chapter 7: fluxtrol induction heating case studies and success stories

© 2006 Fluxtrol, Inc. www.fluxtrol.com

Chapter VII. Chapter VII.

Case StoriesCase Stories


$avings Due to Induction Coil and $avings Due to Induction Coil and Process OptimizationProcess Optimization

• Increased production rates • Lower maintenance costs• Fewer downstream

operations• Reduced part scrap• Shorter change-over times• Energy savings

Everyone must produce “Good Parts”, but there are ways to make them Better, Faster and more Economically!


Case Story 1. Case Story 1.

Design of Stress Relieving Coil and ProcessDesign of Stress Relieving Coil and Process


Problem Description Problem Description

• System they purchased could not meet required production rate

• To achieve marginal parts, they had to run at half the promised speed

• Even at half speed, they did not meet customer specifications because the heated zone was too narrow and the part was experiencing low temperature on seam bottom

• Difficulties were because the machine was built based upon experience in seam annealing, not stress relieving

Problem:

The Customer made contact due to the following:

Inductor for seam annealing on spiral welded big diameter pipe


Step 1: Analysis of Process ConditionsStep 1: Analysis of Process Conditions

• Limited space on spiral welding mill• Power supply and other equipment with frequency 3 kHz

and power 500 kW already exists • Heat tube material: Steel 1040• Wall thickness 12.7 mm• Relatively small temperature window

– Tmax = 650 C– Tmin = 550 C

• Heat Affected Zone (HAZ) 60 mm • Required heating time 16 seconds• Flux concentrator: Laminations on original coil later

replaced with Fluxtrol “A”


Step 2: Simulation of Existing ProcessStep 2: Simulation of Existing Process

Flux 2D program

BASELINECASE_M

250

500

749.999

0 5 10 15

1

2

Temperature evolution in Outside (1) and Inside (2) seam points

Temperature color map at the end of heating

These results are close to experimental data and very far from specifications


Step 3: Initial Inductor Design Using Step 3: Initial Inductor Design Using Power RampingPower Ramping

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 19 Phase (Deg): 0Scale / Color26.79404 / 67.9149367.91493 / 109.03583109.03583 / 150.15672150.15672 / 191.27762191.27762 / 232.39851232.39851 / 273.51941273.51941 / 314.64032314.64032 / 355.76117355.76117 / 396.88208396.88208 / 438.00299438.00299 / 479.12384479.12384 / 520.24475520.24475 / 561.36566561.36566 / 602.48657602.48657 / 643.60748643.60748 / 684.7283999.999

200

300

400

500

600

699.999

0 5 10 15 20

1

2

Results are much better but specifications are still not met

The easiest way to improve temperature distribution is to use power profiling along the coil length. It may be achieved by variation of concentrator geometry.


Step 4: Optimal Process and Coil DesignStep 4: Optimal Process and Coil Design

Temperature color map for new coil and process at the end of heating

Proposed solution:

1. Make central coil leg of two parallel conductors to increase Heat Affected Zone

2. Power ramping and precise holding of maximum temperature

3. Use Fluxtrol concentrator with profile variation


Temperature Evolution with New CoilTemperature Evolution with New Coil

300

399.999

500

600

0 10 20 30 40

99.999

200

300

400

500

600

0 5 10 15 20 25

1

2

Temperature profile along the pipe OD surface at the end of heating

Temperature evolution in Outside (1) and Inside (2) seam points for optimized process

Minimum temperature in HAZ (point 2) reached required value without material exceeding maximum acceptable temperature


New Inductor Sketch (Top and Side Views)New Inductor Sketch (Top and Side Views)

Concentrator has full C-shaped profile at ramping stage. When maximum permissible temperature was reached, concentrator shape started to change by cutting pole length and then complete removal of concentrator.

Top view of concentrator (hatched)

Side view of induction coil

Fluxtrol concentrator

Fluxtrol concentrator


Temperature distribution 6 seconds after heating is finished

• Customer manufactured induction coil according to Fluxtrol suggestions

• Induction user was able to produce parts in specs with desired production rate

• Final customer was completely satisfied

Final ResultsFinal Results

See Seam Anneal Video



Wheel Hub Heat TreatingWheel Hub Heat Treating


Problem DescriptionProblem Description

• Short coil life – (8,000 – 13,000 pieces) resulting in:

– Machine downtime

– Unacceptable personnel time due to extended set-up

– Scrap parts

Note: Inductor repair costs not a problem in this case due to manufacturer warranty

Problem:

The Customer made contact due to the following:

Typical process of induction heating of wheel hubs

Heated part:Single-race wheel hub


Step 1: Analysis of Inductor Failure ModeStep 1: Analysis of Inductor Failure Mode

Coil failure modes:

• Copper cracking under laminations due to overheating

• Lamination degradation

Previously attempted actions:

• Very high water pressure and flow rate did not solve the problem

• Partial installation of Fluxtrol “A” concentrator instead of laminations improved the situation but did not solve the problem

Old coil with partial replacement of laminations with Fluxtrol “A”


Development StrategyDevelopment Strategy

• Step 2.Step 2. (Band Aid)– Complete replacement of laminations with Fluxtrol “A” on

existing coil resulted in:• Lower local concentration of power in copper

• Same heat pattern and machine settings

• Coil Lifetime increased to 15,000 – 25,000 pieces

• Step 3.Step 3. (Optimal Solution)– Design new coil with lower power concentration in copper

using computer simulation– Build and test new coil


Step 3: Development of New Induction Coil Step 3: Development of New Induction Coil Using Computer SimulationUsing Computer Simulation

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 2.5 Phase (Deg): 0Scale / Color20 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 830830 / 1.1E3

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 2.5 Phase (Deg): 0Scale / Color20.0051 / 85.829285.8292 / 151.65332151.65332 / 217.47742217.47742 / 283.30151283.30151 / 349.12561349.12561 / 414.94974414.94974 / 480.7738480.7738 / 546.59796546.59796 / 612.42206612.42206 / 678.24615678.24615 / 744.07025744.07025 / 809.89435809.89435 / 875.71844875.71844 / 941.54254941.54254 / 1.00737E31.00737E3 / 1.07319E3

Predicted hardness patternTemperature distribution in part with new coil design

Flux 2D program


Tests of New CoilTests of New Coil

New coil with Fluxtrol “A” concentrator was designed, manufactured and tested in production line.

Heat pattern and coil performance were very close to parameters predicted by computer simulation.

Coil life and part production increased to >170,000 hits without coil copper failure or concentrator degradation.

New induction coil after 170,000 heating cycles



Surface Hardening of Truck AxleSurface Hardening of Truck Axle


Problem Description and SpecificationsProblem Description and Specifications

In order to get the required depth in the fillet, they had to severely overheat the area (X) just above it resulting in almost through hardening

X

The customer contacted Fluxtrol Inc. for the following reasons:• They were unable to produce good parts for a contract they had won• The problem was the new part had a sharper fillet than other parts they had worked with before


Step 1: Analysis of Results Obtained with Step 1: Analysis of Results Obtained with Existing EquipmentExisting Equipment

• Customer had the necessary equipment (dual-spindle scanner and power supply 400 kW, 1 kHz)

• Customer had 2 standard scan coils that they attempted to use for the process

• Single turn coil came closer to achieving the desired results, but production rate was low during scanning

• Two turn coil could produce parts faster, but could not properly heat the fillet area

Required hardness pattern


Step 2a: Simulation of the Process with Step 2a: Simulation of the Process with Existing Single Turn CoilExisting Single Turn Coil

. Temperature distribution and magnetic field lines at the end of dwelling period

Flux 2D program

Simulation of the process with a single-turn coil without a magnetic flux controller clearly showed overheating of the stem near fillet area (in good agreement with tests results)


Step 2 a: Temperature Distribution at the Step 2 a: Temperature Distribution at the End of the DwellEnd of the Dwell

Results: Marginal depth in the radius, severe overheating in the area above the radius (temperature over 1150 C, which would continue to grow in the beginning of the scanning process)

Temperature range 20 – 1150 C Temperature range 800 – 1150 C

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color20.38531 / 91.0521591.05215 / 161.71899161.71899 / 232.38583232.38583 / 303.05267303.05267 / 373.71951373.71951 / 444.38638444.38638 / 515.05322515.05322 / 585.71997585.71997 / 656.38684656.38684 / 727.05371727.05371 / 797.72058797.72058 / 868.38733868.38733 / 939.0542939.0542 / 1.00972E31.00972E3 / 1.08039E31.08039E3 / 1.15105E3

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color800 / 821.94061821.94061 / 843.88129843.88129 / 865.8219865.8219 / 887.76251887.76251 / 909.70319909.70319 / 931.6438931.6438 / 953.58441953.58441 / 975.52502975.52502 / 997.46564997.46564 / 1.01941E31.01941E3 / 1.04135E31.04135E3 / 1.06329E31.06329E3 / 1.08523E31.08523E3 / 1.10717E31.10717E3 / 1.12911E31.12911E3 / 1.15105E3


Step 2b:Step 2b: Simulation of the Process with Simulation of the Process with Existing Two-Turn CoilExisting Two-Turn Coil

Temperature distribution and magnetic field lines at the end of dwell

Flux 2D program


Temperature Distribution at the End of the Temperature Distribution at the End of the Dwell with Two-Turn CoilDwell with Two-Turn Coil

Results: Insufficient depth in the radius and short pattern on fillet, severe overheating in the area above the radius (temperature over 1130 C, which would continue to grow in the beginning of the scanning process).

Temperature range 20 – 1135 C Temperature range 800 – 1135 C

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color20.40543 / 89.9989489.99894 / 159.59247159.59247 / 229.18597229.18597 / 298.77948298.77948 / 368.37299368.37299 / 437.96652437.96652 / 507.56006507.56006 / 577.15356577.15356 / 646.74707646.74707 / 716.34058716.34058 / 785.93408785.93408 / 855.52759855.52759 / 925.12109925.12109 / 994.7146994.7146 / 1.06431E31.06431E3 / 1.1339E3

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color800 / 820.86871820.86871 / 841.73749841.73749 / 862.6062862.6062 / 883.47498883.47498 / 904.34375904.34375 / 925.21246925.21246 / 946.08124946.08124 / 966.95001966.95001 / 987.81873987.81873 / 1.00869E31.00869E3 / 1.02956E31.02956E3 / 1.05043E31.05043E3 / 1.07129E31.07129E3 / 1.09216E31.09216E3 / 1.11303E31.11303E3 / 1.1339E3


Step 3: Development of Optimized CoilStep 3: Development of Optimized Coil

Temperature distribution and magnetic field lines at the end of dwell

Based on results of computer simulation and previous experience it was decided to design a two-turn coil with Fluxtrol concentrator on the lower turn.

This design provides a favorable temperature distribution with required hardness depth on fillet and well controlled preheating of the stem area by the upper turn.

Optimal temperature distribution was achieved by variation of cross-sections of copper and magnetic concentrator made from Fluxtrol “A”.


Temperature Distribution at the End of the Temperature Distribution at the End of the Dwell with Optimized CoilDwell with Optimized Coil

Results: Good depth in the radius, no significant overheating in the area above the radius (temperature less than 1020 C); high scan speed/production rate

Temperature range 20 – 1020 C Temperature range above 800 C corresponding to a total hardness depth

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color20.26257 / 82.7621582.76215 / 145.26172145.26172 / 207.76129207.76129 / 270.26086270.26086 / 332.76044332.76044 / 395.26001395.26001 / 457.75958457.75958 / 520.25916520.25916 / 582.75873582.75873 / 645.2583645.2583 / 707.75787707.75787 / 770.25745770.25745 / 832.75702832.75702 / 895.25659895.25659 / 957.75616957.75616 / 1.02026E3

Color Shade ResultsQuantity : Temperature Deg. Celsius Time (s.) : 9 Phase (Deg): 0Scale / Color800 / 813.76624813.76624 / 827.53253827.53253 / 841.29877841.29877 / 855.065855.065 / 868.8313868.8313 / 882.59753882.59753 / 896.36377896.36377 / 910.1291910.1291 / 923.89624923.89624 / 937.66254937.66254 / 951.42877951.42877 / 965.19501965.19501 / 978.9613978.9613 / 992.72754992.72754 / 1.00649E31.00649E3 / 1.02026E3



Induction Brazing of Aluminum Heat Induction Brazing of Aluminum Heat ExchangerExchanger


Problem DescriptionProblem Description

Part: Aluminum heat exchanger

Operation: Brazing of a pipe to a short tube previously brazed to the heat exchanger header in a furnace

Equipment: All equipment including 60 kW, 10-30 kHz power supply, part handling system and control system with IR pyrometers already existed

Problems: • Insufficient brazed joint depth• Inconsistent process resulting in leakage and other defects

Variables:• Limited coil type modification• Coil dimensions• Coil positioning• Magnetic controller dimensions

Pipe

Connection block

Brazing joint

Header

Body

Induction coil with magnetic controller

Tube

Experimental brazing in Fluxtrol laboratory


System Geometry DescriptionSystem Geometry Description

Existing inductors had a “horseshoe hairpin” shape for simple part loading. The coils consisted of two hair-pin sections connected in series with diagonal cross-over.

Geometry is clearly 3D with multiple components (pipe, tube, header, coil copper, magnetic controllers). Due to planes of symmetry it was possible to simulate only ¼ of system.

Non-symmetry due to crossovers was neglected. Geometry prepared for simulation using Flux 3D program


Step 1: Simulation of Existing ProcessStep 1: Simulation of Existing Process

Current density distribution in joint area for diagonal crossover.

Left – no electrical contact Right – good electrical contact

Coil currents direction and magnetic generic field lines for horseshoe coil with diagonal crossover

+ +

On a base of process analysis it was assumed that the main non-controlled variable was an electrical contact between pipe and tube.

These components are preliminarily coated with flux and electrical contact between them is unstable.

When the joint is filled with molten filler metal, contact is good.

Computer simulation confirmed this theory. With no electrical contact there is a strong heat concentration on the tube edge closest to the coil. Overheating of this area is a common defect of brazing.

With good contact current flows from tube to pipe resulting in heat pattern change. It is not possible to balance these two patterns achieving stable quality.


Step 2: Development of Induction Coil with Step 2: Development of Induction Coil with Horizontal CrossoverHorizontal Crossover

Current density distribution in joint area for horizontal crossover.

Left – no electrical contact Right – good electrical contact

Coil currents direction and magnetic generic field lines for horseshoe coil with horizontal crossover

+

. +

For a coil with horizontal crossover currents flow mainly inside of each component for good and no contact conditions.

However, computer simulation showed that heating of pipe and tube is very weak compared to the heat exchanger header.

Small heating of header is useful for support of temperature profile in the tube, but main power must be delivered to pipe and tube in a correct proportion.


Final coil design for one of the brazing joints

Optimized power distribution in brazing joint components

Step 3: Optimal Coil and Process DesignStep 3: Optimal Coil and Process Design

Additional magnetic controller was placed between the coil bottom and header surface. Variation of magnetic flux controllers made of Fluxtrol “A” material allowed the process to achieve optimal power distribution between all three heat exchanger components.

Laboratory tests confirmed stable process flow and good quality of brazed parts.

Additional consideration:

It was noticed that change of crossover resulted in much stronger electrodynamic forces. Coil tends to “open” when power is turned ON. To reduce effect of electrodynamic forces, an additional fiberglass connector was installed on the coil.


Final ResultsFinal Results

• Computer simulation helped to understand the factors causing brazing inconsistency and optimized coil design

• Combination of coil copper optimization and proper geometry of magnetic flux controllers solved a problem of brazing quality. There were no more part rejects caused by improper induction heating

• It was found that a variety of products may be brazed by the coils with the same copper by adjusting geometry of magnetic flux controllers

• Brazing cycle reduced by 15-30% when using new coils

chapter 7: fluxtrol induction heating case studies and success stories

Business

coil length

original coil

coil performance

new coil temperature

problem old coil

new coil design flux

coil copper failure

central coil leg