By Le Xie
This blog post is based on a recent paper from my group which describes the locational marginal prices behavior with time-coupled multi-period (a.k.a. look-ahead) dispatch. While there has been increasing discussion about the real-time price volatility under demand response, most of the discussions have been centered around single-period economic dispatch. As the industry moves toward time-coupled look-ahead dispatch, there is a lack of understanding of price behavior in such a multi-period dispatch. This paper is an attempt to pose the formulation and assess the price behavior in multi-period dispatch in comparison with single-period dispatch.
As a first step, this paper does not consider demand flexibility. Loads are considered as inelastic. Another simplifying assumption is that the forecast of net load is assumed to be perfect in the near term. Section III describes a clean formulation of multi-period dispatch. Even with such a simple model, the pricing behavior at the current step (step 0) as compared that of conventional single-period dispatch, is not entirely clear. An interesting observation is that the price volatility as measured by the standard deviation does NOT monotonically decrease with the more look-ahead steps. The least price volatility at step 0 seems to always occur when the multi-period dispatch window size is 2.
Another interesting observation is in general, look-ahead (or multi-period) dispatch tends to reduce the price spikes during heavy load conditions. This is another desirable reason for system operators to adopt look-ahead dispatch. Though this might not be good news for energy traders! ☺
This blog post is based on a recent paper from my group which describes the locational marginal prices behavior with time-coupled multi-period (a.k.a. look-ahead) dispatch. While there has been increasing discussion about the real-time price volatility under demand response, most of the discussions have been centered around single-period economic dispatch. As the industry moves toward time-coupled look-ahead dispatch, there is a lack of understanding of price behavior in such a multi-period dispatch. This paper is an attempt to pose the formulation and assess the price behavior in multi-period dispatch in comparison with single-period dispatch.
As a first step, this paper does not consider demand flexibility. Loads are considered as inelastic. Another simplifying assumption is that the forecast of net load is assumed to be perfect in the near term. Section III describes a clean formulation of multi-period dispatch. Even with such a simple model, the pricing behavior at the current step (step 0) as compared that of conventional single-period dispatch, is not entirely clear. An interesting observation is that the price volatility as measured by the standard deviation does NOT monotonically decrease with the more look-ahead steps. The least price volatility at step 0 seems to always occur when the multi-period dispatch window size is 2.
Another interesting observation is in general, look-ahead (or multi-period) dispatch tends to reduce the price spikes during heavy load conditions. This is another desirable reason for system operators to adopt look-ahead dispatch. Though this might not be good news for energy traders! ☺