A digital management-based energy-saving control method for new energy substations

Abstract

This invention discloses an energy-saving control method for new energy substations based on digital management, belonging to the field of power system automation and energy management technology. The method includes: constructing a multi-source data fusion sensing layer to collect data such as main transformer operating parameters, reactive power compensation status, auxiliary system load, and ambient temperature and humidity; establishing a dynamic energy efficiency evaluation model with the goal of minimizing comprehensive energy consumption, integrating main transformer losses, reactive power switching losses, auxiliary system energy efficiency, and ambient temperature correction factors; generating a coordinated control strategy for main transformer tap positions, reactive power switching, and cooling system start / stop using a rolling optimization algorithm, and verifying safety boundaries and action frequency constraints; executing closed-loop control and online correction strategies based on deviation feedback; and adaptively updating model parameters using the recursive least squares method, combined with an energy efficiency knowledge base to improve response speed. This invention achieves accurate modeling of the entire substation's energy consumption and multi-factor coordinated optimization, significantly reducing the substation's electricity consumption while ensuring safety and new energy consumption.

Description

Technical Field

[0001] This invention relates to the field of power system automation and energy management technology, and in particular to an energy-saving control method for new energy substations based on digital management. Background Technology

[0002] With the rapid growth of installed capacity of new energy power generation, substations, as the core hub connecting distributed power sources and the main power grid, are undergoing a rapid evolution in their operation mode from traditional passive response to proactive intelligent control. Driven by the "dual-carbon" strategic goal, the power system places higher demands on energy efficiency management, urgently requiring the use of digital means to achieve efficient and low-carbon operation of substations throughout their entire lifecycle. Currently, the digital construction of substations has initially covered basic functions such as data acquisition, status monitoring, and remote control, but its core logic still focuses on equipment safety and power supply reliability, and a closed-loop control system oriented towards minimizing comprehensive energy consumption has not yet been formed. Especially in scenarios with high penetration of new energy sources, the intermittent access of power sources such as photovoltaics and wind power leads to frequent fluctuations in power flow within the substation. Key energy-consuming equipment such as main transformers, reactive power compensation devices, station AC systems, and ventilation and cooling units are in suboptimal operating conditions for extended periods, resulting in a large amount of ineffective energy consumption.

[0003] Among these, energy-saving control methods based on digital platforms aim to integrate multi-source operational data and environmental parameters to construct a dynamic energy efficiency optimization mechanism oriented towards the substation itself. The core of this approach lies in breaking through the traditional control paradigm of "emphasizing power consumption while neglecting energy conservation," incorporating variables such as transformer load rate, reactive power distribution, ambient temperature and humidity, and load forecast curves into a unified decision-making framework, driving coordinated equipment actions through real-time energy efficiency assessment. However, existing technologies have not effectively solved the problem of dynamic modeling and execution linkage under multi-factor coupling, resulting in a lack of foresight and adaptability in energy-saving strategies.

[0004] In existing technologies, some solutions introduce digital management platforms, but their functions are limited to renewable energy consumption or static drawing recognition, failing to support deep energy-saving operation of the substation itself. On the one hand, methods such as those described in CN114914929A, which suppress power flow backflow through demand-side response and mobile energy storage, improve renewable energy utilization but fail to model and optimize key energy consumption points within the substation, such as main transformer iron and copper losses, reactive power switching frequency, and auxiliary system start-up and shutdown logic, leading to a disconnect between control targets and energy-saving needs. On the other hand, automatic electrical drawing recognition systems, such as those proposed in CN120635933A, only complete the static digitization of the topology, neither accessing real-time operational data streams nor possessing energy efficiency analysis and control command generation capabilities, making it difficult to form an energy-saving closed loop of "perception-evaluation-decision-execution." Furthermore, existing energy efficiency models mostly use fixed weights or empirical thresholds, failing to establish dynamic evaluation functions coupled with real-time load, ambient temperature, and equipment aging status, resulting in insufficient robustness of control strategies under complex operating conditions, and potentially exacerbating equipment losses due to frequent adjustments or erroneous switching. Summary of the Invention

[0005] The purpose of this invention is to provide a digital management-based energy-saving control method for new energy substations, in order to solve the problems of one-sided energy efficiency assessment, lack of coordination in control, and insufficient model adaptability in the existing technology.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A digital management-based energy-saving control method for new energy substations includes: Step S1: Construct a multi-source data fusion perception layer: Real-time collection of main transformer operating parameters, reactive power compensation device status, station AC system load, ventilation and cooling unit start / stop signals, environmental temperature and humidity data, and new energy output and grid-side load prediction curves. Through unified timestamp alignment and data cleaning, a highly consistent operating dataset is formed. Step S2, establish a dynamic energy efficiency evaluation model: based on the aforementioned operating dataset, and combining the characteristics of main transformer iron and copper losses, reactive power switching loss function, auxiliary system energy efficiency curve and ambient temperature correction factor, construct a multi-dimensional dynamic energy efficiency evaluation function with the goal of minimizing comprehensive energy consumption; Step S3, Generate collaborative energy-saving control strategy: Based on the output results of the dynamic energy efficiency evaluation model, use the rolling optimization algorithm to solve the optimal control sequence of main transformer gear adjustment, reactive power compensation switching combination, and ventilation and cooling system start-up and shutdown sequence, and verify the equipment action frequency constraints and safe operation boundaries; Step S4, execute closed-loop control and feedback correction: send the optimal control sequence to each execution unit, synchronously monitor the actual energy consumption changes and equipment response status, and use the deviation feedback mechanism to make online corrections to the control strategy for the next cycle; Step S5, update energy efficiency model parameters: Based on historical control effects and equipment aging trends, the weight coefficients and loss parameters in the energy efficiency evaluation model are adaptively updated using the recursive least squares method to ensure the accuracy and robustness of the model in long-term operation.

[0007] In step S1, the operating parameters of the main transformer include voltage, current, active power, reactive power and oil temperature on the high-voltage side and low-voltage side. The sampling frequency meets the preset sampling requirements and the data accuracy meets the requirements of IEC 61850-9-2 standard.

[0008] In step S1, the ambient temperature and humidity data are acquired by a multi-point sensor network deployed in the main transformer room, power distribution room and outdoor equipment area. The spatial resolution meets the predetermined distance requirements and the time synchronization error is within the preset time error range.

[0009] In step S2, the characteristics of the main transformer's iron loss and copper loss are expressed using a piecewise linearized function, where the iron loss is proportional to the square of the voltage and the copper loss is proportional to the square of the load current. The proportionality coefficient is calibrated based on the equipment nameplate parameters and the measured no-load short-circuit test data.

[0010] The reactive power switching loss function in step S2 includes two parts: the energy consumption of the switching action and the steady-state reactive power loss. The former is proportional to the capacitor bank capacity and the number of switching operations, while the latter is proportional to the square of the reactive power and the line impedance. The total loss per unit time is expressed as... Where α is the energy consumption coefficient for a single switching operation, and β is the reactive power loss coefficient. Q represents the number of switching operations per unit time, and Q represents the reactive power.

[0011] The dynamic energy efficiency evaluation function in step S2 is defined as follows: ,in Mainly variable energy consumption, For reactive power system energy consumption, To aid in the system's energy consumption, This represents the absolute value of the ambient temperature deviating from the baseline value. to It is a dynamic weighting coefficient, whose value is adjusted in real time according to the load rate, the penetration rate of new energy sources and the health status of equipment.

[0012] In step S3, the rolling optimization algorithm uses a predetermined time period with a time step within a finite time domain, and a preset prediction interval for the prediction time domain. The optimization objective is... t ranges from 1 to 12, and the constraints include the main transformer load rate being within the preset load rate range, the bus voltage deviation not exceeding the preset voltage deviation threshold, the number of main transformer speed adjustments per day not exceeding the preset upper limit of the operating frequency, and the number of times the reactive power compensation device is switched on and off per day not exceeding the preset upper limit of the operating frequency.

[0013] The safe operation boundary verification in step S3 includes power flow calculation verification, short-circuit capacity verification, and equipment thermal stability verification, to ensure that the control strategy still meets the safe operation requirements under N-1 fault conditions.

[0014] In step S4, the deviation feedback mechanism compares the relative error between the actual energy consumption and the model's predicted energy consumption. When the absolute value of the error exceeds a preset error threshold, it triggers the online correction process for the model parameters and delays the issuance of the next cycle control command until the correction is completed.

[0015] In step S5, the forgetting factor of the recursive least squares method is set to a preset forgetting factor value, which is used to balance the contribution weights of historical data and new observation data, and ensure that the model has the ability to track the performance degradation caused by equipment aging.

[0016] The method also includes establishing an energy efficiency control knowledge base, storing the optimal control mode and its corresponding energy consumption index under typical operating conditions, and directly calling historical strategies to reduce computational latency when the real-time operating condition matches the mode in the knowledge base better than a preset matching threshold.

[0017] The method is integrated into the substation integrated monitoring platform and communicates with the intelligent electronic equipment in the station through the IEC 61850 MMS protocol. The transmission delay of the control command is less than the preset transmission delay threshold, and the status feedback cycle is not greater than the preset feedback cycle threshold.

[0018] Compared with the prior art, the beneficial technical effects of the present invention are as follows: This invention breaks through the limitations of traditional methods that only focus on renewable energy consumption or static drawing identification. For the first time, it incorporates main transformer losses, reactive power switching, auxiliary systems, and environmental factors into a unified quantitative framework. It achieves accurate characterization of energy consumption through a multi-dimensional dynamic energy efficiency evaluation function. This model abandons fixed weights and empirical thresholds and introduces a dynamic weight mechanism coupled with real-time operating conditions, making the energy efficiency assessment error lower than the preset error threshold, which is significantly better than the existing schemes that use fixed coefficient models.

[0019] This invention generates a joint control sequence for main transformer switching, reactive power switching, and auxiliary system start-up and shutdown using a rolling optimization algorithm, and embeds equipment operation frequency constraints and safety boundary checks to ensure that the overall energy consumption of the substation is minimized while guaranteeing power supply reliability and renewable energy absorption. Field tests show that this method can significantly reduce the average daily power consumption of the substation, while controlling the number of main transformer switching operations within the allowable range of the equipment, avoiding mechanical wear caused by excessive operation.

[0020] This invention uses recursive least squares to update energy efficiency model parameters online and combines this with an energy efficiency control knowledge base to accelerate strategy generation, effectively addressing complex operating conditions such as equipment aging, environmental changes, and load fluctuations. In long-term operational testing, the model's prediction accuracy remained above the preset accuracy threshold, the effective execution rate of the control strategy reached the preset execution rate threshold, and no safety incidents occurred due to strategy misjudgment. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the specific steps of a digital management-based energy-saving control method for new energy substations proposed in this invention. Figure 2 This is a schematic diagram illustrating the core principle framework for generating the dynamic energy efficiency evaluation model and collaborative energy-saving control strategy in this invention. Detailed Implementation

[0022] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely intended to explain the present invention and not to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present invention by illustrating examples of the invention.

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0024] Step S1 involves constructing a multi-source data fusion sensing layer, specifically including the following operational procedures: First, intelligent electronic devices deployed on the high-voltage and low-voltage sides of the main transformer collect real-time data on voltage, current, active power, reactive power, and top-layer oil temperature. The sampling frequency is set to 4000 Hz, and the data accuracy meets the Class 0.2 requirements of the IEC 61850-9-2 standard. All analog quantities are synchronously sampled via a merging unit and marked with a precise IEEE 1588v2 timestamp. Second, the status information of the reactive power compensation device is periodically reported by its internal controller via GOOSE messages, including the current switching group number, capacitor bank capacitance, switching relay status, and cumulative number of operations. Third, the station's AC system load data is collected by the station's smart meters at 1-second intervals, including three-phase voltage, current, power factor, and harmonic distortion rate, and transmitted via Modbus. The TCP protocol is uploaded to the station control layer; fourth, the start and stop signals of the ventilation and cooling unit are connected to the digital input module through the auxiliary contacts of the contactors of the fan and oil pump. The sampling period is 100 milliseconds, and the signal is converted into a Boolean state quantity after photoelectric isolation; fifth, the ambient temperature and humidity data are acquired by a distributed sensor network. The network has three integrated temperature and humidity sensors at the top, middle and bottom of the main transformer room, one sensor at every 5 meters in the 10 kV distribution room, and industrial-grade sensors with IP67 protection level deployed in the outdoor GIS equipment area at a grid density of 10 meters × 10 meters. All sensors have built-in RTC clocks and are synchronized with the station's main clock through the NTP protocol. The time synchronization error is controlled within ±1 millisecond, and the spatial resolution meets the 5-meter distance requirement; sixth, the new energy output prediction curve and the grid-side load prediction curve are issued by the dispatching master station through the power dispatching data network. The time granularity is 15 minutes, the prediction time domain covers the next 24 hours, and the data format adopts the CIM / XML standard. All the above raw data flows through the station's process layer switch and converges to the data access service module of the integrated monitoring platform. This module performs a unified timestamp alignment operation, using 100 milliseconds as the base time window, and performs linear interpolation or zero-order hold processing on asynchronously arriving data to ensure that all variables within the same time window have strict time consistency. Subsequently, a data cleaning process is performed, including outlier removal (using the 3σ criterion to identify and replace sampling points that exceed the mean ± 3 times the standard deviation), missing value imputation (using the moving average of adjacent time windows for imputation), and data normalization (mapping each physical quantity to the [0,1] interval), finally forming a structured, highly consistent running dataset. This dataset is stored in the form of a time series database, with each record containing a timestamp field and no less than 200 feature variables, and an update period of 100 milliseconds.

[0025] Step S2: Establish a dynamic energy efficiency evaluation model. Based on the highly consistent operating dataset formed in Step S1, and combining the iron and copper loss characteristics of the main transformer, the reactive power switching loss function, the energy efficiency curve of the auxiliary system, and the ambient temperature correction factor, construct a multi-dimensional dynamic energy efficiency evaluation function with the goal of minimizing comprehensive energy consumption. The iron and copper loss characteristics of the main transformer are expressed using a piecewise linearized function: iron loss... Calculated as copper loss Calculated as proportionality coefficient and According to the no-load loss in the equipment nameplate parameters With load loss Calibration is performed, and online correction is made based on on-site measured no-load short-circuit test data. When the load current is lower than 30% of the rated value, the low-load correction factor is activated. To compensate for nonlinear effects; reactive power switching loss function The total loss per unit time includes both switching energy consumption and steady-state reactive power loss. in, This is the energy consumption factor for a single switching operation, and its value is determined by the capacitor bank capacity. Decision, take =0.002 C, The reactive power loss factor is calculated as follows: , The equivalent resistance of the line. Bus voltage To calculate the time window length, The number of cuts per unit time. Reactive power; Auxiliary system energy efficiency curve The value is obtained by integrating the power consumption of the ventilation and cooling unit. It exhibits a non-linear relationship with ambient temperature and is retrieved from a pre-stored energy efficiency characteristic curve using a lookup table method. This curve is provided by the equipment manufacturer and calibrated through on-site testing during commissioning. Ambient temperature correction factor. Defined as the absolute deviation between the current ambient temperature and the reference temperature of 25°C. Based on this, the dynamic energy efficiency evaluation function... Defined as: in, Main transformer energy consumption equals The integral value within the calculation period, The energy consumption of the reactive system is... , To aid in the system's energy consumption, to The dynamic weighting coefficient has the following value selection mechanism: With the main transformer load rate Change, when hour ,when hour ,when hour ; With the penetration rate of new energy negative correlation Defined as the proportion of renewable energy output to total load. ; Affected by equipment health status index modulation, The results are derived from a comprehensive evaluation of state parameters such as insulation resistance and oil chromatography. ; The value is fixed at 0.1 to quantify the impact of environmental temperature control on overall energy efficiency. All weighting coefficients are updated every 5 minutes to ensure that the model reflects the current operating conditions in real time.

[0026] Step S3 generates a collaborative energy-saving control strategy. Specifically, based on the dynamic energy efficiency evaluation model results output in step S2, step S3 uses a rolling optimization algorithm to solve for the optimal control sequence of main transformer gear adjustment, reactive power compensation switching combination, and ventilation and cooling system start-up and shutdown timing. This rolling optimization algorithm is executed within a finite time domain, with a time step set to 15 minutes, and the prediction time domain covers the next 3 hours (i.e., 12 time steps). The optimization objective is to minimize the total comprehensive energy consumption within the prediction time domain. The solution process employs the Sequential Quadratic Programming (SQP) algorithm, with decision variables including the tap position of the main transformer. Reactive power compensation switching group combination vector ( (Number of capacitor banks), start / stop status of ventilation and cooling system. Constraints are strictly embedded in the optimization model: main transformer load rate. It must be within the preset load rate range of 30% to 80%; 10 kV bus voltage deviation The preset voltage deviation threshold is no more than ±5%; the preset action frequency limit is no more than 8 times per day for the main transformer to adjust its speed; and the preset action frequency limit is no more than 20 times per day for the reactive power compensation device to switch on and off. After obtaining the preliminary optimal control sequence, a safety operation boundary verification is performed. This verification includes three sub-processes: A) Verifying the power flow of the entire network based on the DC power flow model to ensure that the power of all branches does not exceed the limit; B) Performing short-circuit capacity verification, calculating the short-circuit current of key nodes and comparing it with the breaking capacity of circuit breakers, with a margin of no less than 15%; C) Conducting equipment thermal stability verification, performing dynamic simulation of the hot spot temperature of the main transformer winding to ensure that the temperature rise rate does not exceed 2℃ / minute and the peak temperature is below 120℃ under N-1 fault conditions. If any verification fails, a constraint relaxation mechanism is triggered, gradually relaxing non-critical constraints until a feasible solution is obtained, and finally outputting the optimal control sequence that satisfies all safety and equipment life constraints.

[0027] Step S4 involves executing closed-loop control and feedback correction. Specifically, step S4 distributes the optimal control sequence generated in step S3 to each execution unit via the station control layer communication service module. Control command transmission uses the IEC 61850 MMS protocol, transmitted via redundant dual networks, with a transmission delay less than a preset transmission delay threshold of 50 milliseconds. Main transformer tap-changing commands are sent to the on-load tap changer controller, reactive power switching commands are sent to the reactive power compensation device intelligent terminal, and ventilation and cooling start / stop commands are sent to the station power control system. Simultaneously with command issuance, the system initiates a monitoring process, collecting actual energy consumption change data and equipment response status at 1-second intervals: actual energy consumption is measured in real-time by high-precision metering, and equipment response status includes the actual tap position of the tap changer, capacitor bank switching confirmation signals, and fan operating current. The deviation feedback mechanism continuously operates to calculate actual energy consumption. Energy consumption predicted by the model relative error .when When the absolute value of the error exceeds a preset error threshold of 5%, the online correction process for model parameters is immediately triggered. This process suspends the generation of the control strategy for the next cycle and calls the parameter update module in step S5 to quickly correct the key parameters in the energy efficiency evaluation model. At the same time, the system delays the issuance of the control command for the next cycle until the correction is completed and the new model passes the internal consistency test. If the error does not exceed the limit, the control cycle for the next cycle is executed normally, forming a complete "decision-execution-monitoring-feedback" closed loop.

[0028] Step S5 updates the energy efficiency model parameters. Specifically, step S5 uses recursive least squares (RLS) to adaptively update the weight coefficients and loss parameters in the energy efficiency evaluation model based on historical control effect data and equipment aging trend indicators. The state vector of the RLS algorithm... Include , , , and to There are a total of 9 parameters, observation matrix The measured values ​​consist of input variables such as voltage, current, and temperature for the corresponding time period. This represents the integral value of actual energy consumption. Forgetting factor. A preset forgetting factor value of 0.98 was set, determined through offline simulation, to achieve a balance between tracking device performance degradation and suppressing measurement noise. After each adjustment cycle, the system collects new input-output data pairs and performs an RLS iterative update: first calculating the gain matrix... Then update the parameter estimates Finally, update the covariance matrix. In addition, the system establishes an energy efficiency control knowledge base, which stores the optimal control modes and their corresponding energy consumption indicators under typical operating conditions. Typical operating conditions are defined by a three-dimensional combination of load rate ranges (e.g., 0-30%, 30%-60%, 60%-100%), renewable energy output levels (low, medium, high), and ambient temperature ranges (<15℃, 15-30℃, >30℃), totaling 27 basic modes. When the matching degree (similarity calculated by Euclidean distance) between the real-time operating condition and a certain mode in the knowledge base is higher than a preset matching degree threshold of 90%, the system directly calls that historical strategy, skipping complex rolling optimization calculations, reducing the strategy generation time from 2 seconds to 200 milliseconds, and significantly reducing computational latency.

[0029] To further detail the implementation of this invention, a specific application example is constructed: A 110 kV renewable energy collection station is equipped with two 50 MVA main transformers, four sets of 10 Mvar capacitors, and two sets of main transformer air-cooling systems. At 14:00 on a typical summer day, the system detected a load rate of 65%, renewable energy output accounting for 40%, and an ambient temperature of 32℃. Step S1 collects data on the high-voltage side of the main transformer: voltage 115 kV, current 280 A, oil temperature 68℃, reactive power compensation currently in operation (two sets), station power load 180 kW, outdoor temperature 32℃, and humidity 60%. Step S2 calculates... , , , Dynamic weights , , , Therefore Step S3 performs rolling optimization. Within the next 3 hours, the algorithm suggests increasing the main transformer's gear from the current 9th to 10th to reduce excitation current and thus iron losses; maintaining two reactive power compensation groups in operation to avoid frequent switching; and keeping the air-cooled system fully operational due to oil temperature approaching the 70℃ warning line. Safety verification confirms the bus voltage will stabilize at 10.3 kV (qualified), short-circuit capacity margin is 22%, and thermal stability meets requirements. Step S4, after issuing the command, shows that within 15 minutes, the actual energy consumption drops to 168 kW, with a relative error of 4.1%, which does not exceed the threshold, therefore no correction is triggered. Step S5, at the end of the day, based on 24-hour data, the RLS algorithm will... The adjustment from 0.98 to 0.99 reflects slight aging of the main transformer core. The entire process resulted in a reduction of station power consumption of approximately 4.2%, and all equipment operations remained within their permissible lifespan.

[0030] Example 2 In another implementation scenario, the method of the present invention is applied to a 220 kV hub substation in a high-altitude area, which faces the challenges of large diurnal temperature differences and drastic load fluctuations. In step S1, the environmental sensor network is enhanced with air pressure monitoring to correct for the impact of air density on heat dissipation efficiency; in the dynamic energy efficiency evaluation function of step S2, the environmental temperature correction factor is expanded to... ,in This refers to the deviation of air pressure from standard atmospheric pressure. The altitude correction factor is used; the rolling optimization algorithm in step S3 introduces load fluctuation rate as an additional constraint. When the predicted load change rate exceeds 10% / 15 minutes, reactive power compensation is preferred over main transformer speed adjustment to avoid mechanical wear; the deviation feedback mechanism in step S4 sets a dual threshold: the error threshold is 5% under normal operating conditions, and relaxed to 8% under extreme weather (such as sandstorms) to avoid frequent corrections; the RLS algorithm in step S5 adopts a variable forgetting factor strategy, during periods of stable load... Periods of dramatic load changes To enhance tracking capabilities. During a sudden load drop on a winter day, the system lowered the main transformer speed by two levels, disconnected one set of capacitors, and shut down one wind turbine within 30 minutes, successfully reducing the station's power consumption from 320 kW to 275 kW, a reduction of 14.1%, while ensuring that the voltage remained stable within the acceptable range.

[0031] Example 3 In the third implementation scenario, the invention is integrated into the provincial power grid dispatch cloud platform to achieve multi-station collaborative energy-saving control. In this case, the data source in step S1 is expanded to all associated substations within the region, acquiring synchronous phasor data through a wide-area measurement system (WAMS); the dynamic energy efficiency evaluation model in step S2 adds a regional coordination term. ,in To address network losses caused by reactive power flow in the region, step S3's rolling optimization was upgraded to a distributed optimization architecture. Each station solved sub-problems locally and exchanged Lagrange multipliers using the ADMM algorithm to achieve global optimality. Step S4's instruction issuance was routed through an encrypted channel of the power dispatch data network, with the transmission delay requirement relaxed to 200 milliseconds but digital signature verification added. Step S5's parameter update introduced a federated learning mechanism, where each station only uploaded model gradients, not the original data, protecting data privacy. During the peak summer season, this model coordinated 10 substations, reducing regional station electricity consumption by 7.8% and simultaneously increasing renewable energy absorption capacity by 3.2%.

[0032] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for energy-saving control of new energy substations based on digital management, characterized in that, Includes the following steps: Step S1: Construct a multi-source data fusion perception layer to collect real-time main transformer operating parameters, reactive power compensation device status, station AC system load, ventilation and cooling unit start / stop signals, ambient temperature and humidity data, and new energy output and grid-side load prediction curves. A highly consistent operating dataset is formed through unified timestamp alignment and data cleaning. Step S2: Establish a dynamic energy efficiency evaluation model. Based on the running dataset, and combining the characteristics of main transformer iron and copper losses, reactive power switching loss function, auxiliary system energy efficiency curve and ambient temperature correction factor, construct a multi-dimensional dynamic energy efficiency evaluation function with the goal of minimizing comprehensive energy consumption. Step S3: Generate a collaborative energy-saving control strategy. Based on the output results of the dynamic energy efficiency evaluation model, use a rolling optimization algorithm to solve for the optimal control sequence of main transformer gear adjustment, reactive power compensation switching combination, and ventilation and cooling system start-up and shutdown timing, and verify the equipment action frequency constraints and safe operation boundaries. Step S4: Perform closed-loop control and feedback correction, distribute the optimal control sequence to each execution unit, synchronously monitor actual energy consumption changes and equipment response status, and use the deviation feedback mechanism to make online corrections to the control strategy for the next cycle. Step S5: Update the energy efficiency model parameters. Based on historical control effects and equipment aging trends, use the recursive least squares method to adaptively update the weight coefficients and loss parameters in the energy efficiency evaluation model.

2. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The operating parameters of the main transformer include voltage, current, active power, reactive power and oil temperature on the high-voltage side and low-voltage side. The sampling frequency meets the preset sampling requirements and the data accuracy meets the requirements of IEC 61850-9-2 standard.

3. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The ambient temperature and humidity data are acquired by a multi-point sensor network deployed in the main transformer room, power distribution room and outdoor equipment area. The spatial resolution meets the predetermined distance requirements and the time synchronization error is within the preset time error range.

4. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The iron and copper loss characteristics of the main transformer are expressed by a piecewise linearized function, in which the iron loss is proportional to the square of the voltage and the copper loss is proportional to the square of the load current. The proportionality coefficient is calibrated based on the equipment nameplate parameters and the measured no-load short-circuit test data.

5. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The reactive power switching loss function includes two parts: switching operation energy consumption and steady-state reactive power loss. The total loss per unit time is expressed as: Where α is the energy consumption coefficient for a single switching operation, and β is the reactive power loss coefficient. Q represents the number of switching operations per unit time, and Q represents the reactive power.

6. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The dynamic energy efficiency evaluation function is defined as follows: ,in Mainly variable energy consumption, For reactive power system energy consumption, To aid in the system's energy consumption, This represents the absolute value of the ambient temperature deviating from the baseline value. to It is a dynamic weighting coefficient, and its value is adjusted in real time according to the load rate, the penetration rate of new energy sources and the health status of equipment.

7. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The rolling optimization algorithm uses a predetermined time period with a finite time step within the time domain, and a preset prediction interval for the prediction time domain. The optimization objective is... t ranges from 1 to 12, and the constraints include the main transformer load rate being within the preset load rate range, the bus voltage deviation not exceeding the preset voltage deviation threshold, the number of main transformer speed adjustments per day not exceeding the preset upper limit of the operating frequency, and the number of times the reactive power compensation device is switched on and off per day not exceeding the preset upper limit of the operating frequency.

8. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The safety operation boundary verification includes power flow calculation verification, short-circuit capacity verification, and equipment thermal stability verification to ensure that the control strategy still meets the operation safety requirements under N-1 fault conditions.

9. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The deviation feedback mechanism compares the relative error between actual energy consumption and model-predicted energy consumption. When the absolute value of the error exceeds a preset error threshold, it triggers an online correction process for model parameters and delays the issuance of the next cycle control command until the correction is completed.

10. The energy-saving control method for new energy substations based on digital management according to claim 1, characterized in that, The method also includes establishing an energy efficiency control knowledge base, storing the optimal control mode and its corresponding energy consumption index under typical operating conditions, and directly calling historical strategies to reduce computational latency when the real-time operating condition matches the mode in the knowledge base better than a preset matching threshold.

Images