(568t) A Novel Modeling and Optimization Based Energy Management Solution for Integrated Steel Plants

Energy efficiency and energy cost are of paramount importance for today’s world economy, and even more crucial for heavy energy consumers like iron and steel making plants. Coal constitutes major energy input to integrated steel plants, apart from other sources of energy such as electricity, oil and natural gas. At the same time, integrated steel plants produce large amount of byproduct gases such as Blast Furnace Gas (BFG), Coke Oven Gas (COG), Corex Gas (CXG) and others. These are high calorific value (CV) gases hence very rich source of the energy and can be used as an alternative fuels. Thus, steel mill by-product gases can be a clean, oil-replacing energy as fuel, for electric power generation. Byproduct gases are consumed by iron- and steel- making processes and also to generate power in captive power plants. Therefore, the efficient management of byproduct gases plays a central role in the energy cost reduction for the iron- and steel-making industries. Thus, it provides great potential to make self sustaining steel making plants in terms of energy usage.

In recent past, Mixed Integer Linear Programming (MILP) approach for optimal distribution of byproduct gases in an integrated iron and steel making plant have been presented by several researchers (Akimoto et al. 1991; Bemporad and Morari 1999; Kong et al. 2010).  Zhang et al. (2011) have presented a prediction method for byproduct gas holder levels using machine learning technique. Jeong et al. (2011) and Zhao et al. (2011) have presented prediction method for various byproduct gas generations based on moving average time-series modeling and echo state network (ESN) modeling, respectively. Thus, there are number of literature available that addresses various aspects of byproduct gas management in an integrated iron and steel plant. However, when it comes to case study, they consider relatively simple byproduct gas network with 2-3 gas generators and 2-3 consumers. On the other hand, in reality, byproduct gas network in a steel plant can be very complex and modeling of this network is a non-trivial task due to existence of different pressure drops in different segment of the gas network and various operational and safety  constraints.

This work is geared towards developing simplified modeling strategy for industrial byproduct gas network and followed by formulating MILP optimization problem using this model. Modeling of gas network starts with understanding P&ID of byproduct gas network and then by simplifying the gas network through partitioning the gas network into sub-networks. The partitioning algorithm uses pressure and calorific value of byproduct gases as its criteria.. Further, the various gas generators and consumers in particular sub-network are rearranged in a systematical manner, thereby increasing the modeling ease without losing its mathematical accuracy with respect to optimal scheduling of byproduct gas problem.

It is observed that the proposed modeling strategy is able to reduce significant number of binary as well as continuous variables at modeling stage itself. This model is then used for formulating MILP scheduling problem for optimal distribution of byproduct gases among various external as well as in-house consumers of these gases. Thus, proposed modeling strategy enables one to formulate a tractable MILP based scheduling problem for very large and complex byproduct gas network and consequently achieve an optimal schedule for distribution of byproduct gases such that it

•    Reduces flaring of byproduct gases hence improving on economical as well as environmental performance

•    Meets the demand of maximum possible consumers of byproduct gases

•    Maximizes the usage of byproduct gases in heat and power generation

•    Minimizes the use of external fuel such as oil and natural gas

Hence increase overall economic performance of the plant.

References:

Akimoto, K., Sannomiya, N., Nishikawa, Y., and Tsuda, T. (1991). An optimal gas supply for a power plant using a mixed integer programming model. Automatica, 27(3), 513-518.

Bemporad, A. and  Morari, M. (1999). Control of systems integrating logic, dynamics, and constraints. Automatica, 35, 407-427.

Jeong, C., Chu, Y., Han, C. and Yoon, E.,S. (2011). Gasholder level control based on time-series analysis and process heuristics. Korean J. Chem. Eng., 28(1), 16-21.

Kong, H., Qi, E., Li, G. and Zhang, X. (2010). An MILP model for optimization of byproduct gases in the integrated iron and steel plant. Applied Energy, 87, 2156-2163.

Zhao, J., Wang, W., Liu, Y. and Pedrycz, W. (2011). A two-stage online prediction method for a blast furnace gas system and its application. IEEE Trans. on Control Systems, 19(3), 507-520.

Zhang, X., Zhao, J., Wang, W., Cong, L. and Feng, W. (2011). An optimal method for prediction and adjustment on byproduct gas holder in steel industry. Expert Systems with Applications, 38, 4588-4599.