A widely recognized reality in industries around the world is that of being consistently innovative. Without it the survival is too difficult and can be only short lived. And so only the engineers working on new products in industrial research labs have been striving to enhance the product value for customer by innovating. However, it has been very difficult task till now as the ways of validating new products have been lengthy and costly. One has mainly resorted to crude or scaled down prototypes or at most modeling or simulation of limited functionality of the product being worked upon. The validation of the prototypes has been very difficult and not very reliable as these did not represent full functionality of the product. Further to iterate over various design choices (design space), one has to depend on large number of prototype variants derived using Design of Experiments theory. The tools like COMSOl and ANSYS have been around for long but been restricted mostly to the researchers investigating various phenomenon. These days however, they have been able to handle multiple domains simultaneously through coupling mechanisms. With ever increasing capability of such tools, we are thus witnessing application of these to real-life products and systems which had been too complex so far for such task. This is more so as the real-life products do not spawn over to single phenomenon or physics but comprise of multiple phenomena over multiple domains. Hence these tools are coming to the aid of the practicing engineers to evolve new products or innovative products. In this presentation, a few examples of products which involve expertise and understanding of multiple physics phenomena and their interactions for their further evolution and enhancement are discussed. How the use of COMSOL Multiphysics has been helpful in solving complex design aspects saving extensive time and effort would also be delved into in some detail.

1. Innovating New Products using
Multiphysics Modeling
Dr. Rajveer S Shekhawat
New Products Development (NPD)
Secure Meters Ltd, Udaipur

2. 22
Content
What is innovation and why it is needed?
What are different ways to innovate new product/designs
Design space constraints: time, effort, cost, complexity,
feasibility
Approaches in vogue
Multi-disciplinary Products
Modeling/Simulation types and tools
COMSOL Multiphysics – using more efficiently
Case Studies:
© Secure Meters Ltd

3. 33
Challenge to New Product Designer
Modern products need to meet very demanding
requirements
– Lowest cost
– Best features
– Smallest size
– Longest life
– Simpler to manufacture
– Faster to manufacture
– Require minimum new parts (Re-use)
– First time right etc.
So has to resort to methods that help in innovation
© Secure Meters Ltd

4. 44
What is innovation?
Is the sweeping change in the product feature(s), cost,
life, size etc.
So It is not incremental improvements. In organizations,
the products and processes are gradually refined. So
improvements are continuous and do not lead to step
change characteristic of innovation.
Innovation is a result of out-of-box thinking. And break-
through thoughts are difficult to come by and are thus
rare.
In industrial R&D, 10% efforts only result in innovative
products and the 90% failures contribute to the
proportion.
copyright (c)
Secure Meters Ltd

5. 55
Nature of innovation
© Secure Meters Ltd
Incremental Change (improvement)
– Simplification of processes
– Increase range of values of some features in a product
– Remove bugs in products
– Lasting for short term
Radical Change (breakthrough improvement)
– More long term and strategic in focus
– Change capabilities of the firm
– Supports to jump start
– Distinguishes from competition

6. 66
Why innovation is needed?
Innovative products in terms of cost, performance and
feature set are essential to maintain presence in market
or thus remain in business (profits? Not always).
Launch of new products to avoid obsolescence or to
increase customer base permits the company to diversify
and ensure long time presence in markets with its share
of profits.
© Secure Meters Ltd

7. 77
Different ways to innovate
Develop product from the idea:
– costly affair if it does not lead to the conceived performance
– Risks are too high as the assumptions may not stand valid
– Management may not support such approach
Build prototypes:
– difficult to make a replica so use scaled down model
– Proto may not exhibit all the product behaviour
– May not afford costs involved for small numbers
– Technology may not be available for the concept to realize in
short time
– Costs involved may not be feasible
So use simulation models:
© Secure Meters Ltd

8. 88
Why modeling/ simulation?
Time to market
First time right
Lower cost of iterations
Faster turn-around times
Larger design space exploration
Faster analysis of results (automated)
more assumptions can be relaxed
Advances in computing allow model of real-life
(complex) products
© Secure Meters Ltd

9. 99
Higher rate of obsolescence
Fast pace of technological advancements
Knowledge is public on the internet
First time right designs – reduce validation
Requirements capture
Design and implementation
Time to Market

10. 1010
Requirements Capture
– Abstractions of requirements
– Executable Specifications
– Rapid Prototypes
Design and Implementation
– Feature Set
– Design space exploration
– Automated synthesis (Hardware/Firmware)
Validation
– Testing
– Coverage
First Time Right Approach

11. 1111
Design Space Exploration
Constraints
– time,
– effort,
– cost,
– complexity,
– working environment- exact or approximate (some prototypes
not feasible or possible)
Methods
– Theoretical Analysis (Exact Formulation and solution)
– Prototyping (actual or scaled)- Design of Experiments
– modeling/simulation- limited representation still
© Secure Meters Ltd

12. 1212
Products are Multi-disciplinary
Products normally involve multiple physical domains
So experimenting or understanding behavior in single
physical domain is incomplete
Multiple physical domains have to be modeled
simultaneously allowing interaction across interfaces
The coupled models are thus essential which allow
integrated behavioral model
© Secure Meters Ltd

13. 1313
Modeling/Simulation
Single physical domain
Multidomain (multiphysics) modeling tools (ANSYS,
COMSOL, COSMOS, Altair etc)
– iterative is time consuming: propose new approach for linear
domain
– Non linear design space: Neural Learning, Genetic Algorithms
Hybrid Approaches are more efficient – COMSOL is leader in
it: LiveLinks with Matlab, SolidWorks, ProEngineer,
Comparison of modeling approaches
© Secure Meters Ltd

14. 1414
Using COMSOL Multiphysics
Solving complex models is limited by machine capacity
Machine capacity –
– Number of CPUs (cores) and clock speeds
– memory: amount of memory
Likely Alternatives:
coarse/fine model
Select suitable solver
Partition the problem (advanced capability?)
– Spatial partition and join (grid computing also can be thought)
– Temporal partition – log intermediate results
– Hybrid (Spatio-temporal)
© Secure Meters Ltd

15. 1515
Practices followed
For complex problems not amenable to Multiphysics
modeling due to various limitations
– Analytical results of part of the model
– Simulation results on part of the model
– Experimental results
– Extend/project analytical and experimental results to compare
with experimental results
Desirable capabilities
– Design parameter sweep
– Optimization of objective function
© Secure Meters Ltd

16. 1616
Case Studies
Acoustics: signal generation, coupling to the
structure, acoustic impedance, propagation in
media, modal investigations
Flow: profiling studies obstructions, flanges,
fillets, flow domains: laminar, transient or
turbulent
Transducers: modeling, coupling to waveguides,
transmission and reception characteristics, fan
beam angle
© Secure Meters Ltd

17. 1717
Acoustics
Validation of model in air
Propagation different media
Propagation in different structures
© Secure Meters Ltd

18. 1818
Acoustic Signal Propagation in air
Experimental Signal Plot
Simulation Signal Plot

19. 1919
Acoustic Signal Propagation in Air and
Methane)
Particulars Comparison parameter
Rise time (#Cycles ) Fall time (#Cycles )
Simulation 9-10 7-8
Experimental 9-10 7-8
S.
No.
Medium Simulation (Relative
Attenuation in dB)
Experimental (Relative
Attenuation in dB)
1 Air 900 2.8
2 Methane 500 (-5.1 dB) 1.5 (-5.42 dB)
Signal Shape Study
Signal Attenuation in
Methane

20. 2020
Acoustic Signal through tube geometry
Experimental Signal
Simulation Signal
Plastic Tube

21. 2121
Flow Modeling
Profiling around obstructions
– Dead zones
Pressure drop
Nature of flow along critical paths:
– fully developed (laminar)
– Transient
– turbulent
© Secure Meters Ltd

22. 2222
Entry/Exit profile
© Secure Meters Ltd

23. 2323
Different fillets
© Secure Meters Ltd
Velocity Profile Comparison at the inlet

24. 2424
Pressure drop
© Secure Meters Ltd

25. 2525
Nature of Flow-1
© Secure Meters Ltd

26. 2626
Nature of Flow-2
© Secure Meters Ltd

27. 2727
Transducers
Design Requirements
– Signal coupling
– Beam optimization
– impedance matching (medium)
– Bandwidth (pulse or frequency method)
Types explored
– Piezo Polymeric: Flexural modes
– Piezo Discs with interface layer
– Piezo discs closed form
© Secure Meters Ltd

28. 2828
Transducer models
© Secure Meters Ltd
S.
No.
Material
Description
Relative
Damping Factor
ξ
Corresponding Stiffness
Damping Parameter (βdk)
1 Aluminum
Casing
0.0004 3.18309 e-9
2 Foam Backing 0.1 7.95775 e-7
3 Silicon Filling 0.05 3.97887 e-7

29. 2929
Transient Behavior
© Secure Meters Ltd

30. 3030
Thanks
Phone:
– +91-294-2499208
Emails:
– rajveer.shekhawat@securetogether.com
Website:
– www.securemeters.com
© Secure Meters Ltd