JP7830005B2 - Power converters, control devices, and distributed power systems - Google Patents

Description

Embodiments of the present invention relate to a power conversion device, a control device, and a distributed power supply system.

In distributed power systems using distributed power sources such as solar power generators, wind turbines, and batteries connected to the power grid, reactive power is injected from the distributed power sources to the grid connection point to compensate for voltage fluctuations at the grid connection point. The injection of reactive power is controlled by a power converter that converts the power from the distributed power sources into power appropriate for the power grid. Furthermore, the distributed power system may include a control device that controls the operation of the power converter.

For example, it is being considered to estimate the apparent system characteristics from the interconnection point of distributed power sources using computational methods such as the least squares method and extended Kalman filters, and then inject reactive power based on the estimation results. This allows for the injection of reactive power corresponding to system impedance, and can suppress voltage fluctuations even when system impedance is high.

System characteristics are estimated based on measured values such as active power, reactive power, and voltage at the connection point with the power system. However, these measured values may be superimposed with noise due to measurement errors, etc. Furthermore, system characteristics can change rapidly and significantly due to factors such as circuit breaker switching within the power grid, or tripping of nearby loads or power sources.

When estimating system characteristics using computational methods robust to noise, there is a concern that it may take a long time to accurately estimate the system characteristics after they have changed. Conversely, when using computational methods highly responsive to changes in system characteristics, there is a concern that the error in estimating system characteristics will increase due to the influence of noise if noise is superimposed on the measured values.

Thus, attempting to improve the noise tolerance of measured values reduces the ability to track changes in system characteristics, and attempting to improve the ability to track changes in system characteristics can lead to oversensitivity to noise superimposed on the measured values, potentially increasing the error in estimating system characteristics.

Therefore, it is desirable that power converters, control devices that manage the operation of power converters, and distributed power systems using these components be designed to improve both the resistance to noise in measured values and the ability to track changes in grid characteristics.

Japanese Patent Publication No. 2017-63525

Nikolaus Hansen. The CMA Evolution Strategy: A Tutorial. 2005. hal-01297037

Embodiments of the present invention provide a power conversion device, a control device, and a distributed power supply system that can improve resistance to noise in measured values while also improving the ability to track changes in system characteristics.

According to an embodiment of the present invention, a power conversion device connects a distributed power source to a power system by converting the power of the distributed power source into AC power corresponding to a power system connected to an infinite bus power system, and supplying the converted AC power to the power system, comprising: a main circuit unit that converts the power of the distributed power source into AC power corresponding to the power system; and a control unit that controls the operation of the main circuit unit, wherein the control unit calculates an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance, and an estimated value of the voltage of the infinite bus power system based on the active power value of the connection point with the power system, the reactive power value of the connection point, and the voltage value of the connection point; and an estimated value calculation unit that calculates an estimated value of the resistance component of the system impedance and the reactive value of the system impedance. A power converter is provided, comprising: a reactive power calculation unit that calculates a reactive power command value for the reactive power supplied to the power system based on an estimated actance component and an estimated voltage value of the infinite busbar power system; and a drive circuit that drives the main circuit unit to output a predetermined active power and reactive power corresponding to the reactive power command value. The estimation unit receives period data including multiple active power values, multiple reactive power values, and multiple voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated resistance component of the system impedance of the power system, an estimated reactance component of the system impedance of the power system, and an estimated voltage value of the infinite busbar power system based on the period data.

A power converter, control device, and distributed power system are provided that can improve resistance to noise in measured values while also enhancing the ability to track changes in grid characteristics.

This is a block diagram schematically representing a distributed power supply system according to the embodiment. This is a schematic block diagram showing a power conversion device for a distributed power supply system according to an embodiment. This is a schematic block diagram showing the estimated value calculation unit according to the embodiment. This is a flowchart that schematically represents the operation of the estimation unit. Figures 5(a) and 5(b) are schematic graphs that represent an example of period data. Figures 6(a) to 6(d) are schematic graphs illustrating an example of the calculation results of the estimation unit. This is a schematic block diagram showing a modified example of the estimated value calculation unit according to the embodiment. This is a schematic block diagram showing a modified example of the estimated value calculation unit according to the embodiment. This is a schematic block diagram showing a modified example of the estimated value calculation unit according to the embodiment. This is a block diagram schematically representing a modified example of the distributed power supply system according to the embodiment.

Each embodiment will be described below with reference to the drawings.
Please note that the drawings are schematic or conceptual, and the relationships between the thickness and width of each part, as well as the ratios of the sizes of the parts, are not necessarily identical to those of reality. Furthermore, even when representing the same part, the dimensions and ratios may differ between drawings.
In this specification and in each figure, elements similar to those described above are denoted by the same reference numerals, and detailed explanations are omitted as appropriate.

Figure 1 is a block diagram schematically representing a distributed power supply system according to an embodiment.
As shown in Figure 1, the distributed power system 2 comprises a power system 4 connected to an infinite busbar power system 3, a distributed power source 6, and a power converter 10. The power of the power system 4 is alternating current power. The power of the power system 4 is, for example, three-phase alternating current power.

The distributed power source 6 is, for example, a solar panel. The power from the distributed power source 6 is direct current (DC) power. The power converter 10 is connected to the distributed power source 6 and also to the power system 4 via transformers 12, 14, etc. The power converter 10 converts the DC power from the distributed power source 6 into AC power corresponding to the power system 4, and supplies the converted AC power to the power system 4, thereby connecting the distributed power source 6 to the power system 4.

The distributed power source 6 is not limited to solar panels; it may also be other generators such as wind turbines or gas turbines. Furthermore, the distributed power source 6 may also be an energy storage device using batteries or capacitors. The power from the distributed power source 6 is not limited to DC power; it may also be AC power. The configuration of the power conversion device 10 may be any configuration capable of converting the power supplied from the distributed power source 6 into AC power compatible with the power grid 4.

Immediately adjacent to power system 4, in addition to distributed power sources 6 and power converter 10, other potential connections include, for example, consumers 16 (loads) and other generators 18. The power converter 10 supplies active power to power system 4 based on the output of distributed power sources 6, and also supplies optimal reactive power to power system 4. This allows the power converter 10 to suppress voltage fluctuations at the connection point LP with power system 4, which are caused by its own active power supply and the influence of consumers 16 and generators 18.

The distributed power system 2 comprises, for example, multiple distributed power sources 6 and multiple power converters 10 corresponding to each of the multiple distributed power sources 6. However, the number of distributed power sources 6 and power converters 10 provided in the distributed power system 2 can be any number. The number of distributed power sources 6 and power converters 10 may be just one. Furthermore, the number of power converters 10 does not necessarily have to be the same as the number of distributed power sources 6. For example, multiple distributed power sources 6 may be connected to a single power converter 10.

Figure 2 is a schematic block diagram showing a power conversion device for a distributed power supply system according to an embodiment.
As shown in Figure 2, the power converter 10 has a main circuit section 40 and a control section 42. The main circuit section 40 converts DC power or AC power supplied from the distributed power source 6 into AC power corresponding to the power system 4. The control section 42 controls the operation of the main circuit section 40.

The main circuit section 40, for example, has multiple switching elements, and performs power conversion by switching the multiple switching elements on and off. The control unit 42 controls the power conversion by the main circuit section 40 by controlling the on/off switching of the multiple switching elements of the main circuit section 40. For example, a well-known inverter circuit is used in the main circuit section 40. The configuration of the main circuit section 40 can be any configuration that can perform the above power conversion.

The distributed power system 2 further includes, for example, measuring devices 20 and 22. The measuring device 20 detects the voltage value Vdc of the DC voltage input from the distributed power source 6 to the power converter 10, and the current value Idc of the DC current input from the distributed power source 6 to the power converter 10, and inputs the detected voltage value Vdc and current value Idc to the control unit 42.

The control unit 42, for example, performs MPPT (Maximum Power Point Tracking) control to make the DC power track the maximum power point of the distributed power source 6. For example, based on the voltage value Vdc and current value Idc detected by the measuring device 20, the control unit 42 extracts the maximum power point (optimal operating point) of the distributed power source 6 and controls the operation of the main circuit unit 40 to supply active power corresponding to the extracted maximum power point to the power system 4.

However, the method for determining the active power supplied from the power converter 10 to the power system 4 is not limited to the MPPT method. The active power supplied from the power converter 10 to the power system 4 may be determined, for example, based on an active power command value input from a higher-level controller. The control unit 42 may control the operation of the main circuit unit 40 to supply active power to the power system 4 according to the input active power command value.

The measuring device 22 detects the active power value P, the reactive power value Q, and the voltage value Vs at the interconnection point LP of the power converter 10 with the power system 4, and inputs the detected active power value P, reactive power value Q, and voltage value Vs to the control unit 42.

The control unit 42 comprises an estimated value calculation unit 50, a reactive power calculation unit 52, and a drive circuit 54. The control unit 42 inputs the active power value P, reactive power value Q, and voltage value Vs, received from the measuring device 22, into the estimated value calculation unit 50.

The estimation unit 50 calculates the estimated value ^R of the resistance component R of the power system 4's system impedance, the estimated value ^X of the reactance component X of the power system 4's system impedance, and the estimated value ^Vr of the voltage Vr of the infinite bus power system 3, based on the active power value P, reactive power value Q, and voltage value Vs input from the measuring device 22. Note that the ^ (hat) symbol representing estimated values such as ^R is normally placed directly above R, as shown in Figure 2, but in this specification, for formatting reasons, it will be offset, as shown ^R.

In other words, the estimation unit 50 estimates the system characteristics of power system 4 based on the active power value P, the reactive power value Q, and the voltage value Vs. In this process, as shown in Figure 2, the estimation unit 50 considers the system model of power system 4 as the simplest system model, consisting only of the resistive component R and the reactance component X of the system impedance. The estimation unit 50 inputs the calculated estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52.

The reactive power calculation unit 52 calculates the reactive power command value Qop to be supplied to the power system 4 based on the estimated values ^R, ^X, and ^Vr input from the estimated value calculation unit 50. The reactive power calculation unit 52 inputs the calculated reactive power command value Qop to the drive circuit 54.

The reactive power calculation unit 52 calculates the reactive power command value Qop from each estimated value ^R, ^X, and ^Vr using, for example, the following equation (1). The reactive power calculation unit 52 calculates the optimal reactive power command value Qop based on each estimated value ^R, ^X, ^Vr and the active power value P. In equation (1), "sign" is the sign function.

The drive circuit 54 receives the reactive power command value Qop calculated by the reactive power calculation unit 52, as well as the voltage value Vdc and current value Idc of the distributed power supply 6 measured by the measuring device 20.

The drive circuit 54 determines the active power by MPPT control based on the voltage value Vdc and the current value Idc, and drives the main circuit unit 40 to output the determined active power and reactive power corresponding to the reactive power command value Qop. The drive circuit 54 supplies the determined active power and reactive power corresponding to the reactive power command value Qop from the main circuit unit 40 to the power system 4 by switching the on and off states of multiple switching elements in the main circuit unit 40.

The measuring device 22 and the estimation unit 50 periodically acquire the active power value P, the reactive power value Q, and the voltage value Vs. Each time the estimation unit 50 acquires the active power value P, the reactive power value Q, and the voltage value Vs, it calculates (predicts) the respective estimated values ^R, ^X, and ^Vr.

The reactive power calculation unit 52 calculates the reactive power command value Qop each time it receives the estimated values ^R, ^X, and ^Vr from the estimated value calculation unit 50. The drive circuit 54 generates a control signal for the main circuit unit 40 each time the reactive power command value Qop is received, and supplies active power and reactive power from the main circuit unit 40 to the power system 4. The control unit 42 repeats the above process, supplying active power to the power system 4 according to the distributed power source 6, and also supplying reactive power to the power system 4 as needed, according to the system characteristics of the power system 4.

In this way, by supplying active and reactive power to the power system 4, it is possible to suppress the supply of active power from the power converter 10 to the power system 4, and to suppress voltage fluctuations at the interconnection point LP caused by the influence of consumers 16 and generators 18. For example, fluctuations in the voltage value Vs at the interconnection point LP can be kept within ±2%.

Figure 3 is a block diagram schematically representing the estimation calculation unit according to the embodiment.
As shown in Figure 3, the estimated value calculation unit 50 includes a data generation unit 60 and an estimation unit 62.

The data generation unit 60 generates period data containing multiple active power values P, multiple reactive power values Q, and multiple voltage values Vs for a predetermined period, based on the active power value P, reactive power value Q, and voltage value Vs input from the measuring device 22. The data generation unit 60 then inputs the generated period data to the estimation unit 62.

The data generation unit 60 generates period data based on the input from the measuring device 22 by, for example, storing the active power value P, reactive power value Q, and voltage value Vs input from the measuring device 22 for a predetermined period.

The measuring device 22 detects, for example, the active power value P, the reactive power value Q, and the voltage value Vs at intervals of several seconds. In this case, the predetermined period is, for example, a period of several minutes, such as 5 minutes to 30 minutes. For example, if the measurement period of the measuring device 22 is 5 seconds and the predetermined period is 30 minutes, the period data will contain 360 data points each for the active power value P, the reactive power value Q, and the voltage value Vs. However, the predetermined period for the period data is not limited to the above and may be any period. The measurement period of the measuring device 22 is not limited to the above and may be any period.

Furthermore, the data generation unit 60 generates period data, for example, by processing a moving window. After generating period data according to the elapsed time of a predetermined period, when the data generation unit 60 receives input of the next active power value P, reactive power value Q, and voltage value Vs from the measuring device 22, it deletes the oldest active power value P, reactive power value Q, and voltage value Vs data and adds the newly input active power value P, reactive power value Q, and voltage value Vs data to generate the next period data. In this case, the data generation unit 60 generates period data each time it receives input of the active power value P, reactive power value Q, and voltage value Vs from the measuring device 22, and inputs the generated period data to the estimation unit 62.

However, the method of generating period data by the data generation unit 60 is not limited to the above. For example, the data generation unit 60 may generate period data at predetermined intervals by starting the accumulation of active power value P, reactive power value Q, and voltage value Vs data from scratch each time period data is generated. The data for active power value P, reactive power value Q, and voltage value Vs may be input to the data generation unit 60 from sources other than the measuring device 22, such as a higher-level controller. The configuration of the data generation unit 60 can be any configuration capable of receiving input of active power value P, reactive power value Q, and voltage value Vs, and generating period data containing multiple active power values P, multiple reactive power values Q, and multiple voltage values Vs for a predetermined period based on the input active power value P, reactive power value Q, and voltage value Vs.

Furthermore, the period data does not necessarily have to be generated by the estimation unit 50 (data generation unit 60). The period data may be generated, for example, by a measuring device 22 or a higher-level controller. The estimation unit 50 may be configured to receive period data input from, for example, the measuring device 22 or a higher-level controller, and input the input period data to the estimation unit 62. Thus, the data generation unit 60 can be provided in the estimation unit 50 as needed and is optional. When period data is input to the estimation unit 50 from an external device such as the measuring device 22 or a higher-level controller, the data generation unit 60 is omitted from the estimation unit 50.

The estimation unit 62 receives period data as input and applies a Covariance Matrix Adaptation Evolution Strategy (CMA-ES) to the input period data. Based on this, it calculates the estimated value of the resistance component R of the power system 4's system impedance, the estimated value of the reactance component X of the power system 4's system impedance, and the estimated value of the voltage Vr of the infinite busbar power system 3, ^Vr. In other words, the estimation unit 62 calculates representative system characteristics (^R, ^X, ^Vr) of the period data by using the Covariance Matrix Adaptation Evolution Strategy.

The covariance matrix adaptive evolutionary strategy is a method for searching for solutions using a multivariate normal distribution N(m, ^σ²C ) that represents the probability distribution of random variable vectors. Here, m is the mean vector, C is the covariance matrix (variance-covariance matrix), and σ is the step size. The mean vector m is a vector representing the mean values of the solution's resistance component R, reactance component X, and voltage value Vr. The step size σ is a scalar value that increases or decreases depending on the distance between the search point and the optimal solution, and it is a parameter that plays a role in increasing the search speed by correcting the covariance matrix C.

Figure 4 is a flowchart schematically illustrating the operation of the estimation unit.
Figure 4 shows, in more detail, an example of the operation of the estimation calculation of phylogenetic characteristics using the covariance matrix adaptive evolutionary strategy by the estimation unit 62.

When the estimation unit 62 performs calculations of phylogenetic characteristics, it first sets the period data input from the data generation unit 60, etc., as input data for the covariance matrix adaptive evolutionary strategy (step S101 in Figure 4).

After setting the period data, the estimation unit 62 initializes each parameter of the covariance matrix adaptive evolutionary strategy (step S102 in Figure 4). The estimation unit 62 initializes the covariance matrix C with the identity matrix. The estimation unit 62 initializes the mean vector m with the initial values _R₀ , _X₀ , and _Vr₀ of the mean values of the resistance component R, reactance component X, and voltage value Vr, respectively. The initial values _R₀ , _X₀ , and _Vr₀ are set to, for example, typical values of the resistance component R, reactance component X, and voltage value Vr that can be taken in the power system 4. The estimation unit 62 initializes the step size σ to the initial value _σ₀ . The initial value _σ₀ is set to any value greater than or equal to 0 depending on the search space (R, X, Vr).

Furthermore, the estimation unit 62 initializes the evolutionary path p _C of the covariance matrix C used in the covariance matrix adaptive evolutionary strategy, and the evolutionary path p _σ with step size σ, to 0. The evolutionary paths p _C and p _σ are used to increase the stability of optimization by accumulating the movement directions of individuals in past generations.

After initializing each parameter, the estimation unit 62 performs calculations for multiple search points x that follow a multivariate normal distribution N(m, ^σ²C ) (step S103 in Figure 4). The estimation unit 62 generates λ search points x through calculations. The number of search points x, λ, can be determined, for example, by the formula λ = 4 + 3 ln N, where N is the input dimension of the objective function. In this example, the input dimension N is 3, consisting of the resistance component R, the reactance component X, and the voltage value Vr, so λ is, for example, 7. That is, the estimation unit 62 generates 7 search points x. However, the number of search points x, λ, is not limited to the above and can be any value.

In calculating the λ search points x, the estimation unit 62 first decomposes the covariance matrix C into the following form. In the equation below, B is an orthogonal matrix. D is a diagonal matrix whose diagonal elements are the square roots of the eigenvalues of the covariance matrix C. T represents the transpose.
C = BDDB ^T
Next, the estimation unit 62 generates a vector z containing a number of random numbers corresponding to the input dimension N from a standard normal distribution N(0,I). In this example, a vector z containing three normally distributed random numbers is generated. After generating the vector z, the estimation unit 62 calculates the search point x using the following formula.
y = BDz
x = m + σy
In the above equation, y represents a linear transformation of the vector z generated from a standard normal distribution. The estimation unit 62 calculates λ search points x by repeating the generation of vector z and the calculation of search points x λ times. Each of the multiple search points x is, in other words, a vector containing randomly generated values of resistance component R, reactance component X, and voltage value Vr.

The estimation unit 62 calculates λ search points x and then calculates the evaluation value for each search point x (step S104 in Figure 4). In calculating the evaluation value, the estimation unit 62 first calculates the calculated voltage Vse of the interconnection point LP using the following equation (2), based on the values of multiple active power values P, multiple reactive power values Q, and the resistance component R, reactance component X, and voltage value Vr included in the period data.

The estimation unit 62 calculates multiple voltage values Vse for a single search point x, corresponding to the number of data points included in the period data. For example, as described above, if the period data includes 360 data points for active power P, reactive power Q, and voltage Vs, the estimation unit 62 calculates 360 voltage values Vse for a single search point x.

The estimation unit 62 calculates the calculated voltage values Vse for each of the multiple data points included in the period data for a single search point x, and then calculates the error (Vs - Vse) between the actual measured voltage value Vs included in the period data and the calculated value Vse.

More specifically, the estimation unit 62 calculates the calculated voltage Vse using equation (2) above, based on the active power value P and reactive power value Q of one data point included in the period data, the resistance component R, reactance component X, and voltage value Vr of one search point x, and then calculates the error between the calculated voltage Vse and the voltage value Vs included in the same data.

The estimation unit 62 calculates the above error for each of the multiple data points included in the period data, and then calculates the sum of the absolute values of the errors between the voltage value Vs and the calculated value Vse as the evaluation value Jv for the voltage of one search point x, as shown in equation (3) below.

Furthermore, the estimation unit 62 calculates the reactive power value Qop using equation (1) above, based on the multiple active power values P included in the period data, and the values of the resistance component R, reactance component X, and voltage value Vr included in the λ search points x.

The estimation unit 62 calculates multiple reactive power values Qop corresponding to each of the multiple data points included in the period data for a single search point x, and then calculates the error (Q - Qop) between the actual measured reactive power value Q included in the period data and the calculated value Qop.

More specifically, the estimation unit 62 calculates the reactive power value Qop using equation (1) above, based on the active power value P of one data point included in the period data, the resistance component R, the reactance component X, and the voltage value Vr of one search point x, and then calculates the error between the reactive power value Q included in the same data and the calculated value Qop.

The estimation unit 62 calculates the above error for each of the multiple data points included in the period data, and then calculates the sum of the absolute values of the errors between the reactive power value Q and the calculated value Qop as the evaluation value Jq for the reactive power of one search point x, as shown in equation (4) below. In equations (3) and (4), n corresponds to the number of data points included in the period data.

The estimation unit 62 performs the calculation of the evaluation value Jv related to the voltage and the calculation of the evaluation value Jq related to the reactive power for each of the λ search points x.

The estimation unit 62 calculates the evaluation values Jv and Jq for each of the λ search points x, and then extracts the top μ search points x with the highest evaluations based on the evaluation value Jv (step S105 in Figure 4).

The number of top-ranking search points x, μ, is determined, for example, by an integer value of λ/2. For example, if λ is 7, then μ is 3. The estimation unit 62 then extracts, for example, three search points x with the highest evaluation from among seven search points x.

The estimation unit 62 gives a higher evaluation to search points x with small evaluation values Jv related to voltage. In other words, the estimation unit 62 gives a higher evaluation to search points x with small errors between the actual measured voltage value Vs and the calculated value Vse. That is, the estimation unit 62 extracts the top μ search points x with small evaluation values Jv from among λ search points x.

The estimation unit 62 extracts the top μ search points x, and then updates each parameter of the covariance matrix adaptive evolutionary strategy based on the extracted μ search points x (step S106 in Figure 4). In other words, the estimation unit 62 updates the multivariate normal distribution N(m, ^σ²C ) based on the extracted μ search points x.

In updating each parameter, the estimation unit 62 first calculates a weighted average y _W of y corresponding to the top μ search points x using the weights w _i , according to equation (5) below.

The weights w _i can be determined, for example, as shown in equation (6) below, such that the sum of the weights w _i of the μ search points x equals 1.

The estimation unit 62 calculates the weighted average y _W and then updates the mean vector m according to the following equation (7). In equation (7), c _m represents the learning rate of the mean vector m. c _m is an arbitrary coefficient, for example, which is recommended to be set to 1.

Next, the estimation unit 62 updates the step size σ. In updating the step size σ, the estimation unit 62 first updates the evolutionary path p _σ of step size σ using the following equation (8).

In equation (8), c _σ represents the learning rate of the evolutionary path p _σ . _{μ eff} is expressed by the following equation (9).

The estimation unit 62 updates the evolutionary path p _σ , and then uses the updated evolutionary path p _σ to update the step size σ according to equation (10) below. In equation (10), d _σ is the decay parameter for updating the step size σ.

The estimation unit 62 then updates the covariance matrix C. In updating the covariance matrix C, the estimation unit 62 first updates the evolutionary path p _C of the covariance matrix C using the following equation (11).
In equation (11), c _C represents the learning rate of the evolutionary path p _{C. h σ} is the evolutionary path, and it serves to stop updating the evolutionary path p _C of the covariance matrix C when the step size σ is too large.

After updating the evolutionary path p _C , the estimation unit 62 updates the covariance matrix C using the updated evolutionary path p _C according to equation (12) below.

In equation (12), c ₁ represents the learning rate for rank-one updating of the covariance matrix C. _{c μ} represents the learning rate for rank-μ updating. δ(h _σ ) can be expressed as δ(h _σ ) = (1 - h _σ )c _C (2 - c _C ). Σw _j is expressed by the following equation (13).

Furthermore, in equation (12), the part of the first term c ₁ δ(h _σ ) - c ₁ - c _μ Σw _j can be added as needed and omitted. In other words, c ₁ δ(h _σ ) - c ₁ - _{c μ} Σw _j can be set to 0.

After updating each parameter of the covariance matrix adaptive evolutionary strategy as described above, the estimation unit 62 determines whether the termination conditions for the estimation of phylogenetic characteristics have been met (step S107 in Figure 4).

If the estimation unit 62 determines that the termination conditions are not met, it returns to the process in step S103. The estimation unit 62 performs a termination determination based, for example, the number of calculations performed from step S103 to step S106 and the evaluation value Jv. For example, if the number of calculations is less than a predetermined number and the evaluation value Jv is greater than or equal to a predetermined value, the estimation unit 62 determines that the termination conditions are not met. Then, the estimation unit 62 determines that the termination conditions are met when the number of calculations reaches the predetermined number or when the evaluation value Jv falls below the predetermined value.

Thus, the estimation unit 62 repeatedly performs the calculations from step S103 to step S107, for example, until the number of calculations reaches a predetermined number, or until the evaluation value Jv falls below a predetermined value and it is determined that the estimation calculation of the system characteristics has sufficiently converged. Here, the evaluation value Jv used for the determination is the evaluation value Jv calculated using the same procedure as for the search point x, based on the average values of the resistance component R, reactance component X, and voltage value Vr included in the average vector m updated in the final calculation, and the period data. Furthermore, the termination conditions are not limited to those described above. The estimation unit 62 may, for example, determine termination based solely on the number of calculations.

If the estimation unit 62 determines that the termination condition has been met, it outputs the average values of the resistance component R, reactance component X, and voltage value Vr included in the average vector m updated in the final calculation as the estimated value ^R for the resistance component R of the power system 4's system impedance, the estimated value ^X for the reactance component X of the power system 4's system impedance, and the estimated value ^Vr for the voltage value Vr of the infinite bus power system 3.

As described above, the estimation unit 62 applies a covariance matrix adaptive evolutionary strategy to the period data to calculate the estimated value of the resistance component R of the power system 4's system impedance, the estimated value of the reactance component X of the power system 4's system impedance, and the estimated value of the voltage Vr of the infinite busbar power system 3, based on the period data.

The estimation unit 62 sets, for example, the resistance component R, the reactance component X, and the voltage value Vr as the mean vector m of a multivariate normal distribution N(m, ^σ²C ), calculates multiple search points x that follow the multivariate normal distribution N(m, ^σ²C ), calculates an evaluation value Jv for each of the multiple search points x, extracts the top few search points x with the highest evaluations based on the evaluation value Jv, updates the multivariate normal distribution N(m, ^σ²C ) based on the extracted top few search points x, and repeatedly performs the calculations from calculating multiple search points x to updating the multivariate normal distribution N(m, ^σ²C ) until a predetermined termination condition is met, and in the final calculation, the updated multivariate normal distribution N(m, σ²C ⁾ The resistance component R, reactance component X, and voltage value Vr included in the mean vector m of C) are calculated as the estimated value ^R of the resistance component R of the system impedance of power system 4, the estimated value ^X of the reactance component X of the system impedance of power system 4, and the estimated value ^Vr of the voltage value Vr of infinite bus power system 3.

Figures 5(a) and 5(b) are schematic graphs that represent an example of period data.
Figures 6(a) to 6(d) are schematic graphs illustrating an example of the calculation results of the estimation unit.
Figure 5(a) schematically shows an example of the active power value P and reactive power value Q included in the period data.
Figure 5(b) schematically shows an example of a voltage value Vs included in the period data.
Figure 6(a) schematically shows an example of the evaluation value Jv and step size σ calculated by the estimation unit 62 based on the period data shown in Figure 5.
Figure 6(b) schematically shows an example of the estimated value ^R of the resistance component R of the power system impedance of the power system 4, calculated by the estimation unit 62 based on the period data shown in Figure 5.
Figure 6(c) schematically shows an example of the estimated value ^X of the reactance component X of the power system impedance of the power system 4, calculated by the estimation unit 62 based on the period data shown in Figure 5.
Figure 6(d) schematically shows an example of an estimated value ^Vr of the voltage value Vr of the infinite busbar power system 3, calculated by the estimation unit 62 based on the period data shown in Figure 5.

Figures 6(a) to 6(d) show an example where the process from steps S103 to S107 in Figure 4 is repeated 2000 times, based on the period data shown in Figure 5. As shown in Figures 6(a) to 6(d), in this example, repeating the process from steps S103 to S107 approximately 800 times results in the estimated values ^R, ^X, ^Vr, and the evaluation value Jv becoming nearly constant, indicating convergence. Furthermore, the step size σ, which corresponds to the distance between the search point and the optimal solution, gradually decreases with increasing calculation counts, approaching zero. The evaluation value Jv represents the evaluation value Jv used for termination determination.

Figure 5(b) shows an example of an estimated voltage value Vr of the infinite busbar power system 3, ^Vr, calculated based on period data generated for each input of active power value P, reactive power value Q, and voltage value Vs through the processing of the moving window. It also shows an example of a calculated voltage Vs value Vsh, calculated based on the same equation as in equation (2) above, using the calculated estimated values ^R, ^X, and ^Vr, along with an example of the voltage value Vs included in the period data.

As shown in Figure 5(b), the calculated value Vsh tracks the measured voltage value Vs, indicating that the covariance matrix adaptive evolutionary strategy appropriately estimates each of the estimated values ^R, ^X, and ^Vr.

As described above, in the distributed power system 2 and power converter 10 according to this embodiment, the estimation unit 62 provided in the control unit 42 of the power converter 10 calculates the estimated values ^R, ^X, and ^Vr by applying a covariance matrix adaptive evolutionary strategy to the period data.

The estimation unit 62 calculates the estimated values ^R, ^X, and ^Vr based on period data including multiple active power values P, multiple reactive power values Q, and multiple voltage values Vs for a predetermined period. This improves noise resistance compared to calculation methods that calculate the estimated values ^R, ^X, and ^Vr based on a single active power value P, reactive power value Q, and voltage value Vs, even when noise is superimposed on the measured active power value P, reactive power value Q, and voltage value Vs.

Furthermore, even when the grid characteristics change, the grid characteristics after the change can be appropriately estimated by acquiring period data after the change. In other words, it is possible to suppress delays in tracking changes in grid characteristics that exceed the length of the period data. Therefore, it is possible to provide a power converter 10 and a distributed power supply system 2 that can improve both the resistance to noise in measured values and the ability to track changes in grid characteristics. The power converter 10 and the distributed power supply system 2 can estimate grid characteristics representative of the period data with high accuracy and reliability.

In the power converter 10 according to this embodiment, for example, the system characteristics of the power system 4 can be estimated more appropriately, and voltage fluctuations at the interconnection point LP can be suppressed more appropriately. In the power converter 10 according to this embodiment, by injecting appropriate reactive power into the interconnection point LP, fluctuations in the voltage Vs at the interconnection point LP can be suppressed more appropriately, and the power factor of the output AC power can be improved, allowing for more efficient supply of active power to the power system 4. This, for example, can increase the amount of active power supplied from the distributed power system 2 to the power system 4, and is expected to increase the electricity sales revenue of the business owner of the distributed power system 2.

Furthermore, calculations using the covariance matrix adaptive evolutionary strategy eliminate the need to calculate gradients and preprocess data filtering. This allows for, for example, suppressing the increased computational load on the estimation unit 62. It also helps to prevent the estimation unit 62 from becoming overly complex due to the use of the covariance matrix adaptive evolutionary strategy.

Furthermore, in the power conversion device 10 according to this embodiment, the estimated value calculation unit 50 includes a data generation unit 60, which generates period data by processing a moving window. This allows for a more accurate estimation of the system characteristics of the power system 4 and a more appropriate suppression of voltage fluctuations at the interconnection point LP.

Figure 7 is a block diagram schematically showing a modified example of the estimation calculation unit according to the embodiment.
As shown in Figure 7, the estimated value calculation unit 50a further includes a selection unit 64. Components that are substantially the same in function and configuration as those in the above embodiment are denoted by the same reference numerals, and detailed explanations are omitted.

In this example, the estimation unit 62 performs calculations using a covariance matrix adaptive evolutionary strategy multiple times on the input period data, thereby generating multiple output data sets from the input period data, which include an estimated value ^R for the resistance component R of the power system 4's system impedance, an estimated value ^X for the reactance component X of the power system 4's system impedance, and an estimated value ^Vr for the voltage value Vr of the infinite busbar power system 3. The generated multiple output data sets are then input to the selection unit 64.

The estimation unit 62, for example, has multiple arithmetic units. It inputs one period of data into multiple arithmetic units and generates multiple output data by performing calculations using a covariance matrix adaptive evolutionary strategy in each of the multiple arithmetic units. However, the estimation unit 62 may also generate multiple output data by repeatedly performing calculations on one period of data using one arithmetic unit. Figure 7 shows an example where the estimation unit 62 generates five output data. The number of output data generated is not limited to five; it can be any number.

Furthermore, the estimation unit 62 includes the evaluation value Jv of each estimated value ^R, ^X, and ^Vr in the output data. In other words, the estimation unit 62 generates output data that includes each estimated value ^R, ^X, and ^Vr and the evaluation value Jv. The estimation unit 62 also includes, for example, the evaluation value Jv used for the termination determination in the output data. In other words, the estimation unit 62 includes, for example, the evaluation value Jv calculated based on the average values of the resistance component R, reactance component X, and voltage value Vr included in the average vector m updated in the final calculation, and the period data, in the output data.

The selection unit 64 selects the optimal single output data from the multiple input output data. Based on the evaluation value Jv contained in each output data, the selection unit 64 selects the output data with the smallest evaluation value Jv among the multiple output data as the optimal single output data. The selection unit 64 inputs the estimated values ^R, ^X, and ^Vr contained in the selected single output data to the reactive power calculation unit 52.

The inventors of this application, after diligent investigation, have found that, even when using data from the same period, the calculation results for each estimated value ^R, ^X, and ^Vr are not unique when using a covariance matrix adaptive evolutionary strategy. This is presumed to be because random numbers are used to generate the λ search points x. In other words, it is presumed that variations in the calculation results for each estimated value ^R, ^X, and ^Vr arise depending on how the random numbers for the λ search points x are generated.

Therefore, the estimation calculation unit 50a generates multiple output data from a single period of data, and selects the output data with the smallest evaluation value Jv from among them as the optimal single output data. This allows for a more accurate estimation of the power system characteristics of the power system 4 and a more appropriate suppression of voltage fluctuations at the interconnection point LP.

Figure 8 is a block diagram schematically showing a modified example of the estimated value calculation unit according to the embodiment.
As shown in Figure 8, in the estimated value calculation unit 50b, the selection unit 64 receives input of multiple output data from the estimation unit 62, as well as input of preset values Rd, Xd, and Vrd for the resistance component R of the power system impedance of the power system 4, the reactance component X of the power system impedance of the power system 4, and the voltage value Vr of the infinite bus power system 3.

The pre-set values Rd, Xd, and Vrd are set to conservative system characteristics that take safety into consideration in power system 4, for example. The pre-set values Rd, Xd, and Vrd may be the same as the initial values _R0 , _X0 , and _Vr0 used in the calculation of the covariance matrix adaptive evolutionary strategy, for example.

The pre-set values Rd, Xd, and Vrd may be, for example, the estimated values ^R, ^X, and ^Vr (previous day's values) output to the reactive power calculation unit 52 during the same time period on the previous day. For example, if the distributed power system 2 is a solar power generation system, the amount of power generated is low during the morning and evening hours, and high during the daytime hours. Consequently, the grid characteristics of the power grid 4 may also exhibit time-dependent trends. Therefore, by using the previous day's values for the pre-set values Rd, Xd, and Vrd, grid characteristics that take time-dependent trends into account can be achieved.

However, the pre-set values Rd, Xd, and Vrd are not limited to those described above. For example, the weekly or monthly average values up to the previous day for each time period may be set as the pre-set values Rd, Xd, and Vrd. The pre-set values Rd, Xd, and Vrd may be any system characteristics that take safety in the power system 4 into consideration.

The control unit 42, for example, has a memory unit (not shown), which stores the preset values Rd, Xd, and Vrd. The preset values Rd, Xd, and Vrd are input from the memory unit to the selection unit 64, for example. The memory unit may be located within the estimated value calculation unit 50b or at another location within the control unit 42. The preset values Rd, Xd, and Vrd may also be input to the selection unit 64 from an external device such as a higher-level controller, or from an operation unit provided on the power converter 10. The method of inputting the preset values Rd, Xd, and Vrd to the selection unit 64 is not limited to the above and may be any method.

The selection unit 64 determines whether each of the estimated values ^R, ^X, and ^Vr included in the multiple output data input from the estimation unit 62 falls within a predetermined range. More specifically, the selection unit 64 determines whether the estimated value ^R falls within the predetermined range for estimated value ^R, whether the estimated value ^X falls within the predetermined range for estimated value ^X, and whether the estimated value ^Vr falls within the predetermined range for estimated value ^Vr.

The selection unit 64 excludes output data from selection if it determines that any of the estimated values ^R, ^X, or ^Vr fall outside the predetermined range. If there is output data where each of the estimated values ^R, ^X, and ^Vr falls within the predetermined range, the selection unit 64 selects the output data with the smallest evaluation value Jv from among the output data where each of the estimated values ^R, ^X, and ^Vr falls within the predetermined range as the optimal single output data.

On the other hand, if any of the multiple output data have any of the estimated values ^R, ^X, or ^Vr outside the predetermined range, the selection unit 64 selects the preset values Rd, Xd, and Vrd and inputs these preset values Rd, Xd, and Vrd to the reactive power calculation unit 52 as the estimated values ^R, ^X, and ^Vr. In other words, if there is no output data in which each of the estimated values ^R, ^X, and ^Vr falls within the predetermined range, the selection unit 64 inputs the preset values Rd, Xd, and Vrd to the reactive power calculation unit 52 as the estimated values ^R, ^X, and ^Vr.

Thus, in the estimated value calculation unit 50b, the pre-set values Rd, Xd, and Vrd are input to the selection unit 64. If any of the multiple output data have an estimated value ^R, ^X, or ^Vr outside the predetermined range, the selection unit 64 inputs the pre-set values Rd, Xd, and Vrd as the estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52.

This prevents the operation of the main circuit section 40 from being controlled based on unrealistic estimated values ^R, ^X, and ^Vr, even if unrealistic estimated values are calculated due to some malfunction. Therefore, voltage fluctuations at the interconnection point LP can be suppressed more effectively.

In the example shown in Figure 8, multiple output data and the preset values Rd, Xd, and Vrd are input to the selection unit 64. However, the configuration is not limited to this; a single estimated value ^R, ^X, ^Vr calculated by the estimation unit 62 and the preset values Rd, Xd, and Vrd may also be input to the selection unit 64.

The selection unit 64 determines whether each of the input estimated values ^R, ^X, and ^Vr falls within a predetermined range. If it determines that they do, it inputs each of the input estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52. If it determines that they do not fall within the range, it may input the pre-set values Rd, Xd, and Vrd as the estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52.

Figure 9 is a block diagram schematically showing a modified example of the estimated value calculation unit according to the embodiment.
As shown in Figure 9, in the estimated value calculation unit 50c, the selection unit 64 receives input of multiple output data from the estimation unit 62, and also receives input of the previous values Rp, Xp, and Vrp of each estimated value ^R, ^X, and ^Vr that were input to the reactive power calculation unit 52 in the previous calculation.

Similar to the example shown in Figure 8, the selection unit 64 selects the previous values Rp, Xp, and Vrp if any of the multiple output data have any of the estimated values ^R, ^X, or ^Vr outside the predetermined range, and inputs these previous values Rp, Xp, and Vrp to the reactive power calculation unit 52 as the estimated values ^R, ^X, and ^Vr. In other words, the selection unit 64 inputs the previous values Rp, Xp, and Vrp to the reactive power calculation unit 52 as the estimated values ^R, ^X, and ^Vr if there are no output data where each of the estimated values ^R, ^X, and ^Vr falls within the predetermined range.

Thus, in the estimated value calculation unit 50c, the previous values Rp, Xp, and Vrp are input to the selection unit 64. If any of the multiple output data show that any of the estimated values ^R, ^X, or ^Vr fall outside a predetermined range, the selection unit 64 inputs the previous values Rp, Xp, and Vrp as the estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52.

This prevents the operation of the main circuit section 40 from being controlled based on unrealistic estimated values ^R, ^X, and ^Vr, even if unrealistic estimated values are calculated due to some malfunction, as shown in the example in Figure 8. Therefore, voltage fluctuations at the interconnection point LP can be suppressed more effectively.

Similar to the example shown in Figure 8, the estimated value calculation unit 50c may also be configured to input one estimated value ^R, ^X, ^Vr calculated by the estimation unit 62, along with the previous values Rp, Xp, Vrp, to the selection unit 64.

The selection unit 64 determines whether each of the input estimated values ^R, ^X, and ^Vr falls within a predetermined range. If it determines that they do, it inputs each of the input estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52. If it determines that they do not fall within the range, it may input the previous values Rp, Xp, and Vrp as the estimated values ^R, ^X, and ^Vr to the reactive power calculation unit 52.

Furthermore, the configurations in Figure 8 and Figure 9 may be combined, and multiple output data or one estimated value ^R, ^X, ^Vr may be input to the selection unit 64, along with the preset values Rd, Xd, Vrd and the previous values Rp, Xp, Vrp.

For example, in the first calculation performed when the control unit 42 starts operating, such as during morning operation of a solar power generation system, there may be no appropriate previous values Rp, Xp, and Vrp. Therefore, the selection unit 64 may select the pre-set values Rd, Xd, and Vrd if there is no output data where each of the estimated values ^R, ^X, and ^Vr falls within a predetermined range when no previous values Rp, Xp, and Vrp exist. Alternatively, it may select the previous values Rp, Xp, and Vrp if there is no output data where each of the estimated values ^R, ^X, and ^Vr falls within a predetermined range when previous values Rp, Xp, and Vrp exist.

Figure 10 is a block diagram schematically showing a modified example of the distributed power supply system according to the embodiment.
As shown in Figure 10, the distributed power supply system 2a further includes a control device 100. The control device 100 controls the operation of each of the multiple power converters 10 provided in the distributed power supply system 2a. The control device 100 may be called, for example, a main site controller. However, the control device 100 is not limited to controlling multiple power converters 10, but may control the operation of a single power converter 10.

The control device 100 comprises a communication unit 102 and a control unit 104. The communication unit 102 communicates with the power converter 10. For example, the communication unit 102 communicates with each of the multiple power converters 10. The communication between the communication unit 102 and the power converter 10 may be via a wired connection or wireless connection.

The control unit 104 includes an estimated value calculation unit 110. The configuration of the estimated value calculation unit 110 can be the same as any of the configurations of the estimated value calculation units 50 to 50c described in the above embodiment. The control unit 104 controls the operation of each power converter 10 so that it outputs reactive power corresponding to each estimated value ^R, ^X, and ^Vr calculated by the estimated value calculation unit 110.

The control unit 104 has, for example, a reactive power calculation unit similar to the reactive power calculation unit 52 described in the above embodiment. By inputting the reactive power command values calculated by the reactive power calculation unit to each power converter 10, the control unit controls the operation of each power converter 10 so that it outputs reactive power corresponding to the estimated values ^R, ^X, and ^Vr.

For example, the estimated values ^R, ^X, and ^Vr may be input from the control device 100 to each power converter 10, and the reactive power calculation may be performed by each power converter 10. The control method for controlling the operation of each power converter 10 to output reactive power corresponding to each estimated value ^R, ^X, and ^Vr can be any method that enables each power converter 10 to output appropriate reactive power corresponding to each estimated value ^R, ^X, and ^Vr.

In the distributed power supply system 2a, the estimated value calculation unit 50 of the control unit 42 of each power converter 10 can be omitted. Furthermore, if a reactive power calculation unit is provided in the control unit 104 of the control device 100, the reactive power calculation unit 52 of the control unit 42 of each power converter 10 can also be omitted.

Thus, the estimated value calculation unit is not limited to the control unit 42 of each power converter 10, but may also be provided in the control unit 104 of the control device 100 that controls the operation of each power converter 10. In this case as well, similar to the above embodiment, it is possible to improve the resistance of measured values to noise while also improving the ability to track changes in system characteristics.

This embodiment includes the following aspects.
(Note 1)
A power conversion device that converts the power of a distributed power source into AC power corresponding to a power system connected to an infinite busbar power system, and supplies the converted AC power to the power system, thereby connecting the distributed power source to the power system,
A main circuit unit that converts the power of the distributed power source into AC power corresponding to the power system,
A control unit that controls the operation of the main circuit section,
Equipped with,
The control unit,
An estimation calculation unit calculates, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance, and an estimated value of the voltage value of the infinite bus power system.
A reactive power calculation unit calculates a reactive power command value for the reactive power to be supplied to the power system based on the estimated value of the resistance component of the system impedance, the estimated value of the reactance component of the system impedance, and the estimated value of the voltage of the infinite bus power system.
A drive circuit drives the main circuit section to output a predetermined active power and reactive power corresponding to the reactive power command value,
It has,
The power converter has an estimation unit which receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data.

(Note 2)
The estimation unit,
The resistance component, the reactance component, and the voltage value are set as the mean vector of a multivariate normal distribution.
The process involves calculating multiple search points that follow the aforementioned multivariate normal distribution,
The evaluation value of each of the aforementioned multiple search points is calculated,
Based on the aforementioned evaluation values, extract the top few search points with the highest evaluations.
Based on the extracted top few search points, the multivariate normal distribution is updated.
The calculations from the calculation of the multiple search points to the update of the multivariate normal distribution are repeatedly performed until the predetermined termination conditions are met.
The power conversion device described in Appendix 1, which calculates the resistance component, the reactance component, and the voltage value included in the mean vector of the multivariate normal distribution updated in the final calculation as an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage value of the infinite bus power system.

(Note 3)
Each of the plurality of search points is a vector containing the values of the resistance component, the reactance component, and the voltage value,
The estimation unit,
Based on the multiple active power values, multiple reactive power values, and the respective values of the resistance component, reactance component, and voltage value included in the period data, a plurality of calculated voltage values are calculated for one of the search points, corresponding to the number of voltage values included in the period data.
The error between the voltage value included in the period data and the calculated value is calculated for each of the multiple voltage values included in the period data.
The sum of the absolute values of the errors between the voltage value and the calculated value is calculated as the evaluation value of one of the search points.
The power conversion device described in Appendix 2, which increases the evaluation of the search points with small evaluation values and extracts the top few search points with small evaluation values from among the plurality of search points.

(Note 4)
The estimated value calculation unit further includes a data generation unit that receives the active power value, the reactive power value, and the voltage value as inputs, generates the period data based on the input active power value, the reactive power value, and the voltage value, and inputs the generated period data to the estimation unit.
The power conversion device according to any one of the appendices 1 to 3, wherein the data generation unit generates the period data according to the elapsed predetermined period, and then, upon receiving the next input of the active power value, the reactive power value, and the voltage value, performs the processing of a moving window to generate the next period data by deleting the oldest data of the active power value, the reactive power value, and the voltage value, and adding the newly input data of the active power value, the reactive power value, and the voltage value.

(Note 5)
The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit performs calculations using the covariance matrix adaptive evolution strategy multiple times on the input period data, thereby generating multiple output data from the input period data, which include an estimated value of the resistance component of the power system's system impedance, an estimated value of the reactance component of the power system's system impedance, an estimated value of the voltage value of the infinite busbar power system, and evaluation values of the estimated resistance component, the estimated reactance component, and the estimated voltage value. The generated multiple output data are then input to the selection unit.
The power conversion device according to any one of the appendices 1 to 4, wherein the selection unit selects the optimal output data from a plurality of input output data based on the evaluation value, and inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage included in the selected output data to the reactive power calculation unit.

(Note 6)
The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage to the selection unit.
The power conversion device according to any one of the appendices 1 to 5, wherein the selection unit receives input of estimated values of the resistance component, estimated values of the reactance component, and estimated values of the voltage from the estimation unit, and also receives input of preset values for the resistance component of the power system impedance, the reactance component of the power system impedance, and the voltage of the infinite bus power system, and if the input estimated values of the resistance component, estimated values of the reactance component, and estimated values of the voltage fall within a predetermined range, the input estimated values of the resistance component, estimated values of the reactance component, and estimated values of the voltage are input to the reactive power calculation unit, and if they do not fall within a predetermined range, the preset values are input to the reactive power calculation unit as estimated values of the resistance component, estimated values of the reactance component, and estimated values of the voltage.

(Note 7)
The power converter described in Appendix 6, wherein the aforementioned pre-set values are the previous day's values of the estimated resistance component, the estimated reactance component, and the estimated voltage value output to the reactive power calculation unit during the same time period on the previous day.

(Note 8)
The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage to the selection unit.
The power conversion device according to any one of the appendices 1 to 7, wherein the selection unit receives input of the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage from the estimation unit, and also receives input of the previous values of the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage that were input to the reactive power calculation unit in the previous calculation, and if the input estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage fall within a predetermined range, the input estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage are input to the reactive power calculation unit, and if they do not fall within a predetermined range, the previous values are input to the reactive power calculation unit as the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage.

(Note 9)
A control device for controlling the operation of a power conversion device that connects a distributed power source to a power system by converting the power of a distributed power source into AC power corresponding to a power system connected to an infinite bus power system, and supplying the converted AC power to the power system, wherein the device connects the distributed power source to the power system,
A communication unit that communicates with the aforementioned power converter,
A control unit that controls the operation of the power converter,
Equipped with,
The control unit has an estimation calculation unit that calculates an estimated value of the resistance component of the power system impedance, an estimated value of the reactance component of the power system impedance, and an estimated value of the voltage of the infinite bus power system, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, and controls the operation of the power converter so that the power converter outputs reactive power corresponding to the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage calculated by the estimation calculation unit.
The control device includes an estimation unit which receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data.

(Note 10)
Distributed power sources,
A power conversion device that converts the power of the distributed power source into AC power corresponding to a power system connected to an infinite busbar power system, and supplies the converted AC power to the power system, thereby connecting the distributed power source to the power system;
Equipped with,
The aforementioned power converter is
A main circuit unit that converts the power of the distributed power source into AC power corresponding to the power system,
A control unit that controls the operation of the main circuit section,
It has,
The control unit,
An estimation calculation unit calculates, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance, and an estimated value of the voltage value of the infinite bus power system.
A reactive power calculation unit calculates a reactive power command value for the reactive power to be supplied to the power system based on the estimated value of the resistance component of the system impedance, the estimated value of the reactance component of the system impedance, and the estimated value of the voltage of the infinite bus power system.
A drive circuit drives the main circuit section to output a predetermined active power and reactive power corresponding to the reactive power command value,
It has,
A distributed power system having an estimation unit that receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data.

While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications are possible without departing from the spirit of the invention. These embodiments and their variations are included within the scope and spirit of the invention, as well as within the scope of the invention and its equivalents as described in the claims.

2, 2a... Distributed power system, 3... Infinite busbar power system, 4... Power system, 6... Distributed power source, 10... Power converter, 12, 14... Transformer, 16... Consumer, 18... Generator, 20, 22... Measuring device, 40... Main circuit section, 42... Control section, 50, 50a-50c... Estimated value calculation section, 52... Reactive power calculation section, 54... Drive circuit, 60... Data generation section, 62... Estimation section, 64... Selection section, 100 Control device, 102 Communication section, 104 Control section, 110... Estimated value calculation section

Claims (10)

A power conversion device that converts the power of a distributed power source into AC power corresponding to a power system connected to an infinite busbar power system, and supplies the converted AC power to the power system, thereby connecting the distributed power source to the power system,
A main circuit unit that converts the power of the distributed power source into AC power corresponding to the power system,
A control unit that controls the operation of the main circuit section,
Equipped with,
The control unit,
An estimation calculation unit calculates, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance, and an estimated value of the voltage value of the infinite bus power system.
A reactive power calculation unit calculates a reactive power command value for the reactive power to be supplied to the power system based on the estimated value of the resistance component of the system impedance, the estimated value of the reactance component of the system impedance, and the estimated value of the voltage of the infinite bus power system.
A drive circuit drives the main circuit section to output a predetermined active power and reactive power corresponding to the reactive power command value,
It has,
The power converter has an estimation unit which receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data. The estimation unit,
The resistance component, the reactance component, and the voltage value are set as the mean vector of a multivariate normal distribution.
The process involves calculating multiple search points that follow the aforementioned multivariate normal distribution,
The evaluation value of each of the aforementioned multiple search points is calculated,
Based on the aforementioned evaluation values, extract the top few search points with the highest evaluations.
Based on the extracted top few search points, the multivariate normal distribution is updated.
The calculations from the calculation of the multiple search points to the update of the multivariate normal distribution are repeatedly performed until the predetermined termination conditions are met.
The power conversion device according to claim 1, wherein the resistance component, the reactance component, and the voltage value included in the mean vector of the multivariate normal distribution updated in the final calculation are calculated as the estimated resistance component of the system impedance of the power system, the estimated reactance component of the system impedance of the power system, and the estimated voltage value of the infinite bus power system. Each of the plurality of search points is a vector containing the values of the resistance component, the reactance component, and the voltage value,
The estimation unit,
Based on the multiple active power values, multiple reactive power values, and the resistance component, reactance component, and voltage values included in the period data, for one of the search points, the calculated voltage values of the multiple interconnection points corresponding to the number of voltage values included in the period data are calculated.
The error between the voltage value included in the period data and the calculated value is calculated for each of the multiple voltage values included in the period data.
The sum of the absolute values of the errors between the voltage value and the calculated value is calculated as the evaluation value of one of the search points.
The power conversion device according to claim 2, wherein the evaluation of the search points with small evaluation values is increased, and the top few search points with small evaluation values are extracted from the plurality of search points. The estimated value calculation unit further includes a data generation unit that receives the active power value, the reactive power value, and the voltage value as inputs, generates the period data based on the input active power value, the reactive power value, and the voltage value, and inputs the generated period data to the estimation unit.
The power conversion device according to claim 1, wherein the data generation unit generates the period data in accordance with the elapsed of the predetermined period, and then, upon receiving the next input of the active power value, the reactive power value, and the voltage value, performs the processing of a moving window to generate the next period data by deleting the oldest data of the active power value, the reactive power value, and the voltage value, and adding the newly input data of the active power value, the reactive power value, and the voltage value. The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit performs calculations using the covariance matrix adaptive evolution strategy multiple times on the input period data, thereby generating multiple output data from the input period data, which include an estimated value of the resistance component of the power system's system impedance, an estimated value of the reactance component of the power system's system impedance, an estimated value of the voltage value of the infinite busbar power system, and evaluation values of the estimated resistance component, the estimated reactance component, and the estimated voltage value. The generated multiple output data are then input to the selection unit.
The power conversion device according to claim 1, wherein the selection unit selects the optimal output data from a plurality of input output data based on the evaluation value, and inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage included in the selected output data to the reactive power calculation unit. The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage to the selection unit.
The power conversion device according to claim 1, wherein the selection unit receives inputs of estimated values for the resistance component, estimated values for the reactance component, and estimated values for the voltage from the estimation unit, and also receives inputs of preset values for the resistance component of the power system impedance, the reactance component of the power system impedance, and the voltage of the infinite bus power system, and if the input estimated values for the resistance component, estimated values for the reactance component, and estimated values for the voltage fall within a predetermined range, the input estimated values for the resistance component, estimated values for the reactance component, and estimated values for the voltage are input to the reactive power calculation unit, and if they do not fall within a predetermined range, the preset values are input to the reactive power calculation unit as estimated values for the resistance component, estimated values for the reactance component, and estimated values for the voltage. The power conversion device according to claim 6, wherein the pre-set values are the previous day's values of the estimated resistance component, the estimated reactance component, and the estimated voltage value output to the reactive power calculation unit during the same time period on the previous day. The aforementioned estimated value calculation unit further includes a selection unit,
The estimation unit inputs the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage to the selection unit.
The power conversion device according to claim 1, wherein the selection unit receives input of the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage from the estimation unit, and also receives input of the previous values of the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage that were input to the reactive power calculation unit in the previous calculation, and if the input estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage fall within a predetermined range, the input estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage are input to the reactive power calculation unit, and if they do not fall within a predetermined range, the previous values are input to the reactive power calculation unit as the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage. A control device for controlling the operation of a power conversion device that connects a distributed power source to a power system by converting the power of a distributed power source into AC power corresponding to a power system connected to an infinite bus power system, and supplying the converted AC power to the power system, wherein the device connects the distributed power source to the power system,
A communication unit that communicates with the aforementioned power converter,
A control unit that controls the operation of the power converter,
Equipped with,
The control unit has an estimation calculation unit that calculates an estimated value of the resistance component of the power system impedance, an estimated value of the reactance component of the power system impedance, and an estimated value of the voltage of the infinite bus power system, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, and controls the operation of the power converter so that the power converter outputs reactive power corresponding to the estimated value of the resistance component, the estimated value of the reactance component, and the estimated value of the voltage calculated by the estimation calculation unit.
The control device includes an estimation unit which receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data. Distributed power sources,
A power conversion device that converts the power of the distributed power source into AC power corresponding to a power system connected to an infinite busbar power system, and supplies the converted AC power to the power system, thereby connecting the distributed power source to the power system;
Equipped with,
The aforementioned power converter is
A main circuit unit that converts the power of the distributed power source into AC power corresponding to the power system,
A control unit that controls the operation of the main circuit section,
It has,
The control unit,
An estimation calculation unit calculates, based on the active power value at the connection point with the power system, the reactive power value at the connection point, and the voltage value at the connection point, an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance, and an estimated value of the voltage value of the infinite bus power system.
A reactive power calculation unit calculates a reactive power command value for the reactive power to be supplied to the power system based on the estimated value of the resistance component of the system impedance, the estimated value of the reactance component of the system impedance, and the estimated value of the voltage of the infinite bus power system.
A drive circuit drives the main circuit section to output a predetermined active power and reactive power corresponding to the reactive power command value,
It has,
A distributed power system having an estimation unit that receives period data including a plurality of active power values, a plurality of reactive power values, and a plurality of voltage values for a predetermined period, and applies a covariance matrix adaptive evolutionary strategy to the input period data to calculate an estimated value of the resistance component of the system impedance of the power system, an estimated value of the reactance component of the system impedance of the power system, and an estimated value of the voltage of the infinite bus power system based on the period data.