A photovoltaic inverter intelligent control method and system based on the Internet of Things

Abstract

The present invention relates to the field of photovoltaic inverter control technology, specifically a photovoltaic inverter intelligent control method and system based on the Internet of Things, comprising the following steps: based on the illumination prediction value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, setting a time window and performing serialized extrapolation of the voltage within the window, and establishing an over-limit risk warning. The present invention sets a time window based on the illumination prediction value and the grid connection point voltage measurement value, and performs serialized extrapolation of the voltage within the window. By performing point-by-point comparison with the voltage threshold, it is possible to identify the over-limit risk time point in advance and improve the pre-warning capability of local voltage anomalies. On this basis, the voltage over-limit amplitude is extracted in combination with the indoor and outdoor temperature information of the building, converted into a voltage regulation demand, and further formed into a quantifiable voltage absorption index, so that the power load regulation is transformed from the traditional static setting to dynamic quantifiable regulation, effectively improving the accuracy of the regulation response.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic inverter control, and in particular to an Internet of Things-based intelligent control method and system for photovoltaic inverters. Background Art

[0002] The field of photovoltaic inverter control technology is a core branch of the integrated development of smart grid and new energy. It mainly studies how to efficiently, safely and stably manage the working status and output behavior of the inverter in the photovoltaic power generation system.

[0003] Existing technologies typically use instantaneous operating conditions such as voltage and current during inverter grid-connected operation as triggers, switching strategies through fixed parameters or static judgment mechanisms. This lacks the ability to predict and judge future voltage trends, making it difficult to provide proactive risk warnings. This can lead to delayed responses when voltage fluctuates dramatically, resulting in temporary overvoltage or undervoltage conditions in the system. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent control method and system for photovoltaic inverters based on the Internet of Things.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent control method for photovoltaic inverters based on the Internet of Things, comprising the following steps:

[0006] Based on the predicted sunlight value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set and the voltage within the window is serially estimated to establish an over-limit risk warning;

[0007] Based on the over-limit risk warning, the predicted voltage over-limit amplitude is extracted, and the voltage regulation demand is calculated in combination with the current indoor and outdoor temperature parameters of the building to obtain a voltage absorption quantitative index. Based on the voltage absorption quantitative index, it is converted into an instruction containing virtual electricity price information and sent to the building automatic control system via the Internet of Things to obtain a flexible load scheduling incentive signal;

[0008] Based on the upper-level substation circuit breaker status signal, determine whether the signal indicates a disconnected state and obtain the islanding operation trigger state; if the islanding operation trigger state is true, calculate the line transient energy and solve the reactive power value required to absorb the energy, and establish the transient reactive feedforward compensation amount;

[0009] Based on the transient reactive feedforward compensation amount, it is set as the control component with the highest priority, and a priority control instruction is immediately generated; when the island operation trigger state is false, the response state of the building automatic control system to the flexible load scheduling excitation signal is received and the voltage correction effect is judged, and the reactive adjustment amount is set according to the residual voltage deviation to form a photovoltaic inverter composite control instruction.

[0010] Preferably, the steps for obtaining the limit-crossing risk warning are:

[0011] Based on the predicted light value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set according to a fixed time length. The real-time voltage measurement value of the grid connection point and the corresponding light prediction value at consecutive moments within the time window are recorded to generate a mapping sequence of voltage measurement value and light prediction value with time tags;

[0012] Based on the mapping sequence of voltage measurement values ​​and light prediction values ​​with time tags, the voltage measurement values ​​are arranged in chronological order to construct a voltage change trend sequence, the voltage increments between adjacent time points are calculated and accumulated in sequence to derive the voltage prediction value in the future time period, and form a voltage prediction sequence;

[0013] Based on the voltage prediction sequence, each predicted voltage value in the voltage prediction sequence is compared one-to-one with the upper and lower limit thresholds of the distribution network voltage, and all predicted time points exceeding the upper threshold or falling below the lower threshold are marked. All marked time nodes and corresponding voltage values ​​are counted to generate an over-limit risk warning.

[0014] Preferably, the steps for obtaining the voltage absorption quantitative index are:

[0015] Based on the above-mentioned over-limit risk warning, the predicted voltage value is extracted from each marked prediction time point, and compared with the upper and lower voltage thresholds of the distribution network respectively. The portion exceeding the upper threshold is selected as the positive voltage over-limit value, and the portion below the lower threshold is selected as the negative voltage over-limit value. The voltage over-limit value sequence is obtained by integrating the values ​​in chronological order.

[0016] Calculating a recommended total regulated power value based on the voltage over-limit value sequence, the building's nominal flexible load power at each time point, the building's indoor temperature parameters, and the building's outdoor temperature parameters;

[0017] Based on the total regulated power recommendation value and combined with the dispatch capability threshold of the load response capability at each time point, it is checked whether the recommended power in each time period is within the load regulation capability range, and a voltage absorption quantitative index is generated.

[0018] Preferably, the steps of obtaining the flexible load scheduling excitation signal are:

[0019] Based on the voltage absorption quantitative index, the recommended voltage regulation power value corresponding to each time point is extracted, and a list of flexible load devices that are currently connected and whose operating time is not less than the minimum response period is counted. The response duration recorded in the last three rounds of historical operation logs is read, and a list of flexible load devices that can participate in regulation at each time point and response status information is generated by sorting according to time tags;

[0020] Calculate the dynamic incentive electricity price for each flexible load device based on the voltage regulation power recommendation value, flexible load device list and response status information;

[0021] Based on the dynamic incentive electricity price of each flexible load device, the dynamic incentive electricity price is matched with the equipment adjustment capability one by one to form a scheduling reference comparison table. The target devices with incentive electricity price greater than the benchmark trigger value are screened and the address, adjustment power instruction and incentive electricity price are encapsulated into an instruction structure and distributed to the corresponding building automatic control system to generate a flexible load scheduling incentive signal.

[0022] Preferably, the steps of obtaining the island operation triggering state are:

[0023] Extracting the status signal encoding content from the upper-level substation circuit breaker status signal, identifying the function bit identifier and logic level field in the encoding, parsing the control bit status indicating whether the circuit breaker is connected or disconnected in the logic level field, and extracting the upper-level substation circuit breaker status signal at the current moment;

[0024] Based on the upper-level substation circuit breaker status signal, determine whether the control bit status field is a disconnected state indicating value; if the status field is consistent with the preset disconnected state value, confirm that the circuit breaker is currently in the disconnected state; otherwise, mark it as not disconnected, and obtain the circuit breaker disconnection determination result;

[0025] Based on the circuit breaker disconnection judgment result, the voltage amplitude fluctuation trend of the current period is read, and when the circuit breaker disconnection judgment result is the disconnected state and the voltage fluctuation range is greater than the off-grid operation judgment threshold, the current moment is marked as the island operation trigger state.

[0026] Preferably, the step of obtaining the transient reactive feedforward compensation amount is:

[0027] If the island operation trigger state is true, read the active output power value corresponding to the current time point, and simultaneously call the unit inductance parameter and unit capacitance parameter recorded in the current power line configuration file, calculate the sum of the corresponding inductance energy term and capacitance energy term, and generate the line transient energy value;

[0028] Based on the transient energy value of the line, combined with the energy change curve trend map under similar operating conditions in the current photovoltaic inverter grid-connected operation history, a sample group with similar initial energy magnitude and system response period is retrieved, and the optimal reactive power matching value corresponding to the transient energy change of the line in the sample group in the same time period is calculated to generate the reactive power value required to absorb the transient energy;

[0029] Based on the reactive power value required to absorb transient energy, the reactive power value is matched with the preset reactive power output gear, and the standard output gear value closest to the required reactive power value is selected as the control instruction target to generate a transient reactive feedforward compensation amount.

[0030] Preferably, the step of obtaining the priority control instruction is:

[0031] Based on the transient reactive feedforward compensation amount, the transient reactive feedforward compensation amount is inserted into the first position of the control component queue, the priority label is set to the highest response level in the current period, and a sorted control component sequence is generated;

[0032] Based on the sorted control component sequence, the control strategy execution path with the highest priority is read, the target output parameter field and the control instruction type field encapsulated in the control component are parsed, and a control signaling data packet that complies with the inverter communication protocol standard is generated to form a priority control instruction.

[0033] Preferably, the steps of obtaining the photovoltaic inverter composite control instruction are:

[0034] When the island operation trigger state is false, the response signal message returned by the building automation system is received, the device response state field and the power change feedback field in the response signal message are parsed, the power change value of each responding device is classified and counted in chronological order, and a response state set of the flexible load scheduling excitation signal is generated;

[0035] Extracting the real-time voltage sampling value of the grid connection point corresponding to each time point according to the response state set of the flexible load scheduling excitation signal, performing difference calculation on the real-time voltage sampling value of the grid connection point corresponding to the target voltage value set before voltage regulation, determining whether the difference is within the voltage deviation tolerance range, and screening the time points and corresponding voltage deviation amplitudes at which the deviation exceeds the tolerance range to generate a residual voltage deviation sequence;

[0036] Based on the residual voltage deviation sequence, the reactive power adjustment gear value corresponding to each voltage deviation interval is obtained, each residual voltage deviation amplitude is mapped to the closest reactive power adjustment gear, and all adjustment gear values ​​are embedded in the current inverter control data structure in chronological order to generate a photovoltaic inverter composite control instruction.

[0037] The present invention also provides a photovoltaic inverter intelligent control system, comprising:

[0038] Voltage risk warning module: Based on the predicted sunlight value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set and the voltage within the window is serialized and estimated to establish an over-limit risk warning;

[0039] Voltage Accommodation Control Module: Based on the over-limit risk warning, the module extracts the predicted voltage over-limit amplitude and calculates the voltage regulation demand in combination with the current indoor and outdoor temperature parameters of the building to obtain a voltage accommodation quantitative index. Based on the voltage accommodation quantitative index, the module converts it into an instruction containing virtual electricity price information and sends it to the building automation system via the Internet of Things to obtain a flexible load scheduling incentive signal.

[0040] Islanding Identification Module: Based on the upper-level substation circuit breaker status signal, it determines whether the signal indicates a disconnected state and obtains the islanding trigger state. If the islanding trigger state is true, it calculates the transient energy of the line and solves the reactive power value required to absorb the energy, and establishes the transient reactive feedforward compensation value.

[0041] Inverter integrated control module: Based on the transient reactive feedforward compensation amount, it is set as the highest priority control component and immediately generates a priority control instruction; when the island operation trigger state is false, it receives the response state of the building automatic control system to the flexible load scheduling excitation signal and judges the voltage correction effect, sets the reactive adjustment amount according to the residual voltage deviation, and forms a photovoltaic inverter composite control instruction.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are:

[0043] The present invention sets a time window based on the predicted illumination value and the measured voltage value at the grid connection point, and performs a serialized calculation of the voltage within the window. By comparing it point by point with the voltage threshold, it can identify the time point of over-limit risk in advance, improving the pre-warning capability of local voltage anomalies. On this basis, the voltage over-limit amplitude is extracted in combination with the indoor and outdoor temperature information of the building, converted into the voltage regulation demand, and further formed into a quantifiable voltage absorption index, so that the power load regulation is transformed from the traditional static setting to dynamic quantifiable regulation, effectively improving the accuracy of the regulation response. The voltage absorption index is converted into information containing virtual electricity prices, and the regulation incentive signal is sent to the building side through the Internet of Things. This establishes a direct interactive relationship between the power generation end and the energy consumption end driven by the voltage state, enabling the flexible load to have the ability to be associated with the dynamic state of the power grid in real time. When the upper-level substation circuit breaker status indication is disconnected, by extracting the active output power of the photovoltaic inverter and the line inductance and capacitance parameters, the transient energy is calculated and the reactive compensation amount is determined accordingly, forming the highest priority control instruction for the temporary response, which can stabilize the system voltage and energy balance in the early stage of island operation. In the non-islanding state, the voltage correction effect is evaluated based on the feedback results of the building response behavior, and the reactive power adjustment value is configured according to the residual deviation to achieve a composite control response. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 Schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0046] See also Figure 1 The present invention provides a technical solution, a photovoltaic inverter intelligent control method based on the Internet of Things, comprising the following steps:

[0047] Based on the predicted sunlight value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set and the voltage within the window is serially estimated to establish an over-limit risk warning.

[0048] Based on the over-limit risk warning, the predicted voltage over-limit amplitude is extracted, and the voltage regulation demand is calculated in combination with the current indoor and outdoor temperature parameters of the building to obtain the voltage absorption quantitative index; based on the voltage absorption quantitative index, it is converted into an instruction containing virtual electricity price information and sent to the building automatic control system through the Internet of Things to obtain the flexible load scheduling incentive signal.

[0049] Based on the status signal of the upper-level substation circuit breaker, it is determined whether the signal indicates the disconnected state and the islanding operation trigger state is obtained. If the islanding operation trigger state is true, the transient energy of the line is calculated and the reactive power value required to absorb the energy is solved to establish the transient reactive feedforward compensation value.

[0050] Based on the transient reactive feedforward compensation, it is set as the highest priority control component and a priority control instruction is generated immediately. When the island operation trigger state is false, the response state of the building automatic control system to the flexible load scheduling excitation signal is received and the voltage correction effect is judged. The reactive adjustment amount is set according to the residual voltage deviation to form a composite control instruction for the photovoltaic inverter.

[0051] The steps to obtain the limit-crossing risk warning are as follows:

[0052] Based on the predicted light value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set according to a fixed time length. The real-time voltage measurement value of the grid connection point and the corresponding light prediction value at consecutive moments within the time window are recorded to generate a mapping sequence of voltage measurement value and light prediction value with time tags;

[0053] Based on the mapping sequence of voltage measurement values ​​and light prediction values ​​with time tags, the voltage measurement values ​​are arranged in chronological order to construct a voltage change trend sequence. The voltage increments between adjacent time points are calculated and accumulated in sequence to derive the voltage prediction value in the future time period, forming a voltage prediction sequence.

[0054] Based on the voltage prediction sequence, each predicted voltage value in the voltage prediction sequence is compared one-to-one with the upper and lower limit thresholds of the distribution network voltage. All predicted time points that exceed the upper threshold or fall below the lower threshold are marked. All marked time nodes and corresponding voltage values ​​are counted to generate an over-limit risk warning.

[0055] Specifically, based on the predicted solar irradiance data obtained by the photovoltaic inverter and the real-time voltage measurement data collected by voltage sensors deployed at the grid-connected point, a time window for analysis is first defined according to a pre-set fixed time length. The length of this time window, for example, is set to 15 minutes, which is determined based on empirical analysis of grid voltage characteristics and photovoltaic output fluctuation cycles. This captures representative voltage variation patterns while avoiding excessive computational load. Within this time window, the system continuously records the real-time voltage measurement values ​​of the grid-connected point at a high frequency, for example, once per second. At the same time, for each voltage measurement time point, the system retrieves the solar irradiance prediction value corresponding to that precise moment or the immediately future period provided by the photovoltaic inverter. These solar irradiance prediction values ​​are typically generated by meteorological data analysis and photovoltaic power generation models integrated within the inverter. Subsequently, each set of data points containing a precise timestamp, a real-time voltage measurement value, and a corresponding solar irradiance prediction value is integrated to ensure that the voltage value is organically associated with the predicted solar irradiance conditions. By organizing these data points in chronological order, a sequence of voltage measurement value and solar irradiance prediction value mappings with precise time tags is ultimately generated.

[0056] Based on the mapping sequence of voltage measurement values ​​and light prediction values ​​with time tags obtained in the previous step, all voltage measurement values ​​are first extracted from the sequence and arranged in strict chronological order to construct a voltage change trend sequence that reflects the recent dynamic changes in voltage. Then, in order to quantify the voltage changes, the system calculates the voltage increment between each adjacent time point in the voltage change trend sequence, that is, the voltage value at the next moment is subtracted from the voltage value at the previous moment to obtain a series of historical voltage increment values. The previously established mapping sequence of voltage measurement values ​​and light prediction values ​​is used to analyze the sensitivity of voltage changes to light intensity changes within the time window. For example, an average voltage change response coefficient can be calculated. This coefficient Represents the voltage change caused by a unit change in light intensity. When calculating, you can select the period when the light intensity in the mapping sequence changes significantly, calculate the ratio of the corresponding voltage change to the light change, and then take the average value, or use the least squares fitting method. Come get ,in is the voltage change, is the change in the predicted illumination value, is the intercept. Next, the system obtains a sequence of minute-by-minute light forecast values ​​provided by the photovoltaic inverter for a period of time in the future, for example, the next 30 minutes. Based on these future light forecast values, the expected light change at each time interval in the future is calculated. , and then use the voltage change response coefficient obtained previously , estimate the voltage increment at each future time interval Finally, starting from the latest real-time voltage measurement value, the future voltage increments estimated one by one are calculated. Accumulate in chronological order, that is ,in is the current real-time voltage, from which a series of voltage prediction values ​​in the future time period are derived, and these prediction values ​​together constitute the voltage prediction sequence.

[0057] Based on the voltage prediction sequence generated in the previous process, the sequence contains a series of future time points and their corresponding predicted voltage values. The system then accurately compares each predicted voltage value in this sequence with the preset upper and lower limit standard thresholds of the distribution network operating voltage. These voltage upper and lower limit thresholds are set according to national or regional power grid guidelines and power supply agreements. For example, for a low-voltage distribution network with a nominal voltage of 230 volts, the upper limit threshold is usually set to +10% of the nominal voltage, that is, 253 volts, and the lower limit threshold is usually set to -10% of the nominal voltage, that is, 207 volts. These threshold parameters are set during system initialization. It is configured in real time and can be adjusted according to the regulations of the local power department. During the comparison process, if a predicted voltage value is higher than the set voltage upper limit threshold, or lower than the set voltage lower limit threshold, the system will specially mark the predicted time point and its corresponding predicted voltage value, indicating that it exceeds the normal operating range. After completing the comparison and marking of the entire voltage prediction sequence, the system will collect information on all marked prediction time points, including the specific time when the limit is exceeded, the predicted voltage value, and the specific situation of exceeding the upper limit or the lower limit. This information will be summarized and sorted to finally generate a structured limit-exceeding risk warning.

[0058] The steps to obtain the voltage absorption quantitative index are as follows:

[0059] Based on the over-limit risk warning, the predicted voltage value is extracted from each marked prediction time point and compared with the upper and lower voltage thresholds of the distribution network respectively. The part exceeding the upper threshold is selected as the positive voltage over-limit value, and the part below the lower threshold is selected as the negative voltage over-limit value. The voltage over-limit value sequence is obtained by integrating them in chronological order.

[0060] According to the voltage over-limit value sequence, the building's nominal flexible load power at each time point, the building's indoor temperature parameters, and the building's outdoor temperature parameters, the recommended total regulated power value is calculated using the following formula:

[0061] ;

[0062] in, is the recommended value of total regulated power (unit: W), For the The nominal flexible load power of the building at a time point (unit: W), For the The voltage exceeds the limit value at a time point (unit: V), is the maximum allowable voltage deviation (unit: V), is the response sensitivity index (dimensionless), For the The outdoor temperature of the building at a certain time point (unit: ℃), For the Indoor temperature of the building at a certain time point (unit: °C), is the reference temperature difference (unit: °C), is the total number of time points in the voltage limit-exceeding sequence;

[0063] Based on the recommended value of the total regulated power and combined with the dispatching capacity threshold of the load response capability at each time point, it is checked whether the recommended power in each period is within the load regulation capability range, and a quantitative indicator of voltage absorption is generated.

[0064] Specifically, based on the over-limit risk warning, the system first traverses each marked prediction time point contained in the warning, and accurately extracts the corresponding predicted voltage values ​​therefrom. These predicted voltage values ​​are obtained by analyzing and calculating the light forecast and historical voltage data in the previous step. Next, the system compares each extracted predicted voltage value with the distribution network voltage upper threshold and distribution network voltage lower threshold pre-configured in the system. The distribution network voltage upper threshold and lower threshold are set according to the national grid operation guidelines or the standards issued by the local power supply department. For example, for a low-voltage distribution network with a nominal voltage of 230 volts, the upper voltage threshold is usually set to +10% of the nominal voltage, that is, 253 volts, and the lower voltage threshold is usually set to -10% of the nominal voltage, that is, 207 volts. These thresholds are used as Key operating parameters are fixed in the system, and authorized operation and maintenance personnel are allowed to adjust them according to the latest power specifications. If a predicted voltage value exceeds the set distribution network voltage upper threshold, the difference between the predicted voltage value and the upper threshold (i.e., predicted voltage value - distribution network voltage upper threshold) is recorded as the positive voltage over-limit value at that time point. Conversely, if a predicted voltage value is lower than the set distribution network voltage lower threshold, the difference between the lower threshold and the predicted voltage value (i.e., distribution network voltage lower threshold - predicted voltage value) is recorded as the negative voltage over-limit value at that time point. After processing all marked predicted time points, the system collects all calculated positive and negative voltage over-limit values ​​and their corresponding time tags, and arranges them in strict chronological order, and integrates them in chronological order to obtain a voltage over-limit value sequence.

[0065] formula: The benefit of the formula is that it takes into account the severity of voltage deviation, the available regulation capacity of flexible loads in the building, and the current indoor and outdoor thermal environment conditions, so that the amount of power required for voltage regulation can be dynamically evaluated and recommended by introducing the response sensitivity index. , can adjust the system's positivity in responding to voltage deviations, and The introduction of the function makes the effect of temperature on regulation potential present a saturated characteristic, avoiding unreasonable regulation suggestions caused by extreme temperature differences, thereby improving the accuracy and economy of voltage regulation;

[0066] For the Nominal flexible load power of the building at a time point (unit: W):

[0067] This parameter represents the At a certain point in time, the total rated power of flexible loads (such as air conditioners, electric water heaters, adjustable lighting, etc.) inside the building that can participate in grid regulation is dynamically summarized by the building automation system (BMS) based on the list of flexible load devices that are currently connected and in a dispatchable state. For example, at a specific point in time, , the building automation system detects that there are 3 air conditioners with a rated power of 2000W and 1 electric water heater with a rated power of 1500W available for dispatch, then .

[0068] For the Voltage over-limit value at each time point (unit: V):

[0069] This parameter indicates the At a certain time point, the absolute difference between the predicted voltage value of the grid connection point exceeding or falling below the normal operating voltage limit of the distribution network is derived from the "voltage limit value sequence" generated in the previous step. For example, if At a certain time point, the upper threshold of the distribution network voltage is 253V, and the predicted voltage is 255V. , if the voltage lower limit threshold is 207V and the predicted voltage is 205V, then .

[0070] is the maximum allowable voltage deviation (unit: V):

[0071] This parameter defines the maximum voltage fluctuation range allowed in the distribution network. It is a fixed reference value set according to the grid operation standard of the relevant region. It is usually given as a percentage of the nominal voltage. For example, if the nominal voltage of the regional grid is 230V, the standard stipulates that the maximum voltage deviation allowed is ,but .

[0072] is the response sensitivity index (dimensionless):

[0073] This parameter is a dimensionless index used to adjust the sensitivity of the total regulated power recommendation value to the voltage over-limit value. Its value affects the enthusiasm of the regulation response. The value of is usually determined by simulation analysis of specific power grid and load characteristics or optimization adjustment based on historical operation data. It is used to balance response speed and system stability. For example, the initial setting is 1.0 to indicate linear response. If the voltage correction speed is found to be slow through system simulation testing, it is gradually increased. If adjustment overshoot or oscillation occurs, reduce the After a series of tests and evaluations, a value such as 1.2 was finally selected to achieve a fast and effective response to voltage deviations while avoiding system instability.

[0074] For the Outdoor temperature of the building at each time point (unit: ℃):

[0075] This parameter indicates the The outdoor ambient temperature of the building at a certain point in time is usually collected in real time by an outdoor temperature sensor connected to the system, or obtained from an authoritative third-party meteorological service platform, and recorded and provided by the building automation system. For example, during a certain forecast period , the outdoor temperature sensor reading is ,but .

[0076] For the Indoor temperature of the building at a time point (unit: ℃):

[0077] This parameter indicates the The average temperature inside the building or the temperature of the key area at a certain time point is collected by the building automation system through multiple temperature sensors deployed inside the building and the average value is calculated. It reflects the current thermal comfort state of the building. For example, in a certain forecast period , the building automation system aggregates the sensor data from each area and obtains the average indoor temperature as ,but .

[0078] is the reference temperature difference (unit: ℃):

[0079] This parameter is a preset reference temperature difference, which is used to normalize the indoor and outdoor temperature difference to reflect the correlation between the adjustable potential of flexible loads (especially temperature control loads) and the current thermal environment. Its setting is based on the building type, energy saving goals and user comfort requirements. For example, for office buildings, if the indoor temperature is allowed to be adjusted during the demand response period, The floating range of Can be set to .

[0080] is the total number of time points in the voltage limit-exceeding sequence:

[0081] This parameter indicates the total number of future time points with voltage over-limit risks identified in the voltage forecast sequence currently being analyzed. It is directly obtained from the "voltage over-limit amplitude sequence" generated in the previous step. For example, if it is predicted that there will be three 15-minute time periods with voltage over-limit in the next hour, then .

[0082] Calculation process:

[0083] Here, there is only one time point in the voltage limit sequence (i.e. , let this time point be ) is used as an example for calculation. The values ​​of the parameters are as follows:

[0084] ;

[0085] ;

[0086] ;

[0087] ;

[0088] ;

[0089] ;

[0090] ;

[0091] The calculation steps are as follows:

[0092] Calculate the voltage over-limit level:

[0093] ;

[0094] Calculate the temperature difference term:

[0095] ;

[0096] Compute the hyperbolic tangent function:

[0097] ;

[0098] Calculate the recommended adjustment power value at a single time point :

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] like ,but For all time point For example, if , and calculated , , , the recommended total regulated power value is .

[0104] The results show that under the current voltage over-limit situation, the available capacity of flexible loads and the thermal environment of the building, the total power amount recommended for voltage regulation, for example, is calculated. This indicates that the system assessment believes that approximately 748.93W of power can be contributed by dispatching flexible loads to help stabilize the grid voltage. The higher this value, the greater the available flexible regulation potential. This recommended total regulation power value is an important basis for subsequently formulating specific load dispatch strategies and generating incentive signals.

[0105] Based on the recommended total adjustment power value obtained in the previous calculation step, the system further combines the evaluation results of the actual response capability of the flexible load in the building at each relevant time point, that is, the dispatching capability threshold, for verification. The dispatching capability threshold does not only refer to the rated power of the load, but refers to the power range that the flexible load can actually safely and reliably adjust up or down at a specific time point, taking into account multiple factors such as the current operating status of the equipment, the user's preset comfort or usage restrictions (for example, the minimum operating temperature of the air conditioner, the insulation requirements of the water heater), the minimum start and stop interval of the equipment, and the safety constraints of the power grid. These dispatching capability thresholds are dynamically calculated and provided by the building automation system (BEMS) based on real-time monitored equipment data and preset operating strategies. For example, for an air conditioner with a rated power of 5kW, at a certain moment it is already running at a higher load. Its dispatching capability threshold for increasing power (absorbing power) may be 1kW, while its dispatching capability threshold for reducing power (reducing output) may be 1kW. The threshold is 3kW. The verification process specifically compares the recommended total regulated power value (if it is an aggregate value, it needs to be broken down into each forecast time period according to a specific strategy) with the actual dispatch capacity threshold for the corresponding time period. If the recommended regulated power (for example, adding 500W of load to absorb excess PV power generation) is less than or equal to the maximum capacity increase dispatch capacity threshold that the flexible load can provide in that time period (for example, 800W), the recommendation is considered feasible, and the actual dispatchable power is 500W. If the recommended regulated power exceeds the dispatch capacity threshold (for example, a 1000W increase is recommended, but the actual maximum increase is 800W), the actual dispatchable power is limited to the dispatch capacity threshold (i.e., 800W). This verification process is repeated for each forecast time period with voltage over-limit risk to ensure that all recommended regulated power values ​​are within the practical range. Finally, these verified and feasible regulated power values ​​are integrated to generate a quantitative voltage absorption index.

[0106] The steps for obtaining the flexible load scheduling excitation signal are as follows:

[0107] Based on the voltage absorption quantitative index, the recommended voltage regulation power value corresponding to each time point is extracted. A list of flexible load devices that are currently connected and whose operating time is not less than the minimum response period is compiled. The response duration recorded in the last three rounds of historical operation logs is read, and a list of flexible load devices that can participate in regulation and response status information is generated at each time point by time tag.

[0108] According to the recommended voltage regulation power value, the list of flexible load devices and the response status information, the dynamic incentive electricity price of each flexible load device is calculated. The calculation formula is:

[0109] ;

[0110] in, For the Dynamic incentive electricity price for flexible load equipment (unit: yuan / kWh), is the basic incentive electricity price (unit: yuan / kWh), is the system response gain coefficient (dimensionless), The recommended value of voltage regulation power at the current time point (unit: kW), The total available power of all devices in the flexible load device list at the current time (unit: kW), For the The cumulative response time of each flexible load device recorded in the last three rounds of operation (unit: min), is the set reference response time (unit: min);

[0111] Based on the dynamic incentive electricity price of each flexible load device, the dynamic incentive electricity price is matched with the equipment adjustment capability one by one to form a scheduling reference comparison table. The target devices with incentive electricity price greater than the benchmark trigger value are screened and the address, adjustment power instruction and incentive electricity price are encapsulated into an instruction structure and distributed to the corresponding building automatic control system to generate a flexible load scheduling incentive signal.

[0112] Specifically, based on the voltage absorption quantitative index generated in the previous step, the system first extracts the specific voltage regulation power recommendation value corresponding to each predicted voltage over-limit time point from the index. This recommendation value clarifies the active power adjustment required to correct the voltage deviation. Then, the system accesses the real-time database of the building automation system (BMS), counts all flexible load devices that are currently connected to the power grid and in a controllable state, and screens out the devices whose current cumulative operating time has reached or exceeded their respective preset minimum response cycles. The minimum response cycle is the shortest continuous operation or shutdown period set for each type of flexible load (such as air conditioners, electric water heaters) according to their physical characteristics and manufacturer recommendations. Time, for example, the air conditioner may need to run for at least 10 minutes before it can stabilize and enter the adjustable state. This period is set to ensure the effectiveness of scheduling and avoid damage to the equipment due to frequent start and stop. For the qualified equipment screened out, the system further retrieves and reads the actual response duration records of these devices in the last three demand response events from the historical operation log of the building automatic control system, and classifies and organizes this information, including device identification, adjustable power range, and historical response time, according to the predicted voltage limit exceeding time point, and finally generates a detailed list of flexible load devices that can participate in regulation and their corresponding complete response status information for each time point that requires regulation.

[0113] formula: The benefit of the formula is that it can calculate the incentive electricity price differentially according to the real-time supply and demand relationship (i.e., the ratio of the system's demand for regulated power to the currently available flexible load resources) and the recent participation history of each flexible load device. When the regulation demand is urgent or the available resources are scarce, the incentive electricity price will be increased accordingly, thereby more effectively guiding the load to participate in the response. This item can prevent a small number of devices from being called too frequently, promote fairness in load participation, and take into account device fatigue and user experience. This refined incentive mechanism helps achieve voltage regulation goals in a more economical and efficient manner.

[0114] The basic incentive electricity price is set (unit: yuan / kWh):

[0115] This parameter represents the basic compensation standard without considering the real-time supply and demand tension and the historical contribution of the equipment. Its setting mainly refers to the ancillary service price of the local power market, the peak period electricity price in the time-of-use electricity price policy, or the benchmark compensation price agreed in the flexible load participation agreement signed with the building operator and the user. For example, if the average compensation provided by the local power company for demand response is 0.7 yuan / kWh, in order to encourage user participation, it can be set It is 0.5 yuan / kWh.

[0116] is the system response gain coefficient (dimensionless):

[0117] This parameter is used to adjust the sensitivity of the incentive electricity price to the imbalance between supply and demand in the system. When the ratio increases (demand far exceeds supply), The larger the value, the greater the increase in incentive electricity prices. The setting is based on regression analysis of historical operation data or simulation platform for different The goal is to find a balance point that can effectively encourage load participation while controlling the overall incentive cost. For example, after a month of simulation operation and data analysis, it was found that When the system can effectively dispatch sufficient flexible loads in most high-demand scenarios, and the total incentive expenditure is within the budget, the selected , the coefficient generally ranges from 0.1 to 1.5.

[0118] Recommended voltage regulation power value at the current time point (unit: kW):

[0119] This parameter indicates the total active power that needs to be absorbed or reduced by the flexible load in order to maintain the grid voltage within the allowable range at the current specific forecast time point. It is directly derived from the extraction result of the "voltage absorption quantification index" in the previous step. This index has been verified and is the actual power value that needs to be adjusted. For example, according to the "voltage absorption quantification index", it is predicted that a 15kW load will be needed to absorb the excess power generated by the photovoltaic system to suppress excessive voltage during the period from 2:00 to 2:15 in the afternoon. At this time, kW.

[0120] The total available power of all devices in the flexible load device list at the current time (unit: kW):

[0121] This parameter represents the total regulation capacity (load increase or load reduction) that can be provided by all flexible load devices that are identified as being able to participate in regulation at the current specific time point. Its value is obtained by summarizing the maximum regulation power currently provided by each device in the "List of Flexible Load Devices that Can Participate in Regulation at Each Time Point". The list is generated by the previous step. For example, if the list shows that there are currently 3 air conditioners that can increase power by 2kW each and 2 water heaters that can increase power by 1.5kW each, then To ensure the accuracy of the calculation, we take kW.

[0122] For the Cumulative response time of flexible load devices recorded in the last three rounds of operation (unit: min):

[0123] This parameter reflects a specific flexible load device The total cumulative response time in the three most recent demand response events is used to assess its recent contribution and fatigue. The data comes from the log records of each device participating in the demand response process in the building automation system (BMS). For example, for air conditioning device AC001, query its BMS operation log and find that when it was called to participate in voltage regulation for the last three times, it responded for 15 minutes, 20 minutes and 10 minutes respectively. min.

[0124] The reference response time is set (unit: min):

[0125] This parameter is a preset benchmark cumulative response time. Close to or exceed When the incentive electricity price is significantly reduced, it can avoid excessive use of equipment and ensure that response opportunities are fairly distributed among different devices. The setting of the demand response time should take into account the equipment type (such as air conditioning, lighting), the design life of the equipment, the user's tolerance for comfort or convenience, and the overall goal of the demand response strategy. For example, for the air conditioning system of an office building, in order to avoid affecting the comfort of the office environment for a long time, the reference cumulative response time within a working day can be set to 120 minutes, that is, min.

[0126] Calculation process:

[0127] To calculate flexible load equipment (For example, the dynamic incentive electricity price of the air conditioner AC001 mentioned above) is used as an example, and the specific values ​​are substituted:

[0128] Yuan / kWh;

[0129] ;

[0130] kW;

[0131] kW;

[0132] min;

[0133] min;

[0134] The calculation process is as follows:

[0135] Calculating the supply-demand ratio :

[0136] ;

[0137] Calculating the Price Gain Factor :

[0138] ;

[0139] Calculate the device's historical engagement ratio :

[0140] ;

[0141] Calculate the historical participation to price decay factor Function Value :

[0142] ;

[0143] Calculate the full historical engagement decay factor :

[0144] ;

[0145] Calculate the final dynamic incentive electricity price :

[0146] ;

[0147] ;

[0148] ;

[0149] ;

[0150] The results show that for device AC001, under the current system status and its historical participation, the calculated dynamic incentive electricity price is approximately 0.3785 yuan / kWh. This price will be used to incentivize the device to participate in the current voltage regulation task. If this price is higher than the participation threshold set for the device, the device will be dispatched.

[0151] Based on the dynamic incentive electricity price calculated for each flexible load device that can participate in regulation in the previous step, the system will then match these dynamic incentive electricity prices with the current specific regulation capabilities of each device (that is, the power value that can be adjusted upward or downward, this information comes from the previously generated "flexible load device list and response status information"), thereby instantly forming a scheduling reference comparison table containing key information such as device identification, dynamic incentive electricity price, available upward power, available downward power, etc. Subsequently, the system compares the dynamic incentive electricity price of each device in this scheduling reference comparison table with its preset benchmark trigger value. The benchmark trigger value is a minimum acceptable incentive price set for each category or even each device. For example, the benchmark trigger value of an air conditioner is set to 0.3 yuan / kWh, which takes into account its operating costs and potential impact on user comfort. After the influence and user participation willingness are determined, the economic parameters set by the building operator or user through the building automation system interface. Only when the calculated dynamic incentive electricity price is greater than or equal to the benchmark trigger value of the device, the device will be preliminarily screened as the target scheduling device. For all target devices that pass the screening, the system extracts its unique communication address in the building automation system, the specific adjustment power instruction value determined according to the current voltage regulation requirement (increase load or reduce load) (for example, "increase power by 2kW"), and the dynamic incentive electricity price calculated for it, and encapsulates this information into a standardized instruction data structure in strict accordance with the requirements of the building automation system communication protocol. Finally, the system distributes these encapsulated instruction structures in batches to the corresponding building automation system execution units through the internal network or Internet of Things channel, thereby generating and issuing flexible load scheduling incentive signals.

[0152] The steps to obtain the island operation trigger status are:

[0153] Extracting the status signal encoding content from the upper-level substation circuit breaker status signal, identifying the function bit identifier and logic level field in the encoding, parsing the control bit status indicating whether the circuit breaker is connected or disconnected in the logic level field, and extracting the upper-level substation circuit breaker status signal at the current moment;

[0154] Based on the status signal of the circuit breaker of the upper-level substation, determine whether the control bit status field is a disconnected state indicating value. If the status field is consistent with the preset disconnected state value, it is confirmed that the circuit breaker is currently in the disconnected state. Otherwise, it is marked as not disconnected, and the circuit breaker disconnection judgment result is obtained;

[0155] Based on the circuit breaker disconnection judgment result, the voltage amplitude fluctuation trend of the current period is read. If the circuit breaker disconnection judgment result is the disconnected state and the voltage fluctuation range is greater than the off-grid operation judgment threshold, the current moment is marked as the island operation trigger state.

[0156] Specifically, the system first parses the received raw signal data packet from the upstream substation to extract the complete status signal encoding content. This encoding content typically follows a specific power system communication protocol, such as the IEC61850 standard or a specific vendor-defined protocol, and may contain multiple information fields. Based on pre-configured protocol parsing rules, the system accurately identifies the function bit identifier in the encoding that specifically indicates the circuit breaker status, such as a specific byte or bit sequence. Next, the system locates the logic level field associated with this function bit identifier. This field represents the actual status of the circuit breaker in binary or specific encoding (such as BCD code). The system then decodes this logic level field and parses the control bit information representing the circuit breaker's "on" (closed) or "off" (disconnected) state. For example, if the protocol specifies that a bit is "1" for disconnected and "0" for connected, the system reads the value of that bit. Through this series of precise extraction, identification, and parsing operations, the system obtains the exact status signal of the upstream substation circuit breaker at the current moment.

[0157] Based on the current upstream substation circuit breaker status signal obtained in the previous step, the system's core task is to determine whether the control bit status field indicated by the signal clearly indicates that the circuit breaker is in the disconnected state. To this end, the system will strictly compare the actual value of the parsed control bit status field with a "disconnected state indication value" preset in the system configuration. This "preset disconnected state value" is defined based on the communication protocol used and the specific circuit breaker model. For example, if the communication protocol stipulates that the logic level "1010" represents an open circuit breaker, the system will compare the status field parsed from the signal with "1010". If the parsed status field is completely consistent with the preset disconnected state value, the system confirms that the upstream substation circuit breaker is currently in the disconnected state and records this result. Conversely, if the parsed status field is inconsistent with the preset disconnected state value, or indicates an on state or other non-disconnected state, the system marks the current state of the circuit breaker as not disconnected. Through this clear comparison and judgment logic, the system ultimately obtains a clear circuit breaker disconnection determination result.

[0158] Based on the circuit breaker disconnection judgment result obtained in the previous step, the system further combines the real-time monitoring data of the voltage amplitude fluctuation trend of the grid connection point in the current period to make a comprehensive judgment. First, the system continuously collects voltage data at high frequency through the voltage sensors deployed at the grid connection point, and calculates the voltage amplitude change within a short time window (for example, the past 5 seconds). Specifically, it calculates the difference between the maximum and minimum voltage values ​​within the time window to obtain the voltage fluctuation range, or calculates the standard deviation of the voltage value to quantify the voltage stability. Next, the system checks whether two key conditions are met at the same time: the first condition is whether the circuit breaker disconnection judgment result clearly indicates that the circuit breaker of the upper-level substation is currently in the disconnected state, and the second condition is whether the currently monitored voltage fluctuation range is greater than a preset "off-grid operation" The "off-grid operation judgment threshold" is set based on experience or obtained through analysis of historical islanding events. It aims to distinguish normal grid voltage fluctuations from severe voltage oscillations caused by disconnection from the main grid. For example, if the voltage fluctuation range during normal grid-connected operation generally does not exceed 2% of the nominal voltage, the off-grid operation judgment threshold can be set to 5% of the nominal voltage, that is, for a 230V system, the threshold is 11.5V. If the system confirms that the circuit breaker is indeed in the disconnected state and at the same time monitors that the voltage fluctuation range of the grid-connected point significantly exceeds the preset 11.5V threshold, for example, if the voltage drops from 220V to 200V and then rises back to 240V in a short period of time, and the fluctuation range reaches 40V, which is much greater than 11.5V, the system will mark the current moment as the islanding operation trigger state.

[0159] The steps for obtaining transient reactive feedforward compensation are as follows:

[0160] If the islanding operation trigger state is true, read the active output power value corresponding to the current time point, and at the same time call the unit inductance parameter and unit capacitance parameter recorded in the current power line configuration file, calculate the sum of the corresponding inductance energy term and capacitance energy term, and generate the line transient energy value;

[0161] Based on the transient energy value of the line, combined with the energy change curve trend map under similar operating conditions in the current PV inverter grid-connected operation history, sample groups with similar initial energy levels and system response periods are retrieved. The optimal reactive power matching value corresponding to the transient energy change of the line in the sample group in the same time period is calculated to generate the reactive power value required to absorb the transient energy.

[0162] Based on the reactive power value required to absorb transient energy, the reactive power value is matched with the preset reactive power output gear, and the standard output gear value closest to the required reactive power value is selected as the control instruction target to generate the transient reactive feedforward compensation amount.

[0163] Specifically, if the island operation trigger state determined in the previous step is true, indicating that the local photovoltaic system has been disconnected from the main power grid, the system will immediately read the active output power value corresponding to the current precise time point from the data acquisition of the photovoltaic inverter. The active output power value reflects the actual power injected by the photovoltaic system to the local load or line at the moment of island formation. At the same time, the system calls the current power line configuration file stored locally or accessible through the network. The configuration file records in detail the electrical parameters of the line connecting the photovoltaic inverter to the local load or grid connection point, including but not limited to the inductance value per unit length (for example, 0.2 millihenry / km) and the capacitance value per unit length (for example, 10 nanofarads / km), as well as the total length of the line. Based on these parameters and the actual operating current of the current line and voltage (These are also measured in real time by the inverter or additional sensors), the system follows Calculate the energy stored in the total line inductance (where is the total line inductance), and according to Calculate the energy stored in the total capacitance of the line (where The two calculated inductive energy terms and capacitive energy terms are added together to generate a line transient energy value representing the total electromagnetic energy stored in the current line.

[0164] Based on the transient energy value of the line calculated in the previous step, the system then analyzes the energy change curve trend map under similar working conditions accumulated in the grid-connected operation history of the photovoltaic inverter stored in the local database. This map is constructed by recording the initial transient energy of the line, the system's own electrical parameters (such as inverter internal resistance, filter parameters, etc.) when multiple islanding events or similar transient disturbance events occur during the long-term operation of the system, and the actual trajectory data of the line energy change over time under different reactive power compensation strategies. The data is constructed through data mining and pattern recognition technology (for example, a historical working condition classification method based on K-means clustering is used to cluster historical data to form a typical working condition library, with each category corresponding to an energy change trend). The retrieval process first selects historical samples with similar initial energy magnitudes in the map based on the currently calculated transient energy value. For example, if the current transient energy value of the line is 50 joules, the records with initial energy in the range of 45 to 55 joules in the historical samples are retrieved. At the same time, the system also considers the inherent response cycle of the current system (for example, the historical working condition classification method based on K-means clustering is used to cluster historical data to form a typical working condition library, with each category corresponding to an energy change trend). For example, the response time constant of the inverter control loop (for example, 20 milliseconds) is used to filter out a sample group whose response cycle characteristics match the current system. For this group of historical samples with similar initial energy levels and system response cycles, the system further analyzes the different reactive power compensation values ​​adopted by these samples within a very short and similar time period after the islanding occurs (for example, within the first 100 milliseconds after the islanding occurs) and the actual absorption or release of corresponding line transient energy. By comparing the control effect of different reactive power values ​​on energy changes in these historical samples, for example, calculating the average energy decay rate under each reactive power compensation strategy, the historical reactive power value that can most quickly guide the line transient energy to a stable state or minimize energy fluctuations is selected as a reference. These reference values ​​are then weighted averaged or solved using an optimization algorithm (such as a particle swarm optimization algorithm to search for the optimal solution in historical data). Ultimately, an optimal reactive power matching value for the current specific line transient energy value is calculated. This value is the reactive power value required to absorb the current line transient energy.

[0165] Based on the theoretical reactive power value calculated in the previous step to effectively absorb the transient energy of the current line, the system needs to convert it into a control instruction that can be actually executed by the photovoltaic inverter. Since the reactive power output of the photovoltaic inverter is usually not continuously adjustable, but is adjusted according to a series of preset discrete gears, for example, the inverter is set to several standard reactive power output gears such as -5kVar, -2.5kVar, 0kVar, +2.5kVar, +5kVar, etc. These gear values ​​are set during system initialization or parameter configuration according to the design capacity and control capability of the inverter. Therefore, the system will match and compare the calculated required reactive power value with these preset reactive power output gear values ​​one by one. The comparison principle is to select the standard output gear value that is closest to the required reactive power value (that is, the absolute value of the difference is the smallest). For example, if the calculation shows that 3kVar of reactive power needs to be absorbed (that is, -3kVar of reactive power needs to be emitted), and the available gears of the inverter include -2.5kVar and -5kVar, the system will select -2.5kVar as the target output because the difference between it and -3kVar (0.5kVar) is less than the difference between -5kVar and -3kVar (2kVar). After selecting this closest standard output gear value, the system sets it as the target reactive power value in the control instruction to be issued to the inverter, thereby generating a specific transient reactive feedforward compensation amount.

[0166] The steps to obtain the priority control instruction are:

[0167] Based on the transient reactive feedforward compensation, the transient reactive feedforward compensation is inserted into the first position of the control component queue, and the priority label is set to the highest response level in the current period to generate a sorted control component sequence;

[0168] Based on the sorted control component sequence, the control strategy execution path with the highest priority is read, the target output parameter field and the control instruction type field encapsulated in the control component are parsed, and a control signaling data packet that complies with the inverter communication protocol standard is generated to form a priority control instruction.

[0169] Specifically, based on the transient reactive feedforward compensation generated in the previous step, which represents the optimal reactive power output gear value calculated to cope with the transient energy of the line under the island operation state, the system first encapsulates this transient reactive feedforward compensation as an independent control component into a data structure containing information such as the target reactive power value and execution time. Then, the system accesses and operates an internally maintained control component queue for managing various control tasks of the photovoltaic inverter. There may be other control components to be executed in this queue, such as conventional reactive power regulation instructions based on voltage deviation or active and reactive power coordinated control instructions based on power factor setting. The system will just encapsulate the transient reactive feedforward compensation. The control component is forcibly inserted at the front end of the control component queue to ensure that it is accessed first in subsequent processing. At the same time, the system sets a special priority tag for the newly inserted control component, which is assigned the highest response level among all available control tasks in the current period. For example, if the priority levels are divided into 1 to 5, with 1 being the highest, then this tag is set to 1. This is to ensure that under emergency island operation conditions, the transient reactive power compensation instruction can override all conventional control instructions and be executed with the highest priority. After the insertion and priority setting are completed, the system generates a reordered control component sequence with transient reactive power feedforward compensation as the primary task.

[0170] Based on the sorted control component sequence generated by the previous process, which already includes the highest priority transient reactive feedforward compensation, the system first reads the control component corresponding to the transient reactive feedforward compensation from the head of the sequence (i.e., the position with the highest priority). Then, based on the highest response level priority label attached to the control component, the system searches and calls the control strategy execution path that matches this highest level from the pre-configured control strategy library. The execution path specifies how to convert the abstract control target into a specific hardware operation instruction. Subsequently, the system analyzes in detail the key information fields encapsulated inside the extracted control component, including the target output parameter field, which specifies the specific value of the reactive power that the inverter needs to output (i.e., the aforementioned selected standard output file). The system has a bit value, for example, -2.5kVar), and a control instruction type field, which indicates the nature of this operation, for example, "reactive power setting" or "emergency reactive support mode switching". After obtaining these core parameters, the system will accurately assemble and encode these parameters in accordance with the data frame format specified by the protocol (including device address, function code, data area length, data content, checksum, etc.) based on the specific communication protocol standard followed by the photovoltaic inverter itself, such as ModbusRTU, ModbusTCP / IP or CAN bus and other industrial communication protocols. Finally, it will generate a control signaling data packet that fully complies with the inverter communication protocol standard and can be directly recognized and executed by the inverter. This data packet constitutes the priority control instruction.

[0171] The steps for obtaining the composite control instructions of the photovoltaic inverter are as follows:

[0172] When the island operation trigger state is false, the response signal message returned by the building automation system is received, the device response state field and the power change feedback field in the response signal message are parsed, the power change value of each responding device is classified and counted in chronological order, and a response state set of the flexible load scheduling excitation signal is generated;

[0173] Based on the response state set of the flexible load dispatch excitation signal, the real-time voltage sampling value of the grid connection point corresponding to each time point is extracted, and the difference between the real-time voltage sampling value and the target voltage value set before voltage regulation is calculated to determine whether the difference is within the voltage deviation tolerance range. The time points and corresponding voltage deviation amplitudes where the deviation exceeds the tolerance range are screened to generate the residual voltage deviation sequence.

[0174] Based on the residual voltage deviation sequence, the reactive power adjustment gear value corresponding to each voltage deviation interval is obtained, each residual voltage deviation amplitude is mapped to the closest reactive power adjustment gear, and all adjustment gear values ​​are embedded in the current inverter control data structure in chronological order to generate a photovoltaic inverter composite control instruction.

[0175] Specifically, when the island operation trigger state is judged to be false, that is, when the system confirms that the photovoltaic inverter is still in the grid-connected operation state, the system begins to actively receive and process the response signal messages returned from the building automation system (BMS) regularly or on demand. These messages are the feedback of the building automation system on the actual response of each flexible load device under its jurisdiction after executing the previously issued flexible load scheduling incentive signal. The system first parses each response signal message received, and accurately extracts the key device response status field according to the predefined message structure and communication protocol (implementing common building automation protocols such as BACnet, Modbus or KNX), which clearly indicates whether a specific flexible load device has successfully responded to the scheduling instruction. At the same time, the system also extracts the power change feedback field from the message, which records the specific value of the actual power consumption or output power change of the device after responding to the scheduling instruction. For example, if the instruction requires an air conditioner to increase the load by 2kW, the feedback field may show that the actual increase is 1.8kW. After completing the parsing of a single message, the system will classify and accumulate the power change feedback values ​​of all flexible load devices participating in this round of scheduling in strict accordance with the precise time sequence of their responses, and finally generate a structured response state set of the flexible load scheduling incentive signal that includes the response details of each device and the overall power regulation effect.

[0176] Based on the response state set of the flexible load dispatch excitation signal generated in the previous step, the set records in detail the actual power adjustment of each flexible load after the dispatch instruction is executed. The system then combines this information with the real-time voltage data of the grid connection point to evaluate the actual effect of flexible load dispatch on voltage correction. Specifically, the system extracts each time point when the flexible load power changes from the response state set, and retrieves the voltage sampling values ​​that correspond precisely to these time points and are collected in real time by the voltage sensors deployed at the grid connection point. Then, the system calculates the difference between these real-time voltage sampling values ​​and the target voltage value set before implementing the voltage regulation strategy. The target voltage value is usually the nominal voltage value of the distribution network (for example, 230V) or a desired ideal operating voltage point. The difference is the voltage deviation at the current moment. Then, the system determines whether the absolute value of the calculated voltage deviation is within a preset "voltage deviation tolerance range". The tolerance range is set according to the grid operation stability and equipment safety requirements. For example, the voltage deviation is allowed to be within the nominal voltage. For a 230V system, the allowable deviation range is If the absolute value of the voltage deviation is less than or equal to 4.6V, it is considered that the current voltage control effect is good and no further adjustment is required. On the contrary, if the absolute value of the voltage deviation exceeds this tolerance range, the system will record the time point and its corresponding specific voltage deviation amplitude (that is, the difference between the actual voltage and the target voltage), and screen and summarize them. Finally, all the time points where the deviation exceeds the tolerance range and the corresponding voltage deviation amplitudes are arranged in chronological order to generate the remaining voltage deviation sequence.

[0177] Based on the residual voltage deviation sequence generated in the previous step, the sequence indicates the time point and degree of deviation when the grid-connected point voltage still fails to fully recover to the ideal state after the initial adjustment of the flexible load. The system then needs to use the reactive power regulation capability of the photovoltaic inverter to further accurately correct these residual voltage deviations. To this end, the system first accesses a pre-configured "voltage deviation-reactive power adjustment gear mapping table". This mapping table is established based on the reactive regulation characteristics of the photovoltaic inverter, the local grid impedance parameters and historical operation data through simulation analysis or empirical rules (for example, every 1% voltage deviation corresponds to a reactive output of 2% of the rated capacity of the inverter). It divides different voltage deviation ranges (for example, the voltage deviation is greater than +4.6V but less than +7V, or the voltage deviation is less than -4.6V but greater than -7V, etc.) is associated with a specific reactive power adjustment gear value that the inverter can output (for example, outputting +1kVar reactive power, or outputting -1.5kVar reactive power, etc.). The system traverses each voltage deviation amplitude in the residual voltage deviation sequence and, based on the mapping table, accurately maps each residual voltage deviation amplitude to the reactive power adjustment gear that is closest to its value or that can most effectively compensate for the deviation. After completing the mapping of all residual voltage deviations to reactive power adjustment gears, the system embeds these selected reactive power adjustment gear values ​​into the corresponding parameter fields (for example, reactive power set value fields) in the control data structure to be executed by the current photovoltaic inverter in strict accordance with their corresponding time sequence, thereby dynamically updating the inverter's operating instructions and ultimately generating and preparing to issue a photovoltaic inverter composite control instruction.

[0178] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A photovoltaic inverter intelligent control method based on the Internet of Things, characterized in that: The following steps are involved: Based on the predicted sunlight value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set and the voltage within the window is serially estimated to establish an over-limit risk warning; Based on the over-limit risk warning, the predicted voltage over-limit amplitude is extracted, and the voltage regulation demand is calculated in combination with the current indoor and outdoor temperature parameters of the building to obtain a voltage absorption quantitative index. Based on the voltage absorption quantitative index, it is converted into an instruction containing virtual electricity price information and sent to the building automatic control system via the Internet of Things to obtain a flexible load scheduling incentive signal; Based on the status signal of the circuit breaker of the upper-level substation, determine whether the signal indicates the disconnected state and obtain the island operation trigger status; If the island operation trigger state is true, the transient energy of the line is calculated and the reactive power value required to absorb the energy is solved to establish the transient reactive feedforward compensation amount; Based on the transient reactive feedforward compensation amount, it is set as the highest priority control component and a priority control instruction is immediately generated; when the island operation trigger state is false, the response state of the building automatic control system to the flexible load scheduling excitation signal is received and the voltage correction effect is determined, and the reactive adjustment amount is set according to the residual voltage deviation to form a photovoltaic inverter composite control instruction; The steps for obtaining the flexible load scheduling excitation signal are as follows: Based on the voltage absorption quantitative index, the recommended voltage regulation power value corresponding to each time point is extracted, and a list of flexible load devices that are currently connected and whose operating time is not less than the minimum response period is counted. The response duration recorded in the last three rounds of historical operation logs is read, and a list of flexible load devices that can participate in regulation at each time point and response status information is generated by sorting according to time tags; Calculate the dynamic incentive electricity price for each flexible load device based on the voltage regulation power recommendation value, flexible load device list and response status information; Based on the dynamic incentive electricity price of each flexible load device, the dynamic incentive electricity price is matched with the equipment adjustment capability one by one to form a scheduling reference comparison table. The target devices with incentive electricity price greater than the benchmark trigger value are screened and the address, adjustment power instruction and incentive electricity price are encapsulated into an instruction structure and distributed to the corresponding building automatic control system to generate a flexible load scheduling incentive signal.

2. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the limit-crossing risk warning are as follows: Based on the predicted light value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set according to a fixed time length. The real-time voltage measurement value of the grid connection point and the corresponding light prediction value at consecutive moments within the time window are recorded to generate a mapping sequence of voltage measurement value and light prediction value with time tags; Based on the mapping sequence of voltage measurement values ​​and light prediction values ​​with time tags, the voltage measurement values ​​are arranged in chronological order to construct a voltage change trend sequence, the voltage increments between adjacent time points are calculated and accumulated in sequence to derive the voltage prediction value in the future time period, and form a voltage prediction sequence; Based on the voltage prediction sequence, each predicted voltage value in the voltage prediction sequence is compared one-to-one with the upper and lower limit thresholds of the distribution network voltage, and all predicted time points exceeding the upper threshold or falling below the lower threshold are marked. All marked time nodes and corresponding voltage values ​​are counted to generate an over-limit risk warning.

3. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the voltage absorption quantitative index are as follows: Based on the above-mentioned over-limit risk warning, the predicted voltage value is extracted from each marked prediction time point, and compared with the upper and lower voltage thresholds of the distribution network respectively. The portion exceeding the upper threshold is selected as the positive voltage over-limit value, and the portion below the lower threshold is selected as the negative voltage over-limit value. The voltage over-limit value sequence is obtained by integrating the values ​​in chronological order. Calculating a recommended total regulated power value based on the voltage over-limit value sequence, the building's nominal flexible load power at each time point, the building's indoor temperature parameters, and the building's outdoor temperature parameters; Based on the total regulated power recommendation value and combined with the dispatch capability threshold of the load response capability at each time point, it is checked whether the recommended power in each time period is within the load regulation capability range, and a voltage absorption quantitative index is generated.

4. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the island operation trigger status are: Extracting the status signal encoding content from the upper-level substation circuit breaker status signal, identifying the function bit identifier and logic level field in the encoding, parsing the control bit status indicating whether the circuit breaker is connected or disconnected in the logic level field, and extracting the upper-level substation circuit breaker status signal at the current moment; Based on the upper-level substation circuit breaker status signal, determine whether the control bit status field is a disconnected state indicating value; if the status field is consistent with the preset disconnected state value, confirm that the circuit breaker is currently in the disconnected state; otherwise, mark it as not disconnected, and obtain the circuit breaker disconnection determination result; Based on the circuit breaker disconnection judgment result, the voltage amplitude fluctuation trend of the current period is read, and when the circuit breaker disconnection judgment result is the disconnected state and the voltage fluctuation range is greater than the off-grid operation judgment threshold, the current moment is marked as the island operation trigger state.

5. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the transient reactive feedforward compensation amount are as follows: If the island operation trigger state is true, read the active output power value corresponding to the current time point, and simultaneously call the unit inductance parameter and unit capacitance parameter recorded in the current power line configuration file, calculate the sum of the corresponding inductance energy term and capacitance energy term, and generate the line transient energy value; Based on the transient energy value of the line, combined with the energy change curve trend map under similar operating conditions in the current photovoltaic inverter grid-connected operation history, a sample group with similar initial energy magnitude and system response period is retrieved, and the optimal reactive power matching value corresponding to the transient energy change of the line in the sample group in the same time period is calculated to generate the reactive power value required to absorb the transient energy; Based on the reactive power value required to absorb transient energy, the reactive power value is matched with the preset reactive power output gear, and the standard output gear value closest to the required reactive power value is selected as the control instruction target to generate a transient reactive feedforward compensation amount.

6. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the priority control instruction are: Based on the transient reactive feedforward compensation amount, the transient reactive feedforward compensation amount is inserted into the first position of the control component queue, the priority label is set to the highest response level in the current period, and a sorted control component sequence is generated; Based on the sorted control component sequence, the control strategy execution path with the highest priority is read, the target output parameter field and the control instruction type field encapsulated in the control component are parsed, and a control signaling data packet that complies with the inverter communication protocol standard is generated to form a priority control instruction.

7. The photovoltaic inverter intelligent control method based on the Internet of Things according to claim 1 is characterized in that: The steps for obtaining the photovoltaic inverter composite control instruction are as follows: When the island operation trigger state is false, the response signal message returned by the building automation system is received, the device response state field and the power change feedback field in the response signal message are parsed, the power change value of each responding device is classified and counted in chronological order, and a response state set of the flexible load scheduling excitation signal is generated; Extracting the real-time voltage sampling value of the grid connection point corresponding to each time point according to the response state set of the flexible load scheduling excitation signal, performing difference calculation on the real-time voltage sampling value of the grid connection point corresponding to the target voltage value set before voltage regulation, determining whether the difference is within the voltage deviation tolerance range, and screening the time points and corresponding voltage deviation amplitudes at which the deviation exceeds the tolerance range to generate a residual voltage deviation sequence; Based on the residual voltage deviation sequence, the reactive power adjustment gear value corresponding to each voltage deviation interval is obtained, each residual voltage deviation amplitude is mapped to the closest reactive power adjustment gear, and all adjustment gear values ​​are embedded in the current inverter control data structure in chronological order to generate a photovoltaic inverter composite control instruction.

8. The photovoltaic inverter intelligent control system according to the photovoltaic inverter intelligent control method based on the Internet of Things according to any one of claims 1 to 7, characterized in that: include: Voltage risk warning module: Based on the predicted sunlight value of the photovoltaic inverter and the real-time voltage measurement value of the grid connection point, a time window is set and the voltage within the window is serialized and estimated to establish an over-limit risk warning; Voltage Accommodation Control Module: Based on the over-limit risk warning, the module extracts the predicted voltage over-limit amplitude and calculates the voltage regulation demand in combination with the current indoor and outdoor temperature parameters of the building to obtain a voltage accommodation quantitative index. Based on the voltage accommodation quantitative index, the module converts it into an instruction containing virtual electricity price information and sends it to the building automation system via the Internet of Things to obtain a flexible load scheduling incentive signal. Island operation identification module: Based on the status signal of the upper-level substation circuit breaker, it determines whether the signal indicates a disconnected state and obtains the island operation trigger status; If the island operation trigger state is true, the transient energy of the line is calculated and the reactive power value required to absorb the energy is solved to establish the transient reactive feedforward compensation amount; Inverter integrated control module: Based on the transient reactive feedforward compensation amount, it is set as the highest priority control component and immediately generates a priority control instruction; when the island operation trigger state is false, it receives the response state of the building automatic control system to the flexible load scheduling excitation signal and judges the voltage correction effect, sets the reactive adjustment amount according to the residual voltage deviation, and forms a photovoltaic inverter composite control instruction.

Images