The recent interest in the production of green hydrogen through water electrolysis is hampered by its high cost when compared to steam methane reforming. To overcome this disadvantage, some studies explore replacing oxygen production with hydrogen peroxide at the anode, which has a higher value. Existing electrocatalysis research primarily focuses on hydrogen peroxide synthesis, neglecting process design and separation. Additionally, hydrogen peroxide’s thermodynamic instability in alkaline conditions and the existence of other ions make the separation difficult. This paper proposes a novel concept for the paired water electrolysis process that can be used to improve green hydrogen production economics through valuable chemical coproductions. Valorizing hydrogen peroxide to sodium percarbonate as the final product was chosen to address hydrogen peroxide separation challenges. An electrolyzer stack of 2 MW was chosen, incorporating a recirculating structure, and a boron-doped diamond anode to enhance the hydrogen peroxide production as the base case. According to the techno-economic analysis, for a 2 MW electrolyzer stack, capital expenditure was calculated as 64.5 M€, operational expenses as 21.6 M€, and revenue was calculated as 2.5 M€, resulting in a negative cash flow of −19.1 M€. Results revealed that the process can be profitable (breakeven point) at a capacity of approximately 308 electrolyzer stacks, which is 616 MW in capacity. A sensitivity analysis was conducted to determine how cost drivers including electricity price, anode price, Faradaic efficiency, price of the products and tax subsidy affect the breakeven point. A breakeven point of 60 electrolyzer stacks (120 MW) was found with a 100% increase in the sodium percarbonate sale price. In comparison, a theoretical 100% Faradaic efficiency in the anode material would result in a breakeven point of 38 electrolyzer stacks (76 MW). Even a more realistic 75% Faradaic efficiency leads to a breakeven plant size of 75 stacks (150 MW). Further, multiple two-parameter sensitivity analyses were conducted to assess the relations between Faradaic efficiency, sodium percarbonate sale price and anode material price. For instance, if sodium percarbonate price increases by 100% and Faradaic efficiency increases to 75%, the breakeven capacity drops down to 13 stacks (26 MW). Despite facing economic challenges for the proposed process design based on available technologies, the techno-economic analysis highlights key targets for future works. It also provides valuable insights into the economic feasibility of simultaneously producing hydrogen and sodium percarbonate through water electrolysis, indicating promising potential for the future.

Process design is a creative task that is currently performed manually by engineers. Artificial intelligence provides new potential to facilitate process design. Specifically, reinforcement learning (RL) has shown some success in automating process design by integrating data-driven models that learn to build process flowsheets with process simulation in an iterative design process. However, one major challenge in the learning process is that the RL agent demands numerous process simulations in rigorous process simulators, thereby requiring long simulation times and expensive computational power. Therefore, typically short-cut simulation methods are employed to accelerate the learning process. Short-cut methods can, however, lead to inaccurate results. We thus propose to utilize transfer learning for process design with RL in combination with rigorous simulation methods. Transfer learning is an established approach from machine learning that stores knowledge gained while solving one problem and reuses this information on a different target domain. We integrate transfer learning in our RL framework for process design and apply it to an illustrative case study comprising equilibrium reactions, azeotropic separation, and recycles, our method can design economically feasible flowsheets with stable interaction with DWSIM. Our results show that transfer learning enables RL to economically design feasible flowsheets with DWSIM, resulting in a flowsheet with an 8% higher revenue. And the learning time can be reduced by a factor of 2.