Energy commodities have significant roles in modern industrial economics. Energy price fluctuations caused by unbalanced supply and demand and other exotic reasons can result in financial risks. To mitigate or avoid the financial risks, risk management is necessary and profitable for energy utility owner/operators. In this dissertation, two problems in energy risk management are considered. One is 'valuation of an energy swap for a flexible fuel process' and the other is 'valuation and optimization of energy commodity storage'.  An energy swap is a contract between energy utility owner/operators and a banker, where the energy utility owner/operators pay the banker a certain amount of money initially and the banker will take care of the energy cost of the energy utility during an agreed period of time. The problem is to decide a 'fair value' for the energy utility owner/operator to pay. To obtain the 'fair value', a self-financing hedging portfolio is constructed consisting of a flexible fuel process, physical ownership of the underlying energy commodities whose stochastic price models are correlated, and a risk-free asset. PDE method is employed to obtain the valuation of the energy swap and the optimal control strategies (the hedging) for the flexible fuel process. Both simulated data with Monte Carlo method and historical data are used for validation. The results can be used either by the process operator directly to reduce financial risk, or by an energy banker to price an 'energy swap' to finance process operation.  As for the storage problem, the purpose is to help energy utility with energy storage obtain the optimal storage operations to lower financial risks and to make more profit. First, a storage model for a single commodity with stochastic price serving a known and fixed demand is proposed. Convenience yield is also included in the model. A 2-dimensional (2D) Hamilton-Jacobi-Bellman (HJB) equation with embedded optimization problem is formulated and solved to obtain the optimal storage strategies. An oil storage problem is given as an example. The valuation result is compared to two other cases with different control methods and ends up performing the best, which demonstrates the benefit to operate the facility following the optimal control strategies. The control problem is found to be a stochastic sliding mode control problem, which appears novel. The storage problem is further extended by including stochastic demand instead of known demand. A similar 3-dimentional (3D) HJB equation is formulated and solved. The results are consistent with the 2D's. The convergence of the 3D computation is verified with the numerical convergence order. This extended 3D model can be easily adapted for other stochastic variable, like stochastic convenience yield, stochastic volatility or another stochastic commodity price. Proper boundary conditions need to be chosen to solve the corresponding 3D problem. This storage model can also be used for general commodity storage optimization problems.