Refrigeration for a Multi-Plant Chemical Complex: 17 Projects, 31 Packages, Seven Consecutive Years

Why a Chemical Complex Buys a Fleet, Not a Chiller

In the second quarter of 2026, Wanhua Chemical's Fujian base completed an MDI revamp that lifted single-site capacity to 1.5 million tonnes per year, taking the group's global MDI capacity to roughly 4.5 million tonnes per year. Capacity headlines travel fast. What travels less is the engineering consequence sitting underneath every one of those tonnes: each new process unit brings a fresh block of process cooling duty that somebody has to absorb.

That is where multi-plant refrigeration becomes a procurement discipline rather than an equipment purchase. An isocyanate train, a polycarbonate line, a bisphenol-A unit and a Rectisol acid-gas removal section do not share an evaporating temperature, a load curve, a hazardous-area classification or a refrigerant. Buy each of them in isolation and, ten years later, the site operates a museum of six brands, four secondary fluids and a spare-parts inventory nobody can rationalise.

Between 2016 and 2023, Lmart supplied refrigeration packages to Wanhua Chemical across multiple bases — 17 projects, 31 packages in total. This article opens that record up, not as a trophy cabinet, but because the pattern inside it is genuinely useful to anyone specifying process refrigeration for a chemical complex: which package architecture goes where, how capacities scaled, and how the refrigerant selection moved over a decade.

A note on our position in these projects. Lmart is an equipment supplier. The plants, the process design and the delivery of these units belong to Wanhua Chemical and to the design institutes — Hualu Engineering, Wuhuan Engineering and CNCEC Huanqiu among them — who set the technical conditions. Our scope was the refrigeration packages and the pressure equipment inside them.

The 2016–2023 Delivery Record

The relationship opened in 2016 with three 980 kW secondary-loop packages on R134a for a PMMA unit. It then ran without a gap:

• 2017 — PC unit: one 900 kW brine package, R134a.
• 2018 — HDI unit: three 925 kW direct-expansion packages at −25 °C, part of the isocyanate integration revamp.
• 2020 — VC, PMMA phase II, HCl units: several brine packages, with the largest at 1,579 kW moving to R507.
• 2021 — TDI unit: three 2,400 kW brine packages, R507.
• 2021 — MDI ethylene tank farm: one 1,300 kW brine package with a 1,400 kW motor.
• 2021 — Rectisol (low-temperature methanol wash): two 1,700 kW ammonia condensing packages, R717.
• 2021 — PC phase II: one 2,114 kW brine package, R507A.
• 2022 — XDI unit: three dual-temperature packages delivering 1,828 kW at −25 °C and 86 kW at −45 °C from a single skid.
• 2022 — BPA unit: two 1,560 kW brine packages, R134a.

Three observations fall straight out of that list.

Unit capacity roughly doubled. The 980 kW package of 2016 became the 2,400 kW package of 2021 — a direct read-out of how far single-train chemical capacity scaled over five years. Anyone re-using a 2018 datasheet for a 2026 revamp is starting from the wrong number.

Refrigerants turned over. R134a and R22 dominate the early projects; from 2020 onwards new packages moved to R507/R507A, and the deep-temperature duty went to ammonia. More on where that heads in a moment.

The secondary fluid never changed. Every brine package uses an ethylene-glycol/water solution. That is not coincidence; it is an owner decision. One secondary fluid across a whole complex means one set of concentration and inhibitor procedures, one freeze-point matrix, one training package for operators. Standardising the boring layer is where multi-plant refrigeration actually pays.

Three Package Architectures and How the Duty Splits

Group the 17 projects and a clean division of labour appears.

Secondary-loop (brine) packages, roughly −10 to −35 °C. A screw compressor package chills glycol, and the glycol serves multiple process users. This is the default wherever the load swings, where several consumers share one cold source, or where the process side must be isolated from refrigerant. PMMA, PC, TDI, BPA and the tank farms all sit here. You pay for the extra heat-transfer step in approach temperature and therefore in efficiency; you buy back thermal inertia, simpler process interfaces and far easier future tie-ins.

Direct-expansion packages, roughly −25 to −40 °C. Refrigerant evaporates in the process exchanger itself. One heat-transfer step disappears, so the compressor sees a higher suction temperature for the same process temperature — meaningful efficiency on a duty that runs flat all year. The price is a larger refrigerant charge, longer refrigerant piping, and a maintenance boundary that crosses into the process area.

Ammonia condensing packages, roughly −33 to −45 °C. Where the duty is both deep and large, R717 is hard to beat on thermodynamics and unbeatable on refrigerant cost and global warming potential. The trade is toxicity and flammability, and therefore a higher safety design tier.

The useful framing for a specifier is that these are not competing products. They are three answers to three different questions, and a well-planned complex will run all three.

Case: −25 °C Direct-Expansion Packages for an HDI Unit

In 2018 the HDI unit within Wanhua's isocyanate integration revamp needed −25 °C process cooling. The design institute, Hualu Engineering, specified direct expansion rather than a glycol loop.

The reasoning is worth spelling out because it generalises. HDI phosgenation is temperature-sensitive, which superficially argues for the buffering effect of a secondary loop. But the load is steady. When the load barely moves, the thermal-inertia argument for glycol loses most of its force, while the efficiency penalty of the extra heat-transfer step is paid every hour of every year. On a flat duty, direct expansion wins on operating cost.

The delivered configuration was three 925 kW direct-expansion packages, two duty and one standby. They remain in service, and — candidly — they are the reason the later XDI and TDI enquiries came to us at all. In this market, the first package you deliver into a complex is a sample; the next sixteen are the order.

Case: Two PC Packages, Four Years Apart

The polycarbonate unit gives an unusually clean before-and-after. In 2017 we supplied a 900 kW brine package on R134a with a 250 kW motor. In 2021, for the PC expansion, the package was 2,114 kW on R507A with a 1,250 kW motor.

Same owner, same process family, four years apart — and almost nothing carried over:

• Capacity 900 kW → 2,114 kW, driven by train size.
• Refrigerant R134a → R507A, driven by a lower required evaporating temperature, where R507A's higher volumetric capacity gives a smaller, cheaper compressor for the same duty.
• Motor 250 kW → 1,250 kW, which cascades into transformer sizing, cable schedules, starter type and switchgear room layout.

The practical warning for EPC and owner engineers: copying the previous package specification into an expansion enquiry is one of the most reliable ways to receive quotations that cannot be compared. Re-derive the duty, re-select the refrigerant, and re-issue the electrical data. The four-year-old datasheet is a starting reference, not a specification.

Case: Ammonia Condensing Packages for a Rectisol Unit

MDI production itself is largely exothermic; the refrigeration demand concentrates in the supporting units. In 2021 we supplied a 1,300 kW brine package for the MDI ethylene tank farm, holding cargo temperature in storage.

The more interesting job that year was the Rectisol (low-temperature methanol wash) unit: two 1,700 kW ammonia condensing packages on R717.

Rectisol needs cooling around the −40 °C class. Running an HFC through that temperature lift means a high pressure ratio and poor coefficient of performance. Ammonia's normal boiling point of −33 °C puts it naturally in range, its latent heat is exceptional, its global warming potential is zero, and it costs a fraction of any synthetic alternative. The objections — toxicity and flammability — are real, but they are objections about context. A chemical complex already operates a full hazardous-materials regime: gas detection, emergency ventilation, permit systems, trained response teams. Adding an ammonia machinery room to that environment is a far smaller marginal step than adding one to a food plant or a commercial cold store. That is precisely why ammonia has stayed the mainstream answer for deep-temperature duty in heavy chemicals, and why its share is growing rather than shrinking.

Case: One Skid, Two Temperature Levels (XDI)

The 2022 XDI project added a variable that pure capacity tables do not capture. The unit needed 1,828 kW at −25 °C and 86 kW at −45 °C.

Specified naively, that is two machines: a large mid-temperature package plus a small deep-temperature one, each with its own condensing side, oil system, control panel, foundation and footprint. For 86 kW, the auxiliaries cost more than the refrigeration.

The delivered solution used staged main and auxiliary circuits sharing the condensing side, so a single skid served both temperature levels — three such packages in total. The general rule we now apply when reviewing this kind of enquiry: once the deep-temperature duty exceeds roughly 30% of total capacity, a separate dedicated package usually becomes the better engineering and commercial answer. Below that, integration wins.

Getting this right requires the supplier to be looking at the whole complex's cooling demand rather than filling in one datasheet at a time. It is the clearest argument for treating multi-plant refrigeration as a programme.

The Refrigerant Roadmap Under the HFC Phase-Down

Trace the refrigerant column across those 17 projects and you are looking at regulation, not fashion. R22 is deep into its phase-out in China. The HFCs that replaced it — R134a, R507A — are themselves on a schedule: under the Kigali Amendment, China froze HFC production and consumption in 2024 and begins its phase-down in 2029. Regulators in other markets continue to adjust the detail of their own phase-down rules, and the transition pace is still debated, but the direction has not changed. The room available to high-GWP refrigerants keeps narrowing.

For a new chemical plant that means two things.

First, selecting R507A today is defensible — but an industrial refrigeration package is designed for twenty years or more of service. The relevant question is not whether the refrigerant is available at commissioning; it is what its supply and price look like in 2035, halfway through the asset's life. Ask for that answer in writing at the enquiry stage.

Second, natural refrigerants will take share in exactly the segment heavy chemicals occupy: large capacity, deep temperature, competent hazardous-materials management, professional operators. The two R717 packages we supplied in 2021 looked like a conventional technical choice at the time. From 2026 they look like the choice that ages well.

A practical recommendation for technical conditions documents: require bidders to state the refrigerant's regulatory status over the package design life, and to describe the retrofit path — compressor suitability, exchanger sizing margin, lubricant compatibility, safety classification change — if a substitution becomes necessary.

What to Verify Before You Award

Because performance claims in this sector are easy to inflate, here is the short list we would want a buyer to apply to us as much as to anyone else:

• Reference records that can be audited line by line — owner, unit, year, capacity, refrigerant, motor rating, secondary fluid. Vague "we have supplied major chemical groups" language is not a reference.
• Three artefacts per claimed project: a contract page or award notice, a site photograph showing the nameplate, and a contactable owner or design-institute engineer. Genuine references produce all three.
• Repeat purchase over years, which is much harder to fabricate than a single flagship job.
• Clear scope boundaries — what the supplier manufactures itself versus what it buys in and integrates. A supplier that is straight about this is easier to work with when a warranty question arrives.
• Pressure equipment certification appropriate to your jurisdiction, evidenced by certificate numbers rather than logos on a brochure.

Manufacturing Scope and Certification

Suzhou Lmart Energy Equipment Co., Ltd. is a member of the Huachang Group. Our core business is industrial refrigeration packages, gas compression packages, pressure vessels and skid integration.

Where the boundary sits, precisely: Lmart manufactures pressure vessels and shell-and-tube heat exchangers — evaporators, condensers, oil coolers, receivers — and integrates them with bought-in oil-injected screw compressor blocks, motors, oil systems, piping and control panels into a tested, skid-mounted refrigeration package. Compressor blocks are purchased, not built in house; centrifugal solutions are sourced and integrated project by project. Carbon steel and stainless steel are fabricated in separate workshops to avoid iron contamination of stainless surfaces.

Certification is held for ASME BPVC Section VIII Div.1 (U-Stamp), PED 2014/68/EU and ISO 9001, together with CCS Type & Works Approval and works approval from DNV, LR, BV, NK and RINA for marine-scope equipment, plus KGS for the Korean market.

Beyond the Wanhua projects, the reference list includes more than 130 refrigeration and compression projects for owners including Zhejiang Petrochemical, Hengli Petrochemical and Satellite Chemical.

If you are preparing technical conditions for a process refrigeration package — or trying to rationalise an existing multi-unit refrigeration fleet — we are happy to share the full project schedule, including design institute, motor rating and secondary fluid detail, so you can verify it line by line.