Marking 10 years of CSIC and looking ahead to the next decade, Professor Mark Girolami, Sir Kirby Laing Professor of Civil Engineering and Royal Academy of Engineering Research Chair at the University of Cambridge, Academic Director for CSIC, and Chief Scientist at The Alan Turing Institute, considers future opportunities that CSIC’s continued collaboration with industry partners across engineering sectors, sciences and disciplines will bring.
The founders of CSIC recognised that only with data-driven insights could design and construction processes deliver resilient, resource-efficient and cost-effective infrastructure. In other engineering sectors that previously recognised the necessity to adopt a data-driven perspective, new market opportunities and new business models have emerged such as for Rolls-Royce in aerospace.
The data-driven management of engines enabled Rolls-Royce, as far back as 1962, to launch the ‘Power-by-the-Hour’ business model where a replacement engine and parts service was provided at a fixed cost per flying hour to the operator. Today, this service now includes Engine Health Monitoring and other monitoring systems, where for example, on-wing performance is assessed using onboard sensors and a global network of maintenance centres are coordinated to minimise downtime. In a similar manner, CSIC has, over the last decade, been pivotal in the comprehensive instrumentation of assets from rail bridges to highways, all now producing data-driven insights into their operating condition. For example, in collaboration with Network Rail, CSIC will soon deliver the first remotely monitored railway bridge on the UK network providing efficiencies in both operation and maintenance. This is a proof of concept that points to a possible transformation in the management of critical infrastructure which would provide safer, more responsive and efficient operation for asset owners and a more resilient service to customers.
Next consider data and how it can be transformed into information and, ultimately, actionable insight. Sir Ronald Fisher, arguably the father of modern-day data science, in paraphrase said that ‘to call in the data analyst after the measurements have been made is like calling for the doctor when the patient has died’. The excitement of the realisation that ubiquitous digital technology was enabling data in its various forms to be gathered at rates and volumes never before imagined gave rise to the Big Data banner. However, as Fisher warned, it is not Big Data that is of value – it is information and insight. It is not enough to just measure. The research area of Data Science is essential for CSIC to develop the new theory, methods and analytic tools which will provide the data-driven actionable insights to meet the scientific and societal challenges our critical infrastructure faces. It will be the enabler for us to transform data into insight, action, control and policy.
The ‘Power-by-the-Hour’ model originally dealt with the management of single assets given the data obtained from each aircraft. However, operators manage fleets of aircraft for which safety and profitability are complex functions of performance of the fleet as a whole. The same applies to our infrastructure, for example Network Rail is reliant upon its ‘fleet’ or ‘population’ of tunnels and bridges to deliver the essential transportation system across the UK. The impact of one bridge being closed cascades across the whole network in a complex, nonlinear manner. Taking a population-based approach by sharing data across assets is one of the next steps in considering the improvement of resilience of our infrastructure. There are deep theoretical challenges to understanding these networks of influence, and how data can be shared to ensure an increase in information (and not an increase in noise). CSIC is instrumental in leading and driving forward this important strand of research and development.
Underlying the design, construction, and operation of many forms of infrastructure is knowledge of the underlying physics, chemistry, and basic science that governs operation and performance. The dynamics of an asset – bridge aerodynamics, energy efficiency of buildings, material properties, structural characteristics, the list goes on – are described by mathematical principles which are well understood in some cases and poorly understood in others. It is essential that we do not lose sight of or neglect the knowledge embodied in the fundamental laws distilled by our predecessors in civil engineering.
The synthesis of both mathematical models describing an asset and informative data from the asset itself is a hugely powerful combination for control and design, with data ‘infilling’ in cases where the underlying science is not well understood, and the science ‘infilling’ where data is sparse – examples are constitutive material models, or turbulent fluid flow. This synthesis of data and models has given rise to the popular moniker of the digital twin. There is much foundational research required to deliver safe and efficient digital twins and this is the case across all sectors of engineering. The power to twin at a single asset level, or indeed using physical or socio-economic theories, models and data measured at the population, urban and city scale, provides great promise.
As CSIC looks to the next 10 years, we will continue to seek inspiration, challenges and insight by collaborating across the various engineering sectors and their associated sciences and disciplines. CSIC will remain at the vanguard of synthesising models and data which will ultimately deliver value to infrastructure construction and operation and support better and sustainable services on which society depends.