Method and system for improving operation efficiency of photovoltaic power station

Abstract

The invention discloses a method and system for improving the operation efficiency of a photovoltaic power station. The system comprises a sensor module, a data processing module, an intelligent control module and a user interface module. By constructing the fault prediction model, the state and future development of the photovoltaic power station are accurately predicted, the operation condition of the photovoltaic power station is better known, problems are found in advance, and corresponding adjustment is made. A data analysis algorithm is constructed, and operation parameters, power generation power and load requirements of a photovoltaic power station are adjusted in real time according to a current state and environment changes so as to maximize system benefits and match the system benefits with power grid consumption capability, so that photovoltaic power generation resources are utilized more efficiently, energy waste is reduced, the energy utilization rate is improved, and the load pressure of a power grid is reduced.

Description

Technical Field

[0001] The present invention relates to the field of distributed photovoltaic technology, and in particular to a method and system for improving the operating efficiency of a photovoltaic power station. Background Art

[0002] Traditional photovoltaic power station monitoring mainly relies on manual inspection and data recording, which is not only labor-intensive and inefficient, but also easy to ignore potential faults and problems. With the continuous expansion and increase in the scale and number of photovoltaic power stations, the demand for monitoring and management is becoming more and more urgent. With the development of Internet of Things technology, it provides technical support for the real-time monitoring, analysis and management of distributed photovoltaic power stations. In the early photovoltaic power station monitoring, sensors and monitoring equipment were used to collect key data of photovoltaic power stations, but the data collection and transmission methods were relatively simple and rough, and wired communication was generally used. With the development of cloud computing technology and big data analysis technology, photovoltaic power station data can be transmitted to cloud servers through wireless networks for storage and processing, realizing real-time monitoring and analysis of large amounts of data. At the same time, by analyzing historical data, photovoltaic power station performance prediction can be achieved, potential problems can be discovered in advance, and corresponding response strategies can be formulated.

[0003] However, although the existing distributed photovoltaic power station monitoring platform can realize real-time data monitoring and improve the performance and efficiency of the entire photovoltaic power station, it still has many technical defects, including at least the following prominent aspects: 1. The existing monitoring platform cannot effectively solve the problem of timely and accurate judgment of equipment damage or power generation abnormalities, resulting in the inability to timely understand the situation and carry out repairs when the equipment and instruments are damaged, increasing downtime and losses, and even affecting the normal operation of photovoltaic power generation devices. 2. Insufficient data processing and analysis of distributed photovoltaic power stations. The current monitoring platform cannot issue the output plan of each photovoltaic power station according to the output value of each photovoltaic power station, and calculate the current power grid's photovoltaic absorption capacity in real time. There are still deficiencies in optimizing the operation of distributed photovoltaic power stations in various regions. 3. Common monitoring systems generally have problems such as poor scalability and single function. Maintenance personnel can only complete some basic monitoring and control operations through a small LCD screen, and the user experience needs to be further improved. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and system for improving the operating efficiency of a photovoltaic power station in view of various deficiencies in the prior art.

[0005] In order to solve the above technical problems, the technical solutions adopted by the present invention are as follows.

[0006] A distributed photovoltaic power station intelligent monitoring system includes the following technical modules: a. a sensor module, which is used to monitor the status of power station equipment in real time; b. a data processing module, which is used to receive data from the sensor module transmitted via the Internet, store the data on a cloud platform, and predict possible failures of power station equipment through a preset learning model; c. an intelligent control module, which automatically controls photovoltaic equipment in the photovoltaic power station through the analysis results of the data processing module; d. a user interface module, which displays the analysis results of the intelligent processing module and prompts maintenance information.

[0007] As a preferred technical solution of the present invention, the sensor module includes a thermal imaging sensor, a vibration analysis sensor and a current monitoring sensor.

[0008] As a preferred technical solution of the present invention, the intelligent control module further has the function of communicating with the grid operator, and is used to adjust the output of the photovoltaic power station according to the analysis results to optimize the grid load.

[0009] As a preferred technical solution of the present invention, the data processing module automatically adjusts the operating parameters of the photovoltaic power station to maximize power generation efficiency and reduce the risk of equipment damage.

[0010] As a preferred technical solution of the present invention, the user interface module is adapted to a variety of devices including mobile phones, tablets and laptops, and displays operation information graphically; it also includes a response mechanism for automatically notifying maintenance personnel, which is triggered when potential equipment problems and failures are detected.

[0011] A distributed photovoltaic power station monitoring method is used to support the implementation of a distributed photovoltaic power station intelligent monitoring system, comprising the following steps: (1) monitoring real-time data of multiple detection points in the power station, and transmitting the data to a centralized cloud computing platform via the Internet; (2) using a fault prediction model stored on the cloud computing platform in combination with historical data to analyze the real-time data to identify potential faults; (3) constructing a data analysis algorithm to perform real-time data analysis, optimize and adjust the output of the photovoltaic power station to match the current power grid's absorption capacity; and (4) displaying the analysis results and the operating status of the equipment to operators through a user interface.

[0012] As a preferred technical solution of the present invention, the fault prediction model is trained to analyze the equipment performance data to predict the damage risk. The specific construction process of the fault prediction model is as follows:

[0013] (1) The performance indicators of photovoltaic modules, including temperature, light intensity, current, voltage, and power output, are set as {X1, X2, …, X}, and the overall health status of the power station can be used as the target variable Y;

[0014] (2) Based on the knowledge of photovoltaic power plants, determine the influence relationship between various performance indicators, construct directed edges between nodes to represent these influence relationships, define the conditional probability table of each node, describe the dependency relationship between different nodes, and the impact of these indicators on the health status, and define each edge X in the structure. i →X, define the conditional probability distribution P(X|X i );

[0015] (3) Learning parameters from historical data. By monitoring the performance indicators and health status at different time points in the historical data, the conditional probability distribution between nodes is estimated. For each node’s conditional probability table, by giving X i When the parent node value of i The conditional probability distribution P(X i |Parents(X i ));

[0016] For each edge X i →X, the parameter in its conditional probability table is θ i , parameter θ i It is estimated by counting the frequency of each combination in historical data and calculated by the following formula:

[0017] θ i =(N(x i =x i ,Parents(X i )=π i )) / (N(Parents(X i )=π i ))

[0018] Where N(·) represents the counting function;

[0019] (4) Based on the real-time monitoring performance data, the posterior probability of power plant failure is calculated based on the Bayesian network. By repeatedly updating the probability distribution, fault prediction and risk assessment can be performed in real time.

[0020] To calculate the posterior distribution P(Y|X1=x1,…,X=x), we use the chain rule and the joint distribution:

[0021] P(Y|X1=x1,...,X=x)=(P(Y,X1=x1,...,X=x)) / (P(X1=x1,...,X=x))

[0022] Use variable elimination method to traverse the decision graph and sequentially calculate the marginal distribution of each variable to provide comprehensive and reliable probability distribution for different fault conditions;

[0023] (5) Convert the posterior distribution P(Y|X1=x1,…,X=x) into a risk score and define the loss function Where Y is the real health status, is the predicted state. In order to minimize the expected loss, the posterior distribution is used to make the optimal health state prediction:

[0024] The obtained health status probability can be used as the basis for risk assessment. According to the probabilities of different fault states and combined with the loss function, the risk of failure of the power station can be evaluated.

[0025] As a preferred technical solution of the present invention, a corresponding fault prediction and maintenance decision-making strategy is formulated based on the fault prediction model and risk index. When the risk index exceeds the threshold, the management personnel are warned and the maintenance process is initiated. At the same time, through continuous monitoring, the equipment performance data is acquired in real time, the energy landscape model is updated, and the risk index is recalculated according to the latest data, dynamic fault prediction and maintenance decisions are achieved.

[0026] As a preferred technical solution of the present invention, the construction process of the data analysis algorithm is as follows:

[0027] (1) Set the time series data {Xt}, where Xt represents the feature vector at time t, and predict the photovoltaic power generation Pv(t) and the grid load demand D(t). The photovoltaic power generation and grid load demand are predicted by the following methods:

[0028] h_t,c_t=LSTM(X_t,h_{t-1},c_{t-1};θ), where h_t is the hidden state at time t, c_t is the cell state at time t, and θ is the network parameters;

[0029] For the output layer, an attention mechanism is introduced to determine the important features for prediction: in and is the predicted photovoltaic power generation and grid load demand, Attention is the function of the attention mechanism, and φ is the parameter of the output layer;

[0030] The model is trained by minimizing loss functions such as mean absolute error and mean squared error;

[0031] (2) Define an optimization problem to balance power generation and grid demand. The objective function is defined as maximizing economic benefits:

[0032]

[0033] The constraints are used to ensure the feasibility of power consumption and operation and maintenance, for all time t:

[0034]

[0035] in represents the predicted photovoltaic power generation at time t, Gmin(t) and Gmax(t) represent the minimum and maximum power generation allowed by the photovoltaic power generation system at time t, respectively. represents the load demand predicted at time t, Grid_Cap(t) is the upper limit of the grid capacity, and Egrid(t) is the energy level of the grid at time t;

[0036] (3) By introducing quantum computing principles to enhance the capabilities of neural networks, using qubits as processing units, when constructing a DQN for predicting the matching problem between grid demand and photovoltaic power generation, the following quantum neural network is designed:

[0037] Yt=DQN(Xt;θ), where Yt is the output layer, predicting future grid demand and photovoltaic power generation, Xt is the input feature vector, and θ is the parameter of the quantum network;

[0038] (4) According to the real-time operation of the photovoltaic power station, data including various sensor data, photovoltaic power generation power, and load demand are continuously collected to form the current state s_t and action a_t; based on the model s_(t+1)~GP(s_t,a_t;Σ,L,σ 2 ),Σ is the covariance matrix, L represents the length scale, σ 2 is the noise variance. The current state and action are used to predict the probability distribution of the next state s_(t+1). The predicted s_(t+1) is compared with the current power grid’s absorptive capacity. If the predicted photovoltaic power generation exceeds the upper capacity limit of the power grid, the action a_t is adjusted to ensure that the input of the photovoltaic power station meets the capacity limit of the power grid. Through the reinforcement learning algorithm, the operation decision strategy is updated and optimized according to the current state s_t, action a_t and the predicted next state s_(t+1) to adapt to the power grid’s absorptive capacity.

[0039] As a preferred technical solution of the present invention, it includes real-time updating of the communication between the power station and the power grid to achieve dynamic power management.

[0040] The beneficial effect of adopting the above technical solution is that the present invention can make more accurate predictions on the status and future development of photovoltaic power stations by constructing a fault prediction model, better understand the operating status of photovoltaic power stations, discover problems in advance and make corresponding adjustments. By constructing a data analysis algorithm for real-time data analysis, the operating parameters of the photovoltaic power station are adjusted in real time according to the current status and environmental changes to maximize the benefits of the system and match the grid's absorption capacity. By adjusting the power generation power and load demand of the photovoltaic power station in real time, the system can more efficiently utilize photovoltaic power generation resources, reduce energy waste, improve energy utilization, and reduce the load pressure of the grid. With the help of reinforcement learning algorithms, the management and monitoring of photovoltaic power stations in the present invention become more intelligent and automated, which reduces the workload of operators, improves the operating efficiency and reliability of the system, and further enhances the user experience. DETAILED DESCRIPTION

[0041] The following examples describe the present invention in detail. Each device used in the present invention is a conventional commercial product, which can be directly obtained by purchasing from the market. In the description of the following embodiments, specific details such as specific system structures and technologies are proposed for illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits and methods are omitted to prevent unnecessary details from hindering the description of the present application. It should be understood that when used in the present application specification and the attached claims, the term "including" indicates the presence of the described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections. It should also be understood that the term "and / or" used in the present application specification and the attached claims refers to any combination of one or more of the items listed in the association and all possible combinations, and includes these combinations.

[0042] As used in the specification of the present application and the appended claims, the term "if" can be interpreted as "when..." or "once" or "in response to determination" or "in response to detection" according to the context. Similarly, the phrase "if it is determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once it is determined" or "in response to determination" or "once [the described condition or event] is detected" or "in response to detection [the described condition or event]" according to the context. In addition, in the description of the specification of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance. The reference "one embodiment" or "some embodiments" described in the specification of the present application means that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Thus, the sentences "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in the specification are not necessarily all referring to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "include", "comprising", "having" and variations thereof all mean "including but not limited to", unless specifically emphasized otherwise.

[0043] Example 1

[0044] A distributed photovoltaic power station intelligent monitoring system comprises a sensor module, a data processing module, an intelligent control module and a user interface module.

[0045] The sensor module includes thermal imaging sensors, vibration analysis sensors and current monitoring sensors. The thermal imaging sensor is used to detect the temperature distribution of power station equipment and identify hot spots, which can indicate electrical connection problems or equipment failures; the vibration analysis sensor is used to monitor the vibration frequency and amplitude of the mechanical parts of the photovoltaic power station to analyze the health status of the mechanical equipment; the current monitoring sensor is responsible for recording the current size of the equipment to detect current anomalies, that is, circuit problems that may cause failures. At the same time, the sensor module also includes light intensity sensors, temperature sensors, etc. These sensors are integrated into the equipment and transmit the monitored data to the data processing module in real time.

[0046] The data processing module develops or applies existing middleware as the receiving point of sensor data to ensure the effective reception and pre-processing of data. The edge computing device or data center is used to collect the data transmitted by the sensor module, and the data is stored on the cloud platform. The distributed database management system is used to ensure data reliability and accessibility. The elastic storage and computing power of cloud computing are used to achieve long-term storage and efficient operation of data, and support big data processing frameworks such as Apache Hadoop or Spark. Machine learning and data analysis techniques are used to analyze historical and real-time data to predict potential failures and equipment performance degradation, and are also used to optimize daily power generation efficiency and set maintenance plans.

[0047] The intelligent control module is connected to each sensor and actuator so that it can make adjustments immediately after receiving instructions from the data processing module. The analysis results of the data processing module can be used to automatically control the photovoltaic equipment in the photovoltaic power station. The output of the photovoltaic power station can be adjusted according to data such as demand peaks and valleys, weather forecasts, and grid requirements. Communication with the grid operator may be carried out through standardized interfaces, such as open communication standards, so that the system can achieve demand response and output regulation. Considering data security and operational stability, the intelligent control module has a good security mechanism.

[0048] The user interface module displays the analysis results and maintenance prompts of the intelligent processing module. Develop cross-platform compatible applications that work on devices such as mobile phones, tablets and laptops to ensure a good user experience on different terminal devices. Design an intuitive graphical user interface to visualize equipment status, alarms and performance statistics. Implement a push notification system to automatically send prompts to the devices of relevant maintenance and operation personnel when equipment anomalies or failures are detected. Provide maintenance log recording and fault reporting functions to facilitate tracking and analysis of historical problems. The componentized design allows functional modules such as information display and maintenance reminders to be used independently or in combination to meet the needs of different users. Provide an API interface for third-party maintenance system access to achieve deeper system integration and automation.

[0049] Example 2

[0050] Collect historical data of photovoltaic power plants, including real-time monitoring data of performance indicators such as temperature, light intensity, current, voltage, power output, etc. of photovoltaic modules, as well as health status data of corresponding power plants. Perform data cleaning and preprocessing including processing missing values, outliers and noise, as well as feature engineering processing including feature selection, feature extraction, feature conversion, etc. on the collected data to better reflect the performance characteristics and overall health status of photovoltaic modules. Use domain knowledge and statistical methods to screen and transform features to improve the expressiveness and predictive ability of features.

[0051] The performance indicator set {X1, X2, …, X} is used as input features, and the overall health status of the power station is used as the target variable Y to establish a prediction model. Based on the knowledge and practical experience of photovoltaic power stations, the possible influence relationships between the performance indicators are determined. For example, temperature affects the current and voltage of photovoltaic modules, and light intensity affects the power output of photovoltaic modules. Directed edges are established between nodes to represent these influence relationships, such as X1→X2 represents the influence of temperature on light intensity, etc.

[0052] For each node X i , define its conditional probability distribution P(X i |PA(X i )), where PA(X i ) indicates X i The parent node set of . For each edge X i →X, define the conditional probability distribution P(X|X i ), indicating that given X i The probability distribution of X under the condition of . These conditional probability tables can be obtained through statistical calculations of historical data, or set based on professional knowledge and experience. For each node X

[0053] , its conditional probability distribution P(X|X i ) can be determined by methods such as maximum likelihood estimation. For example, we can statistically i Under the condition of taking a certain value, X takes the frequency of each possible value, and then normalizes it to get the probability value. For performance indicators that may have discrete values ​​and continuous values, it is necessary to handle the determination of conditional probability distribution separately. When defining the conditional probability table, the dependency between different nodes can be described by conditional probability distribution. For example, if the light intensity has a greater impact on the power output, this dependency will be reflected in the conditional probability distribution.

[0054] Set each edge X i →X, the parameter in its conditional probability table is θ i , which can be expressed as i In the case of , the probability distribution of child node X taking different values ​​is learned from the historical data, and the parameters are estimated by counting the frequency of each combination in the historical data.

[0055] For each node X i , its conditional probability table can be expressed as P(X i |Parents(X i )), where Parents(X i ) indicates X i The parent node set of a node X. i A parent node set πi , and under this parent node set X i The value is x i The number of historical data occurrences N(x i =x i ,Parents(X i )=π i ), the total number of possible values ​​N(Parents(X i )=π i ), the maximum likelihood estimation method can be used to calculate the parameter θ i , specifically, parameter θ i The estimated value of can be calculated using the following formula:

[0056] θ i =(N(x i =x i ,Parents(X i )=π i )) / (N(Parents(X i )=π i )).

[0057] After monitoring the real-time performance data, we can use the Bayesian network model to calculate the posterior probability P(Y|X i =x1,...,X=x).

[0058] According to Bayes' theorem, the posterior probability can be calculated from the joint distribution and the conditional distribution:

[0059] P(Y|X1=x1,...,X=x)=(P(Y,X1=x1,...,X=x)) / (P(X1=x1,...,X=x))

[0060] When calculating the posterior probability, we need to calculate the joint distribution P(Y,X1=x1,...,X=x) and the marginal distribution P(X i =x i,...,X=x). For more complex Bayesian network models, it may be difficult to calculate the joint distribution and marginal distribution. Therefore, the variable elimination method is used to handle these distributions to calculate the marginal distribution of each variable by calculating the order, and finally calculate the posterior probability of the target variable. In the variable elimination method, we traverse the decision graph and calculate the marginal distribution of each variable, and use these distributions to calculate the posterior probability of the target variable. The order of variable elimination has a significant impact on the computational efficiency, and the optimal variable elimination order is selected to minimize the computational complexity. Use heuristic algorithms or search strategies to find the optimal order. As the real-time monitoring data is continuously updated, the probability distribution needs to be updated in a timely manner and the posterior probability needs to be recalculated for real-time prediction and risk assessment. Use an incremental algorithm to update the probability distribution to efficiently update the posterior probability distribution and marginal distribution without repeatedly calculating the probability distribution.

[0061] When conducting risk assessment, we can use the posterior distribution P(Y|X i =x i ,…,X=x) is converted into a risk score by defining the loss function To measure the actual health status Y and the predicted status Then, the posterior distribution is used to make the optimal health state prediction to minimize the expected loss, that is, to find the predicted health state that minimizes the expected loss.

[0062] Define a loss function Used to measure and predict health status and the actual health state Y. The expected loss represents the difference between The expected value of the loss function. We hope to find the predicted health state that minimizes the expected loss. To minimize risk: That is, find the predicted health state that minimizes the expected loss. By calculating the posterior distribution P(Y|X i =x1,…,X=x), we can get the probability distribution of different health states, which can be used as the basis for risk assessment. And the loss function defined The probability of different fault states can be evaluated to derive the risk level of failure in the generating station.

[0063] Example 3

[0064] In the time series data {Xt}, the feature vector Xt at each time point predicts the photovoltaic power generation and grid load demand in the following way: h_t, c_t = LSTM(X_t, h_{t-1}, c_{t-1}; θ), where h_t is the hidden state at time t, c_t is the unit state at time t, and θ is the network parameter; the attention mechanism is introduced to determine the important features of the prediction, and the hidden state h_t is weighted by the attention mechanism function Attention(h_t), thereby highlighting the more important features in the prediction:

[0065] in and is the predicted photovoltaic power generation and grid load demand, and φ is the parameter of the output layer.

[0066] During the training process, the model parameters are optimized by minimizing the loss function, usually using loss functions such as mean absolute error or mean square error to measure the error between the predicted value and the true value. Through optimization algorithms such as back propagation algorithm and gradient descent, the network parameters θ and φ are updated, so that the model can more accurately predict photovoltaic power generation and grid load demand.

[0067] In terms of balancing power generation and grid demand, an optimization problem is defined, the goal is to maximize economic benefits. The objective function represents the goal of maximizing economic benefits, expressed as: in represents the predicted photovoltaic power generation at time t, C(t) represents the electricity price at that time, and Cop(t) represents other costs, such as power generation cost, etc. Therefore, the objective function aims to maximize the benefits brought by photovoltaic power generation while reducing the impact of other costs.

[0068] The constraints are mainly used to ensure the feasibility of power consumption and operation and maintenance, and to ensure the balance between photovoltaic power generation and grid load demand. For all time t, the constraints include:

[0069] Photovoltaic power generation The minimum and maximum power generation allowed by the photovoltaic power generation system must be met, that is,

[0070] Grid load demand Must be less than or equal to the upper limit of the grid capacity Grid_Cap(t);

[0071] The grid energy level Egrid(t+1) is equal to the grid energy level Egrid(t) plus the photovoltaic power generation Subtract load demand Right now

[0072] By introducing quantum computing principles to enhance the capabilities of neural networks and using qubits as processing units, when constructing a DQN for predicting the matching problem between grid demand and photovoltaic power generation, the following quantum neural network is designed:

[0073] Yt=DQN(Xt;θ), where Yt is the output layer, predicting future grid demand and photovoltaic power generation, Xt is the input feature vector, and θ is the parameter of the quantum network.

[0074] According to the real-time operation of the photovoltaic power station, data including various sensor data, photovoltaic power generation power, and load demand are continuously collected to form the current state s_t and action a_t; based on the model s_(t+1)~GP(s_t,a_t;Σ,L,σ 2 ), using the current state and action, predict the probability distribution of the next state s_(t+1); compare the predicted s_(t+1) with the current power grid's absorption capacity. If the predicted photovoltaic power generation power exceeds the upper capacity limit of the power grid, adjust the action a_t to ensure that the input of the photovoltaic power station meets the capacity limit of the power grid; through the reinforcement learning algorithm, according to the current state s_t, action a_t and the predicted next state s_(t+1), update and optimize the operation decision strategy to adapt to the absorption capacity of the power grid.

[0075] In the above embodiments, the description of each embodiment has its own emphasis. For the part that is not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In each embodiment, the hardware implementation of the technology can directly adopt existing intelligent devices, including but not limited to industrial computers, PCs, smart phones, handheld stand-alone machines, floor-standing stand-alone machines, etc. Its input device preferably adopts a screen keyboard, its data storage and calculation module adopts existing memory, calculator, controller, its internal communication module adopts existing communication ports and protocols, and its remote communication adopts existing GPRS networks, World Wide Internet, etc. It can be clearly understood by technicians in the relevant field that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present application. The specific working process of the units and modules in the above system can refer to the corresponding process in the above method embodiment, and will not be repeated here. In the embodiments provided by the present invention, it should be understood that the disclosed device / terminal equipment and method can be implemented in other ways. For example, the device / terminal equipment embodiment described above is only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0076] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for improving the operating efficiency of a photovoltaic power station, characterized in that: The following steps are involved: S1. Use sensor modules to monitor the real-time data of multiple detection points in the power station in real time, and transmit the data to the centralized cloud computing platform through the data processing module; the data processing module develops or applies existing middleware as the receiving point of sensor data to ensure the effective reception and pre-processing of data, and uses edge computing devices or data centers to collect the data transmitted by the sensor module, and the data is stored on the cloud platform; S2. Analyze the real-time data using the fault prediction model stored on the cloud computing platform in combination with historical data to identify potential faults; Based on predicted health status And the loss function defined The probability of different fault states is evaluated to obtain the risk level of failure of the power station, and the health state prediction is: By calculating the posterior distribution P(Y|X1=x1,…,X=x), we can get the probability distribution of different health states, which can be used as the basis for risk assessment. S3. Build a data analysis algorithm to perform real-time data analysis, and use the intelligent control module to optimize and adjust the output of the photovoltaic power station to match the current power grid's absorption capacity; Based on the model s_(t+1)~GP(s_t,a_t;Σ,L,σ 2 ), using the current state and action, predict the probability distribution of the next state s_(t+1); compare the predicted s_(t+1) with the current power grid's absorption capacity to ensure that the input of the photovoltaic power station meets the capacity limit of the power grid; update and optimize the operation decision strategy according to the current state s_t, action a_t and the predicted next state s_(t+1) to adapt to the absorption capacity of the power grid; S4. Display the analysis results and equipment operation status to the operator through the user interface module.

2. The method for improving the operating efficiency of a photovoltaic power station according to claim 1, characterized in that: In S2, the fault prediction model is trained to analyze the equipment performance data, thereby predicting the damage risk. The specific construction process of the fault prediction model is as follows: S201. Collect historical data of photovoltaic power plants, real-time detection data of various performance indicators of photovoltaic modules, and health status data of corresponding power plants, and perform data cleaning and preprocessing, as well as feature engineering processing on the collected data, so as to better reflect the performance characteristics and overall health status of photovoltaic modules; S202. Taking the performance indicator set {X1, X2, …, X} as input features and the overall health status of the power plant as the target variable Y, a prediction model is established to determine the possible influence relationships between the performance indicators, and directed edges are established between the nodes to represent these influence relationships; S203. For each node X i , define its conditional probability distribution P(X i |PA(X i )), where PA(X i ) indicates X i The parent node set of i →X, define the conditional probability distribution P(X|X i ), indicating that given X i The probability distribution of X under the condition of ; For performance indicators that may have discrete values ​​and continuous values, it is necessary to separately handle the determination of conditional probability distribution. When defining the conditional probability table, the dependency relationship between different nodes is described by the conditional probability distribution; For each node X i , its conditional probability table can be expressed as P(X i |Parents(X i )), where Parents(X i ) indicates X i The parent node set πi; Set each edge X i →X, the parameter in its conditional probability table is θ i , indicating that at a given parent node X i In the case of , the probability distribution of child node X taking different values ​​is estimated by counting the frequency of each combination in the historical data, and the parameter θ i The estimated value of is calculated using the following formula: i i =(N(x i =x i ,Parents(X i )=π i )) / (N(Parents(X i )=π i )) Where N(·) represents the counting function; S204. Calculate the posterior probability of power station failure based on the Bayesian network model according to the real-time monitoring performance data P(Y|X1=x1,...,X=x)=(P(Y,X1=x1,...,X=x)) / (P(X1=x1,...,X=x)) Where P(Y,X1=x1,...,X=x) is the joint distribution and P(X1=x1,...,X=x) is the marginal distribution. S205. Convert the posterior distribution P(Y|X1=x1,…,X=x) into a risk score by defining the loss function To measure the actual health status Y and the predicted status In order to minimize the expected loss, the posterior distribution is used to make the optimal health status prediction:

3. The method for improving the operating efficiency of a photovoltaic power station according to claim 2, characterized in that: When calculating the posterior probability in S204, it is necessary to use a variable elimination method to traverse the decision graph and sequentially calculate the marginal distribution of each variable, so as to provide a comprehensive and reliable probability distribution for different fault conditions; As real-time monitoring data is continuously updated, the probability distribution needs to be updated in a timely manner and the posterior probability needs to be recalculated to enable real-time prediction and risk assessment; The probability distributions are updated using an incremental algorithm to efficiently update the posterior and marginal distributions without recalculating the probability distributions.

4. The method for improving the operating efficiency of a photovoltaic power station according to claim 2, characterized in that: The health status probability obtained by the health status prediction in S205 is used as the basis for risk assessment. And the loss function defined The probability of different fault states is evaluated to obtain the risk level of failure of the power station.

5. The method for improving the operating efficiency of a photovoltaic power station according to claim 1, characterized in that: The specific construction process of the data analysis algorithm in S3 is as follows: S301. Set the time series data {Xt}, where Xt represents the feature vector at time t, and predict the photovoltaic power generation and grid load demand in the following way: h_t,c_t=LSTM(X_t,h_{t-1},c_{t-1};θ) Where h_t is the hidden state at time t, c_t is the cell state at time t, and θ is the network parameter; The attention mechanism is introduced to determine the important features of the prediction. The hidden state h_t is weighted through the attention mechanism function Attention(h_t), thereby highlighting the more important features in the prediction: in, and is the predicted PV power generation and grid load demand, φ is the parameter of the output layer; S302. Define an optimization problem to balance power generation and grid demand. The objective function is defined as maximizing economic benefits: in, represents the predicted photovoltaic power generation at time t, C(t) represents the electricity price at that time, and Cop(t) represents other costs; The constraints are used to ensure the feasibility of power consumption and operation and maintenance, and to ensure the balance between photovoltaic power generation and grid load demand. For all time t, the constraints include: Photovoltaic power generation The minimum power generation Gmin(t) and maximum power generation Gmax(t) allowed by the photovoltaic power generation system must be met; Grid load demand Must be less than or equal to the upper limit of the grid capacity Grid_Cap(t) The grid energy level Egrid(t+1) is equal to the grid energy level Egrid(t) plus the photovoltaic power generation Subtract load demand S303. By introducing quantum computing principles to enhance the capabilities of neural networks, using qubits as processing units, when constructing a DQN for predicting the matching problem between grid demand and photovoltaic power generation, the following quantum neural network is designed: Yt=DQN(Xt;θ), where Yt is the output layer, predicting future grid demand and photovoltaic power generation, Xt is the input feature vector, and θ is the parameter of the quantum network; S304. According to the real-time operation of the photovoltaic power station, continuously collect data including various sensor data, photovoltaic power generation power, and load demand to form the current state s_t and action a_t; based on the model s_(t+1)~GP(s_t,a_t;Σ,L,σ 2 ), using the current state and action, predict the probability distribution of the next state s_(t+1); compare the predicted s_(t+1) with the current power grid's absorption capacity. If the predicted photovoltaic power generation power exceeds the upper capacity limit of the power grid, adjust the action a_t to ensure that the input of the photovoltaic power station meets the capacity limit of the power grid; through the reinforcement learning algorithm, according to the current state s_t, action a_t and the predicted next state s_(t+1), update and optimize the operation decision strategy to adapt to the absorption capacity of the power grid.

6. The method for improving the operating efficiency of a photovoltaic power station according to claim 5, characterized in that: During the training process, the model in S301 optimizes the model parameters by minimizing the loss function, and uses the mean absolute error or mean square error to measure the error between the predicted value and the true value; the network parameters θ and φ are updated by the directional propagation algorithm and the gradient descent optimization algorithm, so that the model can more accurately predict the photovoltaic power generation and the grid load demand.

7. The method for improving the operating efficiency of a photovoltaic power station according to claim 1, characterized in that: This includes real-time updates of communications between power stations and the grid to enable dynamic power management.

8. An intelligent monitoring system, characterized in that: The method for improving the operating efficiency of a photovoltaic power station according to any one of claims 1 to 7 comprises: A sensor module, which includes a thermal imaging sensor, a vibration analysis sensor, a current monitoring sensor, a light intensity sensor, and a temperature sensor. These sensors are integrated into the device and transmit the monitored data to the data processing module in real time; A data processing module, which uses edge computing devices or data centers to collect data transmitted by sensor modules, stores the data on a cloud platform, and uses machine learning and data analysis techniques to analyze historical and real-time data to predict potential failures and equipment performance degradation. It is also used to optimize daily power generation efficiency and set maintenance plans; An intelligent control module, which is connected to each sensor and actuator so as to implement adjustments immediately after receiving instructions from the data processing module; automatically controls the photovoltaic equipment in the photovoltaic power station through the analysis results of the data processing module; and adjusts the output of the photovoltaic power station according to demand peaks and valleys, weather forecasts and grid requirements; A user interface module, which displays the analysis results and maintenance prompts of the intelligent processing module; is compatible with a variety of devices including mobile phones, tablets and laptops, and displays operation information in a graphical manner; and also includes a response mechanism for automatically notifying maintenance personnel, which is triggered when potential equipment problems and failures are detected.

9. The intelligent monitoring system according to claim 8, characterized in that: The thermal imaging sensor is used to detect the temperature distribution of power station equipment, identify hot spots, and indicate electrical connection problems or equipment failures; the vibration analysis sensor is used to monitor the vibration frequency and amplitude of the mechanical parts of the photovoltaic power station, thereby analyzing the health status of the mechanical equipment; the current monitoring sensor is responsible for recording the current size of the equipment, thereby detecting current anomalies.

Images