Optimal inverter-based distributed generation in ULP Way Halim considering harmonic distortion

ABSTRACT


THDv
= Total harmonic distortion voltage h max h 0 = The harmonic max and min orders = The initial vector of voltage bus V d,i = Voltage on fundamental frequency I i k+1 = The bus current 1 in iteration k+1 V THD,i = Total harmonic distortion current ε = The tolerance specified V 1 = Slack bus voltage P Loss = Total losses of active power V i = Voltage bus on each bus P Loss i (1) = Fundamental losses of active power V min V max = The standard min and max voltage bus h = Orde of harmonic V rms i = The rms voltage in bus i [V (h) ] = The voltage of harmonic P DG P Load = Active power of DG and load P Loss i

Modelling ULP Way Halim and study case
The ULP Way Halim 88-bus RDS electrical system is located in Bandar Lampung City. Figure 1 shows a map of the location of the system.At ULP Way Halim 88-bus RDS, there are 4 feeders supported by 2 substations.The Rolex feeder (purple), The Bulova feeder (blue) and The Bonia feeder (green) are provided by Sukarame Substation and the Bronze feeder (brown) is provided by Sutami Substation. Figure 1.Maps of ULP Way Halim 88-bus RDS [41] The ULP Way Halim 88-bus RDS with 4 feeders have an electrical power requirement of 17.4 MWatt+j10.783MVAr.Due to the distributed and complex properties of the feeders in this system, modeling is required to determine the dimensionality of the search space when performing simulations.In central load modeling, irregularly distributed loads can be viewed as scattered mass points.Spatial spread of load locations and irregularities in load capacity are not considered in load modeling [42].Centralization of all loads modeling aims to simplify multi-node systems.Centralization of all loads modeling facilitates problem solving in distribution systems [43].Single line diagram of ULP Way Halim 88-bus RDS can be seen in Figure 2.

Objective function and constrain
Multi-objective functions in the form of optimal values are the bare minimum achieved in this study is: The multi objective function is Boundary conditions must be satisfied to make the optimization process more selective.

− Bus voltage limit
The value of bus voltage that must be maintained within operating limits is − Total harmonic distortion limit (THD) The THD of all load bus should be less than the harmonic distortion level allowed by the system.THD value limits refer to IEEE std 95 standards 9 [40].
− The number and value of DG The injection value by DG must not exceed the active power requirement at the load bus.

Optimal with PSO
The steps for using the MOPSO algorithm are [46], [47]: − Starting with initiating a new population of particles with random location and velocity in a search of dimension area.− Evaluate fitness function value in the variable  for all particle.− Comparing the fitness function value of particle with   .If the value of existing is better than   ,   =   .− Identify the particle that give the best result than update the velocity and location of the all particles.− The searching for the fitness function values will ends when the best value is reached in the maximum number of iterations.
The value of, =1, =1, population=100, liter=100,  1 =1 and  2 =1 is used in parameter of MOPSO.In this study, the value of a and b is set equivalent in getting all objective function on the constrain determined.Figure 3 shows the flow of the optimization process with the MOPSO method.By considering the harmonic injection value given by the penetration of the active power of the DG which is placed through the optimization stages using MOPSO in finding the objective function with predetermined constraints.

Integrating DG into RDS
The utilize of DG can bring a better condition in the RDS, such as a improve profile of voltage bus and losses are lowered [48].In this study, PQ is negative for DG integration modeling in the system which can be illustrated in Figure 4.In RDS, the electrical energy flows from the slack bus to the all of load bus.Since the load can absorb real power, flow direction of load current is direct.But, if the DG or power sources injection is on the distribution system, the direction of flow will be foreign to the load.A connection of DG is represented as a negative value in load bus.The equation of DG can be written as in (7).

RESULTS
The harmonic currents injection originating from a nonlinear VFD load at several load bus points in scenario 1 (S-1) generates the spread of harmonic distortion which makes the system performance in maintaining PQ worse.this can be seen by the presence of several load buses that have %THDv values >5%.In scenario 2 (S-2), the placement of several inverter-based DG as many 8 points with a capacity of 45 kW each shows a worsening effect on PQ performance.In scenario 3 (S-3), placement of an inverter-based DG using the MOPSO method provides an increase in PQ performance.The simulation results for each scenario in Table 2, Figures 5 and 6.Table 2 shows a comparison of several parameters from all existing scenarios on changes in PQ performance on the ULP Way Halim 88-bus RDS.In S-1, harmonic current injection due to the use of a nonlinear load in the form of a VFD makes PQ decrease.This can be seen by the presence of several buses that have a value of %THDv >5% and a total loss value of 85.12 kW.In S-2, the addition of a random inverter-based DG placement with a total capacity of 360 kW can reduce losses by up to 2.54 kW or 2.98%.In S-3, optimizing the placement and size of the inverter-based DG with a total capacity of 630 kW using MOPSO was able to reduce losses by up to 12.74 kW or 14.96%.The location and size of the inverter-based DG were optimized using MOPSO can be seen in Table 3.The existence of an active power supply from the DG placement can increase efficiency by reducing total losses in the ULP Way Halim 88-bus RDS. Figure 5 shows a comparison of the bus voltage level values at the ULP Way Halim 88-bus RDS.Changes in the value of the voltage level varies on S-2.The average decrease in the value of the voltage level is up to 0.034%.This is also seen by the increase and decrease in the value of the voltage level at several buses.In contrast to S-3, there was an enhancement in the value of the voltage level across all buses with an average of 0.009%.
Different changes compared to the decrease in the value of losses in S-2, the %THDv value increases in all ULP Way Halim 88-bus RDS buses with an average of 1.24%.this is because the injection of harmonic currents from the inverter-based DG which is placed makes the spread of harmonic distortion increase and worsens PQ has been seen in Figure 6.However, in S-3, the placement of an inverter-based DG optimized with MOPSO was able to reduce the %THDv value by an average of 7.74%.
The spread of harmonics occurring in a RDS is strongly influenced by the location and type of nonlinear load placed on the system as shown in S-1.Primary energy in the form of sunlight which is converted to generate electrical energy in inverter-based DG is growing rapidly.However, the location and size of inverter-based DG should not be placed casually.This is because the inverter-based DG also contributes harmonic currents which can increase the value of the spread of harmonics as seen in S-2.The positive impact in increasing the PQ of the location and size of the inverter-based DG will be optimal if the planning is carried out using an AI-based optimization method that has a predetermined objective function and constraints as shown in S-3.

CONCLUSION
The harmonic current injection originating from nonlinear loads in the form of VFD is able to exacerbate PQ due to the spread of harmonics on the ULP Way Halim 88-bus RDS.Placement of inverter-based DG that is not optimal can reduce total losses but increase the spread of harmonic distortion.The use of MOPSO with the objective function of minimizing total losses and %THDv with specified limits is able to determine the location and size of the inverter-based DG by considering the injection of harmonic currents originating from the DG type.Placement of 24 inverter-based DG points with a total capacity of 690 kW is able to reduce the total losses by 12.74 kW or 14.96%, reduce %THDv by an average of 7.74% to the allowable limit (%THDv<5%) and increase all level values bus voltage with an average increase of 0.009%.Subsequent research will examine the optimization of the combination of inverter-based DG and harmonic filters in industrial electrical systems that use inductive loads optimized with artificial intelligencebased methods.

Figure 2 .
Figure 2. Single line diagram of ULP Way Halim 88-bus RDS

Table 1 .
The harmonic sources value

Table 2 .
The results of simulation for all scenario

Table 3 .
Optimization result of placement inverter-based DG