A Century of Chiller Technology

A convergence of technologies had to occur to make the modern, high-efficiency centrifugal chiller a reality. To appreciate the technology fully, we must go back in history and understand the origins of the air conditioning and refrigeration industry. Along the way, we will find an important diversion in aerospace and the critically important centrifugal compressor. Ultimately, we will find that the modern chiller is a testament to advanced technology that was developed in multiple fields.

Some of the first advances in and applications of modern industrial refrigeration were in the United States. In May 1922, Willis Carrier revealed the “Centrifugal Refrigeration Machine” – a very early incarnation of what we now call a chiller [1]. The first installation went to a Philadelphia candy manufacturer; it’s interesting to know that the birth of modern refrigeration and air conditioning started on a large scale. Back in those days, economy of scale enabled the technology to be developed. It was not until a decade later that the core technology began to be adopted into compact units that could be used in smaller businesses such as boutique shops. It took several more decades for smaller residential air conditioners to take off commercially.

Shown in the photograph below is Carrier’s first centrifugal chiller in his New Jersey factory [1].

First Centrifugal Chiller
Photo from [1]
The size of this machine is evident, as is the fact that its design, at the time, necessitated components be spread out in space for assembly and maintenance. By modern standards, the same footprint space could be used to accommodate a modern chiller with over 500 refrigeration tons in capacity. By comparison the original design has less than 100 refrigeration tons of capacity.

The first refrigeration centrifugal compressor used in Carrier’s design is shown in the next photograph [2]. The machine is large relative to today’s standards for equipment designed for similar fluid flows and pressures. Speaking of fluid, early chillers used ammonia (R717) as the refrigerant. Ammonia is a very efficient refrigerant and is still used today in some industrial refrigeration processes. Ammonia, however, is highly toxic. Would you believe that early ammonia-based air conditioning was used in department stores and theaters? Of course, the refrigeration loop was separated from the water loop, which often was further separated from the occupants via air-to-water heat exchangers. Nevertheless, nowadays acutely toxic refrigerants are not used for commercial and residential refrigeration and air conditioning.

Carrier Centrifugal Compressor Design
Photo from [2]
Many improvements have been made to bring chillers from their crude beginnings to the sleek, efficient, modern machines that they are almost a century after introduction. Among the components and technologies that have been improved are the controls, electronics, heat exchangers, and compressors. Interestingly, other types of compressors have been introduced, including reciprocating, scroll, and screw configurations among other positive displacement technologies. The turbocompressor utilizes dynamic compression as opposed to positive displacement. Centrifugal compressors, which are a subclass of turbocompressors, are the most advanced and efficient when it comes to high-efficiency chillers.

Turbocompressors enjoy their status as class leaders among compression equipment because of their incredible versatility. While most positive displacement equipment excels in very low flow rate applications, radial flow compressors begin to have an advantage at moderately low flow regimes. In high flow regimes, axial compressors are the best choices with regards to efficiency, operational expense, and cost factors. At moderate flow rates, the axial-inlet, radial-outlet centrifugal compressor dominates.  This makes the centrifugal compressor the “Goldilocks” machine of turbomachinery, enabling operation at “just the right conditions” for many applications; just like our heroine from our favorite children’s story prefers her chair neither too big nor too small and her porridge neither too hot nor too cold. In the case of the centrifugal compressor, they can also accommodate a large range of pressure ratios. Multistage machines can develop pressure ratios greater than 100. Conversely, low pressure ratio machines below a pressure ratio of 2 can also be extremely efficient. Furthermore, there is a unique technical aspect of these machines that enables the pressure ratio of a single stage to be quite high – pressure ratios of 6 are practically realizable, while research stages have been made beyond a pressure ratio of 10 for a single stage. By contrast, axial compressor stages are generally limited to a practical per-stage pressure ratio of 2, with most common practical realizations between 1.1 and 1.6.

In addition to high efficiency and high adaptability to various operating regimes of flow and pressure, turbocompressors offer some of the lowest vibration and noise out of a number of possible technologies. They are inherently balanced rotors, and have the fewest moving parts compared to alternative technologies that include reciprocating and scroll machines. For this reason, they are very adept to applications that require relatively long continuous operating times (as opposed to frequent on/off operation) and are reliable and need relatively little maintenance.

It is for all of these reasons that centrifugal turbocompressors find themselves in such a diverse range of applications including refrigeration, stationary power generation, automotive, industrial, and aerospace, among others. And for this reason, the technology to design and manufacture such machines is still evolving to heightened levels of sophistication, after more than a century of practice.

Most laypeople might recognize that centrifugal compressors are used in automotive turbochargers. Some may also know that the gas turbine engines that power turboprop and helicopter engines also use centrifugal compressors. Quite few people will probably know that centrifugal compressors are even found on high bypass turbofan engines designed for commercial aviation. The next photograph shows a demonstration engine for small commercial airliner applications on display. A closeup of the compressor stages is shown, which reveals that the last stage is a centrifugal stage that discharges into the combustion area.

Photo by Author – display by Honeywell at 2019 Turbo Expo

The ability of centrifugal compressors to generate a high pressure ratio in a single stage is quite advantageous in aerospace applications, where weight and volume are prime considerations. Did you know that one of the first ever jet engines, used on the first Allied-produced jet during WWII, the Whittle W.1 was powered by a centrifugal compressor? You can read more about Frank Whittle here.  A photograph of the W.1 engine is shown below. The discharge of the centrifugal stage is easy to identify by the transition pieces that link the diffuser discharge to the combustion chambers.

The Whittle W
The Whittle W.1 [3]
So what makes the centrifugal compressor so unique to be well adapted to so many applications? To tackle this question, one must first tackle what makes dynamic gas compression so challenging. If one looks at a large axial gas turbine engine such as used for large aircraft turbofans or perhaps for stationary power generation, one will immediately notice that while there are perhaps two to four stages of expansion turbines, there are many more stages for compression – typically around 10, and sometimes more than 20. Why is this? The answer to this question is the key to understanding how jet engines became practical machines. Charles Parsons invented the modern steam turbine in 1884, and the technology of turbines had been evolving for five decades prior to practical jet engine realization. Yet one of the technological challenges in achieving a workable Brayton cycle, the thermodynamic cycle upon which jet engines are based, is efficiently compressing the working fluid. If the compressor is not efficient enough, the power consumed by the compressor is more than the power that the turbine can generate, resulting in no useful power and little thrust.

It is comparatively more challenging to design dynamic compression stages compared to expansion (turbine) stages because in an expansion turbine, the fluid flows in the direction of the decreasing pressure gradient, which is where the fluid will naturally travel. In the compressor stage, the fluid flow is by definition made to flow in the direction of increasing pressure gradient. This means that flow separation can increase boundary layer development and result in reverse flow in the compression device – a very severe condition. Such a situation in the expansion case, where the pressure decreases in the flow direction, means a drop in efficiency, but the flow continues in the intended direction. It is for this reason that the flow acceleration and diffusion process has to be gradual and highly controlled in compressor stages, and therefore why axial stages are generally designed for pressure ratios in the neighborhood of 1.4.

There is an important equation that turbomachinery aerodynamicists use called Euler’s Pump and Turbine Equation. It relates to how much power is transferred in a rotor stage between the rotor and fluid, and applies to both energy flow from the rotor to the fluid (compression), or energy flow from the fluid to the rotor (expansion), and is shown below.

Eulers Pump and Turbine Equation

Where is mass flow, U is the rotor velocity, cθ is the fluid tangential velocity component, and the subscripts 1 and 2 denote the leading edge, or rotor inlet, and the trailing edge, or rotor exit, respectively. This is likely to be one of the first equations encountered in any text on turbomachinery. You will notice that for an axial machine, the rotor velocity does not change significantly from the leading to trailing edges, so that it can be factored out and the only parameter that changes is the fluid tangential velocity due to the shape of the blade. However, for a centrifugal compressor impeller, the trailing edge of the blade is at a significantly larger radius than the leading edge. Since the blade velocity is the product of rotational speed multiplied by radius, the value of U2 can be on the order of double the value of U1 for a centrifugal impeller, which means that the work done in a single centrifugal stage can be potentially a lot more than a single axial stage. This is a more “wordy” way of describing the principle of operation from which the centrifugal compressor derives its name – that centrifugal force is used to impart energy on the fluid.

Of course Euler’s turbomachinery equation is just a statement of the physics of the boundary parameters. The underlying fluid dynamics is very complicated indeed. Let’s continue, for the moment, down our little rabbit hole. You may be asking, what about everything you said before about how the acceleration and diffusion of the fluid has to be gradual? Yes indeed it must, and the fundamental shape of the centrifugal impeller geometry enables accomplishing so much in one stage. If we examine the following sketch, a few facts will be revealed. The sketch is a back section projection of the hub blade section of a centrifugal impeller. On the left sketch, all blades are visible and a line has been drawn from the axis of rotation out arbitrarily. The angle with which the blade makes relative to the line at any particular point is usually denoted as the  angle and denotes how much the blade turns the fluid motion relative to purely radial direction. The right side shows only two blade passages and highlights a complete passage. It should be clear in examining this figure that as the flow goes from the inner radius towards the outlet, the equivalent radial flow area increases linearly with the radius.

Sketch by Author
Sketch by Author

For the sake of simplicity in the illustration, we’ll say that the flow in the above image is purely radial. If that is the case, the flow area available to the fluid can be expressed as

Flow area

Where b is the channel width between hub and shroud. Note that I am leaving important details out only to emphasize a point. Notice that axial machines can be similarly described, with the absence of radius as an important parameter. What does this mean for us? Observe that as the blade turns, the effect on the flow area is to decrease it, but as the flow continues radially outward, the effect of radius is to increase flow area, so that the combined effect is still one of increase in flow area. The flow, practically, is diffusing in the rotor at the same time that energy is being transferred into it by the rotation of the blade! This is a phenomenological difference compared to axial machines, where flow is almost dominantly accelerated in the rotor and diffused in the stator.

It is precisely this aspect of centrifugal compressors that enables such a wide range of applications, from high variability in flow rate as well as high pressure ratio capability. By carefully adjusting the curvature and the shape of the blades, as well as the hub and shroud surface, we can control the flow to develop either moderate or high pressure ratios, and low to moderately high flow rates. Of course, if very high flow rates are required, as on large commercial turbofan engines, the economics of multistage axial compressors still dominate. But the centrifugal compressor is truly the king of the compression jungle for many applications. Its preference in so many applications, from automotive, to industrial, aerospace, and refrigeration provided important incentives to develop technologies for cost effective manufacturing and advanced component design of this robust machine. And so, as the centrifugal compressor found itself on early jet engines, it now finds itself used in modern high-efficiency turbofans. So too, the reliable centrifugal compressor finds itself on the most modern high efficiency centrifugal chiller. Nowadays, detailed analytical tools, that utilize complex flow models derived from a century of practice, as well as computational fluid dynamics analysis, in concert with three-dimensional parameterization of the shape of the passages, are key enablers in the design of high-performance machines. The modern chiller takes advantage of many advances in technology since the 1920’s, including heat exchanger surface modification for maximizing heat transfer, compressor performance optimization, controls, and cycle developments. Not to minimize the importance of the compressors, the next figure shows the simulated effect of compressor efficiency on the overall COP of a chiller. The efficient compressor is key in maximizing chiller performance.

Compressor Efficiency
Compressor Efficiency

So in summary, we started with a historic introduction to refrigeration chillers, and noted that the centrifugal compressor was the key component in the original incarnation. We then took an excursion to learn about some of the early applications of centrifugal compressors in aerospace, where developments of high efficiency compression were crucial in realizing modern jet engines. We then took a bit of a dive into the details of what enables centrifugal compressors to have such a wide operating range of application. Their development in aerospace, among other application fields, has led to it being such a high-performance workhorse. And so we are brought back to truly appreciate the advanced technology that makes our buildings comfortable, cools our data centers, provides hot water in heat pump applications, cools industrial processes, and so many more applications for the trusty centrifugal chiller!

Interested in seeing how AxSTREAM can be used to design the most efficient centrifugal compressors for any application quickly? Send us an email at info@softinway.com to set up a demo with our compressor professionals today!