Power inference device, power control system, system stabilization system, model generation device, and power inference method

Abstract

The purpose of the present invention is to provide technology that makes it possible to utilize an existing power storage device for normal operation control in emergency operation control. This power inference device comprises: a generation unit that generates a first inference model on the basis of system information and a charge / discharge power statistical value in main control; an inference unit that infers the charge / discharge power statistical value in main control on the basis of the first inference model and the system information; and a control reserve power inference unit that uses the inferred charge / discharge power statistical value in main control as a basis to infer control reserve power in a predetermined period, namely, charge / discharge power in a predetermined period available for use in auxiliary control.

Description

Power inference device, power control system, power system stabilization system, model generation device, and power inference method

[0001] The present disclosure relates to a power inference device, a power control system, a power grid stabilization system, a model generation device, and a power inference method.

[0002] Various technologies have been proposed for power systems. For example, Patent Literature 1 describes a technology in which the gain of Δf feedback control of a storage battery is changed and a system simulation is performed to predict the future charge / discharge amount or remaining capacity and adjust the gain so that the remaining capacity of the storage battery that performs frequency control does not approach the upper or lower limit.

[0003] Japanese Patent Application Laid-Open No. 2017-99131

[0004] There is a growing need for emergency control using fast-response storage batteries to charge and discharge during grid disturbances, etc. However, because emergency control is only performed when grid disturbances occur and the facility utilization rate is low, installing energy storage devices solely for emergency control is difficult from a cost-effectiveness perspective.

[0005] Therefore, it is conceivable to utilize existing power storage devices that are installed for normal control, which is primary control such as frequency control and load leveling control, also for emergency control, which is secondary control. However, in conventional technology, charging and discharging of a power storage device by multiple controls, such as primary control and secondary control, is not assumed, and therefore there is a problem that existing power storage devices that are installed for normal control cannot be utilized for emergency control.

[0006] Therefore, the present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a technology that allows an existing power storage device for main control to be utilized for secondary control as well.

[0007] The power inference device according to the present disclosure includes an acquisition unit that acquires the charge / discharge power during main control of a power storage device that charges and discharges power to and from a power system through a combination of main control and secondary control that is performed with a lower priority than the main control, and acquires system information that is information about the power system; a generation unit that generates a first inference model that can infer the charge / discharge power statistical value during the main control from the system information based on a charge / discharge power statistical value that is a statistical value of the charge / discharge power during the main control and the system information; an inference unit that infers the charge / discharge power statistical value during the main control based on the first inference model and the system information; and a control margin inference unit that infers the control margin for a predetermined period, which is the charge / discharge power during a predetermined period that can be used for the secondary control, based on the inferred charge / discharge power statistical value during the main control.

[0008] According to the present disclosure, a first inference model is generated based on charge / discharge power statistics and grid information during primary control, charge / discharge power statistics during primary control are inferred based on the first inference model and grid information, and a control margin for a predetermined period that can be used during secondary control is inferred based on the inferred charge / discharge power statistics during primary control. With this configuration, an existing power storage device for primary control can also be used for secondary control.

[0009] The objects, features, aspects and advantages of the present disclosure will become more apparent from the following detailed description and the accompanying drawings.

[0010] FIG. 1 is a diagram illustrating a configuration of a power system according to a first embodiment. FIG. 2 is a diagram illustrating a configuration of a power inference device according to the first embodiment. FIG. 3 is a diagram illustrating an overview of the operation of the power inference device according to the first embodiment. FIG. 4 is a flowchart illustrating a learning phase process of the power inference device according to the first embodiment. FIG. 5 is a flowchart illustrating an inference phase process of the power inference device according to the first embodiment. FIG. 6 is a diagram illustrating a configuration of a power system according to a second embodiment. FIG. 7 is a diagram illustrating a configuration of a power system stabilizing device according to the second embodiment. FIG. 8 is a flowchart illustrating a process of the power system stabilizing device according to the second embodiment. FIG. 9 is a flowchart illustrating a process of the power system stabilizing device according to the third embodiment. FIG. 10 is a diagram illustrating a configuration of a power inference device according to a fourth embodiment. FIG. 11 is a flowchart illustrating an inference phase process of the power inference device according to the fourth embodiment. FIG. 12 is a diagram illustrating a configuration of a power system according to a modified example. FIG. 13 is a block diagram illustrating a hardware configuration of a power inference device according to an example configuration of a computer system. FIG. 14 is a block diagram illustrating a hardware configuration of a power inference device according to an example configuration of a computer system.

[0011] <First Embodiment> In this specification, for example, at least one of A, B, C, ..., and Z means any one of all combinations of one or more types extracted from the group of A, B, C, ..., and Z. Furthermore, in this specification, acquisition includes acquisition as a result of input.

[0012] FIG. 1 is a diagram showing the configuration of a power inference device according to the first embodiment and other components. The configuration of FIG. 1 includes a load 1, a power source 2, an inverter power source 3, a measuring device 4, a control device 5, a power storage device 6, a power control device 7, and a power inference device 8.

[0013] The load 1, power source 2, inverter power source 3, and power storage device 6 are connected to a power grid 9. The load 1 includes, for example, a consumer device that uses power supplied from the power grid 9 or the inverter power source 3. The power source 2 includes at least one generator, such as a nuclear power generator, a thermal power generator, or a hydroelectric power generator. The inverter power source 3 includes at least one device with an inverter, such as a solar power generator or a wind power generator.

[0014] The measuring device 4 sequentially measures, as measurement information, the load amount (active power, reactive power) which is the power consumed by load equipment installed at consumers, etc., the power generated (active power, reactive power) which is the generator output and inverter power supply output, the node frequency and phase, the node voltage, the line current, and the power flowing through the line (active power, reactive power) for the load 1, the power source 2, the inverter power source 3, and the power system 9.

[0015] Instead of the load, the residual demand (also called apparent load or apparent demand) obtained by subtracting the amount of power generation from the load may be measured, or both the load and the residual demand may be measured. The load (or residual demand) is the total value of the load (or residual demand) in any area unit. Area units include, for example, the entire system under the jurisdiction of a power company (general transmission and distribution business operator), a microgrid unit, or a consumer unit, but are not limited to these.

[0016] The measurement device 4 sequentially transmits the measurement information of the measurement device 4 to the control device 5 and the power inference device 8 via the communication network 10. The measurement device 4 may or may not store the measurement information.

[0017] The control device 5 controls the power of the load 1, the power source 2, and the inverter power source 3 based on the measurement information of the measurement device 4 acquired via the communication network 10. If the control device 5 can acquire control information of another control device 5 and measurement information of the power control device 7 via the communication network 10, the control device 5 may control the power of the load 1, the power source 2, and the inverter power source 3 taking this information into consideration.

[0018] The power storage device 6 includes, for example, at least one of a storage battery, a flywheel battery, an electric double layer capacitor, and a FACTS (Flexible AC Transmission System) device, and performs charging and discharging between the power storage device 6 and the power grid 9. The charging and discharging includes at least one of charging from the power grid 9 to the power storage device 6 and discharging from the power storage device 6 to the power grid 9, and the charging and discharging power includes at least one of active power of the charging power, reactive power of the charging power, active power of the discharging power, and reactive power of the discharging power. Note that the units of charging and discharging power (W, Var) are different from the units of the amount and capacity of charging and discharging power (Wh, Varh).

[0019] The power control device 7 sequentially measures and stores the charging / discharging power of the power storage device 6 as measurement information, and sequentially transmits the measurement information of the power control device 7 to the power inference device 8 via the communication network 10. The power control device 7 controls the charging / discharging performed by the power storage device 6 between the power grid 9 based on the measurement information of the power control device 7. Note that, if the power control device 7 can acquire the measurement information of the measurement device 4 and the control information of the control device 5 via the communication network 10, it may also take this information into consideration when controlling the charging / discharging performed by the power storage device 6 between the power grid 9.

[0020] The power control device 7 controls the charging and discharging of the power storage device 6 by issuing commands for charging and discharging to and from the power grid 9 through a combination of a primary control and a secondary control that is executed with a lower priority than the primary control.

[0021] In the following description, the primary control is normal control that continuously charges and discharges the power storage device 6, and the secondary control is emergency control that charges and discharges the power storage device 6 for only a certain period of time. However, the primary control and secondary control are not limited to normal control and emergency control, respectively. The normal control may include frequency control that charges and discharges the power storage device 6 so that the grid frequency reaches a reference value (e.g., 50 or 60 Hz), and load leveling control that charges the power storage device 6 during time periods when the load is low and discharges the power storage device 6 during time periods when the load is high. The load amount is the total value of the load amounts in any of the above-mentioned area units. The emergency control may include stabilization control that charges and discharges the power storage device 6 to improve transient stability and frequency stability when a grid disturbance occurs due to a ground fault, a short-circuit fault, a power supply drop, a load drop, or the like.

[0022] The power inference device 8 infers the control margin for a predetermined period, which is the charge / discharge power of the power storage device 6 that can be used for emergency control, based on measurement information from the measurement device 4 and the power control device 7, and charge / discharge power statistics that are statistics of the charge / discharge power of the power storage device 6 required for normal control obtained from time series analysis and statistical processing of the measurement information. The configuration of the power inference device 8 will be described in detail below. The power control device 7 and the power inference device 8 are provided in a power control system 450.

[0023] Fig. 2 is a block diagram showing the configuration of the power inference device 8 according to the present embodiment 1. The power inference device 8 in Fig. 2 includes an acquisition unit 21, a recording unit 22, a generation unit 23 that generates an inference model, a charge / discharge power statistical value inference unit 24 that is an inference unit, a control margin inference unit 25, and an output unit 26.

[0024] <Acquisition unit 21> The acquisition unit 21 acquires measurement information from the measurement device 4 and the power control device 7 at multiple points in time, and thereby acquires the charge / discharge power under normal control of the power storage device 6 and system information, which is information on the power system 9. The charge / discharge power under normal control of the power storage device 6 is included in the measurement information of the power control device 7, for example. Hereinafter, the charge / discharge power under normal control of the power storage device 6 may be abbreviated as "normal charge / discharge power."

[0025] The system information includes at least one of facility information, measurement information, and information on systems used for power system operation. The systems used for power system operation include, for example, at least one of an energy management system (EMS), a supervisory control and data acquisition system (SCADA), and a distribution automation system (DAS). The facility information includes, for example, at least one of a generator rated capacity (VA), an inverter power supply rated output (W), a power storage device rated capacity (Wh), and a power storage device rated output (W).

[0026] The rated capacity corresponds to the amount of power, the output corresponds to the power, and the rated output of the power storage device includes, for example, the rated output of the storage battery and the rated output of the PCS. The measurement information is, for example, information included in the measurement information of the measurement device 4 and the power control device 7, and includes at least one of the following: generator output, inverter power output (W, Var), node frequency (Hz) and phase, node voltage (V), line current (A), power flowing through the line (W, Var), remaining capacity of the power storage device (Wh or %), load (or residual demand) (W, Var), and power supply information obtainable via an information transmission module (ITM) or cyclic digital data transmission (CDT) of the power station. The power station includes, for example, a substation, a power plant, and a switchyard.

[0027] System information used for power system operation includes, for example, at least one of the system frequency (Hz), local demand (W), generator EDC command value (W), and generator LFC command value (W). The local demand represents the load change amount within the power system 9 and is expressed, for example, as "system constant × system capacity × frequency change amount + interconnection line power flow change amount." The system constant is the change amount of generator output with respect to frequency change, the system capacity is the sum of the rated capacities of the power sources in operation, and the interconnection line power flow change amount is the change amount of power flowing through the transmission lines connecting the inside and outside of the jurisdiction area. Economic Load Dispatching (EDC; Economic Dispatch Control, ELD) is a control that determines the output allocation of each generator taking into account economic efficiency with respect to the required power generation amount, and the generator EDC command value is the output command value for the generator determined by EDC. Load frequency control (LFC; Automatic Frequency Control; AFC) is a control that increases or decreases generator output using local demand as an index in order to balance supply and demand in the power grid, and the generator LFC command value is an output command value for the generator determined by LFC.

[0028] In addition, the system information may include information other than the above as long as it is information about the power system 9, or may include information that can be derived from the above information (for example, residual demand obtained by subtracting inverter power supply output from load amount).

[0029] The acquisition unit 21 stores the acquired normal charging / discharging power and grid information in the recording unit 22 .

[0030] <Generation unit 23> The generation unit 23 generates an inference model, which is a first inference model that can infer normal-time charging / discharging power statistical values ​​from system information, by modeling the relationship between normal-time charging / discharging power statistical values ​​and system information through learning, based on normal-time charging / discharging power statistical values ​​calculated from normal-time charging / discharging power and system information.

[0031] For example, the generation unit 23 calculates normal-state charge / discharge power statistics values ​​used to generate an inference model by performing time series analysis and statistical processing on instantaneous values ​​of normal-state charge / discharge power at multiple points in time that were previously acquired by the acquisition unit 21 and stored in the recording unit 22. The normal-state charge / discharge power statistics values ​​according to the first embodiment include upper-side power, which is a charge / discharge power statistics value on the discharge side under normal control, and lower-side power, which is a charge / discharge power statistics value on the charge side under normal control.

[0032] The upper power may be, for example, the maximum value of the normal charging / discharging power during a predetermined period, the upper 2.5% percentile value of the normal charging / discharging power during a predetermined period, or an average value equal to or greater than the threshold value of the normal charging / discharging power during a predetermined period. The lower power may be, for example, the minimum value of the normal charging / discharging power during a predetermined period, the lower 2.5% percentile value of the normal charging / discharging power during a predetermined period, or an average value less than the threshold value of the normal charging / discharging power during a predetermined period. Alternatively, the normal charging / discharging power statistical value may be the difference between the upper power and the lower power. Note that the above-mentioned percentile value is not limited to 2.5% and may be other values. Furthermore, the above-mentioned percentile value and the threshold value of the normal charging / discharging power during a predetermined period may be determined by the generation unit 23 performing a past probability statistical analysis or may be set by the user. The interval (period) for calculating the normal state charge / discharge power statistics may be fixed in accordance with the calculation cycle of a grid stabilization device or the like, or may be changed depending on the application.

[0033] When the normal state charging / discharging power statistics include the maximum and minimum values ​​of normal state charging / discharging power in a predetermined period, it is possible to ensure a sufficient amount of charging / discharging power required for normal state control.When the normal state charging / discharging power statistics include percentile values ​​or average values ​​of normal state charging / discharging power in a predetermined period, it is possible to ensure a looser amount of charging / discharging power required for normal state control.

[0034] The system information used to generate the inference model is system information at multiple points in time that was previously acquired by the acquisition unit 21 and stored in the recording unit 22. The system information may include, for example, a local demand, a frequency deviation, a power flow in a grid connection, a generator output, a generator EDC command value, and a generator LFC command value, or may include other information. The following describes a case where the inference model generated by the generation unit 23 includes an upper-side power inference model that infers upper-side power and a lower-side power inference model that infers lower-side power. Note that the inference model generated by the generation unit 23 is not limited to this, and may also be, for example, a model that infers a statistical value range, such as the difference between a maximum value and a minimum value over a predetermined period.

[0035] The generation unit 23 may generate the inference model using, for example, a statistical method that generates a statistical model, or may generate the inference model using machine learning (training) that generates a trained model. The machine learning may be, for example, supervised machine learning, unsupervised machine learning, or reinforcement learning. The generation unit 23 may also use a statistical method, an optimization method, or the like. The supervised machine learning may be, for example, a decision tree, a regression tree, a random forest, a random forest regression, a support vector machine, a support vector regression, a neural network, deep learning, or the like, or other methods. The machine learning may also be used as a classification method or a regression method.

[0036] To generate the inference model, the generation unit 23 may use all of the system information acquired by the acquisition unit 21, may use a portion of the system information selected from all of the system information, or may use data obtained by processing all or a portion of the system information. Methods for selecting a portion of the system information from all of the system information include, but are not limited to, a selection method based on a method for determining a normal control command value of the power storage device 6, a selection method based on qualitative characteristics of the inference accuracy of the inference model, and a selection method specified by the user.

[0037] The inference of the normal-state charging / discharging power statistical value may include predicting a future normal-state charging / discharging power statistical value, or may include estimating a past or present normal-state charging / discharging power statistical value. When generating an inference model for predicting a future normal-state charging / discharging power statistical value from current grid information, the generation unit 23 generates the inference model using normal-state charging / discharging power at a point in time in the future than the point in time when the grid information is acquired.

[0038] The generation unit 23 stores the generated inference model in the recording unit 22.

[0039] <Charge / Discharge Power Statistical Value Inference Unit 24> The charge / discharge power statistical value inference unit 24 infers normal-state charge / discharge power statistical values ​​based on the inference model generated by the generation unit 23 and stored in the recording unit 22 and on system information. For example, the charge / discharge power statistical value inference unit 24 infers normal-state charge / discharge power statistical values ​​by inputting the same type of system information as the system information used by the generation unit 23 into the inference model and acquiring normal-state charge / discharge power statistical values ​​output from the inference model. Note that the system information used to infer the normal-state charge / discharge power statistical values ​​may be system information stored in the recording unit 22 or may be system information acquired by the acquisition unit 21 at the time of inference.

[0040] The charge / discharge power statistical value inference unit 24 stores the inferred normal state charge / discharge power statistical value in the recording unit 22 .

[0041] <Control margin inference unit 25> The control margin inference unit 25 infers a control margin for a predetermined period, which is the charge / discharge power of the power storage device 6 that can be used for emergency control during the predetermined period, based on the normal-time charge / discharge power statistical value inferred by the charge / discharge power statistical value inference unit 24 and stored in the recording unit 22, and the power storage device rated output. The power storage device rated output used to infer the control margin for the predetermined period may be the power storage device rated output included in the system information acquired by the acquisition unit 21, or may be a default power storage device rated output that is set in advance in the control margin inference unit 25.

[0042] When a default rated output of the power storage device set in advance in the control margin inference unit 25 is used to infer the control margin, the control margin inference unit 25 infers the control margin for a predetermined period substantially based on the normal charging / discharging power statistical values ​​inferred by the charging / discharging power statistical value inference unit 24. When the inferred normal charging / discharging power statistical values ​​include an upper power inference value that is an inferred value of the upper power, the upper power inference value is used to infer the discharge-side control margin for a predetermined period. When the inferred normal charging / discharging power statistical values ​​include a lower power inference value that is an inferred value of the lower power, the lower power inference value is used to infer the charge-side control margin for a predetermined period.

[0043] The control margin inference unit 25 may infer the control margin for a predetermined period by subtracting the normal charge / discharge power statistical value inferred by the charge / discharge power statistical value inference unit 24 from the rated output of the power storage device and taking the absolute value of the value obtained thereby. For example, assume that the rated output of the power storage device is −300 kW to 300 kW, positive and negative values ​​represent discharge and charge, respectively, and the upper power inference value such as the maximum value and the lower power inference value such as the minimum value of normal charge / discharge power for the predetermined period are 100 kW and −90 kW, respectively. In this case, the control margin inference unit 25 infers 200 kW as the discharge-side control margin for the predetermined period and 210 kW as the charge-side control margin for the predetermined period.

[0044] Note that the estimation of the control margin for a predetermined period is not limited to the above. For example, in order to strictly evaluate the control margin, the control margin estimation unit 25 may estimate the control margin for a predetermined period by adding a margin, which is a fixed amount specified by a user, to the upper estimated power value and subtracting it from the lower estimated power value. Conversely, in order to loosely evaluate the control margin, the control margin estimation unit 25 may estimate the control margin for a predetermined period by subtracting a margin from the upper estimated power value and adding it to the lower estimated power value. Furthermore, the estimation of the control margin for a predetermined period may include predicting the control margin for a future predetermined period, or may include estimating the control margin for a past or present predetermined period.

[0045] The control margin inference unit 25 stores the inferred control margin for a predetermined period in the recording unit 22 .

[0046] <Output unit 26> The output unit 26 transmits the control margin for a predetermined period inferred by the control margin inference unit 25 to the power control device 7 via the communication network 10, and the power control device 7 controls the charging and discharging of the power storage device 6 taking into account the transmitted control margin for the predetermined period.

[0047] The output unit 26 is not limited to transmitting the control margin for a predetermined period to the power control device 7, but may display the control margin for a predetermined period on a display device (not shown) or transmit the control margin for a predetermined period to other devices and systems via the communication network 10. Furthermore, the output unit 26 may display or transmit not only the control margin for a predetermined period, but also an inference model, normal charging / discharging power statistics, and measurement information used for the inference.

[0048] <Operation> Fig. 3 is a diagram showing an outline of the operation of the power inference device 8 according to the present embodiment 1. As shown in Fig. 3, the power inference device 8 performs a process in a learning phase and a process in an inference phase.

[0049] 4 is a flowchart showing the learning phase processing of the power inference device 8. The learning phase processing is performed before the inference phase processing, but does not have to be performed for each inference phase processing. The learning phase processing may be performed when the power inference device 8 is installed, or may be performed periodically, such as once every few days, and the periodic period is not limited.

[0050] In step S1, the acquisition unit 21 acquires grid information at multiple past points in time and normal-state charging / discharging power at multiple past points in time, and stores the information in the recording unit 22. In step S2, the generation unit 23 acquires normal-state charging / discharging power and grid information at multiple points in time required for learning from the recording unit 22. In step S3, the generation unit 23 calculates normal-state charging / discharging power statistics by performing time series analysis and statistical processing on the acquired normal-state charging / discharging power. In step S4, the generation unit 23 generates an inference model based on the acquired grid information and the calculated normal-state charging / discharging power statistics. In step S5, the generation unit 23 stores the generated inference model in the recording unit 22. Thereafter, the processing of FIG. 4 ends.

[0051] FIG. 5 is a flowchart showing the processing of the inference phase of the power inference device 8. In step S11, the charge / discharge power statistical value inference unit 24 acquires system information used to infer the normal state charge / discharge power statistical value. In step S12, the charge / discharge power statistical value inference unit 24 inputs the acquired system information into an inference model and acquires the normal state charge / discharge power statistical value output from the inference model, thereby inferring the normal state charge / discharge power statistical value. In step S13, the control margin inference unit 25 infers the control margin of the power storage device 6 for a predetermined period based on the normal state charge / discharge power statistical value inferred by the charge / discharge power statistical value inference unit 24 and the rated output of the power storage device. In step S14, the output unit 26 outputs the inferred control margin for the predetermined period. Thereafter, the processing of FIG. 5 ends.

[0052] Summary of First Embodiment According to the power estimation device 8 of the first embodiment described above, an inference model is generated based on charge / discharge power statistics, which are statistics of charge / discharge power under normal control, and grid information acquired as measurement information, the charge / discharge power statistics under normal control are inferred based on the inference model and the grid information, and a control margin for a predetermined period that can be used for emergency control is inferred based on the inferred charge / discharge power statistics under normal control. This allows the power storage device 6 to be charged / discharged by multiple controls, such as normal control and emergency control, and therefore allows an existing power storage device for normal control to be used for emergency control.

[0053] Generally, when the charge / discharge power during normal operation contains highly random short-period components, it is difficult to properly estimate the instantaneous value of the charge / discharge power during normal operation. However, because the charge / discharge power statistics during normal operation have a high correlation with the load amount contained in the grid information, it is possible to properly estimate the charge / discharge power statistics during normal operation even when the charge / discharge power during normal operation contains highly random short-period components.

[0054] Furthermore, when the generation unit 23 generates an inference model using machine learning, it can appropriately generate the inference model and thereby infer the control margin for a predetermined period of time.

[0055] 6 is a diagram showing the configuration of a power inference device and other components according to the second embodiment. In the following, among the components according to the second embodiment, components that are the same as or similar to the components described above are given the same or similar reference numerals, and different components will be mainly described.

[0056] The power inference device and other components in Figure 6 are similar to the power inference device and other components in Figure 1 with a power system stabilization device 13 added, and the power inference device 8 according to this embodiment 2 is similar to the power inference device 8 according to embodiment 1.

[0057] The power grid stabilization device 13 is capable of communicating with the measurement device 4, the control device 5, the power control device 7, and the power inference device 8 via the communication network 10. As will be described below, the power grid stabilization device 13 determines at least one of a power limiting policy (power limit targets and power limit amount) for generators connected (in parallel, interconnected) to the power grid 9, and a load limiting policy (load limit targets and load limit amount) for consumers, etc. connected to the power grid 9, based on power grid information and control margin for a predetermined period. The power limit amount is a limit amount of output for generators connected to the power grid 9, and the load limit amount is a limit amount of power consumption for consumers, etc. connected to the power grid 9. The power inference device 8 and the power grid stabilization device 13 are provided in a power grid stabilization system.

[0058] Power supply limitation, also called power control, is a control to suppress generator output or stop generators for the purpose of stabilizing the power grid 9, and affects power costs. Load limitation, also called load control, is a control to suppress the load of consumers or cut off loads to suppress power consumption for the purpose of stabilizing the power grid 9, and affects power quality.

[0059] Fig. 7 is a block diagram showing the configuration of a power system stabilizing device 13 according to Embodiment 2. The power system stabilizing device 13 in Fig. 7 includes an acquisition unit 41, a recording unit 42, a power system stabilization calculation unit 43, and an output unit 44.

[0060] <Acquisition unit 41> The acquisition unit 41 acquires information such as measurement information from the measurement device 4 and the power control device 7 at multiple points in time, thereby acquiring system configuration information, equipment information, measurement information, control conditions, and control margin for a predetermined period inferred by the power inference device 8.

[0061] The system configuration information includes, for example, at least one of the system topology and the on / off state of the switches of the power system 9. The facility information includes, for example, at least one of the rated capacity of the generator, the rated output of the inverter power supply, the rated capacity of the power storage device, the rated output of the power storage device, the capacity of the transformer, and information on the transmission line (or distribution line). The information on the transmission line (or distribution line) includes, for example, at least one of the line type, length, impedance, and admittance of the transmission line (or distribution line). The line type of the transmission line (or distribution line) includes, for example, at least one of the thickness and material of the transmission line (or distribution line).

[0062] The measurement information is information included in the measurement information of the measurement device 4 and the power control device 7, for example, and includes at least one of the following: load amount (active power, reactive power) which is the power consumed by load devices installed in consumers, etc.; power generation amount (active power, reactive power) which is the generator output and inverter power supply output; node frequency and phase; node voltage; line current; and power flowing through the line (active power, reactive power). The control conditions include, for example, a control target indicating whether the generator and load connected to the power grid 9 are subject to control / restriction, and equipment information such as a controllable amount when subject to control / restriction. This information may be input by a user to the power grid stabilization device 13 and acquired by the acquisition unit 41, or may be acquired from the equipment to be controlled. Furthermore, this information may be acquired from systems used for power grid operation, such as an energy management system (EMS), a supervisory control and data acquisition system (SCADA), and a distribution automation system (DAS).

[0063] The information acquired by the acquisition unit 41 is not limited to the above, as long as it includes measurement information and control margin for a predetermined period. The acquisition unit 41 stores the acquired information in the recording unit 42.

[0064] <Grid stabilization calculation unit 43> The grid stabilization calculation unit 43 determines a power limiting policy and a load limiting policy for the power grid 9 based on information including measurement information and control margin for a predetermined period. The power limiting policy for the power grid 9 is a limiting policy for generators connected (in parallel, interconnected) to the power grid 9, and includes, for example, power limiting targets and power limit amounts. The load limiting policy for the power grid 9 is a limiting policy for consumers, etc. connected to the power grid 9, and includes, for example, load limiting targets and load limit amounts. The information used to determine the power limiting policy and the load limiting policy may be information stored in the recording unit 42, or may be information acquired by the acquisition unit 41 when the power limiting policy and the load limiting policy are determined.

[0065] As an example of determining a power limiting policy and a load limiting policy, first, a model is defined in which a model simulating a power storage system including the power storage device 6 and the power control system 450 is added to a simulation system model used for detailed stability calculation. The system stabilization calculation unit 43 determines a power limiting policy and a load limiting policy by repeatedly performing a simulation (detailed stability calculation) by inputting measurement information and control margin for a predetermined period into the defined model. Note that the method for determining a power limiting policy and a load limiting policy is not limited to this, and may be a method of determining the policy by optimization calculation such as combinatorial optimization, or a method of determining the policy by simulating all candidates.

[0066] The system stabilization calculation unit 43 stores the system simulation model and the determined results of the power limiting policy and the load limiting policy in the recording unit 42 .

[0067] <Output unit 44> The output unit 44 transmits the results of the power limiting policy (power limit targets and power limit amount) and the load limiting policy (load limit targets and load limit amount) determined by the system stabilization calculation unit 43 to the equipment to be controlled via the communication network 10. The equipment to be controlled is connected to the power system 9, and performs power limiting and load limiting as appropriate based on the transmitted determination results.

[0068] <Operation> Fig. 8 is a flowchart showing the processing of the power grid stabilization device 13 according to the second embodiment. In step S21, the acquisition unit 41 acquires information including measurement information and a control margin for a predetermined period, and stores the information in the recording unit 42. In step S22, the power grid stabilization calculation unit 43 determines a power limiting policy (power limit targets and power limit amount) and a load limiting policy (load limit targets and load limit amount) based on the information including the measurement information and the control margin for a predetermined period, and stores the information in the recording unit 42. In step S23, the output unit 44 transmits the power limiting policy and the load limiting policy determination results determined by the power grid stabilization calculation unit 43 to the equipment to be controlled via the communication network 10. Thereafter, the processing in Fig. 8 ends.

[0069] Summary of Second Embodiment According to the system grid stabilization device 13 of the second embodiment as described above, a power supply limiting policy and load limiting for the power grid 9 are determined based on system information and the control margin for a predetermined period. This makes it possible to perform appropriate power supply limiting and load limiting for the power grid 9 when the power storage device 6 uses the control margin for a predetermined period in emergency control. In the above description of the second embodiment, the system grid stabilization calculation unit 43 determines both the power supply limiting policy and the load limiting policy for the power grid 9, but it may also determine at least one of the power supply limiting policy and the load limiting policy.

[0070] <Embodiment 3> The power inference device according to this embodiment 3 and other configurations are the same as the power inference device according to embodiment 2 and other configurations (i.e., the configuration in FIG. 6 ). Of the components according to this embodiment 3, components that are the same as or similar to the components described above are given the same or similar reference numerals, and different components will be mainly described.

[0071] The power inference device 8 according to the third embodiment is similar to the power inference device 8 according to the first embodiment. The power system stabilization device 13 according to the third embodiment is similar to the power system stabilization device 13 according to the second embodiment, except that the power system stabilization device 13 according to the third embodiment determines a control amount for stabilizing the power system of the power storage device 6 based on information including measurement information and control margin for a predetermined period. Below, components of the power system stabilization device 13 according to the third embodiment that are different from those of the power system stabilization device 13 according to the second embodiment, specifically, the power system stabilization calculation unit 43 and the output unit 44 in FIG. 7 will be described.

[0072] The grid stabilization calculation unit 43 determines a power limiting policy (power limit targets and power limit amount) and a load limiting policy (load limit targets and load limit amount) for the power grid 9 based on information including the measurement information and the control margin for a predetermined period, and also determines the control amount of the power storage device 6. The control amount of the power storage device 6 is a control amount for controlling the power storage device 6 within the range of the control margin for a predetermined period, and is determined so that power limiting and load limiting are performed appropriately. Note that the grid stabilization calculation unit 43 may determine the control amount of the power storage device 6 by performing a simulation similar to that used when determining the power limiting policy and the load limiting policy. The grid stabilization calculation unit 43 stores the grid simulation model, the power limiting policy, the load limiting policy, and the determined control amount of the power storage device 6 in the recording unit 42.

[0073] The output unit 44 transmits the power supply limiting policy and the load limiting policy determined by the system stabilization calculation unit 43 to the equipment to be controlled via the communication network 10. The output unit 44 also transmits the control amount of the power storage device 6 determined by the system stabilization calculation unit 43 to the power control device 7 via the communication network 10. The power control device 7 controls the charging and discharging of the power storage device 6 in consideration of the transmitted control amount.

[0074] <Operation> Fig. 9 is a flowchart showing the processing of the grid stabilization device 13 according to the third embodiment. In step S31, the acquisition unit 41 acquires information including measurement information and a control margin for a predetermined period, and stores the information in the recording unit 42. In step S32, the grid stabilization calculation unit 43 determines a power limiting policy (power limit targets and power limit amount) and a load limiting policy (load limit targets and load limit amount) based on the information including the measurement information and the control margin for the predetermined period, and also determines the control amount of the power storage device 6, and stores the control amount in the recording unit 42. In step S33, the output unit 44 transmits the power limiting policy and the load limiting policy determination results determined by the grid stabilization calculation unit 43 to the equipment to be controlled via the communication network 10. In addition, the output unit 44 transmits the control amount of the power storage device 6 determined by the grid stabilization calculation unit 43 to the power control device 7 via the communication network 10. The processing of Fig. 9 then ends.

[0075] Summary of Third Embodiment According to the grid stabilization device 13 according to the third embodiment as described above, the control amount of the power storage device 6 is determined based on the grid information and the control margin for a predetermined period. As a result, when the power storage device 6 uses the control margin for a predetermined period in emergency control, it is possible to appropriately charge and discharge the power storage device 6 while performing appropriate power supply and load limitations on the power grid 9. Note that in the above description of the third embodiment, the grid stabilization calculation unit 43 determines both the power supply limiting policy and the load limiting policy for the power grid 9, but it may also determine at least one of the power supply limiting policy and the load limiting policy.

[0076] <Fourth Embodiment> The configuration of a power inference device according to a fourth embodiment is the same as that of the power inference device according to the first embodiment (i.e., the configuration of FIG. 1). Fig. 10 is a block diagram showing the configuration of a power inference device 8 according to the fourth embodiment. Hereinafter, among the components of the power inference device according to the fourth embodiment, components that are the same as or similar to the components described above will be given the same or similar reference numerals, and different components will be mainly described.

[0077] The configuration of the power inference device 8 according to the fourth embodiment shown in Fig. 10 is the same as that of the power inference device 8 according to the first embodiment shown in Fig. 2, except that the charge / discharge power statistical value inference unit 24 is replaced with a charge / discharge power statistical value / instantaneous value inference unit 24a, which is included in the concept of the inference unit, that is, the charge / discharge power statistical value inference unit 24. The following mainly describes, among the components of the power inference device 8 according to the fourth embodiment, the components that are different from those of the power inference device 8 according to the first embodiment, specifically, the generation unit 23, the charge / discharge power statistical value / instantaneous value inference unit 24a, and the control margin inference unit 25.

[0078] The generation unit 23 generates a statistical value inference model and an instantaneous value inference model based on the normal-state charging / discharging power and the grid information. The statistical value inference model corresponds to the inference model according to the first embodiment. That is, the statistical value inference model is a first inference model that models, through learning, the relationship between the normal-state charging / discharging power statistical value calculated from the normal-state charging / discharging power and the grid information, and is capable of inferring the normal-state charging / discharging power statistical value from the grid information. The instantaneous value inference model is a second inference model that models, through learning, the relationship between the normal-state charging / discharging power instantaneous value, which is the instantaneous value of the normal-state charging / discharging power, and the grid information, and is capable of inferring the normal-state charging / discharging power instantaneous value from the grid information.

[0079] The normal state charge / discharge power statistical values ​​used to generate the statistical value inference model are, for example, statistical values ​​calculated by applying time series analysis and statistical processing in the generation unit 23 to the normal state charge / discharge power instantaneous values ​​at multiple points in time that were previously acquired by the acquisition unit 21 and stored in the recording unit 22. The normal state charge / discharge power instantaneous values ​​used to generate the instantaneous value inference model are, for example, the normal state charge / discharge power instantaneous values ​​at multiple points in time that were previously acquired by the acquisition unit 21 and stored in the recording unit 22.

[0080] The generating unit 23 stores the generated statistical value inference model and instantaneous value inference model in the recording unit 22 .

[0081] The charge / discharge power statistical value / instantaneous value inference unit 24a infers normal-time charge / discharge power statistical values ​​based on the statistical value inference model generated by the generation unit 23 and stored in the recording unit 22, and on system information. The charge / discharge power statistical value / instantaneous value inference unit 24a also infers normal-time charge / discharge power instantaneous values ​​based on the instantaneous value inference model generated by the generation unit 23 and stored in the recording unit 22, and on system information. The charge / discharge power statistical value / instantaneous value inference unit 24a stores the inferred normal-time charge / discharge power statistical values ​​and normal-time charge / discharge power instantaneous values ​​in the recording unit 22.

[0082] The control margin estimation unit 25 calculates the normal state charge / discharge power estimation value based on the normal state charge / discharge power statistical value and the normal state charge / discharge power instantaneous value estimated by the charge / discharge power statistical value / instantaneous value estimation unit 24a and stored in the recording unit 22.

[0083] The estimated normal charging / discharging power value may be calculated by using the normal charging / discharging power statistical value as a confidence interval for the instantaneous normal charging / discharging power value. For example, if the upper estimated power value of the normal charging / discharging power statistical value is an estimated maximum value and the lower estimated power value is an estimated minimum value, the upper estimated power value and the lower estimated power value can be used as a confidence interval within which the true value of the instantaneous normal charging / discharging power value lies. For example, if the upper estimated power value is an estimated value of the upper 2.5% percentile of the normal charging / discharging power and the lower estimated power value is an estimated value of the lower 2.5% percentile of the normal charging / discharging power, the upper estimated power value and the lower estimated power value can be used as a confidence interval within which the true value of the instantaneous normal charging / discharging power value lies with a 95% probability.

[0084] The control margin estimation unit 25 may use a confidence interval based on the normal-state charging / discharging power statistical value as reference information to calculate an estimated normal-state charging / discharging power value according to the user's intention based on the confidence interval and the normal-state charging / discharging power instantaneous value. For example, if the estimated normal-state charging / discharging power instantaneous value is within the confidence interval, the control margin estimation unit 25 may set the estimated normal-state charging / discharging power instantaneous value as the true value as the normal-state charging / discharging power estimated value, and if the estimated normal-state charging / discharging power instantaneous value is not within the confidence interval, the control margin estimation unit 25 may set a value within the confidence interval that is closest to the estimated normal-state charging / discharging power instantaneous value as the normal-state charging / discharging power estimated value. Furthermore, the control margin estimation unit 25 may use the estimated normal-state charging / discharging power statistical value and the normal-state charging / discharging power instantaneous value to calculate the normal-state charging / discharging power estimated value using a method other than the above. For example, if the inferred instantaneous value of normal charging / discharging power is not within the range of the confidence interval, the control margin inference unit 25 may use the inferred statistical value of normal charging / discharging power as the inferred value of normal charging / discharging power instead of the inferred instantaneous value of normal charging / discharging power.

[0085] The control reserve inference unit 25 infers the control reserve instantaneous value, which is the instantaneous value of the charging and discharging power that can be used for emergency control of the power storage device 6, by subtracting the normal charging and discharging power inference value calculated as described above from the rated output of the power storage device.

[0086] Note that the inference of the instantaneous value of control reserve capacity is not limited to the above. For example, in order to strictly evaluate the instantaneous value of control reserve capacity, the control reserve capacity inference unit 25 may infer the instantaneous value of control reserve capacity using a value obtained by adding a margin, which is a fixed amount specified by the user, to the inferred value of normal charging / discharging power. Conversely, in order to loosely evaluate the instantaneous value of control reserve capacity, the control reserve capacity inference unit 25 may infer the instantaneous value of control reserve capacity using a value obtained by subtracting a margin from the inferred value of normal charging / discharging power. Furthermore, inference of the instantaneous value of control reserve capacity may include predicting the instantaneous value of control reserve capacity in the future, or may include estimating the instantaneous value of control reserve capacity in the past or present.

[0087] The control margin inference unit 25 stores the inferred control margin instantaneous value in the recording unit 22 .

[0088] <Operation> The power inference device 8 performs a learning phase process and an inference phase process. The learning phase process according to the fourth embodiment is the same as the flowchart showing the learning phase process according to the first embodiment in FIG.

[0089] In step S1, the acquisition unit 21 acquires grid information at multiple past points in time and normal-state charging / discharging power at multiple past points in time, and stores the acquired grid information in the recording unit 22. In step S2, the generation unit 23 acquires, from the recording unit 22, the normal-state charging / discharging power and grid information at multiple points in time required for learning. In step S3, the generation unit 23 performs time series analysis and statistical processing on the acquired normal-state charging / discharging power to calculate normal-state charging / discharging power statistics. In step S4, the generation unit 23 generates a statistical value inference model based on the acquired grid information and normal-state charging / discharging power statistics, and generates an instantaneous value inference model based on the acquired grid information and normal-state charging / discharging power instantaneous values. In step S5, the generation unit 23 stores the generated statistical value inference model and instantaneous value inference model in the recording unit 22. The processing of FIG. 4 then ends.

[0090] 11 is a flowchart showing the processing of the inference phase of the power inference device 8 according to the fourth embodiment. In step S11, the charge / discharge power statistical value / instantaneous value inference unit 24a acquires system information used to infer the normal-state charge / discharge power statistical value and the normal-state charge / discharge power instantaneous value. In step S12a, the charge / discharge power statistical value / instantaneous value inference unit 24a inputs the acquired system information into a statistical value inference model and an instantaneous value inference model, and infers the normal-state charge / discharge power statistical value and the normal-state charge / discharge power instantaneous value by acquiring the normal-state charge / discharge power statistical value and the normal-state charge / discharge power instantaneous value output from the respective models. In step S13a, the control margin inference unit 25 infers the control margin instantaneous value of the power storage device 6 based on the normal-state charge / discharge power statistical value and the normal-state charge / discharge power instantaneous value inferred by the charge / discharge power statistical value / instantaneous value inference unit 24a and the rated output of the power storage device. In step S14a, the output unit 26 outputs the estimated control margin instantaneous value, and then the process of FIG.

[0091] Summary of Fourth Embodiment According to the power estimation device 8 of the fourth embodiment described above, a statistical value inference model and an instantaneous value inference model are generated based on the charge / discharge power statistical values ​​under normal control, the charge / discharge power instantaneous values ​​under normal control, and grid information acquired as measurement information, the charge / discharge power statistical values ​​and the charge / discharge power instantaneous values ​​under normal control are inferred based on the statistical value inference model and the instantaneous value inference model and the grid information, and an instantaneous value of control reserve capacity usable in emergency control is inferred based on the inferred charge / discharge power statistical values ​​and the charge / discharge power instantaneous values ​​under normal control. This allows the power storage device 6 to be charged / discharged by multiple controls such as normal control and emergency control, and therefore the ever-changing instantaneous value of control reserve capacity can be fully utilized for emergency control using an existing power storage device for normal control without restricting the behavior of the normal control.

[0092] Furthermore, when the generation unit 23 generates an inference model using machine learning, it can appropriately generate the inference model and infer the instantaneous value of the control margin.

[0093] <Modification> Fig. 12 is a diagram showing the configuration of a power inference device and other components according to this modification. As shown in Fig. 12, the power inference device 8 may include, in addition to the components shown in Fig. 2, a power grid stabilization unit 28 having the same function as the power grid stabilization device 13. For example, the power grid stabilization unit 28 may determine at least one of a power limiting policy (power limit targets and power limit amount) and a load limiting policy (load limit targets and load limit amount) for the power grid 9 based on power grid information and a control margin for a predetermined period. In this case, the same effect as in the second embodiment can be obtained. Furthermore, for example, the power grid stabilization unit 28 may determine the control amount of the power storage device 6 based on the power grid information and a control margin for a predetermined period. In this case, the same effect as in the third embodiment can be obtained.

[0094] Furthermore, some of the components of the power inference device 8 described above may be configured as separate devices. For example, the acquisition unit 21 and the generation unit 23 may configure a model generation device that generates an inference model. Alternatively, the power inference device may be configured by an acquisition unit that acquires the inference model generated by the model generation device, and a charge / discharge power statistical value inference unit 24 that infers charge / discharge power statistical values ​​in main control based on the inference model and grid information.

[0095] <Configuration Example of Computer System> The acquisition unit 21, generation unit 23, charge / discharge power statistical value inference unit 24, and control margin inference unit 25 shown in Fig. 2 described above will be referred to as the "acquisition unit 21, etc." The acquisition unit 21, etc. are realized by a processing circuit 81 shown in Fig. 13. That is, the processing circuit 81 includes: an acquisition unit 21 that acquires charging / discharging power during main control of a power storage device that charges / discharges power to / from a power grid through a combination of main control and secondary control, which is performed with a lower priority than the main control, and acquires grid information that is information about the power grid; a generation unit 23 that generates a first inference model that can infer charging / discharging power statistics during main control from the grid information based on charging / discharging power statistics that are statistics of charging / discharging power during main control; a charge / discharge power statistical value inference unit 24 that infers charging / discharging power statistics during main control based on the first inference model and the grid information; and a control margin inference unit 25 that infers control margin for a predetermined period, which is charging / discharging power available for secondary control during a predetermined period, based on the inferred charging / discharging power statistics during main control. The processing circuit 81 may be implemented using dedicated hardware, or may be implemented using a processor that executes a program stored in a memory. Examples of the processor include a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, and a DSP (Digital Signal Processor).

[0096] When the processing circuitry 81 is dedicated hardware, the processing circuitry 81 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. The functions of each unit such as the acquisition unit 21 may be realized by a circuit in which processing circuits are distributed, or the functions of each unit may be realized by a single processing circuit.

[0097] When the processing circuitry 81 is a processor, the functions of the acquisition unit 21 and the like are realized in combination with software and the like. Note that software and the like may include, for example, software, firmware, or software and firmware. The software and the like are written as programs and stored in memory. As shown in FIG. 14 , the processor 82 applied to the processing circuitry 81 realizes the functions of each unit by reading and executing programs stored in memory 83. That is, the power inference device 8 includes a memory 83 for storing a program that, when executed by the processing circuit 81, results in the following steps: acquiring charge / discharge power under main control of a power storage device that charges / discharges to / from a power grid through a combination of main control and secondary control that is performed with a lower priority than the main control, and acquiring grid information that is information about the power grid; generating a first inference model that can infer the charge / discharge power statistical value under main control from the grid information based on charge / discharge power statistical values ​​that are statistics of charge / discharge power under main control and the grid information; inferring the charge / discharge power statistical value under main control based on the first inference model and the grid information; and inferring a control margin for a predetermined period, which is charge / discharge power for a predetermined period that can be used for secondary control, based on the inferred charge / discharge power statistical value under main control. In other words, this program can be said to cause a computer to execute the procedures and methods of the acquisition unit 21, etc. Here, the memory 83 may be, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically Erasable Programmable Read Only Memory), a HDD (Hard Disk Drive), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disc), a drive device for any of these, or any storage medium to be used in the future.

[0098] The above describes a configuration in which each function of the acquisition unit 21, etc. is realized either by hardware or software, etc. However, this is not limited to this, and a configuration in which part of the acquisition unit 21, etc. is realized by dedicated hardware and another part is realized by software, etc. For example, the function of the acquisition unit 21 can be realized by a processing circuit 81 as dedicated hardware, and the other functions can be realized by the processing circuit 81 as a processor 82 reading and executing a program stored in a memory 83.

[0099] As described above, the processing circuit 81 can realize the above-mentioned functions by hardware, software, or a combination of these. Note that the components of the power system stabilization device 13 in Fig. 6 and the power system stabilization unit 28 in Fig. 12 are similar to those described above.

[0100] It should be noted that the embodiments and modifications may be freely combined, and the embodiments and modifications may be modified or omitted as appropriate.

[0101] The above description is illustrative in all respects and is not restrictive. It is understood that countless variations not illustrated can be envisioned.

[0102] 6 Power storage device, 7 Power control device, 8 Power inference device, 9 Power system, 13 System stabilization device, 21 Acquisition unit, 23 Generation unit, 24 Charge / discharge power statistical value inference unit, 24a Charge / discharge power statistical value / instantaneous value inference unit, 25 Control margin inference unit, 28 System stabilization unit, 450 Power control system.

Claims

1. A power inference device comprising: an acquisition unit that acquires the charge / discharge power during main control of a power storage device that charges and discharges power to and from a power grid through a combination of main control and secondary control that is implemented with a lower priority than the main control, and acquires system information that is information about the power grid; a generation unit that generates a first inference model that can infer the charge / discharge power statistical value during the main control from the system information based on a charge / discharge power statistical value that is a statistical value of the charge / discharge power during the main control and the system information; an inference unit that infers the charge / discharge power statistical value during the main control based on the first inference model and the system information; and a control margin inference unit that infers the control margin for a predetermined period, which is the charge / discharge power during a predetermined period that can be used for the secondary control, based on the inferred charge / discharge power statistical value during the main control.

2. An electric power inference apparatus according to claim 1, wherein the primary control is a normal control and the secondary control is an emergency control.

3. An electric power inference apparatus according to claim 1 or 2, wherein the generation unit generates the first inference model using machine learning.

4. A power inference device as described in claim 1 or claim 2, wherein the first inference model includes an upper power inference model that infers upper power, which is the charge / discharge power statistical value of the discharge side of the main control, and a lower power inference model that infers lower power, which is the charge / discharge power statistical value of the charge side of the main control.

5. A power inference device according to claim 1 or claim 2, further comprising a system stabilization unit that determines at least one of a power supply limiting policy and a load limiting policy for the power system based on the system information and the control margin for the predetermined period.

6. An electric power inference device according to claim 5, wherein the system stabilization unit further determines the control amount of the power storage device based on the system information and the control margin for the predetermined period.

7. A power control system comprising: a power inference device according to any one of claims 1 to 6; and a power control device that controls charging and discharging of the power storage device based on the control margin during the predetermined period.

8. A power system stabilization system comprising: a power inference device according to claim 1 or claim 2; and a power system stabilization device that determines at least one of a power source limiting policy and a load limiting policy for the power system based on the power system information and the control margin for the predetermined period.

9. A grid stabilization system according to claim 8, wherein the grid stabilization device further determines the control amount of the power storage device based on the grid information and the control margin for the predetermined period.

10. An electric power inference device as claimed in any one of claims 1 to 6, wherein the generation unit further generates a second inference model based on the charge / discharge power in the main control and the system information, the second inference model being capable of inferring a charge / discharge power instantaneous value, which is the instantaneous value of the charge / discharge power in the main control, from the system information; the inference unit further infers the charge / discharge power instantaneous value in the main control based on the second inference model and the system information; and the control margin inference unit infers a control margin instantaneous value, which is the instantaneous value of the charge / discharge power available for the secondary control, based on the inferred charge / discharge power instantaneous value in the main control and the charge / discharge power statistical value.

11. A model generation device comprising: an acquisition unit that acquires charging / discharging power in the main control of an energy storage device that charges / discharges to / from an electric power system by combining a main control with a secondary control that is implemented with a lower priority than the main control, and acquires system information that is information about the electric power system; and a generation unit that generates an inference model that can infer the charging / discharging power statistical value in the main control from the system information, based on a charging / discharging power statistical value that is a statistical value of the charging / discharging power in the main control and the system information.

12. An electric power inference device comprising: an acquisition unit that acquires an inference model that can infer a charge / discharge power statistical value, which is a statistical value of the charge / discharge power under the main control, from the system information, generated based on the charge / discharge power under the main control of an electric power storage device that charges / discharges to / from an electric power system by combining a main control with a secondary control that is implemented with a lower priority than the main control; and an inference unit that infers the charge / discharge power statistical value under the main control based on the inference model and the system information.

13. A power inference method comprising: acquiring charge / discharge power during main control of a power storage device that charges / discharges to / from a power grid through a combination of main control and secondary control that is implemented with a lower priority than the main control; acquiring system information that is information about the power grid; generating an inference model that can infer the charge / discharge power statistical value during the main control from the system information based on a charge / discharge power statistical value that is a statistical value of the charge / discharge power during the main control and the system information; inferring the charge / discharge power statistical value during the main control based on the inference model and the system information; and inferring a control margin for a predetermined period that is the charge / discharge power that can be used for the secondary control during the predetermined period based on the inferred charge / discharge power statistical value during the main control.

Images