ACTIVE POWER FILTER IMPACT ON POWER QUALITY IMPROVEMENT FOR INDUSTRUAL PLANT SUPPLIED BY PHOTOVOLTAIQUE STATION

This work deals on industrial electrical power quality improvement using an active power filter (APF) based on fuzzy logic controller. The studied system is a PV conversion chain used to supply linear and non-linear industrial loads of 200 kw plant. The PV station is modeled then simulated under Matlab Simulink, with standard conditions consideration in the first time then under real irradiation and sun power variation during the day hours, to obtain the true parameters design of the PV station to be able to give the needed power during the climatic conditions changing. APF parameters design and control loop structure are well studied to have an optimal control. MPPT technique is used in PV station control with the P&O algorithm. Also PV panel fault diagnosis has been done by using alarm for fault indicator and protection over under power. Simulation results are analyzed and discussed comparison with real signals, produced by the FLUK 6001B reference source with the use of the PRS 600 and the Teledyne 3,5 GHz-40Gs/s oscilloscope as measurement instrument, to improve the effectiveness of APF on electrical power quality, THD, amelioration and harmonic minimization in the industrial installation.


Introduction
The study presented in this paper is a starting point for an integration of supervising, control and data acquisition (SCADA) system relating to electrical energy quality and active power filters (APF) efficiency, integrated into an industrial area for various administrative companies and industrial production workshops.The main idea is based on PLCs (Programmable Logic Controllers) installed for the control of the APF and the supervision of the quantities affecting the energy quality (Harmonics) of particles (workshops) and all the companies in the industrial zone, This system can be extended, later, to supervise several industrial zones at the same time, figure (1)

represent a block diagram of the proposed global system
The idea is inspired from the study presented by Pedro Esteban in May 2020 and which is summarized by the principle diagram of the APF control using a PLC connected to an HMI which is connected by a Ethernet/Modbus TCP/IP bus shown in figure (2) [1] [2] [3].This configuration type is based on the general principle of remote control and supervision presented in the global diagram block of SCADA communication system shown on figure (3) below [3]:   When The purpose of this work is to integrate a PV system into the power supply system of an industrial company use power of 200 kw by exploiting the company recorded data, presented in Table (1) with these utilization factors daily, based on the detailed operating principle.Also in order to ensure the quality of the energy produced by the PV system [4] [5-8] dedicated to supplying the company's equipment, an APF will be installed after its sizing with these different control parts and data acquisitions as shown in Figure ( 4) below [9][10][11][12] : Modeling of PV system and mainly the solar panels in ideal conditions and simplified after sizing and the deduction of its characteristics make the study of MPPT commands influence on energy quality possible (the output voltage delivered by the proposed PV structure) [5] [13].
The required power to supply the plastic articles manufacturing company, which includes electrical machines of different powers, is summarized in the following table (1): Generally the company start operation at 6h to 8h by elements (11 resistors) which require a power of 15KW, after, a 34KW asynchronous motor will be started, then after 5 min the 2.4KW motor and that of 0.4KW will be started until at 9 a.m or motors of powers 18.5kw, 5.5kw, 1.7kw and 18.5kw and 2.2kw start operation until 1 p.m or motors of powers 18.5kw, 5.5kw, 1.7kw and 18.5kw and 2.2kw are stopped until 5 p.m then resume its operation another time until 7 p.m, from 7 p.m. until 9 p.m, we are in operation with a power of about 60kw.So the PV system must ensure at least the same power necessary for the operation of the company for the same duration, this requires according to the variation of the irradiation and the temperature during the day an equivalent number of panels which must ensure the same power demand, as shown on figure 5 (a), (b),(c)

PV system plant 3.1. PV system parameters Sizing
The PV system parameters sizing requires the study of the main quantities influence on the PV operation ( Ir ; T )

Low illuminance and temperature influence
The efficiency of many commercial modules is not constant and depending on the illumination.This phenomenon is very often neglected by the manufacturers, who give no indication of the behavior of their module at low illumination [13]; the following curves in Figure (6) shows the variation in power values given by a 250 Wp PV panel under standard conditions at a fixed temperature T equal to 25°c variable during the variation in irradiation (a) then T variable with 1000w/m² (b).The power delivered by a PV depends on the ambient temperature, the wind speed, the mounting of the module (integrated in the roof or ventilated) and all these parameters change according to the chosen site for modules installation.In addition, the coefficients linked to the temperature differ according to the materials used for the manufacture of the module [13] [14].
Temperature is a very important parameter in the behavior of PV cells [15].

Choice and calculation of panels according to daily consumption
Figure (8) shows the maximum operating point of PV between 6-7h which gives a maximum power of 43W Or the irradiation is about 200 W/m² with temperature 15°c while the machines requiring a power of 15KW (figure 5.a) this makes it possible to calculate the number of panels necessary to ensure such power which is: 15000/43=384 panels which gives 250 W under STC conditions.The curve in figure (9) summarizes the variation of the maximum operating point of the PV (without panels fault) according to the daily variation of temperature and irradiation.The following table (2) summarizes the general state of the study with consideration of Faults on PV panels So if we calculate the power lost on the power supplied by the panel, we obtain the number of panels needed to add to compensate the fault and preserve Company's equipment saves.For example in the 2 end case: 580/128=5 panels added to compensate lost power.
For the further cases, the same number of panels is obtained because of the power supplied by the panel with the presence of a fault, which has decreased in all cases to half the power of the panels without faults.
Solution: PV panel must be added is =5 panels to compensate the lost power.The study illustrate the influence of the meteorological parameters daily variation, temperature and solar irradiation, on the power produced by the PV system already proposed, sized and modeled, as well as the appearance of a fault at the level of the solar panel.In order to propose, after results analysis, an adequate solution with the diagnosis carried out on the system.

PV Diagnosis system Simulation:
Figure (10) presents the simulation block diagram of an autonomous PV system based on polycrystalline solar panels with a power of 250W under standard temperature and irradiation conditions (25˚C, 1000W/m²), which supplies a load via a DC-DC-AC converter.
. -When the PV system has a fault less than 5%, an indication by a pilot lamp (S1) is triggered.
-When the fault becomes more than 5% and less than 10% with the indication of the pilot lamp (S1), an Alarm buzzer (S2) will be activated by the diagnostic program in order to draw attention to the fault increase and the risk of loss of a significant power of that produced by the PV system.
-When the fault exceeds 15% with the lamp (S1) and the buzzer alarm (S2) automatic intervention becomes mandatory in order to protect the equipment, this is ensured by tripping a circuit breaker controlled by the diagnostic program.
There are several topologies of DC-DC converters.They are categorized according to whether the topology is isolated or non-isolated.Isolated topologies employ an isolation transformer operating at high frequency [5] [16], they are very often used in switching power supplies.The most popular topologies in the majority of applications are flyback, half-bridge and full-bridge.In photovoltaic (PV) applications, hybrid systems often employ these types of topologies when electrical isolation is preferred for safety reasons [17]- [23].

Simulation results Analysis and interpretation: 3.3.1 Case 1 :
Initially, the global system of figure (10) was simulated without the presence of a fault using 250W panels subjected to similar magnitudes of temperature and irradiation.Figure (11) left represents the simulation result on MATLAB/Simulink of an irradiation model (left) where it varies according to the hours of the day from 200W/m² from 6-7h then 600 W/m² from 7-9h and 1000 W/ m² from 9 a.m. to 5p.m. and 800 W/m² from 5 p.m. to 7 p.m. and take a value of 400 W/m² from 7 p.m. to 9 p.m.Where the figure (11) right represents the simulation result on MATLAB/Simulink of a temperature model where it varies according to the hours of the day from 15˚C from 6-7 a.m. then 15˚C from 7-9 a.m. and 15˚C, 35˚C from 9 a.m. to 5 p.m. and 32˚C from 5 p.m. to 7 p.m. and take a value of 25˚C from 7 p.m. to 9 p.m.This variation in temperature and irradiation allows the PV system to deliver a power corresponding to voltage and current depending on implanted load.

Case 2:
In order to follow the influence of the appearance of a fault on the PV panels, we created on the simulation block a set of faults appearing at 7 a.m a fault of 4 defective panels and another fault of 8 panels appearing at 1 p.m and in order to have a precision we have create a greater fault of 16 panels at 15h.The states S1, S2 and S3 of the fault detection program during the presence of a fault are summarized as follows: The state of the control signal of the lamp fault presence indicators or note that the lamp is on (S1=0) until the fault is present or the lamp lights up and (S1=1) take the value one.Also note that the alarm is off (S1=0) until the presence of a fault or the alarm is triggered when (S1=1) takes the value one.Also the breaker circuit is closed (S1=0) until the fault presence or the circuit breaker trips when (S1=1) takes the value one.

Comparaison
In order to compare the results, the curves of these results are superimposed to visualize the power variation due to the presence of a fault, which is presented in the following figures (12).
It is noticed that the value of the output voltage of the panel during the presence of a fault of 4 defective panels is less than that given before the fault presence, this voltage difference can be visualized better when the fault becomes bigger (8,16 panels).
Figure 12: The panel output current during the presence and absence of a fault It is noted that the value of the power delivered by the panels during the presence of a fault of 4 defective panels is less than that given before the presence of the fault, figure (12), this voltage difference can be visualized better when the fault becomes bigger (8,16 panels) from 230W up to 210W for example.The alarm program based on the information delivered by the diagnostic and fault detection program which sends a signal comprises the number of panels suffered and has a fault then reacts depending on the case or S1 the control signal of the indicator lamp of fault detection becomes one (1) even when there is a fault of less than 5 faulty panels.In the event that the faults become more than 5 and less than 15 faulty panels with the indicator lamp, a buzzer alarm controlled by S2 is activated.If the fault becomes larger (more serious) with more than 15 faulty panels, a circuit breaker controlled by the S3 signal must be tripped in order to isolate the workshop and ensure the protection of the equipment until the fault is maintained.

Solution
We propose to use a set of panels with power at least able to compensate the lost power when a serious fault appears (more than 15 faulty 250W panels) in order to ensure continuity of service and efficiency energy of the Photovoltaic system and supply the company equipment with the requested power until defective panels repaired.

S4 = 1 S4 = 0
The proposed solution consists of developing an algorithm Figure (13) integrates with the fault diagnosis program and controlling a circuit breaker connecting all the reserve panels to the global PV system then ensuring disconnection when the PV system maintenance is carried out.

PV Power quality improvement using APF 4.1. Proposed structure
Simulation bloc is shown on figure 9 below Figure 14: simulation bloc of PV load with APF The simulation under MATLAB of a photovoltaic structure composed mainly of a set of panels connected in series and in parallel to form a PV field which ensures the supply of an electrical load (company) via a DC-AC converter [10].The APF is connected to ensure the harmonic detection and the filtering current injection to minimize de THD value of the V and I.For harmonic detection and APF reference compensation currents, many methods can be used, the most several are summarized on figure (15) [19]- [26] Figure 15: different methodes for APF reference current calculation The synchronous reference frame method [10] is the best one to have a quickly calculation and the best adapted method with SCADA system because of minimal data needed to have acceptable results, it is based on the transfer of the three phases load current i abc to dq synchronous reference current using park transformation, then the dq components are fed to low pass filter to eliminate harmonic components from fundamental current [11].
The Park invers transformation, is used to convert instantaneous dq component of current into three phase components as shown on figure 11 below  Where e(t )is the chosen membership functions for error, de(t) for error variation and u(t) for control output.
The FLC intervals are chosen to have the best fuzzification and defuzzification values.

Simulation Results discussion
Simulation results shown that the installation of the APF on the network helps on harmonic minimization as improve the harmonic specter of figure (18).(19), be more near to the sinusoidal wave and the THD value [6,7] decrease from 23,6% to less than 7,46% value.The Dc voltage of APF capacitor is stabilized by a Fuzzy logic 3x3 controller, as shown on figure (23), where the response time is near to 0.09 second.The overtaking and response time of APF, V dc voltage regulator, using FLC is 0.09s.Where the 5, 7, 11, 17, 19 and 23 harmonics are reducing to less than 0.03 A from 1.6 A before the use of APF and the THD has decries to less than 7.46%.

Conclusion
In this work a Simulation under MATLAB of a photovoltaic structure composed mainly of a set of panels connected in series and in parallel to form a PV field which ensures the supply of an electrical load (company) via a DC-DC-AC converter Controlled in MPPT based on the P&O technique during the presence of certain faults of varying degree monitored by a diagnostic system based on a fault detection algorithm.The obtained results analysis confirm the efficiency of the proposed diagnostic system and the compensation of the power lost during fault appearance, which ensures energy efficiency of the photovoltaic system in the company's power supply.
The PV power sizing is based on statistic studies of sun power and temperature variation during climatic changing over the day, Faults can be appeared on PV panel is Also resolved by the proposition of a Diagnostic algorithm connected to indicator lamp when fault is les then 5 Panel and a sound Alarm with the lamp if the fault greater than 5 panel but when fault gone bigger a fault controlled breaker trigger to isolate the plant from PV station and protect equipment from the under voltage and power messing.
The PV plant connected to the network via a rectifier connected to an inverter supply linear and nonlinear loads of 200kw industrial plant voltage and currents are analyzed using the FFT method for power quality supervise and control; results are compared with the reference signals generated by the FLUK 6001B reference source with the use of the PRS 600 and the wavepro735Zi-A oscilloscope, 4 channels 3.5 GHz with band width 40 Gs/s, as measurement instrument Harmonic due to the presence of Nonlinear loads has been decrease by the use of a three phase shunt APF controlled by an FLC 3x3 rules.The APF installation on the network has decrease the THD of the network current from 23.6% to less than 7.46%.
Finely we hope applied more intelligent techniques to the studied system with hybrid electrical power of renewable energy to have more suitable results for industrial factory and achieve the power efficacy of renewable energy plants.

Figure 1 :
Figure 1: SCADA configuration for APF control and supervising application

Figure 4 :
Figure 4: PV system diagrams with APF supplies 200 kw industrial companyThe study of each part of the installation leads us to determine the operating system at each hour of the day and the year; the power output in use is a function of the system elements characteristics, data of solar irradiance and temperature.Modeling of PV system and mainly the solar panels in ideal conditions and simplified after sizing and the deduction of its characteristics make the study of MPPT commands influence on energy quality possible (the output voltage delivered by the proposed PV structure)[5][13].The required power to supply the plastic articles manufacturing company, which includes electrical machines of different powers, is summarized in the following table (1):Table1: Company's various equipment power demands

Figure 5 :
Figure 5: Company machines operating duration (a), Ir (b) and T (c) daily variation Curves

Figure 6 :
Figure 6: I, P(V) characteristic for different values of irradiance / temperature.

Figure 7 :
Figure 7: I, P (V) Characteristic for different values of temperature and different values of fixed irradiation.

Figure 9 :
Figure 9: PV Maximum operating point variation

Figure 10 :
PV and Diagnosis system simulation BlocWhere in Figure(10) presented also the simulation model for fault diagnosis in a PV panel represented by its single diode model and composed from a programming block on Matlab Function for temperature T variation and irradiations G, and another diagnostic block includes the fault and alarm presence indication program or:

Figure 11 :
Figure 11: Irradiation and temperature variation curves according to the day hours.

Figure 16 :
Figure 16: APF reference current calculation by SRF using Park Transformation 4.2.APF Control by FLC APF control loop is based on FLC controller where the membership's function of input and output are represented on figure 17 below:

Figure 18 :
Figure 18: (a) Harmonic specter with APF use FLC, (b) zoom This let network current, shown on figure(19), be more near to the sinusoidal wave and the THD value[6,7] decrease from 23,6% to less than 7,46% value.

Figure 19 :
Figure 19: Load, APF and Is currents after APF installation on NetworkThe APF current is shown on figure(20) below with the load current and PV current after APF connecting.

Figure 20 :
Figure 20: Load, APF & supply delivered Current.So, APF installation on network has decrease the harmonic presence and this make the PV plant more reliable and regulation system working well as if the THD is more important as it's shown on picture 21 given by the Fluk6100B.

Figure 22 :
Figure 22: spectre harmonic of a real voltage generated by the fluk 6100B

Table 1 :
Company's various equipment power demands

Table 2 :
power and number of PV corresponding to each period of operation of the company 21h

Table ( 3
(4)nd table(4)summarize the values of power lost due to panels Fault occurs when short circuit in 10 or 20 panel cells happen:

Table 4
The controller work based on rules table summarized on table 4: