FOLLOW US Twitter CONTACT US FTJ Email address Phone number
 
TEN ISSUE ANNUAL SUBSCRIPTION FROM JUST £210

Past, Present, Future – Induction Technology

Former managing director of Inductotherm Europe Ltd, Graham Cooper considers the history of induction melting.

It all began one hundred years ago on a warm summer afternoon in 1918 in the Palmer Laboratory at Princeton University in the USA.  Present that day was Dr Edwin Northrup, an electro-physicist, Mr Rich, an examiner of the US Patent Office and Malcolm McLaren, the head of Princeton’s Electrical Engineering Department. They all watched a piece of metal in an earthenware pot which had several turns of lamp wire wound around it. The power was switched on. The metal glowed red, then white and then collapsed into a molten pool. The coreless induction furnace was born (fig.1).  

This was the culmination of two years of work and one patent application that was rejected on the basis that it described “an inoperable device”. The newspapers of the day called it the “fireless-wireless”. Dr G W Piece of Harvard University claimed that it was “a very interesting toy but would never amount to anything of value”.

Dr Northrup used a 20KHz spark gap converter as his first high frequency power supply. The next development in the induction furnace in 1922 was the application of motor generators which were able to develop a single phase high frequency output. General Electric and later Westinghouse produced MG sets (fig.2). As an induction furnace power supply, the MG set remained in use until the mid 1970s.

During the depression and war years there were no developments and it was not until the mid 1940s that the next change came when two European companies developed the ‘mains frequency’ power supply (fig.3). Operating at 50 or 60 Hertz, this furnace allowed manufacturers to build larger, more powerful furnaces for non-ferrous and ferrous melting applications.

The next development, in the late 1950s, was the culmination of work done by Fransesco Spinelli in Italy and Dr P B Biringer at Toronto University which led to the first commercial static frequency multiplier. These units operated at 150 or 180 Hertz and were the advent of the ‘medium frequency’ induction furnace (fig.4). This frequency allowed higher kW per kg power densities compared to mains frequencies. In the mid 1960s the first triple ‘Triplers’ operating at 450 or 540 Hertz were introduced to the market and were an immediate hit with steel foundries because even higher power density could be achieved at reduced stirring. This workhorse of the 1960s survived into the early 1970s.

The most significant event in the history of the induction furnace since Dr Northrup’s initial invention in 1918 took place in October 1966 in Rancocas, New Jersey, when Henry M Rowan, the owner and founder of Inductotherm Corporation, met with Ernie Goggio and Phil Landis, who had an idea for a solid state inverter that could be incorporated in an induction melting furnace power supply (fig.5).

Rowan gave them an order for a 50kW 3,000Hz unit and eleven months later, incorporated into an Inductotherm power supply, the first Mk 1 VIPâ (variable induction power) was humming away in Inductotherm’s R&D laboratory. 

The world of induction melting had changed forever. Inductotherm launched the VIPâ range at the 1968 American Foundry Show in Cleveland.   

The first 50kW VIPs used 64 SCRs (fig.6) in a series parallel configuration. The Mk1 VIP’s inverter was a current fed unit with the furnace capacitors connected in parallel to the coil. Fixed capacitors were switched in or out of the power circuit to ‘tune’ a furnace coil but once power was applied the power to the coil was controlled by varying the frequency; a seamless automatic control that allowed you to apply 100 per cent or five per cent of the available power simply by turning a knob. Switching capacitors to correct the furnace coil power factor was a thing of the past.

The limiting factor for these VIPâ power supplies was the rating of the SCRs that were available at the time but the eight to ten per cent increase in efficiency over MG sets was the ruling factor.

In 1976 an engineer at Inductotherm put forward a concept for a voltage fed series connected inverter where the inverter, capacitors and coil were connected in ‘series’. Inductotherm saw some significant advantages for the voltage fed ‘series’ circuit (fig.7) over the current fed ‘parallel’ circuit (fig.8). In 1976 the main disadvantage of the ‘series’ circuit was the maximum power that could be achieved. Using the SCRs available at the time, the maximum power a voltage fed inverter could control was 175kW.   Using the same SCRs in a current fed parallel inverter, the power rating was over 500kW.  

What made Inductotherm choose to develop the Voltage Fed Series circuit was the further ten per cent improvement in efficiency, its ability to deliver maximum power throughout the melt cycle, an uncontrolled rectifier that gave a constant 0.95 line power factor and a belief that SCRs would get bigger. The VIPâ Power-Trakâ had arrived. The efficiency of the Power-Trakâ was further improved in 1984 with the release of the Primary Isolated (PI) VIPâ (fig.9).

For the next ten years the development in power supplies was dictated by the development of the ‘Hockey Puck’ SCR (fig.10). By the late 1980s-90s power supplies rated at 8,000kW were being shipped. Installed in the Netherlands in 1988 (fig.11) this was one of a new breed of iron melting furnace – a ‘batch’ melter. The concept of ‘batch melting’, where the furnace is poured empty at the end of every melt, was first introduced to foundries in an Inductotherm paper published in Foundry Management & Technology in September 1987. Up to this time the practice with large furnaces was ‘heel’ melting where 10-15 per cent of the furnace capacity was always retained in the furnace. With ‘batch’ melting, if metal was required in 10-tonne batches a 10-tonne furnace and power supply was purchased that would provide a 35 to 45-minute melt cycle and the furnace was always poured empty. The result was smaller furnaces, better refractory life and the added benefit of a power boost while the charge was magnetic.

The key to any induction melting system is the ‘utilisation’ of the available power. Foundries purchase a 7,000kW power supply with an 8-tonne furnace. With it they can melt eight tonnes of iron to 1,460°C in 35 minutes but it will typically take between 15 and 35 minutes to pour the furnace empty and recharge ready for the next melt cycle – a ‘utilisation’ of 50 to 70 per cent of the actual power available. In 1993 Inductotherm obtained a worldwide patent to a power supply with ‘plural’ outputs (fig.12). The dual power supply was the most significant development in induction melting since the introduction of the solid state power supply in 1966. This power supply was able to supply power to two furnaces simultaneously. At the end of the melt cycle in one furnace the power is transferred to a second furnace and melting continued. Sufficient power is applied to the first furnace to maintain temperature as it is poured empty while the balance is used by the second furnace to melt the next batch (fig.13). Using this type of power supply, ‘utilisation’ of the available power can be as high as 95 per cent.

These dual output power supplies were quickly followed by multiple output power supplies able to energise up to four furnaces simultaneously. With this configuration 100 per cent utilisation of the available power was easily achievable (fig.14).

At the turn of the century there were still two distinct types of power supplies – the current fed inverter and voltage fed inverter. Inductotherm had built a power supply able to control a 42MW load and there were coreless furnaces in the field able to hold 80 tonnes of iron. What did the future hold?

The SCR (silicon control rectifier) or thyristor had been the switching device of choice for induction melting inverters at all power levels. In the early 2,000s ‘brick type’ IGBT (insulated gate bipolar transistors) (fig.15) became more common, particularly in higher frequency applications. A range of IGBT power supplies was introduced (fig.16). As IGBTs able to handle high currents become readily available the use of these units will undoubtedly increase. At this point in time, however, the SCR still remains the most reliable and cost-effective device for high powered, low frequency applications.

A new power supply based on IGBTs is now being built by Inductotherm and units are currently in the field. The concept behind this unit is to combine the advantage of the current fed inverter where the inverter SCRs only have to control 30 per cent of the full load current and the voltage fed inverter with its uncontrolled rectifier, high line power factor and low harmonics. Look out for the VIPâ-I S Plus.

There have been some interesting developments in recent times in the area of stirring. Conventional stirring creates two distinct zones in the upper and lower halves of the molten bath (fig.17a and b) and the intensity of the stirring is varied by changing the frequency. The lower the frequency, the stronger the stirring. By building a furnace with two coils and energising each coil from a separate power supply, Inductotherm has been able to generate uni-directional stirring, up the centre and down the side (fig.17c) or up the side and down the centre (fig.17d). 

Uni-directional stirring can be applied to vacuum melting applications to improve the degassing cycle and for master alloy manufacturers who want to ensure the homogeneity of their product, it is an ideal application.

Inducotherm’s drive for the future is to improve the total system reliability and efficiency. A melting system comprises of one or more transformers, a power supply, one or more furnaces, a cooling system and, when required, a charging system. Inductotherm is constantly evaluating new components, new materials and new manufacturing procedures in an effort to reduce cost, improve reliability and increase the efficiency.

Automation in the melt shop has always been a goal for Inductotherm. The company has developed and refined computer control and monitoring systems starting in the late 1980s with a DOS based system (fig.18) and the latest design (fig.19) that provides and records information for management, monitoring alarm points, lining condition and tools for more efficient melt cycles.

Back tilting furnaces were introduced to simplify slag removal and the ARMSâ (automated robotic melt shop) system was conceived to take the man off the melt deck (fig.20).  

The new generation coins phrases like the ‘Industrial Internet of Things (IIoT)’ and ‘Industry 4.0’, which focuses on the fourth generation of industrial development (fig.21). 

These advanced and innovative networks can be broken into three widely understood levels. A device network level (level 1), where devices within the foundry or melt shop communicate with each other. From there, systems can interact with business level networks and information, such as enterprise resource planning (ERP), product data management (PDM) and customer relationship management (CRM). This information can then interact with and communicate with the outside world (level 3), such as other suppliers, plants and equipment.

The phone line modems of the 1990s have gone. Today we have fully integrated networked systems (fig.22) with the potential to link each piece of equipment so that the automated pouring unit pour can communicate to the furnace when it requires metal.

The feeder is linked to a charge make up system to ensure the most cost-effective charge. The cooling system is linked to the power supply to reduce power if the performance of the cooling changes. While the foundry manager can look at real time trend graphs to ensure the plant is running well – that is no longer required. Using today’s advanced analytics, the plant can report on its own utilisation. The maintenance engineer can look at a screen and see the hydraulic oil level on furnace two is low or transformer three has a high oil temperature but that could be too late. Advancements in lIoT and Industry 4.0 now make it possible for the equipment to have enough intelligence to report when something is not running under optimal conditions. These reporting features are critical to prevent operational shutdowns. The operator can then check trend graphs to interpret the data. There are data logging systems that record and store data which can be recalled for analytical purposes.  The maintenance engineer and the foundry manager can recall the ‘alarms’ list to see what caused furnace three to trip out during the nightshift and loose production.

There are screens that can be remotely accessed that display the furnace’s real-time operating condition. Inductotherm has developed a package called ‘iSenseÔ technology’. When someone is engrossed in their mobile device, they may not be scrolling through social media. Using the iSenseÔ technology (fig.23) they could actually be reviewing data from the real time operation of their furnace.

About the Author

Graham Cooper is a former managing director of Inductotherm Europe Ltd. A long time Inductotherm employee, he joined the start-up company Inductotherm Australia in 1968. He currently works as a consultant providing training, on-site technical support and sales support to Inductotherm companies.

This article is based on a paper given by Graham Cooper at the Indian Foundry Congress in January 2018.  For copies of the figures, refer to the full printed article in the April 2018 issue of Foundry Trade Journal.