(537b) A Flexible Design Framework for Process Systems with Demand Response Objectives

Liu, Y., University of California, Davis
Palazoglu, A., University of California, Davis
El-Farra, N. H., University of California, Davis
The fluctuation of electricity prices due to the changing power demand and generation profiles during a specific period is already playing a significant role in industrial energy management strategies, and is likely to continue with the increasing power grid penetration of renewable resources. The challenge of mitigating the impact of volatile electricity price profiles evoked the concept of demand response (DR), where electricity consumption is adjusted depending on its prices or events triggered by grid operators during a specific period. While higher electricity prices encourage consumers to lower their power demand, low prices offer an incentive to increase it. As a result, fluctuating electricity price profiles are smoothed benefiting both the energy suppliers and consumers.

Demand response has been studied extensively in the context of chemical process systems [1]; however, the primary focus of current research efforts is on the process scheduling and control aspects of the problem; which are typically addressed by increasing or decreasing production rates during the day or by shifting production load among specific units. In our previous work [2], we developed a proactive operational reconfiguration strategy for heat exchanger networks (HENs) operating under varying electricity prices and subject to intermittent renewable energy resource availability to satisfy DR objectives. However, the design of DR-enabled process networks has received comparatively little attention, and has mainly focused on air separation units [3], where variable capacity operations are accounted for during the process design step. The design of a process network, allowing varied operating levels as well as enabling optimal process flowsheet reconfiguration under different DR scenarios is a novel concept. Intuitively, reconfiguration suggests the purchase of more (redundant) process equipment which generally results in a higher capital investment. Whether the economic benefits from DR operation can compensate for the higher investment and thereafter reduce the overall cost remains an open-research question.

In this work, a classical pump network problem [4] is used as a motivating example to illustrate the flexible design concept utilizing the reconfiguration. We first reformulate the pump network problem to incorporate different electricity prices and production demand as uncertain parameters, design decision and reconfiguration decision. The design problem is then cast as a stochastic mixed-integer nonlinear programming problem, with mixed-integer recourse variables. Li et al. [5] have first developed decomposition algorithm to solve two-stage mixed-integer convex nonlinear stochastic programs with mixed-integer variables in both the first and second stages; however, as the resulting pump network model is nonconvex, a heuristic based progressive hedging algorithm is used to solve the problem. We demonstrate the reconfiguration of the pump network to satisfy different DR-related scenarios, showing the potential benefits of a reconfigurable design under demand response.

[1] Harjunkoski,I., Scholtz, E., and Feng, X., 2014. Industry meets the smart grid. Chemical. Engineering Progress., 110, pp.45–50.

[2] Wang, X., El-Farra, N.H. & Palazoglu, A., 2015. Proactive Reconfiguration of Heat-Exchanger Supernetworks. Industrial & Engineering Chemistry Research, 54(37), pp.9178–9190.

[3] Pattison, R.C. & Baldea, M., 2014. Optimal Design of Air Separation Plants with Variable Electricity Pricing. Computers & Chemical Engineering, 34, pp.393–398.

[4] Westerlund, T., Pettersson, F. & Grossmann, I.E., 1994. Optimization of pump configurations as a MINLP problem. Computers & Chemical Engineering, 18(9), pp.845–858.

[5] Li, C. & Grossmann, I.E., 2018. An improved L-shaped method for two-stage convex 0-1 mixed integer nonlinear stochastic programs. Computers & Chemical Engineering, 112, pp.165–179.