A method and system for carbon assessment and optimization of pumped storage power stations

Abstract

This invention discloses a carbon assessment and optimization method and system for pumped storage power stations, relating to the field of energy storage equipment control technology. The method includes collecting operational data from sensors deployed and integrated into the pumped storage power station and its surrounding environment, preprocessing the operational data to obtain formatted data, constructing a deep reinforcement learning model based on the formatted data and employing a deep deterministic policy gradient algorithm, and re-inputting the sensor operational data into the deep reinforcement learning model to obtain current and future carbon emission assessment results. This invention, through a novel optimization method combining deep reinforcement learning and sensor networks, solves the problems of traditional pumped storage power station optimization methods neglecting external environmental influences, lacking foresight, and exhibiting insufficient dynamic adaptability.

Description

Technical Field

[0001] This invention relates to the field of energy storage equipment control technology, and in particular to a method and system for carbon assessment and optimization of pumped storage power stations. Background Technology

[0002] With the increasing global demand for renewable energy, pumped storage hydroelectric power stations, as an important component of energy storage technology, are playing an increasingly prominent role. Pumped storage hydroelectric power stations effectively regulate the grid load by converting excess electrical energy into potential energy and storing it in an upper reservoir during periods of low electricity demand, and then releasing it as electrical energy during peak demand periods.

[0003] Existing optimization technologies for pumped storage power plants have improved their operational efficiency to some extent, but these methods generally have some limitations. First, traditional methods often focus on the direct control of internal equipment, neglecting the impact of external environmental factors (such as temperature and humidity) on plant operation, resulting in incomplete optimization solutions. Second, existing evaluation models mostly employ static analysis methods, unable to dynamically adapt to constantly changing operating conditions, which limits prediction accuracy and response speed. Furthermore, current optimization strategies are typically based on historical data analysis, lacking foresight and unable to effectively predict future carbon emission trends. Finally, existing technologies rarely consider the uncertainties and complexities of specific operations when implementing optimization strategies, which may lead to poor optimization results or even counterproductive effects. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a carbon assessment and optimization method and system for pumped storage power stations. It is a novel optimization method that combines deep reinforcement learning and sensor networks, which can monitor and adjust the operating parameters of the power station in real time, effectively reduce carbon emissions, and improve overall operating efficiency. This solves the problems of existing pumped storage power station optimization methods that ignore the impact of the external environment and fail to predict future carbon emission trends and adjust optimization strategies in real time.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention provides a method for carbon assessment and optimization of pumped storage power stations, comprising:

[0007] By deploying and collecting operational data from sensors in and around the pumped storage power station, and preprocessing the operational data, formatted data is obtained.

[0008] Based on the formatted data, a deep reinforcement learning model is constructed using the deep deterministic policy gradient algorithm.

[0009] The operational data from the sensor is re-inputted into the deep reinforcement learning model to obtain current and future carbon emission assessment results;

[0010] Based on the carbon emission assessment results, specific optimization strategies are formulated to obtain and implement optimized solutions for reducing carbon emissions.

[0011] The implementation effect of the optimization scheme is monitored, and the monitoring data is fed back to the deep reinforcement learning model.

[0012] As a preferred embodiment of the carbon assessment and optimization method for pumped storage power stations described in this invention, wherein:

[0013] The process involves collecting operational data from sensors deployed in and around the pumped-storage power station and its surrounding environment, and preprocessing the operational data to obtain formatted data. The specific steps are as follows:

[0014] Multiple sensors are installed in the pumped storage power station, each equipped with a LoRaWAN wireless communication module and transmitting operational data to a central server.

[0015] The runtime data is filtered using an algorithm to remove abnormal data points, and the timestamps of all runtime data are synchronized using a GPS clock.

[0016] Convert the timestamped runtime data into a unified JSON format to obtain runtime data in a unified format;

[0017] The unified format runtime data is deeply cleaned, missing values ​​are filled and deviations are corrected to obtain formatted data.

[0018] As a preferred embodiment of the carbon assessment and optimization method for pumped storage power stations described in this invention, wherein:

[0019] The specific steps for constructing a deep reinforcement learning model based on the formatted data and using the deep deterministic policy gradient algorithm are as follows:

[0020] The pressure reading P and temperature reading Temp are extracted from the formatted data as a state vector S. current Components;

[0021] The value Q(S) of state-action pairs is calculated using an Actor-Critic structure. current (Act), leading to the following formula:

[0022] Q(S current ,Act)=w1·P+w2·Temp+w3·Flow+w4·Act-w5·e -λ·τ (1)

[0023] Where: w1 is the weighting coefficient for adjusting pressure P, w2 is the weighting coefficient for adjusting temperature Temp, w3 is the weighting coefficient for adjusting water flow rate Flow, w4 is the weighting coefficient for adjusting action Act, w5 is the weighting coefficient for adjusting the influence of time decay term, λ is the weighting coefficient for controlling future rewards, and τ refers to the current moment.

[0024] The value Q(S) of the state-action pair is calculated. current After setting the Act values, initialize the weights of the Actor and Critic networks, and construct a deep reinforcement learning model using the deep deterministic policy gradient algorithm.

[0025] As a preferred embodiment of the carbon assessment and optimization method for pumped storage power stations described in this invention, wherein:

[0026] The specific steps for re-inputting the sensor's operational data into the deep reinforcement learning model to obtain current and future carbon emission assessment results are as follows:

[0027] The pressure reading P, temperature reading Temp, and water flow rate are extracted from the sensor's operating data and combined into a state vector S. current = [P, Temp, Flow];

[0028] Using the state vector S current The optimal action Act is generated using the Actor network, and then substituted into formula (1) to obtain Q(S) again. current ,Act);

[0029] By re-obtaining Q(S) current Substituting Act into the random forest algorithm yields current and future carbon emission assessment results.

[0030] As a preferred embodiment of the carbon assessment and optimization method for pumped storage power stations described in this invention, wherein:

[0031] The specific steps for formulating and implementing a specific optimization strategy based on the carbon emission assessment results are as follows:

[0032] By comparing the carbon emission values ​​over the time period based on the current and future carbon emission assessments, the maximum value is taken.

[0033] Based on the time period of maximum carbon emission values, identify high carbon emission operations or conditions, and analyze and extract pressure reading P, temperature reading Temp, and water flow rate;

[0034] By modifying the corresponding operation values ​​based on the obtained pressure reading P, temperature reading Temp, and water flow rate Flow, an optimized scheme for reducing carbon emissions can be obtained.

[0035] Modify and implement the optimized scheme for reducing carbon emissions.

[0036] As a preferred embodiment of the carbon assessment and optimization method for pumped storage power stations described in this invention, wherein:

[0037] The specific steps for monitoring the implementation effect of the optimization scheme and feeding back the monitoring data to the deep reinforcement learning model are as follows:

[0038] Obtain pressure reading P, temperature reading Temp, and water flow rate from the sensors after the optimization scheme is implemented;

[0039] The pressure reading P, temperature reading Temp, and water flow rate after implementation will be reassembled into S. current = [P, Temp, Flow];

[0040] The reconstructed S current =The [P,Temp,Flow] values ​​are passed through the Actor network to generate a new optimal action Act, and the new optimal action Act and S are then compared. current =[P,Temp,Flow] are input into a deep reinforcement learning model to calculate the value Q(S) of the state-action pair. current ,Act);

[0041] Based on the new data and the calculated value Q(S) of the state-action pairs current The gradient descent algorithm is used to update the weights of the Actor and Critic networks.

[0042] Repeat the above process periodically to continuously optimize the parameters and strategies of the deep reinforcement learning model, ensuring that the system is always in the best operating state.

[0043] Secondly, the present invention provides a carbon assessment and optimization system for pumped storage power stations, comprising:

[0044] The data collection and preprocessing module collects operational data from sensors deployed and collected from the pumped storage power station and its surrounding environment, and preprocesses the operational data to obtain formatted data.

[0045] The deep reinforcement learning model building module constructs a deep reinforcement learning model based on the formatted data and in conjunction with the deep deterministic policy gradient algorithm.

[0046] The carbon emission assessment module re-inputs the operational data from the sensors into the deep reinforcement learning model to obtain current and future carbon emission assessment results.

[0047] The optimization strategy formulation and implementation module formulates specific optimization strategies based on the carbon emission assessment results, obtains and implements optimized solutions for reducing carbon emissions;

[0048] The monitoring and feedback module monitors the implementation effect of the optimization scheme and feeds back the monitoring data to the deep reinforcement learning model.

[0049] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the carbon assessment and optimization method for pumped storage power stations as described in the first aspect of the present invention.

[0050] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the carbon assessment and optimization method for pumped storage power stations as described in the first aspect of the present invention.

[0051] The beneficial effects of this invention are as follows: By combining deep reinforcement learning with sensor networks, real-time monitoring and dynamic response to external environmental factors (such as temperature and humidity) are achieved, improving the comprehensiveness and accuracy of the optimization scheme. The use of dynamic analysis instead of traditional static methods significantly improves prediction accuracy and response speed, enhancing the system's foresight and adaptability. In particular, this invention fully considers the uncertainties and complexities in operation, effectively avoiding the problem of poor optimization results. In summary, this invention not only significantly improves power plant operating efficiency but also effectively reduces carbon emissions, supports the efficient utilization of clean energy, and has significant technological advancements and broad application prospects. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart of the carbon assessment and optimization method for pumped storage power stations in Example 1. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0057] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a method for carbon assessment and optimization of pumped storage power stations, including the following steps:

[0058] S1. By deploying and collecting operational data from sensors in and around the pumped storage power station, and preprocessing the operational data, formatted data is obtained.

[0059] Specifically, multiple sensors are installed in the pumped storage power station, each equipped with a LoRaWAN wireless communication module and transmitting operational data to a central server.

[0060] It should be noted that, based on the power station's operational parameter requirements, sensors capable of measuring key indicators such as pressure, temperature, and water flow rate were selected. These sensors were strategically deployed within and around the pumped storage power station to ensure coverage of all critical areas. Each sensor was equipped with a LoRaWAN wireless communication module to enable data transmission with the central server. Parameters such as the LoRaWAN module's operating frequency and transmit power were configured to ensure stable and reliable long-distance data transmission. By employing LoRaWAN technology, comprehensive monitoring of the power station environment was achieved, guaranteeing the real-time performance and reliability of data transmission. This is crucial for dynamically adjusting the power station's operational status, especially in complex terrain and widely distributed power station environments. LoRaWAN provides a low-power, long-distance data transmission solution, reducing wiring costs and complexity, and improving data acquisition efficiency.

[0061] Specifically, filtering algorithms are used to remove abnormal data points from the running data, and the timestamps of all running data are synchronized using GPS clocks.

[0062] It should be noted that algorithms such as Kalman filtering or median filtering are used to remove outliers from the raw data, ensuring data accuracy. Using GPS clocks to synchronize the timestamps of all sensor data ensures consistency across sensors. The pre-processed data undergoes initial cleaning and conversion to a unified format. Precise data cleaning and time synchronization significantly improve the foundational quality of subsequent analysis, reduce decision-making errors caused by data inaccuracies, and enhance the overall system's reliability and accuracy. For example, when building deep reinforcement learning models, high-quality data input can significantly improve the model's learning performance and prediction accuracy.

[0063] Specifically, the runtime data with timestamps is converted into a unified JSON format.

[0064] It should be noted that a JSON format template suitable for power plant operation data storage and exchange was designed, including field names, types, and hierarchical relationships. Preprocessed data was mapped and transformed according to the defined JSON template to generate a unified format data file. The transformed data was stored on a central server and backed up regularly to prevent data loss. This unified data format simplifies data management and sharing processes, enhances system compatibility, and enables seamless integration of data from different sources, improving work efficiency. Especially in the process of multi-source data integration, the unified data format reduces the complexity of data conversion, speeds up processing, and thus accelerates the response speed of the entire optimization process.

[0065] Specifically, the uniformly formatted runtime data is deeply cleaned, missing values ​​are filled and deviations are corrected to obtain formatted data.

[0066] It should be noted that by identifying the locations of missing values ​​in the dataset through data analysis tools, and then filling in the missing values ​​using methods such as linear interpolation or mean imputation based on trends in adjacent time periods or other relevant variables, and by comparing historical data with current data and using statistical methods (such as standard deviation) to detect and correct potential biases, the deep cleaning process further improves data quality. This ensures that the data input into the model is as accurate as possible, thereby enhancing the effectiveness of optimization strategies developed based on this data. For example, in adjusting power plant operating parameters, optimization strategies developed based on high-quality data will be more precise and effective, ultimately achieving the goal of reducing carbon emissions. Furthermore, high-quality data is crucial for improving the performance of deep reinforcement learning models, as it directly affects the model's predictive ability and reliability.

[0067] S2. Based on formatted data, a deep reinforcement learning model is constructed using a deep deterministic policy gradient algorithm.

[0068] Specifically, pressure reading P and temperature reading Temp are extracted from the formatted data as a state vector S. currentComponents of.

[0069] It should be noted that pressure and temperature readings are selected from the preprocessed formatted data as part of the state vector. These parameters are key factors affecting power plant operating efficiency and carbon emissions. The selected data points (such as pressure and temperature readings) are standardized to ensure that data of different magnitudes can be uniformly input into the model. Standardization can be achieved by subtracting the mean and dividing by the standard deviation. The standardized data points are combined into a unified state vector for subsequent input into the deep reinforcement learning model. By selecting pressure and temperature readings as the core components of the state vector, this invention directly models the most critical factors in power plant operation. This selection not only simplifies the model's complexity but also improves its effectiveness. Existing technologies often neglect the impact of external environmental factors on power plant operation, resulting in incomplete optimization schemes. This invention, by introducing these key parameters, achieves real-time monitoring and dynamic response to the external environment, thereby formulating more accurate optimization schemes.

[0070] Specifically, the value Q(S) of the state-action pair is calculated using the Actor-Critic structure. current (Act), leading to the following formula:

[0071] Q(S current ,Act)=w1·P+w2·Temp+w3·Flow+w4·Act-w5·e -λ·τ ,

[0072] Where: w1 is the weighting coefficient for adjusting pressure P, w2 is the weighting coefficient for adjusting temperature Temp, w3 is the weighting coefficient for adjusting water flow rate Flow, w4 is the weighting coefficient for adjusting action Act, w5 is the weighting coefficient for adjusting the influence of time decay term, λ is the weighting coefficient for controlling future rewards, and τ refers to the current moment.

[0073] It should be noted that after calculating the value of the state-action pair according to the above formula, the weights of the Actor and Critic networks are initialized. This step ensures that the model can learn from the initial state and gradually optimize its predictive ability. Historical data is used for initial training of the model, and the weight coefficients are adjusted to adapt it to the actual application scenario. Utilizing the Actor-Critic structure, this invention can provide optimal action suggestions while evaluating the current state. This method overcomes the problems of slow response speed and low prediction accuracy of traditional static analysis methods. Traditional optimization methods are usually based on historical data analysis and lack foresight. This invention significantly improves the foresight and adaptability of the system by dynamically calculating the value of the state-action pair, making the optimization strategy more flexible and efficient.

[0074] Specifically, the value Q(S) of the state-action pair is calculated. current After initializing the values ​​of Act, the weights of the Actor and Critic networks are determined, and a deep reinforcement learning model is constructed in conjunction with the deep deterministic policy gradient algorithm.

[0075] It should be noted that the DDPG algorithm was used to train the initially constructed model. DDPG is a reinforcement learning method that combines an Actor-Critic architecture with deterministic policy gradients, and is particularly suitable for dealing with problems in continuous action spaces.

[0076] Parameter tuning: During training, model parameters, including weight coefficients and learning rate, are continuously adjusted to improve the model's accuracy and stability. The model's effectiveness is verified through simulation experiments or actual test data to ensure its accurate prediction of the power plant's operating status and carbon emissions. The application of the DDPG algorithm enables this invention to maintain efficient decision-making capabilities even in complex power plant operating environments. This algorithm can not only handle high-dimensional state spaces but also effectively address the challenges of continuous action spaces. Existing technologies rarely consider the uncertainties and complexities of specific operations when implementing optimization strategies, which may lead to poor optimization results or even counterproductive effects. This invention, by introducing the DDPG algorithm, fully considers the uncertainties and complexities of operations, effectively avoiding these problems and improving optimization performance.

[0077] S3. Re-input the sensor's operating data into the deep reinforcement learning model to obtain current and future carbon emission assessment results.

[0078] Specifically, pressure reading P, temperature reading Temp, and water flow rate are extracted from the sensor's operating data and combined into a state vector S. current =[P,Temp,Flow].

[0079] It should be noted that pressure readings, temperature readings, and water flow rates are selected as the core components of the state vector from the preprocessed formatted data. These parameters are key factors affecting power plant operating efficiency and carbon emissions. The selected data points are standardized to ensure that data of different magnitudes can be uniformly input into the model. Standardization can be achieved by subtracting the mean and dividing by the standard deviation to generate the state vector: the standardized pressure readings, temperature readings, and water flow rates are combined into a unified state vector. This state vector will serve as the input to the deep reinforcement learning model. By selecting pressure readings, temperature readings, and water flow rates as the core components of the state vector, this invention directly models the most critical factors in power plant operation. This selection not only simplifies the model's complexity but also improves its effectiveness. Existing technologies often neglect the impact of external environmental factors on power plant operation, resulting in incomplete optimization schemes. This invention, by introducing these key parameters, achieves real-time monitoring and dynamic response to the external environment, thereby formulating more accurate optimization schemes.

[0080] Specifically, using the state vector S current The optimal action Act is generated using an Actor network, and then substituted into the above formula to obtain Q(S) again. current Act).

[0081] It should be noted that the generated state vector is input into the Actor network, which is responsible for generating optimal action suggestions. Based on the input state vector, the Actor network outputs a continuous action value. This action value represents the best action to be taken in the current state (such as adjusting pressure, temperature, or water flow rate). Substituting the generated action value into the previously defined formula, the value of the state-action pair is recalculated.

[0082] Q(S current ,Act)=w1·P+w2·Temp+w3·Flow+w4·Act-w5·e -λ·τ ,

[0083] In this method, w1 represents the weighting coefficient for adjusting pressure P, w2 represents the weighting coefficient for adjusting temperature Temp, w3 represents the weighting coefficient for adjusting flow rate Flow, w4 represents the weighting coefficient for adjusting action Act, w5 represents the weighting coefficient for adjusting the influence of the time decay term, λ represents the weighting coefficient for controlling future rewards, and τ represents the current moment. The optimal action is generated using an Actor network. This method overcomes the problems of slow response speed and low prediction accuracy of traditional static analysis methods. By dynamically calculating the value of state-action pairs, the system's foresight and adaptability are significantly improved. Traditional optimization methods are usually based on historical data analysis and lack foresight. This invention significantly improves the system's foresight and adaptability by dynamically calculating the value of state-action pairs, making the optimization strategy more flexible and efficient.

[0084] Specifically, by Q(S) current Substituting Act into the random forest algorithm yields current and future carbon emission assessment results.

[0085] It should be noted that the calculated value Q(S) of the state-action pair current The random forest model, using features (Act) as one of its input features, outputs current and future carbon emission assessments. These results help decision-makers understand the impact of current operations on carbon emissions and formulate optimization strategies accordingly. By combining deep reinforcement learning models and the random forest algorithm, this invention can not only dynamically adjust the power plant's operating status but also accurately predict carbon emission trends. This method fully utilizes the advantages of both models, providing a more comprehensive optimization solution. Existing technologies rarely consider the uncertainties and complexities in specific operations when implementing optimization strategies, which may lead to poor optimization results or even adverse effects. This invention, by introducing the random forest algorithm, fully considers the uncertainties and complexities in operations, effectively avoiding these problems and improving optimization results. Furthermore, the random forest algorithm has advantages in handling high-dimensional data and nonlinear relationships, providing more accurate carbon emission predictions, thereby supporting more effective optimization strategy formulation.

[0086] S4. Based on the carbon emission assessment results, formulate specific optimization strategies, obtain optimized solutions for reducing carbon emissions, and implement them.

[0087] Specifically, the carbon emission values ​​over the time period are compared using the current and future carbon emission assessment results, and the maximum value is taken.

[0088] It should be noted that the current and future carbon emission assessment results obtained from the random forest algorithm are typically presented in time series form. The entire assessment period is divided into multiple fixed time periods (e.g., hourly, daily, or weekly) for segmented analysis. Within each time period, the total carbon emissions are calculated. This can be achieved by accumulating all carbon emission data points within that time period, comparing the carbon emission values ​​across different time periods, identifying the maximum value and its corresponding time period, and through detailed analysis of carbon emissions in each time period, high-carbon emission periods can be accurately identified, providing clear targets for subsequent optimization strategies. Existing technologies often rely on historical data analysis and lack detailed analysis of specific time periods. This invention, by comparing carbon emissions across different time periods, can more accurately identify high-carbon emission periods, thereby enabling the development of more targeted optimization strategies.

[0089] Specifically, high-carbon-emission operations or conditions are identified based on the time period of maximum carbon emission values, and pressure readings P, temperature readings Temp, and water flow rates are analyzed and extracted.

[0090] It should be noted that, based on the time period of maximum carbon emissions determined in the previous step, further analysis of the operational data within this period is conducted. Key parameters such as pressure readings, temperature readings, and water flow rates are extracted from this data. These parameters are the main factors affecting the power plant's operating efficiency and carbon emissions. Combined with the power plant's operation logs and other relevant information, the specific operations or conditions leading to high carbon emissions are analyzed, such as excessive equipment load or cooling system malfunctions. By analyzing the specific operations and conditions during periods of high carbon emissions in detail, the root causes of high carbon emissions can be accurately identified. This fine-grained analysis helps to formulate more targeted optimization measures. Existing technologies typically rely on global data analysis, making it difficult to accurately identify the causes of localized high carbon emissions. This invention, by focusing on the specific operations and conditions during periods of high carbon emissions, can more accurately discover the root causes of problems, thereby formulating more effective optimization strategies.

[0091] Specifically, by modifying the corresponding operation values ​​based on the obtained pressure reading P, temperature reading Temp, and water flow rate Flow, an optimized scheme for reducing carbon emissions is obtained.

[0092] It should be noted that a clear carbon emission reduction target is defined, such as a 10% reduction. Based on extracted key parameters (pressure readings, temperature readings, and water flow rate), specific adjustment suggestions are proposed. For example, appropriately reducing pressure or increasing cooling water flow rate. Simulation tools or models are used to verify the proposed adjustment schemes, ensuring they can effectively reduce carbon emissions and maintain the normal operation of the power plant. Based on the simulation results, detailed optimization schemes are generated, including specific adjustment parameters and implementation steps. This method, which proposes specific optimization schemes based on adjustments to key parameters, is not only targeted but also ensures the effectiveness and feasibility of the optimization measures. Existing technologies often lack specific optimization measures, leading to poor practical application results. This invention, through detailed analysis and simulation verification, can generate practical optimization schemes, significantly improving the optimization effect.

[0093] Specifically, the equipment settings are modified and implemented according to the optimized plan for reducing carbon emissions.

[0094] It should be noted that, based on the generated optimization scheme, a detailed implementation plan is formulated, including specific steps for equipment adjustment, personnel arrangements, and timelines. The settings of relevant equipment are gradually adjusted according to the plan, such as regulating pressure valves and adjusting the cooling system. During implementation, the operating status and carbon emissions of the power plant are continuously monitored, and potential problems are promptly identified and addressed. Based on the monitoring data, the actual effectiveness of the optimization scheme is evaluated, and further adjustments and improvements are made as needed. Through systematic implementation and real-time monitoring, the optimization measures are ensured to be successfully implemented and achieve the expected results. This closed-loop control mechanism can continuously optimize and improve the optimization scheme, ensuring long-term effectiveness. Existing technologies rarely consider the uncertainties and complexities in specific operations when implementing optimization strategies, which may lead to poor optimization results or even adverse effects. This invention, through real-time monitoring and feedback mechanisms, can effectively address these uncertainties and ensure the successful implementation of the optimization scheme.

[0095] S5. Monitor the implementation effect of the optimization scheme and feed the monitoring data back to the deep reinforcement learning model.

[0096] Specifically, the pressure reading P, temperature reading Temp, and water flow rate are obtained from the sensors after the optimization scheme is implemented.

[0097] It should be noted that after the optimization scheme is implemented, sensors installed in and around the pumped storage power station collect key parameters such as pressure, temperature, and water flow rate in real time. The collected data is transmitted to a central server using a LoRaWAN wireless communication module, ensuring the real-time nature and accuracy of the data. Preliminary processing of the transmitted raw data is performed, including removing outliers and synchronizing timestamps, to ensure the quality of subsequent analysis. By collecting and transmitting operational data after the optimization scheme is implemented in real time, this invention can promptly understand the actual operating status of the power station, providing a basis for subsequent feedback and adjustments. Existing technologies often lack real-time monitoring of the effectiveness of optimization schemes, leading to the inability to promptly identify and correct potential problems. This invention, through real-time data acquisition, ensures the effectiveness of the optimization scheme and enables adjustments to be made immediately.

[0098] Specifically, the pressure reading P, temperature reading Temp, and water flow rate after implementation will be recombined into S. current =[P,Temp,Flow].

[0099] It should be noted that pressure readings, temperature readings, and water flow rates are selected as core components of the state vector from the initially processed data. The selected data points are standardized to ensure that data of different magnitudes can be uniformly input into the model. Standardization can be achieved by subtracting the mean and dividing by the standard deviation. The standardized pressure readings, temperature readings, and water flow rates are combined into a unified state vector for subsequent input into the deep reinforcement learning model. By reconstructing the state vector, this invention can accurately reflect the actual operating state after the optimization scheme is implemented, providing precise data support for subsequent feedback and adjustments. Existing technologies often neglect the impact of external environmental factors on power plant operation, resulting in incomplete optimization schemes. This invention, by introducing these key parameters, achieves real-time monitoring and dynamic response to the external environment, thereby formulating more accurate optimization schemes.

[0100] Specifically, the reconstructed S current =The [P,Temp,Flow] values ​​are passed through the Actor network to generate the optimal action Act, and the new optimal action Act and S are then compared. current =[P,Temp,Flow] are input into a deep reinforcement learning model to calculate the value Q(S) of the state-action pair. current Act).

[0101] It should be noted that the generated state vector is input into the Actor network, which is responsible for generating optimal action suggestions. Based on the input state vector, the Actor network outputs a continuous action value. This action value represents the best action to be taken in the current state (such as adjusting pressure, temperature, or water flow rate). Substituting the generated action value into the previously defined formula, the value Q(S) of the state-action pair is recalculated. current By using an Actor network to generate optimal actions and combining this with the value calculation of state-action pairs, this invention can dynamically adjust the operating state of a power plant, improving the system's adaptability. Traditional methods are usually based on fixed rules for optimization, lacking flexibility. This invention significantly improves the system's foresight and adaptability by dynamically calculating the value of state-action pairs, making the optimization strategy more flexible and efficient.

[0102] Specifically, based on the new data and calculation results, the gradient descent optimization algorithm is used to update the weights of the Actor and Critic networks.

[0103] It should be noted that, based on new data and calculation results, the error between the predicted and actual values ​​of the Actor and Critic networks is calculated, and weights are assigned to minimize the prediction error. This process is repeated, continuously updating the network weights until the predetermined accuracy requirement or number of training iterations is reached. By using a gradient descent optimization algorithm to update the network weights, this invention can continuously improve the model's predictive ability and stability, ensuring its high efficiency even in complex environments. Existing technologies often struggle to maintain model stability and accuracy in complex and ever-changing operating environments. This invention effectively addresses this challenge and enhances the model's robustness by continuously optimizing the network weights.

[0104] Specifically, the above process is repeated periodically to continuously optimize the parameters and strategies of the deep reinforcement learning model, ensuring that the system is always in the best operating state.

[0105] It should be noted that the time period for regularly repeating the above process is determined, such as evaluating and adjusting once per hour, day, or week. Through automated scripts or programs, steps such as data acquisition, state vector generation, optimal action calculation, and network weight updates are executed automatically on a regular basis. Within each cycle, the deep reinforcement learning model parameters and optimization strategies are continuously optimized to ensure the system is always in its optimal operating state. By regularly repeating the above process, this invention enables the system to achieve self-optimization and continuous improvement, ensuring its high-efficiency operation over the long term. Existing technologies often lack continuous optimization mechanisms, making it difficult to maintain optimization effects in the long term. This invention, through regular repetition and continuous optimization, ensures the long-term stability and high efficiency of the system, significantly improving the overall operating efficiency of the power plant.

[0106] This embodiment also provides a carbon assessment and optimization system for pumped storage power stations, including:

[0107] The data collection and preprocessing module collects operational data from sensors deployed and collected from the pumped storage power station and its surrounding environment, and preprocesses the operational data to obtain formatted data.

[0108] The deep reinforcement learning model building module constructs deep reinforcement learning models based on formatted data and in conjunction with the deep deterministic policy gradient algorithm.

[0109] The carbon emission assessment module re-inputs the sensor's operational data into the deep reinforcement learning model to obtain current and future carbon emission assessment results.

[0110] The optimization strategy formulation and implementation module formulates specific optimization strategies based on the carbon emission assessment results, obtains optimized solutions for reducing carbon emissions, and implements them.

[0111] The monitoring and feedback module monitors the effectiveness of the optimization scheme and feeds the monitoring data back to the deep reinforcement learning model.

[0112] This embodiment also provides a computer device applicable to the carbon assessment and optimization method for pumped storage power stations, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the carbon assessment and optimization method for pumped storage power stations as proposed in the above embodiment.

[0113] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0114] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the carbon assessment and optimization method for pumped storage power stations as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0115] In summary, this invention achieves real-time monitoring and dynamic response to external environmental factors (such as temperature and humidity) by combining deep reinforcement learning with sensor networks, thereby improving the comprehensiveness and accuracy of the optimization scheme. By employing dynamic analysis instead of traditional static methods, it significantly improves prediction accuracy and response speed, enhancing the system's foresight and adaptability. In particular, this invention fully considers the uncertainties and complexities of operation, effectively avoiding the problem of poor optimization results. In conclusion, this invention not only significantly improves power plant operating efficiency but also effectively reduces carbon emissions, supports the efficient utilization of clean energy, and has significant technological advancements and broad application prospects.

[0116] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for carbon assessment and optimization of pumped storage power stations, characterized in that, include: By deploying and collecting operational data from sensors in and around the pumped storage power station, and preprocessing the operational data, formatted data is obtained. Based on the formatted data, a deep reinforcement learning model is constructed using the deep deterministic policy gradient algorithm. The operational data from the sensor is re-inputted into the deep reinforcement learning model to obtain current and future carbon emission assessment results; Based on the carbon emission assessment results, specific optimization strategies are formulated to obtain and implement optimized solutions for reducing carbon emissions. The implementation effect of the optimization scheme is monitored, and the monitoring data is fed back to the deep reinforcement learning model. The specific steps for constructing a deep reinforcement learning model based on the formatted data and using the deep deterministic policy gradient algorithm are as follows: The pressure reading P and temperature reading Temp are extracted from the formatted data as a state vector S. current Components; The value Q(S) of state-action pairs is calculated using an Actor-Critic structure. current (Act), leading to the following formula: Q(S current ,Act)=w1·P+w2·Temp+w3·Flow+w4·Act-w5·e -λ·τ (1) Where: w1 is the weighting coefficient for adjusting pressure P, w2 is the weighting coefficient for adjusting temperature Temp, w3 is the weighting coefficient for adjusting water flow rate Flow, w4 is the weighting coefficient for adjusting action Act, w5 is the weighting coefficient for adjusting the influence of time decay term, λ is the weighting coefficient for controlling future rewards, and τ refers to the current moment. The value Q(S) of the state-action pair is calculated. current After setting the Act values, initialize the weights of the Actor and Critic networks, and construct a deep reinforcement learning model using the deep deterministic policy gradient algorithm.

2. The carbon assessment and optimization method for pumped storage power stations as described in claim 1, characterized in that: The process involves collecting operational data from sensors deployed in and around the pumped-storage power station and its surrounding environment, and preprocessing the operational data to obtain formatted data. The specific steps are as follows: Multiple sensors are installed in the pumped storage power station, each equipped with a LoRaWAN wireless communication module and transmitting operational data to a central server. The runtime data is filtered using an algorithm to remove abnormal data points, and the timestamps of all runtime data are synchronized using a GPS clock. Convert the timestamped runtime data into a unified JSON format to obtain runtime data in a unified format; The unified format runtime data is deeply cleaned, missing values ​​are filled and deviations are corrected to obtain formatted data.

3. The carbon assessment and optimization method for pumped storage power stations as described in claim 1, characterized in that: The specific steps for re-inputting the sensor's operational data into the deep reinforcement learning model to obtain current and future carbon emission assessment results are as follows: The pressure reading P, temperature reading Temp, and water flow rate are extracted from the sensor's operating data and combined into a state vector S. current = [P, Temp, Flow]; Using the state vector S current The optimal action Act is generated using the Actor network, and then substituted into formula (1) to obtain Q(S) again. current ,Act); By re-obtaining Q(S) current Substituting Act into the random forest algorithm yields current and future carbon emission assessment results.

4. The carbon assessment and optimization method for pumped storage power stations as described in claim 3, characterized in that: The specific steps for formulating and implementing a specific optimization strategy based on the carbon emission assessment results are as follows: By comparing the carbon emission values ​​over the time period based on the current and future carbon emission assessments, the maximum value is taken. Based on the time period of maximum carbon emission values, identify high carbon emission operations or conditions, and analyze and extract pressure reading P, temperature reading Temp, and water flow rate; By modifying the corresponding operation values ​​based on the obtained pressure reading P, temperature reading Temp, and water flow rate Flow, an optimized scheme for reducing carbon emissions can be obtained. Modify and implement the optimized scheme for reducing carbon emissions.

5. The carbon assessment and optimization method for pumped storage power stations as described in claim 4, characterized in that: The specific steps for monitoring the implementation effect of the optimization scheme and feeding back the monitoring data to the deep reinforcement learning model are as follows: Obtain pressure reading P, temperature reading Temp, and water flow rate from the sensors after the optimization scheme is implemented; The pressure reading P, temperature reading Temp, and water flow rate after implementation will be reassembled into S. current = [P, Temp, Flow]; The reconstructed S current =The [P,Temp,Flow] values ​​are passed through the Actor network to generate a new optimal action Act, and the new optimal action Act and S are then compared. current =[P,Temp,Flow] are input into a deep reinforcement learning model to calculate the value Q(S) of the state-action pair. current ,Act); Based on the new data and the calculated value Q(S) of the state-action pairs current The gradient descent algorithm is used to update the weights of the Actor and Critic networks. Repeat the above process periodically to continuously optimize the parameters and strategies of the deep reinforcement learning model, ensuring that the system is always in the best operating state.

6. A carbon assessment and optimization system for pumped storage power stations, based on the carbon assessment and optimization method for pumped storage power stations described in any one of claims 1 to 5, characterized in that, include: The data collection and preprocessing module collects operational data from sensors deployed and collected from the pumped storage power station and its surrounding environment, and preprocesses the operational data to obtain formatted data. The deep reinforcement learning model building module constructs a deep reinforcement learning model based on the formatted data and in conjunction with the deep deterministic policy gradient algorithm. The carbon emission assessment module re-inputs the operational data from the sensors into the deep reinforcement learning model to obtain current and future carbon emission assessment results. The optimization strategy formulation and implementation module formulates specific optimization strategies based on the carbon emission assessment results, obtains and implements optimized solutions for reducing carbon emissions; The monitoring and feedback module monitors the implementation effect of the optimization scheme and feeds back the monitoring data to the deep reinforcement learning model.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the carbon assessment and optimization method for pumped storage power stations as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the carbon assessment and optimization method for pumped storage power stations as described in any one of claims 1 to 5.

Images