Economical photovoltaic inverter group control method

Abstract

An economic photovoltaic inverter group control method, comprising calculating photovoltaic inverter group loss sources, photovoltaic inverter group losses including: IGBT power devices, filter inductance, DC bus capacitor and the like; reactive power setting through droop control; photovoltaic inverter group comprehensive operation loss minimum model; constraint condition of objective function; solving proportional coefficient through adaptive simulated annealing particle swarm optimization algorithm; obtaining photovoltaic inverter group reactive power proportional coefficient of photovoltaic inverter group at current applicable environment and operation loss minimum according to above results, through setting reactive power proportional coefficient, making photovoltaic inverter transmit and receive reactive power according to proportional coefficient in the process of droop control according to total reactive power requirement, achieving loss minimum result under the condition of guaranteeing normal grid-connected point voltage. The application realizes reasonable control of photovoltaic inverter to meet the demand of grid-connected point voltage regulation under the premise of photovoltaic inverter operation loss minimum.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of new energy power generation control, and particularly relates to a photovoltaic inverter group control method considering economy. BACKGROUND

[0002] With the increasing proportion of household photovoltaic access to the power grid year by year, multiple photovoltaic inverters are often connected in parallel on a single line to form a parallelly-operated photovoltaic inverter group. In this distributed photovoltaic power generation system with basically the same inverter model specifications, each inverter can work independently according to the irradiance intensity, and when the photovoltaic inverter array is not operated at full power, there is a reactive power output control problem of the inverter group. When there is renewable energy such as photovoltaic in the parallel system, the active power distribution principle based on the traditional droop control method will cause waste of photovoltaic energy; the dynamic characteristics of the droop control inverter adopting the traditional double-loop control structure are limited by the droop control loop bandwidth, and due to the use of linear controllers, the output voltage dynamic response is slow and the disturbance suppression ability is poor when external disturbances occur, and in severe cases, the system protection may be triggered. Because the common method in the prior art is droop control, it will damage the interests of users to some extent (voltage exceeding the dead zone will reduce the output, and stop the output)

[0003] The existing method usually sets the same reactive power output strategy for all photovoltaic inverters, and does not consider the loss and reactive power margin of the photovoltaic inverter, so it is difficult to optimize the economy of the photovoltaic inverter group.

[0004] How to reasonably control the photovoltaic inverter to meet the demand of grid-connected point voltage regulation under the premise of minimum operation loss of the photovoltaic inverter is a difficult problem to be solved. Therefore, it is of great significance to develop a photovoltaic inverter group control method considering economy. SUMMARY

[0005] The purpose of the present application is to overcome the above-mentioned shortcomings of the prior art, and to provide a photovoltaic inverter group control method considering economy, which reasonably controls the photovoltaic inverter to meet the demand of grid-connected point voltage regulation under the premise of minimum operation loss of the photovoltaic inverter.

[0006] The technical scheme of the present application includes the following steps:

[0007] Step 1, calculating the loss source of the photovoltaic inverter group

[0008] The loss of the photovoltaic inverter group includes IGBT power devices, filter inductors, DC bus capacitors, fuses, contactors, circuit breakers, and transformers.

[0009] Step 2: reactive power setting by droop control

[0010] Step 3: minimum comprehensive operation loss model of the photovoltaic inverter group

[0011] Step 4: constraint condition of objective function;

[0012] Step 5: solving proportional coefficient by adaptive simulated annealing particle swarm optimization algorithm (SA-PSO);

[0013] Step 6: obtaining reactive power proportional coefficient of photovoltaic inverter group at the time of minimum operation loss of photovoltaic inverter group in current applicable environment according to the above results, setting the reactive power proportional coefficient, making the photovoltaic inverter transmit and receive reactive power according to the proportional coefficient in the process of droop control, and achieving the result of minimum loss under the condition of ensuring normal voltage of the grid-connected point.

[0014] The proportional coefficient is obtained, then the photovoltaic inverter group outputs power according to the proportional coefficient, the minimum loss is achieved, and the voltage regulation range is limited in step 2, so that the proportional coefficient is obtained, and the photovoltaic inverter is reasonably controlled to meet the demand of grid-connected point voltage regulation.

[0015] The IGBT power device loss in step 1 is:

[0016] The IGBT power device includes IGBT and the anti-parallel diode above, so the power loss is divided into conduction loss P on and total switching loss E sw ; the conduction loss in one cycle includes the conduction loss P con_T of IGBT and the conduction loss P con_D of the anti-parallel diode; the total switching loss E sw includes the loss E on of IGBT opening once, the loss E off of IGBT closing once, and the loss E rr of the diode closing once.

[0017] The conduction loss P con_T of IGBT in one cycle and the switching loss E sw_T of IGBT in one cycle can be represented by the following formula:

[0018]

[0019] In the formula, C con represents a temperature coefficient, which is 2.5 at normal temperature, T j represents the junction temperature of IGBT, r on is the opening equivalent impedance, which is 0.816 m ohm, i(t) is the conduction current, U on is the conduction voltage drop at the opening moment of the device, B con is a fitting curve constant, which is usually 1.63.

[0020]

[0021] where A sw is the switching energy loss, B sw represents the current pending coefficient, usually 1.28; C sw represents the bus voltage pending coefficient in the switching state, usually 0.54, D sw represents the IGBT junction temperature pending coefficient in the switching state, usually 3.45;

[0022] When the number of IGBT switching is enough, formula (2) can be shown by the following formula:

[0023]

[0024] where f s is the carrier frequency, I m represents the output current peak value, ΔE on , ΔE off represents the IGBT on and off energy loss per unit current, π is 3.14;

[0025] The on-state loss P con_D of the anti-parallel diode and the switching loss E sw_D of the anti-parallel diode can be represented by the following formula:

[0026]

[0027] where R D is the on-state resistance of the anti-parallel diode, V F represents the on-state voltage drop of the anti-parallel diode, represents the phase of the output current lagging behind the output voltage, ma represents the modulation degree, usually less than 1,

[0028]

[0029] where ΔE rr is the anti-parallel diode off energy loss per unit current, f s represents the carrier frequency.

[0030] The on-state loss P on of the inverter power device and the total switching loss E sw ,

[0031]

[0032] where P con_T1 , P con_D4 is the IGBT on-state loss and the anti-parallel diode on-state loss, which can be obtained according to the switching control angle and formula (1) (4), E sw_T1 , Esw_D4 For the loss of IGBT conduction and the conduction loss of anti-parallel diode, the switching control angle and formula (2) (5) can be obtained.

[0033] The loss of the DC bus capacitor according to the application is:

[0034] The equivalent resistance R of the known capacitor sL , according to the parameter table of the inverter shown in Figure 1 4.3mΩ, and according to the loss value under the ripple current, the DC bus capacitor loss value can be obtained, and the capacitor side ripple value ΔI can be calculated by the following formula:

[0035]

[0036] I m The peak value of the output inverter side phase current, m is a parameter determined by the modulation mode (SPWM or SVPWM, according to the modulation degree in the selected mode, generally less than 1), The phase difference of the inverter output current lagging voltage.

[0037] The loss of the filter inductance in step 1 according to the application is:

[0038] The loss of the filter inductance can be divided into filter inductance iron loss P Fe And filter inductance copper loss P Cu , which can be calculated by the following formula:

[0039]

[0040] In the formula, i k Indicates the resistance of the inductance under each harmonic, R k Indicates the resistance of the current under each harmonic. (300 times or more, which may not have a fixed value)

[0041]

[0042] In the formula, k is a frequency-related coefficient, ΔU is the difference between the DC side voltage and the effective value of the transformer primary voltage, R cz Indicates the equivalent resistance of the magnetic hysteresis loss of the inductance core;

[0043] The loss ΔP T of the transformer running in step 1 can be calculated by the following formula:

[0044]

[0045] In the formula, ΔR s Indicates the equivalent internal resistance of the transformer, S NS represents the apparent power of the output of the inverter, and ΔP0 represents the no-load loss of the transformer.

[0046] The loss of the fuse, the contactor and the circuit breaker in step 1 is calculated by I 2 R, and the loss value is related to the input current on the DC side and the output current on the AC side.

[0047] After the loss sources of the inverter are clear, the control method of the inverter of the photovoltaic grid-connected system needs to be improved with the minimum loss as the target, and then in the next step, the traditional grid-connected inverter control method is considered, and the constraint conditions and related parameters that need to be considered are obtained. The result of this step is used for the construction of the minimum loss model in step 3.

[0048] The step 2 of the application comprises the following steps:

[0049] The reference voltage (220V, which changes according to the access position) u ref of the voltage control point is obtained by the Q(U) control strategy based on the voltage amplitude of the grid-connected point. ref As shown in the following formula.

[0050]

[0051] In the formula, u pcc represents the real-time detected voltage, m1 and m2 represent the coefficients of the droop control, and are usually-1 (both coefficients can be taken, but in practice, no constant number is usually taken, and generally, the coefficients are adjusted according to the implementation effect), u1 and u4 represent the maximum adjustable voltage (this needs to consider the capacity and reactive coefficient of the inverter, and also needs to consider the local policy, and generally is near the reference voltage of 0.8 and 1.2 times), u2 and u3 represent the dead zone voltage (in this interval, the strategy considers that the voltage normally fluctuates and does not act), and u i =k i u ref Generally, 1.07 >= k i >= 0.93.

[0052] The inverter can adapt to the voltage requirements of the grid-connected point through the droop control, and after the reactive output in different scenarios is determined, it is used as an important boundary quantity, which can reduce the comprehensive loss of the inverter group in the scenario in the next step. It is used for the construction of the constraint condition in step 4.

[0053] The step 3 of the application is that the comprehensive operation loss minimum model of the inverter group is:

[0054] There are n single-phase inverters working in parallel for power supply, and the proportion coefficient of each inverter output reactive power is respectively: alpha 1, alpha 2,... alpha n According to the relationship between reactive power and current, the size of the current in each inverter is obtained:

[0055]

[0056] Where P jmax represents the rated power of each grid-connected inverter, and U represents the rated voltage of the inverter; (Qref is explained above)

[0057] Here, in constructing the minimum loss model, the variables need to be unified for subsequent solution.

[0058] Then the comprehensive loss minimum model of the photovoltaic inverter group can be represented by the following formula:

[0059]

[0060] The first half is related to i, and the second half is a constant value.

[0061] The step 4 of the application is the constraint condition of the objective function:

[0062] According to the characteristics of the power grid, the objective function has the following equality constraints and inequality constraints:

[0063]

[0064] In the formula, lambda represents the power factor of a single grid-connected inverter, Q max represents the total reactive power of the inverter group.

[0065] The constraint condition is used for solving the optimal solution in step 5, and the solution is converged through the constraint condition, otherwise it is out of range and cannot be solved.

[0066] In step 5 of the application, the multi-objective SA-PSO algorithm is used to solve the proportion coefficient [alpha 1, alpha 2,... alpha n ] of the photovoltaic inverter group. First, initialize the loss parameters, such as frequency f s , switching time, tube voltage drop U on , working period T, modulation mode parameter ma, initialize algorithm parameters and position and speed of population particles, calculate the fitness of each particle, find the current optimal solution, and constantly update the temperature, and calculate the new target value and optimal value of each particle under the current temperature, if formula (14) is satisfied, the result is output, the value of current coefficient [beta 1, beta 2,... beta n ] can be obtained, and according to formula (12), the proportion coefficient [alpha 1, alpha 2,... alpha n], otherwise update the temperature and the velocity and position of the particles and recompute.

[0067] The adaptive simulated annealing particle swarm optimization algorithm (SA-PSO) in step 5 is used to solve the proportional coefficient, and the steps are as follows:

[0068] (a) Initialize parameters: including inertia weight w, learning factor c1, c2 and annealing speed σ.

[0069] (b) Randomly generate a population in the solution domain, containing m particles, and randomly initialize the velocity and position of each particle.

[0070] (c) Calculate the fitness of each particle i, Fs, which is the fitness function, that is, the objective function mentioned above, and record the position of each (i-th) particle Pid, the global optimal position Ppd, the fitness F(Pid) and the global optimal fitness F(Ppd).

[0071] (d) Calculate the initial temperature T of the annealing algorithm according to the global optimal fitness F(Ppd).

[0072] (e) Record the function fsa(x) as the function for calculating the fitness of the annealing algorithm, and the annealing algorithm fitness of Pid at the current temperature T can be calculated

[0073] (f) Apply the idea of roulette wheel selection algorithm to select one from the individual optimal position Pid to replace the global optimal position Ppd, denoted as Prd.

[0074] (g) Replace Ppd with Prd and bring it into the particle swarm formula to update the velocity of each particle

[0075] (h) Recalculate the fitness of each particle, update the optimal position Pid of each particle and the population optimal position Ppd

[0076] (i) Perform annealing operation: T = σt

[0077] (j) Determine whether the termination condition set in advance is met, if yes, stop searching and output the return value, if not, go to step (e).

[0078] Particle swarm simulated annealing is an optimization algorithm that continuously updates the solution according to the set boundary conditions to obtain the convergence value, which is not an exact solution.

[0079] The method of the present application can be applied to the system control of rural distributed photovoltaic. The rural household photovoltaic may have a more serious overvoltage phenomenon due to insufficient consumption capacity. It can also be deployed on a photovoltaic inverter controller or controlled by a scheduling instruction from a cloud platform. The power grid can remotely control the instruction. The photovoltaic inverter controller deployed with the method of the present application calculates the output proportion of the household photovoltaic according to the real-time situation and issues instructions to adjust the output of the photovoltaic inverter. The method of the present application considers the loss of the inverter, can adjust multiple inverters, ensures the inverter to adjust within the set range, and at the same time considers the loss. In the case of long-term application, the service life of the photovoltaic inverter can be significantly improved.

[0080] The present application starts from the economy of reactive power regulation when the reactive power capacity of the photovoltaic inverter group is sufficient, analyzes the multiple types of sources of photovoltaic inverter loss, and considers the voltage control demand to perform reactive power optimization control on the photovoltaic inverter group of the same type, to ensure the normal operation of the photovoltaic system and prolong the service life of the photovoltaic inverter. Compared with the existing control method, the control method provided by the present application can determine the proportional coefficient of the reactive power issued by each photovoltaic inverter according to the economic demand of the photovoltaic inverter system, and can better adapt to the development demand of distributed photovoltaic.

[0081] The present application has the following advantages:

[0082] 1. High-power photovoltaic inverters are important components of photovoltaic power station systems, and their efficiency is one of the important indicators for measuring grid-connected power generation systems. The present application can improve the efficiency of photovoltaic inverter control.

[0083] 2. If the loss of the inverter is too large, it will affect the service life of its internal components. The present application can control the loss generated during operation and prolong the service life of the inverter.

[0084] 3. On the basis of the same period of research, the loss model of the photovoltaic inverter is more comprehensively and comprehensively sorted out, and the loss of the inverter group is innovatively optimized. Finally, the simulated annealing particle swarm algorithm is used to solve the proportional coefficient, which has faster convergence speed and more accurate solution than the particle swarm algorithm.

[0085] 4. Under the premise of minimizing the operating loss of the photovoltaic inverter, the photovoltaic inverter is reasonably controlled to meet the demand of voltage regulation at the grid-connected point. BRIEF DESCRIPTION OF DRAWINGS

[0086] Figure 1 It is a structural schematic diagram of a three-phase grid-connected photovoltaic inverter. DETAILED DESCRIPTION

[0087] Figure 1 The structure of the three-phase grid-connected photovoltaic inverter shown, which has two IGBTs and anti-parallel diodes on the A phase, and the B and C phases are the same as the A phase.

[0088] The specific embodiments of the present application are as follows:

[0089] Step 1: Calculate the loss source of photovoltaic inverter

[0090] The group loss of photovoltaic inverter mainly includes: IGBT power device, filter inductance, filter capacitance, DC bus capacitance, fuse, contactor, circuit breaker, transformer, etc.

[0091] The power device in the photovoltaic inverter is IGBT and the anti-parallel diode above it, so its power loss is divided into on-state loss P on and switching loss E sw . The switching loss includes the loss E on of IGBT turning on once, the loss E off of IGBT turning off once, and the loss E rr of diode turning off once.

[0092] The on-state loss P con_T and the switching loss E sw_T can be represented by the following formula:

[0093]

[0094] In the formula, C con represents the temperature coefficient, which is 2.5 at normal temperature, T j represents the IGBT junction temperature, r on is the on-state equivalent impedance, which is 0.816 m ohm, i(t) is the on-state current, U on is the on-state voltage drop at the moment of device turning on, B con is the fitting curve constant, which is 1.63;

[0095]

[0096] In the formula, A sw is the switching energy loss, B sw represents the current undetermined coefficient, which is usually 1.28; C sw represents the bus voltage undetermined coefficient in switching state, which is 0.54, D sw represents the IGBT junction temperature undetermined coefficient in switching state, which is 3.45;

[0097] (The above IGBT loss formula is fitted through a large number of experiments, which is closely combined with the actual and has strong accuracy. However, according to different models, the parameters need to be re-consulted or fitted and calculated according to the technical manual)

[0098] When the number of IGBT switching is enough, formula (2) can be represented by the following formula:

[0099]

[0100] where f s represents carrier frequency, I m represents output current peak value, ΔE on , ΔE off represents energy loss of IGBT turn-on and turn-off under unit current.

[0101] on-state loss P con_D of anti-parallel diode and switching loss E sw_D can be represented by the following formula:

[0102]

[0103] where R D is on-state resistance of anti-parallel diode, V F represents on-state voltage drop of anti-parallel diode reference, represents phase of output current lagging output voltage, ma represents modulation degree.

[0104]

[0105] where ΔE rr is energy loss of anti-parallel diode turn-off under unit current, f s represents carrier frequency.

[0106] Photovoltaic inverters usually adopt different modulation modes, and one modulation mode often contains many carrier cycles. Taking a three-phase SPWM two-level inverter as an example, as shown in Figure 1 , the A phase has two IGBTs and anti-parallel diodes, the B and C phases are the same as the A phase, so the on-state loss P on and total switching loss E sw of the power devices of the inverter can be obtained:

[0107]

[0108] where P con_T1 , P con_D4 are loss of IGBT turn-on and loss of anti-parallel diode turn-on, which can be obtained according to switching control angle and formula (1) (4), E sw_T1 , E sw_D4 are loss of IGBT turn-on and loss of anti-parallel diode turn-on, which can be obtained according to switching control angle and formula (2) (5).

[0109] The loss of DC bus capacitor is easy to calculate. Given the equivalent resistance R sL of the capacitor, according to Figure 1The parameter table of the shown inverter takes 4.3mΩ, and according to the loss value size under the ripple current, the DC bus capacitor loss value can be obtained, and the capacitor side ripple value ΔI can be calculated by the following formula:

[0110]

[0111] I m is the peak value of the output inverter side phase current, m is a parameter determined by the modulation mode, is the phase difference of the inverter output current lagging voltage.

[0112] The loss of the filter circuit in the photovoltaic inverter mainly includes the filter capacitor loss and the filter inductor loss, the filter capacitor loss is usually small and can be ignored, while the filter inductor loss can be divided into filter inductor iron loss P Fe and filter inductor copper loss P Cu , which can be calculated by the following formula:

[0113]

[0114] In the formula, R k represents the resistance of the current at each harmonic.

[0115]

[0116] In the formula, k is a coefficient related to the frequency (k=6.94×10 Figure 1 in the photovoltaic inverter of the shown model - 4 f 1.603 , the magnetic loss coefficient equation of different manufacturers is different), ΔU is the difference between the DC side voltage and the effective value of the transformer primary voltage, R cz represents the equivalent resistance of the hysteresis loss of the core in the inductor (0.37mΩ); see the inductor specification book of the inverter.

[0117] The loss ΔP T of the transformer when running can be calculated by the following formula:

[0118]

[0119] In the formula, S N represents the capacity of the transformer, S represents the apparent power of the inverter output, and ΔP0 represents the no-load loss of the transformer.

[0120] Other losses such as fuses, contactors, circuit breakers, etc. can usually be calculated by I 2 R, and the loss value is related to the size of the DC side input current and the AC side output current.

[0121] Step 2: Reactive power setting by droop control

[0122] The reference voltage u of the voltage control point is approximately obtained by a Q(U) control strategy based on the voltage amplitude of the grid-connected point ref The required reactive power reference Q ref As shown in the following formula.

[0123]

[0124] In the formula, u pcc represents the real-time detected voltage, m1 and m2 represent the coefficients of the droop control, u1 and u4 represent the maximum adjustable voltage u2 and u3 represent the dead zone voltage, and u i =k i u ref Generally, 1.07≥k i ≥0.93.

[0125] Step 3: Inverter group comprehensive operation loss minimization model

[0126] There are n single-phase inverters working in parallel for power supply, and the proportional coefficients of the reactive power output of each inverter are respectively: α1, α2, … α n According to the relationship between reactive power and current, the size of the current in each inverter can be obtained:

[0127]

[0128] Where P jmax represents the rated power of each grid-connected inverter, and U represents the rated voltage.

[0129] Then the comprehensive loss minimization model of the photovoltaic inverter group can be represented by the following formula:

[0130]

[0131] Step 4: Constraint conditions of the objective function

[0132] According to the characteristics of the power grid, the objective function has the following equality constraints and inequality constraints:

[0133]

[0134] In the formula, λ represents the power factor of a single grid-connected inverter, Q max represents the total reactive power of the inverter group.

[0135] Step 5: Self-adaptive simulated annealing particle swarm optimization algorithm (SA-PSO) to solve the proportional coefficient

[0136] The proportional coefficient of the photovoltaic inverter group is solved by using the multi-objective SA-PSO algorithm. Firstly, the loss parameters such as frequency, switching time, tube voltage drop, duty cycle, modulation mode parameters, algorithm parameters, and the position and speed of the population particles are initialized. The fitness of each particle is calculated, and the current optimal solution is found. The temperature is constantly updated, and the new target value and optimal value of each particle at the current temperature are calculated. If formula (14) is met, the result, i.e. the proportional coefficient of the photovoltaic inverter reactive power output [α1, α2, … α n ], is output. Otherwise, the temperature and the speed and position of the particles are updated for re-computation.

[0137] Step 6: According to the above results, the photovoltaic inverter group reactive power proportional coefficient of the photovoltaic inverter group at the minimum operating loss under the current applicable environment is obtained.

Claims

1. A method for controlling a group of photovoltaic inverters taking economy into account, characterized in that Comprise: Step 1, calculate photovoltaic inverter group loss source: Photovoltaic inverter group loss includes: IGBT power device, filter inductance, DC bus capacitor, fuse, contactor, circuit breaker, transformer; Step 2: reactive power setting by droop control; Step 3: photovoltaic inverter group comprehensive operation loss minimum model; Step 4: constraint condition of objective function; Step 5: adaptive simulated annealing particle swarm optimization algorithm is used to solve the proportion coefficient of photovoltaic inverter reactive power output; Step 6: according to the above results, the proportion coefficient of photovoltaic inverter reactive power output is obtained when the photovoltaic inverter group is in the current applicable environment and the operation loss is minimum, by setting the proportion coefficient of photovoltaic inverter reactive power output, the photovoltaic inverter can receive and send reactive power according to the proportion coefficient of photovoltaic inverter reactive power output in the process of droop control, and the minimum loss result is achieved under the condition of ensuring the normal voltage of grid-connected point; The step 2 includes: The reference voltage u of the voltage control point is approximately obtained by a Q(U) control strategy based on the amplitude of the grid-connected point voltage ref The required reactive reference quantity Q ref As shown in the following formula: In the formula, u pcc represents the voltage detected in real time, m1 and m2 represent the coefficients of droop control, u1 and u4 represent the maximum adjustable voltage; u2 and u3 represent the dead-zone voltage, and there are u i = k i u ref ; The step 3: the comprehensive operation loss minimum model of inverter group is: There are n single-phase inverters working in parallel to supply power, and the proportional coefficients of reactive power output of each photovoltaic inverter are: α1, α2, … αn n According to the relationship between reactive power and current, the size of the current in each inverter is obtained: where P jmax represents the rated power of each grid-connected inverter, and U represents the inverter rated voltage; Then the comprehensive loss minimum model of photovoltaic inverter group is represented by the following formula: The step 4: constraint condition of objective function: According to the characteristics of power grid, the objective function has the following equality constraints and inequality constraints: In the formula, λ represents the power factor of a single grid-connected inverter, Q max represents the total reactive power of the inverter group; In step 5, the multi-objective SA-PSO algorithm is used to solve the reactive power output proportion coefficient [α1, α2,... α n ] of the photovoltaic inverter, first, the loss parameters are initialized, including the frequency f s , the switching time, the tube voltage drop U on , the working cycle T, the modulation mode parameters ma, the algorithm parameters and the position and speed of the population particles are initialized, the fitness of each particle is calculated, the current optimal solution is found, the temperature is constantly updated, the new target value and the optimal value of each particle under the current temperature are calculated, if formula (14) is satisfied, the result is output, the current coefficient [β1, β2,... β n ] is obtained, according to formula (12), the reactive power output proportion coefficient [α1, α2,... α n ] of the photovoltaic inverter can be obtained, otherwise, the temperature and the speed and position of the particles are updated and recalculated.

2. The method of claim 1, wherein The IGBT power device loss in the step 1 is: The IGBT power device includes an IGBT and an anti-parallel diode above the IGBT, so that the power loss of the IGBT power device is divided into conduction loss P on , and total switching loss E sw ; the conduction loss in one period includes the conduction loss P con_T of the IGBT and the conduction loss P con_D of the anti-parallel diode; the total switching loss E sw includes the loss E on of the IGBT turning on once, the loss E off of the IGBT turning off once, and the loss E rr of the diode turning off once; P on = Vce * Ic con_T E on = Vce * Ic sw_T is represented by the following equation: where C con represents the temperature coefficient, T j represents the IGBT junction temperature, r on is the on-state equivalent impedance, i(t) is the on-state current, U on is the on-state voltage drop at the moment of device turn-on, B con is the fitting curve constant; In the formula, A sw B is the switching energy loss sw C represents the current pending coefficient sw D represents the bus voltage pending coefficient in the switching state sw E represents the IGBT junction temperature pending coefficient in the switching state When the number of IGBT switching is enough, formula (2) is shown by the following formula: where f s represents the carrier frequency, I m represents the output current peak value, ΔE on , ΔE off represents the energy loss of IGBT turn-on and turn-off under unit current; On-state loss P of the anti-parallel diode con_D and switching loss E of the anti-parallel diode sw_D is represented by the following equation: wherein R D is the on-state resistance of the antiparallel diode, V F represents the reference on-state voltage drop of the antiparallel diode, represents the phase of the output current lags behind the output voltage, ma represents the modulation factor; wherein ΔE rr is the energy loss at unit current and with the antiparallel diode off, f s denotes the carrier frequency; Conduction losses P of the power devices of the inverter on and the total switching losses E sw ; where P con_T1 , P con_D4 is the loss of IGBT and the loss of anti-parallel diode when conducting, which is obtained according to the switching control angle and formula (1) (4) sw_T1 , E sw_D4 is the loss of IGBT and the loss of anti-parallel diode when conducting, which is obtained according to the switching control angle and formula (2) (5) 3. The method of claim 1, wherein The DC bus capacitor loss in the step 1 is: The equivalent resistance R of the capacitor is known sL According to the loss value under the ripple current, the DC bus capacitor loss value is obtained, and the capacitor side ripple value ΔI is calculated by the following formula: I m m is a parameter determined by the modulation method for the peak value of the inverter-side phase current, is the phase difference of the inverter output current lagging voltage.

4. The method of claim 1, wherein The filter inductance loss in the step 1 is: The filter inductance loss is divided into filter inductance iron loss P Fe and filter inductance copper loss P Cu and is calculated by the following formula: where i k represents the resistance of the inductance at each harmonic, R k represents the resistance of the current at each harmonic; In the formula, k is a frequency-dependent coefficient, ΔU is the difference between the DC side voltage and the effective value of the primary voltage of the transformer, R cz represents the equivalent resistance of the magnetic hysteresis loss of the core of the inductor; the loss ΔP of the transformer in operation in step 1 T is calculated from the formula: where ΔR s represents the equivalent internal resistance of the transformer, S N represents the capacity of the transformer, S represents the apparent power of the output of the inverter, and ΔP0 represents the no-load loss of the transformer.

5. The method of claim 1, wherein The loss of the fuse, contactor, and circuit breaker in step 1 is calculated by I 2 R, and the loss value is related to the size of the input current on the DC side and the size of the output current on the AC side.

6. The method of claim 1, wherein The proportion coefficient is solved by the adaptive simulated annealing particle swarm optimization algorithm in the step 5, and the steps are: (a) initialize parameters: including inertia weight w, learning factor c1, c2 and annealing speed sigma; (b) randomly generate a population in the solution domain, including m particles, and randomly initialize the speed and position of each particle; (c) calculate the fitness Fs of each particle i, Fs is the fitness function, that is, the objective function of the step 3 inverter group comprehensive operation loss minimum model, and record the position Pid of each particle, the global optimal position Ppd, the fitness F(Pid) and the global optimal fitness F(Ppd); (d) calculate the initial temperature T of annealing algorithm according to the global optimal fitness F(Ppd); (e) the function fsa(x) is the function for calculating the fitness of annealing algorithm, and the fitness of annealing algorithm of Pid at the current temperature T can be solved; (f) the idea of roulette wheel selection algorithm is applied, one is selected from the individual optimal position Pid to replace the global optimal position Ppd, which is recorded as Prd; (g) replace Ppd with Prd, and update the speed of each particle by bringing it into the particle swarm formula; (h) calculate the fitness of each particle again, and update the optimal position Pid of each particle and the population optimal position Ppd; (i) annealing operation: T = sigma t; (j) judge whether the termination condition set in advance is met, if yes, stop searching and output the return value, if not, go to step (e).

Images