Can Your Thermal Network Analyzer do This?

The ancestral SINDA code was developed a long time ago, and there are various SINDA-like codes available. SINDA/FLUINT is, however, the most featured and complete thermal/hydraulic network solver available. More person-years have been invested in its development than in any other SINDA-like code, and it shows.

Comparing SINDA/FLUINT to other codes? You should!

Here's a few items to consider. The following is not a complete list of the features in SINDA/FLUINT. Rather, it is a list of SINDA/FLUINT's features that cannot be matched by any other thermal network analyzer.

Table of Contents

• Integral Spreadsheet: Complete Model Variability
• Optimization and Goal Seeking
• Test Data Correlation
• Thermal Modeling Features
• Fluid Modeling Features
• Dedicated Graphical User Interface
• Complete Thermal Radiation Analyzer
• CAD/FEM Integration
• Solution Routines
• Reliability Engineering

Integral Spreadsheet: Complete Model Variability

C&R Technologies first introduced the concept of an integral spreadsheet almost a decade ago, and this has proven to be very popular and powerful. The integral spreadsheet feature was further advanced in Version 4.1 to include conditional operators and access to processor variables. These expansions have tremendously increased the ease of use of the code, rendering many common uses of logic blocks archaic. The built-in spreadsheet ...

• makes models easier to read, easier to inherit (more self-documenting), and easier to change and maintain consistently.
• enables sweeping changes while the model is running, facilitating parametric and sensitivity analyses on-the-fly.
• allows model building to precede final dimensions and properties on fast turn-around projects.
• Enables multiple designs and design cases to be stored within a single model
• permits complicated interrelationships to be defined between inputs, and even between inputs and outputs
• permits access to even higher level solutions, such as goal seeking, optimization, and automatic data correlation.

This feature is so popular and powerful that competitors are trying to mimic its features. Still, none can match the following features in SINDA/FLUINT, which distinguish truly integral spreadsheet-like functionality, from a hastily added patchwork:

• The ability to use user-defined variables ("registers") almost anywhere within the model, including array data, fluid data, property data, initial conditions, temperature- and time-dependent options, etc. etc. In fact, about the only thing that can't be algebraically defined is the name of a node.
  
• The ability to make indirect references to other registers and variables. For example, registers can be defined in terms of each other (e.g., OuterDia = InnerDia - 2*WallThck). Indirect references are resolved iteratively in true spreadsheet fashion, even while the processor is executing.
  
• The ability to refer to "processor" (response or output) variables such as "T22" within expressions, including those used to define registers and any other parameter.
  
• The ability to use built-in functions, constants, conditional (IF/THEN/ELSE) operators, etc.
  
• The ability to use indirect references to make expressions reusable, and to eliminate rework or errors when node names changes. For example, the capacitance of a node can be defined in terms of its own temperature "T#this", such that the expression remains valid (or can be copied and pasted to other nodes) even if the node ID changes.

Back to Table of Contents.

Reliability Engineering

Overdesign is common and expensive. In large scale projects, each discipline (thermal, structural, power, etc.) communicates worst-case requirements to other disciplines leading to designs that are heavier and more costly than they need to be, and in some cases does not even result in a safer or more reliable design. Only when meeting an extreme stack-up of margins and uncertainties becomes impossible does a renegotiation of adequate margin begin, and such renegotiations are seldom based on any mathematical rigor or true knowledge of the underlying risk.

As an alternative to stacking up worst-case margins, uncertainties, the engineer can combine these factors statistically to yield information about the degree of confidence (“reliability”) in a particular point design. Combined with the Solver, the user can synthesize a design that meets reliability requirements up front, intelligently balancing cost against risk

Using our Reliability Engineering model, the engineer can

• generate not just a single performance prediction but also a distribution of performance predictions with associated probabilities of occurrence.
• consider tolerances in design parameters, uncertainties in environments, uncertainties in application, and variations in manufacturing as the stochastic phenomena that they are
• calculate the probability that a design will achieve its required performance (i.e., the reliability)
  • providing an assessment of risk or confidence in the design
  • quantifying the amount of over- or under-design present in a given system
• identify variables as "uncertain" and allow them to vary over a prescribed range based on their probability distribution
  • uniform, arbitrary, normal distributions
• select from wide range of statistical analysis routines
  • Monte Carlo, descriptive, and gradient
• For more information on the Reliability Engineering module built into SINDA/FLUINT please review the following white paper:

Reliability Engineering and Robust Design: New Methods for Thermal/Fluid Engineering 

Optimization and Goal Seeking

Some people think "optimization" means manually tweaking a design until it seems better. SINDA/FLUINT features fully automated design optimization and goal seeking, with little or no user programming required. Using a built-in nonlinear programming module, the user

• lists an arbitrary number of design parameters to be adjusted
  • Examples: dimensions, properties, boundary conditions ...
• provides a figure of merit to distinguish what is "better or worse" about a design or correlation
  • Examples: weight (minimize), efficiency (maximize), error between test and predictions (minimize) ...
• defines arbitrarily complicated criteria that separate a useful design from a useless one, based on design parameters, design performance, etc.
  • Examples: keep nodes within temperature limits, don't accept a pressure drop too high ...
• supplies an evaluation procedure, which may invoke any SINDA/FLUINT analysis routine or method available in any combination
  • Example: a simple steady state, a series of transient events ...

SINDA/FLUINT then autonomously seeks and reports the best design, or the values of uncertainties that best match available test data.

This module may also be used to invert the normal input/output sequence for almost any parameter. For example, instead of providing a conductivity and getting back a temperature, the user can provide a temperature and get back a conductivity.

This feature is like programming a series of old SINDA runs to perform some design or analysis task automatically. In addition to reducing engineering labor, this "Solver" module transcends the prior usage of SINDA as a steady or transient analyzer of point-designs. This module provides higher order functionality that lets the thermal engineer contribute to preliminary designs rather than always playing catch-up.

Back to Table of Contents.

Test Data Correlation

Correlation goes by many names: "calibration," "model adjustment," and even "solving the reverse problem." It is a critical, labor-intensive, and until recently procedurally ill-defined task. SINDA/FLUINT offers completely automated data correlation tools that avoid guessed adjustments, iterative runs, and subjective comparisons such as visual plot matching. The user has complete control over:

• the uncertainties to vary, and within what range
• the number and type of comparisons to be made
• the frequency of comparisons, from a single steady state, to multiple steady states, to several comparisons within multiple transient runs, etc.
• the order in which uncertainties are resolved (if all uncertainties are not solved simultaneously)
• the mathematical basis for making a comparison (perhaps least squares, or minimized maximum error, etc.)

Auxiliary routines are available for handling the large amounts of data typically required for these comparisons. These routines help prepare data, make comparisons, and report the results of the comparisons.

Back to Table of Contents.

Thermal Modeling Features

Nodes and conductors are the backbone network elements to almost all SINDA-like thermal analyzers. However, this doesn't mean that all analyzers provide the same modeling tools to the users.

"SINDA '85", the precursor to SINDA/FLUINT, introduced the concept of submodels almost 15 years ago. This powerful feature allows:

• multiple models to be easily combined into a single model without concern for node number collisions, differences in solution techniques or control parameters or logic, etc.
• improved organization
• higher-level solution schemes that can treat each submodel semi-independently, for increased speed and customization
• dynamic model manipulations, such as switching in and out components, changing boundary conditions and environments or materials, etc.

SINDA/FLUINT offers multiple steady-state and transient solution methods not just routines. We are continuously improving these methods, and that includes obsolescing less efficient methods to avoid confusion and to facilitate training. For example, at least four completely different acceleration schemes are available in the two steady-state routines. An addition each submodel can be solved with a different method based upon different criteria. This allow complete customization of the solution. Packaging multiple solution options within a few routines not only makes SINDA/FLUINT easier to learn by avoiding a matrix of complex control options (some of which only work in some routines or which change meaning in different routines), it allows the user to adjust a model dynamically.

Back to Table of Contents.

Fluid Modeling Features

SINDA/FLUINT features true thermodynamic/thermohydraulic fluid networks, not just traditional SINDA nodes and conductors masquerading as "pipes" and "pressure nodes," and barely able to solve single-phase steady incompressible piping networks. The fluid analysis features in SINDA/FLUINT ("FLUINT") are designed to solve the unique problems exhibited by real, complex thermal/fluid systems involving full conservation of mass (including species mass), momentum, and energy in multiple phases.

Thermal control systems increasingly use fluid systems (pumped loops, heat pipes or capillary loops, thermosyphons, air circulators, vapor compression cycles) to meet design objectives. But because SINDA/FLUINT offers the most powerful general-purpose hydraulics code available anywhere in any industry, it is the tool of choice not only for thermal engineers but for specialists in cryogenic systems, climate control systems, refrigeration, oil and gas delivery and storage systems, fuel and exhaust systems, etc. FLUINT has been applied to everything from windshield wipers to nuclear reactors. In fact, many SINDA/FLUINT users use the fluid modeling features exclusively, not needing the traditional SINDA heat transfer networks. But they're there if those users ever change their minds.

The fluid modeling capabilities in SINDA/FLUINT include:

• arbitrary 1D networks, including thermal stratification in 2D and 3D control volumes
• conjugate heat transfer
• steady or transient, from quasi-steady (e.g., thermally-dominated) transients to full hydrodynamic transients such as waterhammer
• fluid independence: user-defined fluid properties of arbitrary complexity
• single- or two-phase flows
• incompressible or compressible working fluids
• pure substances or mixtures of gases, liquids, or both
• pumps, valves, ducts, etc.
• custom control systems and fluid components

The two-phase analysis capabilities in SINDA/FLUINT are extensive, featuring

• boiling, flashing, condensation, cavitation, etc.
• homogeneous or slip flow
• complete mechanistic flow regime mapping
• equilibrium or nonequilibrium phases
• capillary models (wicks, vapor barriers, liquid acquisition/control/separators, etc., including many modeling tools such as evaporator-pumps for capillary pumped loops and loop heat pipes - CPLs, LHPs)
• mixtures of condensible/volatile substances and oils and/or gases
• dissolution/evolution of noncondensible gases

Back to Table of Contents.

Dedicated Graphical User Interface

Since 1992, C&R has offered a complete GUI for SINDA/FLUINT: Sinaps®. Sinaps eliminates the need to work with ASCII inputs and ASCII outputs, allowing visual manipulation and control of the entire SINDA/FLUINT model (fluid, thermal, or both).

While it provides very complete preprocessing and postprocessing options, Sinaps is more than a GUI. It allows the user to pass models as documentation to unlicensed users. These models can be used to review networks, or to generate new results, or even as stand-alone tools.

Sinaps, like SINDA/FLUINT, is nongeometric: it is a sketchpad-like circuit design tool for thermal and/or fluid circuits. When the user needs to work with geometric entities in the construction or postprocessing of a model, then C&R also offers geometry-based GUIs such as RadCAD® and the Thermal Desktop®.

Back to Table of Contents.

Complete Thermal Radiation Analyzer

No other supplier of thermal network analysis software also provides a complete thermal radiation package, much less a CAD-compatible tool that is faster and friendlier than any alternative. RadCAD®, a modern thermal radiation pre- and post-processor to SINDA/FLUINT, is the first CAD-based and FEM-compatible tool of its kind. It features fast new algorithms, and powerful model building and maintenance concepts such as property aliases, analysis groups, and articulators.

Back to Table of Contents.

CAD/FEM Integration

[SINDA/FLUINT is not a finite difference code, it is an equation solver. It can solve lumped parameter, finite difference, and finite element equations all at the same time, and with equal ease.]

Thermal designers can no longer afford to be spectators to the concurrent engineering revolution, and they no longer need be shoe-horned into a structural code in order to participate in a design team.

C&R's Thermal Desktop can be thought of as a geometric GUI to SINDA/FLUINT, but it is more than that: it is the first tool designed specifically for thermal engineers that operates in a CAD environment and that is fully compatible with FEM methods, without sacrificing good thermal modeling practices. Thermal Desktop allows thermal engineers to work directly with CAD designers and structural engineers on their project, without having to be trained in their methods nor being forced to learn and use tools that are inappropriate for thermal design tasks.

For an explanation of why RadCAD and the Thermal Desktop represent truly revolutionary advances in concurrent thermal engineering, follow this link.

Solution Routines

Often we are asked why SINDA/FLUINT has only one steady state and two transient routines, while other codes have many more. This is an ILLUSION! SINDA/FLUINT features many methods and variations built into our routines, while other codes just offer different routines. We believe that too many routines just adds confusion to the user, by forcing the user to decide which routine is appropriate for each of their models. It is like having 24 clocks and not knowing what time it is, and having to look at all 24 before you know which ones you can trust with EACH problem you try. We instead develop our routines to work for EVERY model. SINDA/FLUINT gives the user the default methods which normally work fine, and THEN the user can play around with different options if you wish. In fact, the methods themselves can be varied submodel by submodel, which cannot be done in any other code. In other words, SINDA/FLUINT is easier to use and STILL more flexible.

For example, WITHIN our "one" steady state routine, the user can choose PER SUBMODEL to have matrix methods or iterative methds, to have Aitkin's acceleration or overrelaxation/damping or several types of automatic overrelaxation etc. etc. So in fact, if you have only two submodels, the user could create "solution routines" that do not exist in other codes that offer no submodels. In other codes, the user must pick ONE method and ONE set of criteria and apply them to the WHOLE model blindly. So in fact, our code offers almost an infinite number of methods or routines.

Consider two people combining models. If one used matrix method and one used an iterative method, they are easily combined in SINDA/FLUINT: each submodel can continue to use its own methods. But in other codes, one method must "win" and an another must "lose", and logic, control constants, etc. will all have to be revised to try to get the "loser" to work in the new routine. This is IN ADDITION to all the work needed to renumber the nodes, combine logic, etc.

Within SINDA/FLUINT's two transient routines the use can choose iterative or matrix methods, variable calls to temperature-dependent logic routines, different accuracy levels, etc etc all PER SUBMODEL.

Double precision is also handled very easily in SINDA/FLUINT as an option that can be easily turned on or off (even while the code is executing!) with no changes to the model or to any logic.

Back to Table of Contents.

About Us | Products | Services | Support | What's New | Resources
Home | Request a Quote | Site Map | Feedback

Copyright ® 1999-2008 Cullimore & Ring Technologies, Inc. All rights reserved.