2015-08-21
558
Steven Spielberg’s father and a small publicly owned Welsh steel firm seem an improbable combination. Yet forty years ago their close encounter transformed lives with the first large-scale marriage between mass production and digitization. By the end of the 1960s, South Wales had one of the densest concentrations of process control computers in the world. Italian and US steel firms also played lead roles in this drama. The early history of process control computers has lessons for innovation today.
Digitisation is now in full swing: Familiar consumer products such as cameras, videos and mobile phones have gone digital in the past ten years. Yet, the digital revolution in manufacturing is of long standing, especially in chemicals, steel and electric power. Comprehensive factory automation is forty years old. Full computer control of a large scale manufacturing plant, the hot strip mill at Spencer Steel Works, Llanwern in South Wales, UK commenced operation in October 1964.
Mass production is not new: Another father of a famous son, Marc Isambard Brunel, used machine tools for mass production of wooden blocks for sailing ships at Portsmouth Naval Dockyard at the end of the eighteenth century. But, the real upsurge in mass production is associated with the USA in the 1920s. The trend was sufficiently well established to be satirized by Charlie Chaplin in ‘Modern Times’, released in February 1936.
The steel industry pioneered the use of computers for automation. By the mid1960s, a fifth of the world’s process control computers were installed in the steel industry. Steelmakers exploited second-generation computers, which used ttansistors and magnetic core memory storage as the basis for reliable mainframe computing. Ceneral Electric was the leading computer supplier to steel. By 1965, Ge had 25% of the world market for process control computers in metals. So the state-owned Welsh steel firm Richard Thomas and Baldwins turned to America for the first complete computer control of its new mill at Spencer Steel Works, Llanwern, near Newport in South Wales.
The designer of the GE412computer chosen for the task was Arnold Spielberg, who has since become better known for coaching his son Steven in the art of making home movies. Starting him in a career in Hollywood where one of his best known films was about the extraterrestrial ‘ET’. Arnold Spielberg was a workaholic computer designer, recruited from RCA to head the Process Control Engineering section of GE.
Llanwern was the first successful use of a computer for complete mill control worldwide, with direct digital control of the whole process from entry to exit. The functions of the process control computer were initial set-up of the mill, active operation and adjustment during rolling – including automatic gauge control (AGC), sequence control of slabs and coils through the mill and logging of production.
Llanwern represents a ‘computing of age’ of the modern automated factory. Pioneering developments in process control provided many important lessons for innovators forty years later.
Development of process control on wide strip mills show the importance of the principle that industries ‘learn together’. In the 1960s, the Steel Company of Wales at Port Talbot Learned directly from RTB’s computerization experience at nearby Llanwern.
The Llanwern process control computerization was replicated at Port Talbot, again using a GE412. Full computer control commenced there in September 1966. The same team of a computer personnel then progressed to the new tinplate works of the Steel Company of Wales at Trostre, South Wales. There was an interchange of expertise and personnel. Knowledge of process control techniques and programming skills spread rapidlyThere was an interchange of expertise and personnel. Knowledge of process control techniques and programming skills spread rapidly in this environment as key personnel moved from job-to-job. In the jargon of economics, a high level of ‘tacit knowlegde’ developed among control engineers and software programmers, much of it self-taught and developed on the job.
New growth theory in economics stresses the role of collective learning in promoting growth. Knowledge spills over from one firm to another through trade journals, conferences, scientific publications and standards committees and through informal mechanisms such as job moves, discussions with equipment and component suppliers and gossip over drinks. This is why governments promote industry clusters to encourage innovation. The huge success of the motor sport industry cluster in the UK is partly due to social interaction between engineers, many of whom have a background in aerospace. Rapid turnover of staff between motor sport companies, information leakage through component suppliers, observation in the pit lane, gossip and rumour and more formal contacts through technical committees all help disseminate knowledge of new technologies in ‘motor sport valley’.
The motor sport example shows that to make innovation work, it is important to have access to ‘networks’ of diverse know-how across a range of technologies, mindsets and manufacturing practice. After all, much innovation involves recombinations of existing technologies.
The UK built Ferranti Argus became a mainstay of process control computing, selling some 200 by 1977. Yet this was a computer designed for the Bloodhound ground to air guided missile. Ferranti took a computer designed for missile launch control and control uses. An early application was control of electricity supplies to Steel Peech & Tozer’s Templeborough plant, one of the largest electric arc melting shops in the world in the early 1960s Ferranti’s success shows there is scope for using existing technology in new ways.
Innovations are often complex in so far as they draw upon knowledge of more than one fundamental technology. The breakthrough in computer control at Llanwern was to use digital control of analogue devices to control mechanical items. No one expert could understand it all. Indeed, the original wide strip mill at Butler, Columbia Steel Company in Pennsylvania worked in 1928 because it combined new electrical engineering practice in the form of direct-current motors supplied through Ward-Leonard controls, with new mechanical engineering features such as back-up rolls and roller bearings. It follows that successful innovation involves teams drawn from different disciplines. Knowledge needed for innovation may lay outside your industry. One reason for the success of process plant suppliers has been their ability to combine diverse technologies to meet customer needs.
The spread of computer process control within individual steel firms is just as important as adoption of computer control by new firms. Intra-firm diffusion is as important as inter-firm diffusion. Computer control spread within the Steel Company of Wales from the hot strip mill at Port Talbot to tinplate mills in Troster and Velindre and back to a second attempt to computerize the cold mill at Port Talbot. Similar pattern of intra-firm diffusion can be seen at Jones and Laughlin and Inland Steel in the USA. Demonstration of success encourages replication elsewhere.
Success in adopting large-scale computerization was often associated with other changes in plant configuration. Adoption of computer control on the wide hot strip mill at Port Talbot was the pretext for significant investment in new equipment to improve yield, product quality and width. Computer hardware for the Port Talbot hot strip mill was only around 6% of the $2.4M budget for the whole scheme with the bulk of the money allocated for up-grading the mill with a new edger, loopers, screwdowns and drives to improve mill control.
So, investment is the fundamental vehicle for translating new ideas into new products or process technologies. But, it does not follow that new equipment always embodies the latest technology. The marked difference between Italian and French rates of adoption of process control computers in the 1960s is striking. Both steel industries pursued state sponsored expansion schemes characterized by heavy investment in new steelworks. The publicly owned Italian steelmaker, Italsider played a pioneering role using computers for process control from Italian, French and American suppliers. Whereas in France, apart from isolated installations, steelworks there lagged behind in computer of steel processes, despite progress in computer control in other French industries such as power generation.
