Skip to main content

Using AI to optimise design of propulsion systems

CASE STUDY: SMALLSPARK SPACE SYSTEMS

How Smallspark collaborated with the University of Southampton to developing new software to improve the performance of its aerostructures and propulsion systems

Cardiff-based Smallspark Space Systems is a UK aerospace Research, Development & Innovation start-up venture. The primary goal of the company is to further mature and commercialise hybrid rocket motor technology to significantly reduce the cost/kg of access to low earth orbit for dedicated small launch vehicles.

The company has focused on developing new software to improve the performance of its aerostructures and propulsion systems. To help to achieve these goals, Smallspark Space Systems signed up to SPRINT to access funding for a major project in collaboration with SPRINT partner, the University of Southampton.

Pathway to next generation of rocket engines

The project used artificial intelligence (AI) to assist in the design of its low-cost rocket engines for high performance and cost-effective small satellite launches in the UK.

Through a combination of its engine design capabilities and the mathematical expertise of the University of Southampton’s Operational Research group, Smallspark developed its own artificial intelligence (named Moore AI after sculptor Henry Moore) to assist in the design of its novel combustion chamber architecture.

The SPRINT project enabled Smallspark to develop tools that allow for the optimisation and design of the next generation of low-cost, ecologically safe, rapid response rocket engines for the MoD, and those seeking to maintain telecom constellations.

Smallspark also used its new software capabilities to assist in the design of a variety of its other launch systems, working with industry partners to assist in improving the cost and efficiency of their aerosystems.

Outcomes of the SPRINT project included:

  • Development of Smallspark’s moore.Ai (www.mooreai.co.uk), and more recently, its moore.Ai:Online tool (www.mooreai.online), which is launching its closed beta (a testing arrangement in which a beta test version of software is distributed to a small group of testers)
  • The Engineering and Physical Sciences Research Council has granted Smallspark and Southampton additional funding of £23,000 to continue cutting-edge research into implementing AI into advanced Engineering Systems Design along with securing significant development funding from Venture Capital
  • Accepted into the world leading Interplay Incubation Program, backed by VC firm Interplay Ventures
  • Further support from Welsh Government – £120,000 funding to further iterate its designs through aggressive testing before significantly scaling up its engines

Novel architecture for rocket engines

Joe Ward, Chief Executive Officer of Smallspark Space Systems said: “In the design of both our engines and our AI software, we took a fundamentally different approach. Aerospace and satellite companies are looking to decrease costs for small satellite launches whilst optimising the robustness and thrust of the rocket engines. This, and the geometry of combustion chambers present a series of unique challenges for engine manufacturers.

“The expertise in mathematical research at the University of Southampton enabled us to focus driving up performance while driving down our costs.

“The project has also enabled us to attract further investment from the likes of the Engineering and Physical Sciences Research Council, The Welsh Government & Deepbridge Capital LLP to support further engine development and testing.

Joe Ward, Smallspark Space Systems

“This collaboration is the start of a long-term project that will go beyond SPRINT and that will position Smallspark as a leader for low-cost launches as the only company using architecture like this in the rocket engine market.”


Understanding and resolving real-world problems

Professor Joerg Fliege, Head of Operational Research within Mathematical Sciences at the University of Southampton added: “Smallspark presented us with a very challenging and unique problem; namely how we can use AI and algorithms to shape objects that will improve the design of their rockets.

“We use mathematics to understand real-world problems and to have a real-world impact. By combining algorithmic optimisation and mathematical modelling, we helped to develop a solution that will help them to drive growth in the UK satellite launch market.”

Professor Joerg Fliege, University of Southampton