**1. Introduction**

Under the increasing installed power of renewable energy sources (RES), the purpose of distribution grids is shifting from passive energy delivery only, to energy delivery and energy production, increasing the role and imposing new duties for Distribution System Operators (DSOs) [1]. One of these duties imposes an obligation on DSOs for the elaboration of development plans, which is a key measure for increasing distribution systems' capacity host for increasing RES penetration. Presently, the location of the renewable connection with the distribution network relates to local weather conditions or mounting capabilities. Therefore, wind turbines are usually located in distant places from household areas, while photovoltaics (PVs) are commonly installed on the top of roofs. As a result, energy generation and consumption are not correlated and power flows are increased in both directions. This is especially visible in the case of residential prosumers equipped with PV systems. In the summer, PV peak generation occurs from 12 p.m. to 2 p.m., and it falls to zero when the sun is setting. This is in opposition to the energy consumption pattern in households where demand peaks in the evening and from 12 p.m. to 2 p.m. are low, because usually, most of the consumers are away from their houses. Consequently, the PV energy is not consumed locally, and instead it is injected to the grid, to be transmitted and consumed, i.e., by commercial and

industrial loads, which are often distant from residential areas. This power flow inflicts additional losses, because energy is lost twice: In the evening, due to increased demand, and midday, because of PV generation. This leads to grid infrastructure overdevelopment and stranded cost problems for DSOs and power balancing issues for Transmission System Operators (TSO) and utilities which, even now, face the problem of dynamic load changes (residual load curve variation) resulting in increased re-dispatching costs and reduced power system reliability [1]. These aspects are usually omitted in the research of DG allocation, which usually focuses on power loss minimization or the improvement of voltage stability.

The above-presented problems can be inexpensively solved by the proper allocation of the RES and energy storages (ES) in the distribution system and its operation. If the generation profile of the energy source is adjusted to the demand profile, the majority of the DG generation is self-consumed locally and allows for the avoidance of power losses. An example of such a principle can be seen in an office equipped with a rooftop PV system. Peak demand for commercial type load is correlated with PV generation, due to the working hours and intensity of devices, such as air conditioners.

This paper presents a method for the minimization of distribution system development costs through the optimal allocation of renewable power units, energy storages and the operation of them, including operational features of various types of generation units and consumption of versatile load types. Strong points lie in the complex approach for solving the local problem by:

the modelling of loads and power-generating units by energy consumption/generation profiles reflecting operational conditions;

the modelling of optimization variables allowing for the selection of the type of RES, energy storages (ES), its size and connection point in the grid and its operation.

The paper consists of seven sections, starting with the introduction, followed by the state of the art on methods of RES allocation in power systems. The third chapter presents the method proposed by the author, while the fourth section includes assumptions and fifth section includes simulation scenarios. Simulation results and discussions are presented in the sixth section. The paper ends with conclusions and recommendations for further research and Appendix A which displayed capacity structure for each scenario.

## **2. State of the Art on the Distribution System Development**

This section provides an overview of methods for DG allocation under the following criteria: objective function formulation, variables selection, time horizon, resolution, approach (technical, economic, environmental, etc.), problem formulation and used solving methods.

Power loss minimization is one of the most common objective function formulations implemented for DG allocation problems [2–7]. Equally frequently applicable methods take into account aspects of the quality of supply also, including voltage improvement in an objective function [8–16] or setting constraints on voltage distortion [17,18] while others include the reduction of harmonics distortions as well [19,20]. Power loss reduction and the improvement of voltage are also presented in [21], however these measures are represented by sensitivity factors. Another approach is more energy consumer-oriented maximizing welfare [22] (including minimisation of energy costs and carbon emissions) or minimizing costs of power losses, load not supplied and DG installation O&M costs [23,24]. A similar method—minimizing power losses and energy generation costs—is presented in [25], while in [24,26], air pollutants are minimized as well. Some of the papers present methods using DG allocation for the improvement of supply reliability [27–29]. In [30], the authors propose a restoration matrix that presents the priority of supply from DG. Reference [31] also includes a reduction of the grid upgrade costs. Some papers present a multistep approach, hence, in the first step of the method presented in [12], optimization maximizes the benefits from the node point of view, while in the second step, the optimization maximizes the benefits from the entire considered network. In [32], the first step determines localization only, while the second provides an optimal capacity of each DG. The original approach is presented in [33], where authors maximize system loadability (usage of the network infrastructure). The objective of the optimization in [34] is nodal voltage stabilization. Nevertheless, the most common approach focusses on power loss minimization and voltage improvement, which are optimized in many approaches. This is presented in [35], where a much wider review on the optimization of DG allocation methods is provided.
