When identifying and comparing forecasting models, there may be a risk that poorly selected criteria could lead to wrong conclusions. Thus, it is important to know how sensitive the results are to the selection of criteria. This contribution aims to study the sensitivity of the identification and comparison results to the choice of criteria. It compares typically applied criteria for tuning and performance assessment of load forecasting methods with estimated costs caused by the forecasting errors. The focus is on short-term forecasting of the loads of energy systems. The estimated costs comprise electricity market costs and network costs. We estimate the electricity market costs by assuming that the forecasting errors cause balancing errors and consequently balancing costs to the market actors. The forecasting errors cause network costs by overloading network components thus increasing losses and reducing the component lifetime or alternatively increase operational margins to avoid those overloads. The lifetime loss of insulators, and thus also the components, is caused by heating according to the law of Arrhenius. We also study consumer costs. The results support the assumption that there is a need to develop and use additional and case-specific performance criteria for electricity load forecasting.

Energy storage enables modification of the customer load profile from the grid perspective without leading to a decrease in comfort level. To meet the future challenges of the energy sector, distribution system operators (DSOs) in Finland have recently discussed power-based distribution tariffs (PBDTs) for small customers. The current distribution tariffs of small customers do not respond well to current challenges, and they require reform. The peak powers of the customers can be decreased with energy storage, and via incentives included in the PBDTs, energy storage can prove to be profitable for household customers. In this paper, we study the profitability of electrical energy storage under different pricing structures. The study simulates changes in customer consumption with a modeled storage system. Simulations consider the requirement of load forecasting in the storage control system. The results of the study show that energy storage can be profitable if the consumption includes only a small number of high peak loads during the pricing period when applying new distribution tariff structures. However, the price level of the tariffs and the tariff mechanisms affect the results.

The profitability of domestic battery energy storage systems has been poor and this is the main barrier to their general use. It is possible to increase profitability by using multiple control targets. Market price-based electricity contracts and power-based distribution tariffs alongside storage of surplus photovoltaic energy make it possible to have multiple control targets in domestic use. The battery control system needs accurate load forecasting so that its capacity can be utilized in an optimally economical way. This study shows how the accuracy of short-term load forecasting affects cost savings by using batteries. The study was conducted by simulating actual customers’ load profiles with batteries utilized for different control targets. The results of the study show that knowledge of customers’ load profiles (i.e., when high and low peaks happen) is more important that actual forecast accuracy, as measured by error criteria. In many cases, the load forecast based on customers’ historical load data and the outdoor temperature is sufficient to be used in the control system, but in some cases a more accurate forecast can give better cost savings.

Distributed electric power production by small-scale customers is increasing continuously. Photovoltaic production is a popular method of producing self-energy for customers. Additionally, power systems require more flexibility when weather-dependent renewable energy production increases. Small-scale customers can increase the self-consumption of self-produced energy by using batteries or a demand response operation. However, batteries require high investment, and demand response operations induce a loss of comfort. Customers who heat their buildings using electric heaters are a good target for demand response operations because their heating can be controlled with limited changes in the indoor temperature. The demand response potential of a building can be defined by simply using customer load profiles and knowledge of the outdoor temperature. Any other information is not required in the proposed novel method. A tolerable variation in indoor temperature corresponds to considerably smaller battery capacity, though it is still a significant amount. With an optimally sized photovoltaic system, it is possible to use both methods simultaneously to increase self-consumption. Maximal benefits can be attained from both methods if the battery system is used as a primary control and the demand response is used as a secondary control. The defined novel method for determining the demand response potential of small-scale customers can also be used when estimating the flexibility of a large customer group. Small-scale customers together can provide significant flexible capacity when their electrical heating is centrally controlled.

Accurate load and response forecasts are a critical enabler for high demand response penetrations and optimization of responses and market actions. Project RESPONSE studies and develops methods to improve the forecasts. Its objectives are to improve 1) load and response forecast and optimization models based on both data-driven and physical modelling, and their hybrid models, 2) utilization of various data sources such as smart metering data, weather data, measurements from substations etc., and 3) performance criteria of load forecasting. The project applies, develops, compares, and integrates various modelling approaches including partly physical models, machine learning, modern load profiling, autoregressive models, and Kalman-filtering. It also applies non-linear constrained optimization to load responses. This paper gives an overview of the project and the results achieved so far.