By Ryan Hardesty and Jeff Jelinek
Weil-McLain, a leading manufacturer of residential and commercial boilers used to provide hot water and steam heat, made the decision to develop a new three-pass, oil fired, horizontal-flue water boiler. Because of the lead-time involved in building the mold for the boiler casting, it takes about six months to build and test a prototype. The company's first design iteration -- or version -- met all its performance requirements but was difficult to manufacture. As it embarked on the second iteration, the company's engineers chose to use computational fluid dynamics (CFD) to simulate the performance of the design.
Simulation highlighted performance issues in the second design and engineers closely examined the results to determine their cause. Engineers then used CFD as their guide in developing a series of virtual prototypes that led to creation of an extremely efficient, economical design. Use of simulation saved $150,000 to $300,000 in prototyping expenses and made it possible to bring the new product to market six to 12 months faster than if traditional build-and-test methods were used.
Water boilers with vertical flues have comprised the bulk of the U.S. market for years. In this design, gases are combusted in a chamber, go through multiple parallel vertical passageways arrayed with pins used to transfer heat to the water, and finally pass out the stack. Weil-McLain decided to develop a new multipass design in which combustion gases go through three separate horizontal passages, arrayed with fins, to increase heat transfer efficiency and reduce combustion gas pressure drop. Another advantage of the horizontal design -- it's easier to clean the horizontal fins since they can be accessed simply by opening a door at the front of the unit.
The First Two Designs
Weil-McLain engineers started the design of the new water boiler using their traditional design methods. The prototype of the first design met all performance requirements, but the prototyping process showed it would be difficult to manufacture. Passageways between the fins in the design were so narrow that very tight manufacturing tolerances were required to avoid high scrap rates. It proved impossible to hold these tolerances consistently with a process that would meet the company's cost targets.
Engineers then developed a new design that overcame the manufacturing issues by opening up many of the passageways to the point interferences wouldn't be a concern. Still, they were anxious to avoid prototyping another design only to discover it also wasn't meeting efficiency targets. They already were looking at CFD technology as a possible tool to improve their design process. A CFD simulation provides fluid velocity, temperature and distribution of other quantities throughout the solution domain for systems with complex geometries. As part of the analysis, a designer may change system geometry or boundary conditions, and view the effect on fluid flow patterns and heat transfer effectiveness. For these reasons, CFD makes it possible to visualize equipment problems far more comprehensively than physical experiments. It also allows the analyst to evaluate performance of a wide range of different configurations in a shorter amount of time and at a lower cost.
CFD Design Simulations
Since this was their first experience with CFD, Weil-McLain engineers decided to initially work with a consulting organization. They selected ANSYS Inc., of Canonsburg, PA, because it's the developer of the leading CFD software and was able to demonstrate the ability of its software to accurately simulate similar products. Weil-McLain engineers and ANSYS consultants worked closely together on the project, which allowed it to move along at maximum speed while helping Weil-McLain engineers to quickly come up to speed on CFD technology. The decision was made to model a more manufacturable design with the simulation being driven by laboratory data of the initial prototype.
Weil-McLain engineers provided the consultants with a solid model of the design that was developed using Pro/ENGINEER computer aided design software from Parametric Technology Corp., of Needham, MA. The consultants imported the geometry into ANSYS' Gambit pre-processor and simplified it to prepare it for analysis. While mechanical design is primarily concerned with the solid areas of the part, CFD is mostly concerned with the passages where fluids flow. By slightly simplifying the solid details, they could build a fluid flow model that would save on computational time.
Guiding the Design Process
The simulation results showed that the performance of this second design had fallen considerably in relation to the first prototype. Viewing the simulation results, engineers determined that there was not enough surface area for the primary combustion gases to contact the fins and move heat to the circulating water. This showed up in the simulation as a relatively high flue outlet temperature. Obtaining this information from the simulation saved the need to build and test another prototype at a cost of approximately $150,000 as well as the six months that would have otherwise been spent waiting for the prototype to be built.
The simulation of the second prototype highlighted the benefits of CFD and convinced Weil-McLain to use it to guide a series of rapid design iterations intended to achieve the company's design objectives, particularly high efficiency and low manufacturing cost. Engineers used the results of the simulations, particularly the temperature in each of the three passageways and at the flue outlet, to determine how individual areas of each heat exchanger performed. Data at this level of detail was impossible to obtain from physical testing because of the difficulties in measuring combustion gas flow patterns, temperature and pressure in internal passages and on the physical limits on the number of sensors that could be used.
A Passage at a Time
The approach engineers found to be most useful was to focus their attention on the first passage of the hot combustion gas from the burner through the heat exchanger. Relative efficiency of each passage could be easily determined by taking the difference between the inlet and outlet temperatures. Engineers evaluated reasons for the performance of a particular design by analyzing at combustion gas flow through these passageways. Specifically, they set out to maximize surface area use in each passage by loading each one evenly. Loading for a particular passage could easily be determined by plotting combustion gas mass flow rate across a cross-section of a flue passage. The goal was to evenly distribute the mass flow rate throughout the flue passage.
In optimizing each passageway's performance, engineers primarily tried changing the width, length and height of the fins and the passages' cross-sectional area. Results provided considerable understanding of sensitivity of the design to changes in these critical design parameters -- for example, the effect of making the fins taller and thinner on the flow velocity and convective heat transfer. Engineers also considered the question of whether or not to add baffles to the third passageway. Baffles are used in many vertical flue designs to increase combustion gas velocity, i.e., fixed mass flow rate and reduced cross sectional area yields higher velocities. But by individually optimizing passages in the new design, they were able to eliminate the need for baffles, which reduced manufacturing costs, easing installation and reducing maintenance labor. Overall size of the casting is limited by manufacturing process considerations, so engineers -- as a final step in the design process -- distributed the available volume and weight among the three passes for further increases in efficiency while minimizing weight.
Using this iterative process guided by simulation results, Weil-McLain engineers and ANSYS consultants were able to increase heat exchanger efficiency to a near-industry-leading 86+% -- 87% is the efficiency ceiling with this type of product due to the condensing point of corrosive flue gases -- while keeping manufacturing costs low enough that the unit could be offered competitively priced with other high-efficiency water boilers. Simulation provided far more information on performance of various design iterations than physical testing, yet each iteration took much less time and expense than building and testing a physical prototype. As a result, simulation made it possible to increase design performance to a much higher level than would be possible with traditional design methods. The cost and lead time involved in developing the product were also substantially reduced. In the three months since introduction, market reception of the new product has been outstanding, exceeding all expectations.
Acknowledgment: The authors would like to thank ANSYS Inc. (www.ansys.com), of Canonsburg, PA, for its assistance on this article.
About the Authors: Ryan Hardesty is area product engineers for Weil-McLain, a Michigan City, IN, company specializing in cast iron boiler manufacturing. Jeff Jelinek, former chief engineer, is no longer with the company. Contact: 219-879-6561 or email@example.com