start reading
CHAPTER 1

Fundamentals: what is a chilled water system

Chiller definition: producing chilled water to air-condition — 30 to 60% of a building's electricity bill

A chilled water system (CWS, or chiller) is a thermodynamic machine that produces chilled water ; typically between 5 and 18°C depending on use, and sometimes down to -10°C in industrial settings via a glycol loop. This water circulates through a hydraulic network to supply AHU cooling coils, radiant ceilings, fan coil units, or industrial process heat exchangers. Unlike direct expansion systems (VRF), it places an intermediate fluid , water ; between the refrigerant circuit and the terminal units. This concentrates the F-Gas regulated charge within the machine itself and enables distribution over several hundred metres.

12%of UK commercial electricity use
£430Mannual UK market
20 yearstypical service life

1.1 The refrigeration cycle in four stages

The core of a chiller is a thermodynamic cycle with four stages, identical in principle to a domestic refrigerator but at a vastly different power scale: from a few tens of kilowatts of cooling (kWc) to several megawatts for a datacentre or district cooling network chiller.

Figure 1: The four-stage refrigeration cycle of a chiller
Refrigeration cycle of a chiller : the four stages Compressor LP cold gas → HP hot gas 80–95% of electricity use Condenser Heat rejection → air/water HP gas → HP liquid Expansion valve Electronic EEV HP → LP, T°↓↓ Evaporator Absorbs heat from chilled water LP liquid → LP gas HP discharge, ~80°C Two-phase LP mixture, ~5°C Cold side (LP): absorbs heat from building water — Hot side (HP): rejects it to air or a cooling circuit

1.2 Three structural benefits of a chilled water system

  • Concentrated F-Gas charge: the refrigerant remains confined within the machine in the plant room, whereas VRF distributes it throughout the entire building. With the revised F-Gas Regulation 2024 indirectly pricing CO₂-equivalent charge, this is a growing economic and regulatory advantage.
  • Long-distance distribution: a chilled water loop can serve several hundred metres of pipework with ease, whereas VRF is limited to around 150 m. This enables sharing across campus buildings, hospital floors, and factory workshops.
  • Rich instrumentation: ΔT, flow rate, setpoint, pressures : all become control levers accessible to the building management system. This is precisely what makes chillers the primary target of data-driven HVAC control in 2026.
📌 Key takeaway: Chillers become the preferred solution once cooling loads exceed 100 kWc, distribution distances are long, centralised operation is required, or F-Gas regulatory constraints weigh against the VRF option. In 2026, chillers are also the technology benefiting most from the data-driven revolution in HVAC: no other equipment family exposes as many useful monitoring points.

1.3 Market overview and three simultaneous transitions

The European chiller market represents approximately €3 billion annually. In 2026, the sector is undergoing three converging transitions: refrigerant transition (low GWP, natural refrigerants), energy transition (part-load modulation, electrification of heating via reversible chillers) and digital transition (systematic IoT monitoring, applied artificial intelligence). Three segments structure the market: small chillers up to 100 kWc (scroll inverter, R32 and R454B), medium chillers 100 kWc to 1 MWc (screw or scroll inverter), large chillers above 1 MWc (magnetic bearing centrifugal, R1234ze or natural refrigerants).

CHAPTER 2

Components and hydraulic architecture

Five major compressor technologies share the market in 2026. The choice of modulation determines 80% of real annual efficiency and should take precedence over nominal EER in specifications.

2.1 The compressor: the energy core of the machine

TechnologyPower rangeModulationTypical use 2026
Scroll inverter20–600 kWcContinuous 25–100%Small-medium commercial, R32/R454B
Screw inverter150–1500 kWcContinuous 15–100%Medium-large commercial, industrial
Magnetic centrifugal500 kWc–5 MWcContinuous 10–110%Large commercial, datacentre, R1234ze
Classic centrifugal1–20 MWcIGV + variableDistrict cooling networks
Piston / ammonia screw100 kWc–10 MWcStaged or variableIndustrial refrigeration, food & beverage

Variable speed modulation (inverter) continuously adjusts compressor speed and is the most efficient solution at part load, where the vast majority of operating time is spent. Initial premium of 10–25% recovered in under 5 years in typical commercial operation. Magnetic levitation takes this logic to its extreme by eliminating mechanical bearings: no friction, instant start, continuous modulation down to 10% load, extended service life. This is now the benchmark for datacentres and large new or retrofit commercial builds.

2.2 Variable primary hydraulic architecture

A chilled water plant is organised around two or three hydraulic loops: the primary loop links chillers and terminals (or decoupling point), the secondary loop supplies the terminals, and an optional condenser loop rejects heat via a cooling tower or dry cooler. Modern architecture eliminates the decoupling vessel in favour of variable primary flow, which is simpler and more energy-efficient.

Figure 2: 2-chiller plant in variable primary with differentiated circuits
Variable primary hydraulic architecture Chiller 1 Inverter / magnetic Evaporator Chiller 2 N+1 redundancy Evaporator P P VFD pumps AHU + cooling coils Modulating 2-way valves Radiant ceilings 14 / 18°C Fan coil units 6 / 12°C Supply chilled water (~7°C) Warmed return (~13°C) Target ΔT: 6–10°C depending on use

2.3 Reference formulae for sizing

Q (kW) = m (kg/s) × Cp (kJ/kg·K) × ΔT (K)
Flow rate (m³/h) ≈ Capacity (kW) ÷ (1.16 × ΔT)

For a 500 kWc chiller with ΔT of 5°C, the primary flow rate is approximately 86 m³/h. Increasing ΔT to 10°C halves the flow rate and reduces pumping energy by a factor of eight (cubic law for pressure losses in turbulent flow). This is one of the most cost-effective design levers available.

Interactive Simulator: Chilled Water Loop ΔT Analyser
Animated chilled water loop Chiller Inverter Evaporator P AHU coils Radiant ceilings Fan coil units ΔT 7°C 7°C → ← 14°C Supply Return

Animation speed reflects actual flow rate

7 °C
3° Critical7° Good10°15° Optimal
500 kW
60 %
0.18 £/kWh
2000 h
Primary flow rate
m³/h
Pumping power
kW
Pumping cost / year
£
Estimated EER
Cubic law: pumping power vs ΔT (reference: ΔT = 5°C)

Wattnow measures your EER in real time

Our IoT sensors (MID-class calorimetric chain, class 0.5S electricity meters, paired PT100 temperature probes) calculate EER and equivalent observed IPLV hour by hour. A continuous monitoring algorithm triggers an alert if the evaporator approach temperature drifts (early fouling signal) or if sub-cooling decreases (refrigerant leak alert). Data-driven monitoring of the cooling plant is the only way to secure your long-term energy performance trajectory and CSRD reporting over 20 years.

Exclusive content

Everything you need to know
to optimise and decarbonise

Refrigerants 2026, sizing, IoT BACnet/MQTT monitoring, ROI, datacentre/hospital/food industry case studies, F-Gas and energy regulations.

Chapter 3: Typologies and 2026 RefrigerantsR32, R454B, R1234ze, R290, R744, R717 — PFAS uncertainty
Chapter 4: SizingThree pitfalls leading to 50–80% oversizing
Chapter 5: IoT ControlBACnet, edge+cloud, calorimetric chain, AI
Chapter 6: Optimisation10 levers ranked by declining ROI
Chapter 7: Predictive MaintenanceEarly warning signals, water quality, modern contracts
Chapter 8: ROI & TCOEnergy incentives 15–40%, supervision ROI 18–36 months
Chapter 9: Case StudiesDatacentre PUE 1.52→1.29, hospital, HQE commercial
Chapter 10: Regulations & FAQF-Gas, energy decrees, BACS, building regs, CSRD
Immediate & free access

For commercial and industrial sector professionals

Instant access — No credit card required

Access confirmed!

CHAPTER 3

Typologies and 2026 Refrigerants

3.1 Condensation type

Three options structure the plant room architecture and 25 to 40% of annual efficiency: air-cooled condensation (standard commercial, sensitive to heatwaves, nominal EER 3.2–4.2), water-cooled condensation (+30 to 40% efficiency, requires cooling tower or dry cooler), evaporative condensation (compromise, strong growth in datacentres). At equivalent capacity, water-cooled condensation reduces annual electricity consumption by a factor of 1.5 to 2, but adds capital and operating costs for the secondary circuit.

3.2 Target configurations by segment

Commercial Offices
  • <100 kWc: air-cooled, scroll inverter, R454B or R290
  • 100–500 kWc: air + free-cooling, screw inverter
  • 500 kWc–2 MWc: water-cooled, magnetic, R1234ze
  • Target ΔT 6–7°C
  • BACnet/IP monitoring required in specs
Datacentre
  • Water + extended free-cooling, N+1 or 2N
  • Magnetic centrifugal R1234ze
  • Water regime 15/21 or 18/24°C (ASHRAE)
  • ΔT 8–10°C or higher
  • Real-time PUE, sizing for 40–42°C ambient
Industrial / Process
  • Water or evaporative, industrial screw
  • R717 ammonia or R717/R744 cascade
  • Glycol on process side at -5/-10°C
  • Condenser heat recovery
  • SCADA, 2000+ data points

3.3 Refrigerants: the 2026 landscape

The revised European F-Gas Regulation 2024/573 imposes an accelerated phase-down and caps GWP by segment. The situation is rapidly evolving:

RefrigerantGWPSafety class2026 status
R134a1430A1Being phased out, replaced by R1234ze
R410A2088A1Banned in new small-medium equipment since 2025
R32675A2LCurrent standard for small-medium, limited horizon
R454B466A2LMain transition refrigerant 2024–2030
R1234ze<1A2LDurable reference >200 kWc
R513A631A1R134a substitute, still high GWP
R290 (propane)3A3Strong growth, limited outdoor charge
R744 (CO₂)1A1Transcritical, commercial refrigeration, hotels
R717 (ammonia)0B2LIndustrial standard, training required
PFAS uncertainty: major strategic risk. HFO refrigerants (R1234ze, R454B, R513A) are classified as PFAS under OECD definitions. The European restriction dossier led by ECHA since 2023 could, in its strictest form, lead to the progressive prohibition of HFOs by 2027–2030. In that scenario, only natural refrigerants (R290, R744, R717) would meet the decarbonisation pathway. This must be factored into any long-term planning, particularly for buildings with a service life exceeding 20 years.

3.4 Free-cooling: the most powerful efficiency lever

Free-cooling produces chilled water without running the compressor, by exploiting cool outdoor air. Three modes: direct (dedicated heat exchanger, chiller off), indirect (condenser cooling, modulated chiller), integrated (add-on module on modern air-cooled chiller). On a UK datacentre at a 15/21°C water regime, free-cooling accounts for 50 to 75% of annual running hours, representing 30 to 50% energy savings on cooling production. It is now a standard feature, not an option, in this segment.

CHAPTER 4

Sizing and Common Pitfalls

The golden rule: size for the real load profile, not the theoretical peak. On a well-instrumented commercial estate, the average annual load factor of chillers sits between 30 and 50% of rated capacity. Yet current practices systematically penalise performance through excessive margins.

Three recurring pitfalls in chiller sizing.
Pitfall 1 - margin stacking:
Design calculation margin + engineer's margin + manufacturer's margin + installer's margin + project manager's margin = chiller oversized by 50 to 80% versus actual need. Result: permanent short-cycling, degraded efficiency throughout the system's lifetime.
Pitfall 2 - low ΔT syndrome:
A design ΔT of 7°C collapses to 3–4°C in operation due to poorly adjusted control valves, oversized coils or unintended bypasses. The number one pathology in commercial building stock, severely degrading efficiency and causing chiller short-cycling.
Pitfall 3 - underestimated load profile:
Sizing driven by peak load alone, when 80% of operating time is spent at less than 60% of capacity.

4.1 Choosing ΔT: a strategic trade-off

Target ΔTWater regimeApplicationPumping impact
5°C7/12°CHistorical standard, existing stockReference 100%
7°C6/13°CModern commercial, good compromise-30%
10°C6/16°CDatacentre, industrial-70%
15+°C18/33°CHigh-temp datacentre ASHRAE-85%

4.2 Diversity factor and modularity

The diversity factor reflects that not all loads peak simultaneously. Typical coefficients: 0.85–0.95 (single orientation, single use), 0.75–0.85 (multi-orientation), 0.60–0.75 (mixed use), 0.50–0.70 (complex campus). Ignoring this mechanically overstates sizing by 20 to 40%.

Rather than a single 1,000 kWc chiller, prefer two 500 kWc units or three 350 kWc units: fine modulation, partial redundancy at marginal cost, fault tolerance, part-load optimisation, easier maintenance. Common configurations: N+1 (standard for Tier III datacentres, hospitals, pharma), 2N (Tier IV datacentre, high criticality), 2(N+1) (exceptional criticality).

CHAPTER 5

IoT Control and Energy Monitoring

This is the heart of the 2026 transformation. Energy monitoring of a chiller is no longer a luxury but the prerequisite for achieving contractual, operational and regulatory performance. Three benefits: energy performance (10–25% savings), reliability (30–60% fewer unplanned breakdowns), regulatory compliance (CSRD, energy regulations, BACS).

5.1 Target architecture: edge + cloud

The reference architecture separates two functional layers: edge (on-site, autonomous, real-time regulation, works without internet) and cloud (multi-site, analytics, AI, reporting). No critical function depends on the cloud.

Figure 3: The four layers of chiller energy monitoring
Monitoring architecture: edge + cloud Layer 1: Sensors & instrumentation T° / Δp probes Class A paired Flow meters Ultrasonic / electromagnetic Electricity meters M-Bus / Modbus Water quality pH, conductivity, redox Vibration / acoustic Predictive compressor Layer 2: Local automation (edge) Chiller controller PID, safety, sequences BACnet / Modbus Local BMS / BAS Central regulation BACnet/IP, BACnet/SC Edge gateway Buffering, security, compute → MQTT / TLS — edge / cloud boundary — Layer 3: Monitoring & data storage (cloud) Database Time-series, 5–10 years Alarm engine Thresholds, drifts, anomalies Dashboards Web + mobile API / exports CSV, BI, portability Layer 4: Analytics & value creation AI anomaly detection EER drift, leaks Continuous commissioning Model vs actual CSRD Reporting Scope 1+2 Benchmarking Multi-site

5.2 Protocols to specify in tender documents

  • BACnet/IP between BMS and HVAC equipment (chillers, controllers). BACnet/SC for new data flows (TLS-secured).
  • Modbus RTU/TCP for electricity meters and variable speed drives.
  • MQTT + TLS for edge-to-cloud data transfer.
  • OPC UA for integration with host industrial systems (SCADA, MES).

5.3 Essential KPIs per chiller

KPIFormula / observationAlert thresholdCorrective action
Instantaneous EERQcooling / Pelec<80% nominal EERLoad audit, fouling check
Observed IPLVWeighted EER on actual profile<spec valueSequencing optimisation
Evaporator approachWater outlet T° - evaporation T°Rising >2KChemical cleaning of evaporator
Sub-coolingSat T° - liquid T° at condenserFalling >-2KRefrigerant leak alert
Primary ΔTReturn T° - supply T°<4°C (target 6–7)Valve audit, bypasses
CyclingStarts per hour>4/hReview sequencing

OT cybersecurity: a structural issue

As chillers connect to the cloud and internet, the attack surface grows. In a datacentre, hospital or factory, a targeted attack on cooling production can paralyse the site within hours and generate costs in the millions. Reference: IEC 62443 standard. Principles: OT/IT segmentation via dedicated VLAN, encrypted BACnet/SC, remote access via VPN + MFA, centralised logging, CVE management via maintenance contracts. OT cybersecurity is no longer a side issue — it is an operational requirement.

CHAPTER 6

Part-Load Optimisation: 10 Levers Ranked by ROI

In commercial buildings, more than half of all energy is consumed at less than 40% load. This is where everything is decided. The key indicator is no longer the nominal EER but the IPLV (AHRI 550/590) or SEER (EN 14825).

Figure 4: 10 optimisation levers, ranked by ROI
10 optimisation levers, ranked by declining ROI Immediate ROI (no CAPEX) Light CAPEX (instrumentation, configuration) Heavy CAPEX (equipment, redesign) 1 Audit the chilled water setpoint Raise +1 to +3°C based on comfort. 5–15% gain, zero investment. 2 Diagnose primary ΔT If <4°C: stuck valves, bypasses. Low ΔT syndrome — #1 UK issue. 3 Enable T° reset and HP reset Based on outdoor temp or demand. Head pressure optimisation. 4 Enable free-cooling Auto switch based on outdoor T°. 30–75% of annual hours. 5 Multi-chiller sequencing Load balancing, rotation. Optimised algorithm. 6 Secondary network audit Balancing, 2-way valves, bypasses. Critical Δp demand-based control. 7 Fine occupancy scheduling Night / weekend shutdowns. Integration with access control, sensors. 8 Thermal storage + tariff Chilled water or ice storage. Demand shifting, off-peak tariffs. 9-10 Heat recovery + AI Desuperheater, DHW, heating. Anomaly detection.

6.1 Project methodology in 5 phases

1
Initial audit (2–4 weeks). Equipment inventory, instrumentation audit, initial analysis of existing data. Diagnostic of deviations from best practice. Quantification of optimisation potential.
2
Quick wins (1–3 months). Optimisations requiring no capital investment: setpoint adjustments, reset activation, enabling existing functions. Typical gain: 5–15%.
3
Light investments (3–6 months). Supplementary instrumentation, software updates, advanced algorithm configuration. Additional typical gain: 5–10%.
4
Heavy investments (6–18 months). Variable speed pumps, free-cooling module, hydraulic network redesign. Additional typical gain: 5–15%.
5
Continuous operation. Continuous commissioning, monthly review, ongoing improvement. Maintaining gains and identifying new opportunities.
📊 Realistic cumulative target over 18–24 months: 20 to 40% savings on the cooling energy of a moderately optimised installation at the outset. On a poorly performing system, the first pass can reveal up to 40% in savings. Reference standards to include in specifications: ASHRAE Guideline 36 (control sequences), ISO 50001 (energy management system), EN 15232 class A or B (minimum BMS requirement).
CHAPTER 7

Data-Driven Predictive Maintenance

Evolution in three generations: reactive (to be avoided) → systematic preventive (current standard) → data-driven predictive (target for 2026+). Long-term coexistence, as some checks remain legally mandated at fixed frequencies (F-Gas, Legionella).

7.1 Early warning signals to monitor continuously

  • Rising evaporator approach temperature → waterside fouling, plan chemical cleaning
  • Falling sub-cooling → refrigerant leak, urgent leak test required
  • Unstable superheat → EEV (electronic expansion valve) failure
  • Changed vibration spectrum → compressor bearing wear (magnetic centrifugal: monitor magnetic bearing position)
  • Drifting water conductivity → corrosion in progress, review water treatment
  • Cumulative hours and starts → mechanical wear, plan overhaul
  • High discharge temperature → degraded oil, compression ratio too high

7.2 Water quality: underinvested but critical

Widespread corrosion on a network can lead to full re-tubing within 5–10 years at massive cost: £45,000 to £260,000 depending on size. A biologically fouled plate evaporator loses 30% of its heat transfer coefficient within 2–3 years, equating to 10–15% overconsumption. Remediation is difficult; replacement is very expensive (£9,000–£45,000). Inline probes (pH, conductivity, redox) at £1,300–£3,500 per loop offer rapid payback through prevention.

7.3 Modern maintenance contracts: key clauses

  • Precise SLA: response times, minimum availability, penalties
  • Data clause: monitoring access, data ownership and portability, quarterly report
  • Cybersecurity clause: CVE notification, remote access via VPN + MFA, annual audit
  • Obsolescence clause: spare parts available for 10–15 years, approved retrofit kits
  • Performance commitment: minimum measured EER, minimum IPLV, availability rate
CHAPTER 8

ROI, TCO and Investment Trade-Offs

TCO 20 years = CAPEX + Σ (energy OPEX + maintenance OPEX + other OPEX + hidden costs)

The Total Cost of Ownership (TCO) of a chiller over 20 years represents 3 to 5 times its initial CAPEX, dominated by energy (50–70%). Every investment decision must be analysed in TCO terms, never on CAPEX alone.

8.1 Typical cost breakdown : 500 kWc chiller

ItemAmount over 20 years% of TCO
CAPEX (machine + installation + integration)£515k30%
Energy OPEX (167 MWh/year × £0.18/kWh)£515k30%
Maintenance OPEX (3% CAPEX/year)£310k18%
Other OPEX (water, refrigerants, inspections, monitoring)£172k10%
Hidden costs (downtime, water quality, carbon)£206k12%
Total TCO£1.72M100%

8.2 ROI of energy monitoring

IndicatorMid-size commercialCritical site
Monitoring CAPEX (500 kWc chiller)£17k–£43k£26k–£69k
Annual OPEX (licences, instrument maintenance)£2.5k–£5k£4k–£9k
Annual energy savings£2.5k–£6.5k£8.5k–£26k
Downtime avoidance (annual)£0–£4k£17k–£86k
Typical ROI18–36 months6–18 months

8.3 Available incentives and funding in 2026

  • Energy efficiency grants and schemes: various national and regional programmes covering 15–40% of eligible costs for high-efficiency chillers, BMS upgrades and heat recovery projects.
  • Heat networks investment: for projects integrating heat recovery or district energy coupling, significant grant support available.
  • Industrial decarbonisation support: specific grants for industrial refrigeration modernisation projects.
  • Energy performance contracts (EPC): an ESCO finances the works and is repaid from verified savings. Avoids upfront investment; requires rigorous baseline measurement.
  • Accelerated depreciation: available for qualifying energy performance equipment in many jurisdictions.
📌 Replace or optimise? Replace if the chiller is over 15 years old + refrigerant is being phased out (R22, R407C, R134a) + seasonal EER is irrecoverably below 3 + frequent breakdowns. Optimise if under 10 years + future-proof refrigerant (R32, R454B, R1234ze) + reliable machine. Intermediate case: retrofit (compressor + controls + refrigerant conversion) at 40–60% of new replacement cost, gaining 8–12 years of service life.
CHAPTER 9

Sector Case Studies

🏢 Regional Datacentre : 5 MW IT load

Context. 4 magnetic centrifugal water-cooled chillers of 1.5 MWc, N+1, R134a refrigerant, 12/18°C water regime. PUE of 1.52 deemed inadequate.
24-month action plan: chilled water setpoint reset from 12°C to 16°C driven by IT load, extended indirect free-cooling (outdoor T° <14°C), optimised multi-chiller sequencing, partial migration R134a → R513A → R1234ze in 2027, monthly continuous commissioning. Results: PUE 1.52 → 1.29, energy saving 2.8 GWh/year = £420k, ROI 14 months on £414k investment, £78k in energy incentives captured, 196 tCO₂eq/year avoided.

🏥 800-bed Hospital

Context. 3 water-cooled centrifugal chillers (2 × 1.2 MWc + 1 × 800 kWc), closed-circuit hybrid cooling tower, R1234ze refrigerant, N+1 architecture. Measured IPLV 5.1 vs specified 6.5, small chiller short-cycling, formal complaint to manufacturer under consideration.
18-month detailed audit: incorrect variable primary flow tuning (ramp rate too high causing superheat control oscillations), 0.4°C calorimetric chain drift, underused free-cooling.
Results at 12 months: IPLV 5.1 → 6.4 (target achieved), condenser heat recovery activated for winter heating (320 MWh/year), total savings £82k/year.

🏗️ HQE-Certified Commercial HQ : 18,000 m²

Context. 2 air-cooled screw inverter chillers R454B 350 kWc, reversible geothermal heat pump (4 boreholes × 150 m), HQE Excellent and BREEAM certifications. Energy target not met despite certification.
Audit reveals: AHU control valves all in heating mode in winter, geothermal heat pump underused, summer free-cooling deactivated for simplicity, no cross-control between heating and cooling.
Action plan: chilled water setpoint reset from constant 7°C to weather-compensated 7–14°C, free-cooling enabled on extended range, geothermal heat pump recovery coupled in mid-season, BMS integrated with room booking system.
Results at 18 months: combined heating + cooling consumption -32%, long-term energy trajectory secured, ROI 28 months on £95k investment.

Common finding across all cases. Monitoring systematically reveals drifts invisible to routine operations: low ΔT, poorly controlled valves, drifted calorimetric chains, sub-optimal sequencing, unactivated features. Without structured continuous commissioning, gains degrade within 12–18 months. Change management is as important as the technical work: without team buy-in and clear governance, data stays in servers.
CHAPTER 10

Regulatory Framework 2026–2035 and FAQ

10.1 Structural regulations

ReferenceSubjectKey date
F-Gas Regulation 2024/573GWP caps by segment, GWP<150 on chillers <100 kW, leak checks2027 (GWP), ongoing
EU Energy Performance of Buildings Directive (EPBD)Mandatory renovation passports, near-zero energy targets, BMS requirements2025–2030
BACS DecreeMinimum class B BMS (EN 15232) on existing commercial >290 kW HVAC2025 / 2027
Building Regs / Part LRefrigerant GWP limits, energy efficiency requirements for new buildsIn force
CSRDAudited Scope 1 (leaks) and Scope 2 (electricity) reporting : ESRS E12024–2026
EU TaxonomyTop 15% energy performance threshold for green asset qualificationIn force
PFAS RestrictionECHA dossier, could affect HFO refrigerantsDecision 2025–2027

10.2 Roadmap 2026–2035

Figure 5: 2026–2035 trajectory for commercial and industrial operators

Roadmap 2026–2035 for a commercial / industrial operator

Short term 1–2 years
  • Full estate audit
  • Monitoring >100 kW (BACS)
  • Continuous commissioning
  • Urgent R22, R407C migration
Medium term 3–5 years
  • Replace units >15 years old
  • Refrigerant GWP <150
  • Free-cooling + heat recovery
  • CSRD reporting operational
Long term 5–10 years
  • Full migration GWP <150
  • Electrical grid flexibility
  • Heating decarbonisation
  • Anticipate PFAS restriction
Vision 2035+
  • Net zero energy targets
  • Natural refrigerants dominant
  • Magnetocaloric mature?
  • Smart grid integrated
2026
2028
2030
20332035
Convergence of obligations: F-Gas → CSRD → Energy regulations → BACS → Building Regs → EU Taxonomy

All of these requirements demand the same thing in practice: measure, optimise, decarbonise, document.

Instrumentation and monitoring become the common regulatory infrastructure, no longer just a performance tool.

10.3 FAQ

EER or IPLV? IPLV (or SEER per EN 14825) reflects real annual performance, since >95% of operating time is at part load. Specify it in tender documents, not nominal EER.
What ΔT to target in new builds? 6–7°C in commercial buildings, 8–10°C in datacentres. Higher ΔT = lower flow rate = lower pumping energy (cubic law).
Variable primary or decoupling vessel? Variable primary in new builds: better efficiency, simpler. Decoupling vessel for specific cases (water quality concerns, conventional design).
Which refrigerant in 2026? R454B or R1234ze depending on capacity and manufacturer. Anticipate PFAS by favouring natural refrigerants (R290, R744, R717) for long-term and critical projects.
BACnet or Modbus? BACnet/IP for HVAC equipment (chillers, BMS), Modbus for peripherals (meters, drives). Specify BACnet/SC on new equipment.
Edge or cloud? Both. Regulation and safety at edge (full autonomy), data history, analytics, AI, multi-site at cloud. No critical function should depend on the cloud.
What ROI for monitoring? 18–36 months in mid-size commercial, 6–18 months on critical sites. Covered by energy efficiency incentives of 15–40%.
Replace or optimise? Replace if >15 years + phased-out refrigerant + irrecoverable EER. Optimise otherwise. Retrofit as intermediate option (40–60% of new cost).
How many monitoring points per chiller? 35 minimum, 60 ideal. Plus: calorimetric chain with <3% accuracy, communicating electricity meter, water quality probes on larger installations.
How to get started? (1) Audit, (2) no-CAPEX quick wins 1–3 months, (3) deploy monitoring, (4) monthly continuous commissioning. Realistic target: 20–40% savings over 18–24 months.

10.4 Essential glossary

  • Approach temperature: gap between refrigerant and water temperatures at the heat exchanger. Monitor continuously — a rising trend indicates fouling.
  • BACS: building automation and control system regulation requiring minimum class B BMS on commercial buildings >290 kW HVAC.
  • Calorimetric chain: temperature probes + flow meter + calculator to measure thermal energy.
  • EER / IPLV / SEER: instantaneous / weighted part-load / seasonal efficiency per EN 14825.
  • Free-cooling: production of chilled water without the compressor, using cool outdoor air.
  • F-Gas: European regulation on fluorinated gases, revised in 2024.
  • GWP: Global Warming Potential of a refrigerant (CO₂ = 1).
  • Low ΔT syndrome: effective ΔT much lower than design ΔT — a common pathology in commercial building stock.
  • PFAS: group of chemical substances subject to European restriction, affects HFO refrigerants.
  • PUE: Power Usage Effectiveness — datacentre efficiency indicator (total energy / IT energy).
  • Reset: control strategy that adapts a setpoint based on conditions (outdoor temperature, load).
  • Sub-cooling: temperature difference between saturation and liquid refrigerant at the condenser outlet.
❄️

Get a personalised assessment for your chiller estate

Our experts audit your chiller plant, design the optimal solution (replacement, retrofit, monitoring), support you through energy compliance and CSRD requirements, and deploy BACnet/MQTT IoT monitoring compliant with ISO 50001.

Book a free assessment