Torrent Research Centre Case Study: Passive Downdraft Cooling (Nimish Patel and Parul Zaveri)

The Torrent Research Centre is a pharmaceutical research campus at Bhat, near Ahmedabad, designed by Nimish Patel and Parul Zaveri of Abhikram with the British environmental designer Brian Ford. Its laboratory blocks are cooled for most of the year by passive downdraught evaporative cooling (PDEC): outside air is cooled by a fine water spray at the top of towers, sinks through a central concourse and leaves through stacks on the outer walls. Ford described it in 2001 as the first large-scale application of PDEC.

Who designed the Torrent Research Centre, and for whom

Abhikram was set up in 1979 by Nimish Patel and Parul Zaveri. IA&B's July 1998 feature by Uttara Chauhan traces the Torrent design to the practice's smaller experiments with natural ventilation, including row houses in Ahmedabad (1983) and Madras (1986) and the Anokhi garment factory in Jaipur. Nimish Patel died on 12 December 2018, aged 70 (World Architecture). The same report says Parul Zaveri runs the practice as co-principal.

The PDEC idea came with Brian Ford, an architect and environmental consultant who had studied traditional evaporative cooling in north India with Mark Hewitt. According to the 1998 IA&B article, the design of a typical laboratory block was evolved jointly by Brian Ford Associates and Abhikram. Abhikram then applied the concept to two more laboratory blocks and designed the much larger administrative building and central core itself.

Patel was clear about the label. "We don't call it passive cooling," he told IA&B: "We call it a human comfort solution."

The brief: clean rooms and chemistry labs without full air conditioning

Torrent Pharmaceuticals wanted a research centre that covered almost every discipline of drug research. Abhikram's project page describes the range: from clean areas needing a Class 10,000 atmosphere to the dirtiest areas, which emit obnoxious gases. Ford's 2001 paper in arq lists chemical synthesis labs at one end and tissue culture, molecular biology and drug design at the other, and says conventional wisdom would put more than half of such a building on refrigerant-based air conditioning.

The case was partly economic. In the same paper Ford writes that Indian construction cost about a tenth of European costs, while energy was relatively expensive and unreliable enough that companies bought diesel generators.

The client took some convincing. Zaveri told IA&B that "we were warned that eventually, we would have to install air-conditioners." Dr Dutt said the team was sceptical at first, but that the extra construction cost, set against running and maintaining plant for almost 200 tons of refrigeration (TR), made the trial attractive.

Torrent Research Centre Ahmedabad: site and climate

The campus sits on the outskirts of Ahmedabad on a site of about 30 acres. The site plan published in IA&B in 1998 shows the laboratory blocks grouped around the central core in landscaped grounds with water bodies.

Ahmedabad's climate suits evaporative cooling for part of the year only. Ford's 2001 paper describes the seasons the building has to handle: a hot, dry summer from February to June, when the system runs, and a monsoon from July to September, when the air is already humid and evaporation gives little cooling. The design therefore had to work in more than one mode: IA&B reported in 1998 that operating strategies had been worked out for the three main seasons, summer, monsoon and winter.

Concept: let cooled air fall through the building

Ford traces evaporative cooling from ancient Egypt and Iran to Mughal north India and Moorish Spain, and to the misting "cool towers" at Seville's Expo 92. The physics is simple. When water evaporates into hot, dry air, the air becomes cooler and denser. Dense air sinks, so a tall shaft with water sprayed at the top creates a downward flow without fans. Ford notes that in a well-insulated building the room air may be only 1 to 2°C warmer than the air leaving the cooler.

At Torrent this became a sealed building with chosen openings. Abhikram's 1999 presentation, "Human Comfort System", describes "a climatically sealed environment" with only designated inlets and outlets, where a fine spray of water cools the air at the point of entry. Shaft locations, sizes and heights were set by simulation.

Torrent Research Centre plan and zoning

IA&B reported in 1998 that of the built-up area, about 42% was cooled by PDEC, about 31% was air-conditioned, 9% used forced ventilation and 18% was naturally ventilated. Leena Thomas and George Baird, who surveyed the users in 2006, describe six laboratory and office blocks, four with PDEC, with air conditioning kept to the two equipment-intensive laboratories.

Inside a typical PDEC block, Ford's 2001 paper describes laboratories and offices on three levels on either side of an open concourse. The concourse lets people move between spaces, and it is also the duct: cooled air is introduced to the rooms at each level from the middle and exhausted through stacks on the perimeter. Abhikram's list of aims adds one more planning rule: place all laboratory spaces to take steady light from the north.

This is a relationship diagram, not a floor plan; see Visiting below for the published drawings.

Section and structure: towers in the middle, stacks at the edge

The section is the system, and it gives the buildings their skyline of towers. Outside air enters at the top of the inlet towers above the concourse, is cooled by the water spray, falls down the concourse, crosses the laboratories and offices on each floor, and leaves through the perimeter stacks. Ford's 1997 sketch for Architecture Today marks two exits from the stacks: at high level when it is windy, and at low level in still air.

The structure is conventional. IA&B describes an RCC frame with brick infill walls. To keep downstand beams out of the clear spaces, the floor slabs were cast as coffered RCC slabs over hollow concrete blocks laid on flat formwork. No section dimensions or tower heights are given in the sources we read, so the diagram shows none.

Materials and construction: insulation and dust control

Abhikram's project page says the design first cut heat gain by insulating the building fabric, and then added the evaporative system. The buildings are insulated with vermiculite, which Abhikram chose as a natural mineral rather than what the practice called "chemically incongruent" materials. IA&B adds that vermiculite was used together with a cement and brick-bat waterproofing technique.

Because a research building with open inlets could easily become a dusty one, much of the detailing is about dust:

• internal plaster surfaces are curved or have rounded corners, so dust and cobwebs do not collect at junctions
• universal louvres, insect screens and motor-operated shut-off louvres control the openings
• half-round ceramic pipes on the outer faces of the inlet and exhaust shafts create local turbulence that keeps larger dust particles out

Light, climate and services: what was measured

Measured figures from named studies:

• April 1997, ten days (Ford, Architecture Today, July 1997): internal temperatures 10°C below the peak outside air temperature in the first completed laboratory building, with the system still being commissioned.
• April to December 1997 (Ford, Patel, Zaveri and Hewitt, 1998): in the first summer, inside temperatures 10 to 15°C below peak outside air temperature, with no discomfort complaints from staff (AIVC abstract).
• April 1998 (Ford, arq, 2001): peaks of 27°C in the ground-floor laboratory and 29°C on the first floor when outside reached 38°C, with 9 air changes an hour on the ground floor and 6 on the first floor.
• 2006 user survey (Thomas and Baird): using the Building Use Studies questionnaire, occupant satisfaction in both the PDEC and air-conditioned blocks was well above the international benchmarks, and the PDEC blocks used less energy (UTS record). The full paper is not open access, so we could not check its detailed figures.

Other numbers are claims by the architects or the client, not published measurements. Abhikram's project page says summer temperatures stabilise between 28 and 32°C, with 6 to 9 air changes even in a chemistry laboratory, an inside swing of 3 to 4°C over 24 hours against 14 to 17°C outside, and humidity of no more than 65 to 70% in summer. Its 1999 presentation gives 29 to 31°C inside with 43°C outside. A 2000 talk notice from the Mohile Parikh Center gives 28 to 31.5°C inside with up to 44°C outside.

Daylight is part of the claimed saving: Abhikram's figures include not using artificial light during the day.

What it cost and what it saves

No source we found gives a current running cost. The figures below are from the late 1990s, at the tariffs of the time:

Figure	Value	Source and status
Extra civil works, including insulation	About 13% of a conventional building's civil cost (12 to 13% in IA&B)	Abhikram 1999; claim
Air-conditioning plant avoided	About 200 tons, about 7% of civil cost	Abhikram 1999; claim
Annual electricity saving, including daylight	About ₹50 lakh at then-current tariffs	Abhikram 1999; claim
Annual electricity saving	About ₹60 lakh	Mohile Parikh Center talk notice, 2000; claim
Electricity use in the first year	About 64% less than an equivalent air-conditioned building	Ford 2001, "reported"; not shown as a measurement
Spending on mechanical and electrical plant	36% lower	Ford 2001

Abhikram's presentation also puts the payback of "the total capital cost" from electricity savings alone at a little over one year, and the payback of the cost of the entire complex, from electricity and plant replacement savings, at around 15 years. Treat these as the design team's figures.

Reception: results, problems and the monsoon

The early reports were candid. Dr Dutt, quoted in IA&B in 1998, said that in the summer of 1997 staff worked in the laboratories without fans, and the labs did not smell the way chemistry labs usually do. He also said the monsoon was muggy, that fans had to be installed for two to three months, and that on windy days some air flowed backwards, from the exhausts to the inlets. Ford (2001) confirms that the system is not run in the monsoon.

Use has grown. Thomas and Baird's chapter in Sustainable Buildings in Practice (2010) says the centre was designed for 150 staff and accommodates over 300, some working in shifts; Abhikram's project page says it served 600 users as of 2005. The two figures differ and we could not resolve them.

A 2012 review by Andrew Corney (ACEEE Summer Study), which cites Torrent, lists the general risks of downdraught cooling: water inside the building, mould and legionella, and poor performance when outside air is too hot or humid.

What designers in India can take from Torrent

Our reading, not the architects':

1. Sort spaces by how much control they really need. Torrent air-conditioned the equipment-heavy labs and cooled the rest with PDEC, fans or open windows. A cooling-mode plan is a useful first sketch.
2. Put the air path in the first section. Ford noted in 1997 that PDEC needs transitional space to distribute cool air. It cannot be added after the plan is fixed.
3. Design for the season it fails. Evaporative cooling stops helping in Ahmedabad's monsoon, and wind can reverse the flow. Plan for fans, controllable louvres and a clear monsoon mode from the start.
4. Measure and publish. Torrent is still cited because it was monitored and surveyed. Budget for loggers and a user survey. For a field method, see our guide to documenting a building as a case study.

Related reading: green building certification and the energy code in India, building materials for Ahmedabad and Gujarat and B.Arch thesis topics by typology. Our Sangath and IIM Ahmedabad case studies show other Ahmedabad answers to heat.

Visiting and seeing the drawings

The Torrent Research Centre is a private pharmaceutical research site. We found no official visitor programme, so do not plan a site visit without written permission from the company.

Abhikram's articles page hosts scans of the July 1998 IA&B feature, with the site plan, a section and a first-floor plan of a laboratory block, and of Ford's July 1997 Architecture Today article, with his air-path sketch and a typical floor plan. Ford's 2001 arq paper has a site plan and a section of a typical block.

FAQ

Who is the architect of the Torrent Research Centre?

Nimish Patel and Parul Zaveri of Abhikram, Ahmedabad. Brian Ford developed the PDEC design for the typical laboratory block with them, and Dr C L Gupta of Solar Agni International vetted Abhikram's designs for the other blocks.

Where is the Torrent Research Centre?

At Bhat, near Ahmedabad in Gujarat, on a site of about 30 acres. It was built for Torrent Pharmaceuticals and is a private research site, not a public building.

When was the Torrent Research Centre built?

The first laboratory block was being commissioned in April 1997. Sources give different completion dates: 1997 (Thomas and Baird), 1998 (Ford) and 2002 (Abhikram's project page).

How does the passive downdraught cooling work at Torrent?

Hot, dry outside air enters towers above a central concourse, is cooled by a fine water spray, sinks because it is denser, flows through the laboratories on each floor and leaves through stacks on the outer walls. It is not used in the monsoon, when ceiling fans take over.

How much cooler is it inside?

Ford measured peaks of 27°C on the ground floor and 29°C on the first floor in April 1998, when it was 38°C outside. The first-summer monitoring paper reported inside temperatures 10 to 15°C below the outside peak.

Planning an institutional building in a hot climate? Find an architect on MOA.

Sources

• Brian Ford, arq (Cambridge University Press): Passive downdraught evaporative cooling: principles and practice, vol. 5 no. 3, 2001: PDEC principles, the brief and floor area, the concourse and stacks, April 1998 measurements, reported energy and plant savings, completion 1998, monsoon operation (accessed 4 Nov 2026)
• B Ford, N Patel, P Zaveri, M Hewitt: Cooling without air conditioning, Renewable Energy vol. 15, 1998, pp. 177 to 182, record on IDEAS/RePEc: 14,000 m² figure, first-year monitoring (accessed 4 Nov 2026)
• AIVC (Air Infiltration and Ventilation Centre): Cooling without air conditioning. The Torrent Research Centre, Ahmedabad, India, abstract: April to December 1997 monitoring, 10 to 15°C below peak outside temperature (accessed 4 Nov 2026)
• Brian Ford, Architecture Today: Climate of change, July 1997, scan hosted by Abhikram: ten-day April 1997 results, concourse and perimeter stacks, exhaust at high and low level, airflow modelling credit (accessed 4 Nov 2026)
• Uttara Chauhan, Indian Architect & Builder: Rites of Initiation, July 1998, scan hosted by Abhikram: team roles, client, quotes from Patel, Zaveri and Dr Dutt, area split by cooling mode, structure, vermiculite, dust detailing, costs (accessed 4 Nov 2026)
• Abhikram: Human Comfort System, presentation at the Ambuja Lecture Series, Chandigarh, November 1999: aims, PDEC credits, ten buildings and 19,700 m², 30-acre site, claimed temperatures and economics, client statement (accessed 4 Nov 2026)
• Abhikram: Torrent Research Centre, Ahmedabad, project page: client names, area, completion 2002, awards, claimed performance and occupancy (accessed 4 Nov 2026)
• Abhikram: Articles, archive of published features (accessed 4 Nov 2026)
• L E Thomas and G Baird, University of Technology Sydney: Post-occupancy evaluation of passive downdraft evaporative cooling and air-conditioned buildings at Torrent Research Centre, ANZAScA conference, 2006, record and abstract (accessed 4 Nov 2026)
• Architecture Media, Acumen: Post occupancy evaluation of evaporative cooling, Leena Thomas and George Baird, 30 November 2006, abstract: six blocks, four PDEC, satisfaction above benchmarks (accessed 4 Nov 2026)
• L E Thomas and G Baird, University of Technology Sydney: Torrent Research Centre, Ahmedabad, Gujarat, India, chapter in Sustainable Buildings in Practice (Routledge, 2010), record: completion 1997, awards, occupancy (accessed 4 Nov 2026)
• Mohile Parikh Center: Torrent Research Centre, Ahmedabad, notice of a talk by Parul Zaveri and Nimish Patel, 19 October 2000 (accessed 4 Nov 2026)
• Andrew Corney, ACEEE Summer Study 2012: Passive Downdraft Systems: A Vision for Ultra-Low Energy Heating, Cooling and Ventilation: general design risks of downdraught cooling (accessed 4 Nov 2026)
• World Architecture, Pappal Suneja: Veteran conservation architect Nimish Patel from India passed away at 70, 12 December 2018 (accessed 4 Nov 2026)