CN118194699A - On-line evaluation method and system for excitation system of thermal power generating unit - Google Patents

Abstract

The invention discloses an on-line evaluation method and system of an excitation system for a thermal power unit, which relate to the field of power engineering and comprise the steps of collecting data of the excitation system for the thermal power unit for preprocessing; extracting features of the preprocessed integrated data, and establishing a thermal power unit excitation system performance evaluation model for performing performance analysis; optimizing model parameters, carrying out online evaluation on faults of the excitation system, and outputting evaluation results; and analyzing the evaluation result to judge whether the performance of the excitation system is declined or not, so as to realize early detection of faults. According to the invention, feature extraction is carried out by combining dynamic time warping DTW and mean square error MSE, so that the accuracy of fault prediction is improved; the DTW analysis and performance degradation evaluation are used, the mutual information value is adjusted by considering the performance degradation rate, the decision tree algorithm dynamically adapts to the change of the system state, and the adaptability to the newly-appearing fault mode is improved; and the method is adjusted and optimized according to different excitation system characteristics and operating conditions of the thermal power generating unit, so that the operating efficiency is improved and the risk is reduced.

Description

A method and system for online evaluation of excitation system for thermal power unit

Technical Field

The invention relates to the field of electric power engineering, and in particular to an online evaluation method and system for an excitation system for a thermal power unit.

Background technique

The excitation system of a thermal power unit is a key component to ensure stable voltage and frequency output. Traditional excitation systems are usually based on methods such as PID controllers to maintain steady-state operation of the motor, but they cannot evaluate the performance of the excitation system under different loads, environments, and working conditions, especially in nonlinear systems or load mutations.

Modern thermal power unit excitation system assessment methods use advanced sensor technology, data analysis and artificial intelligence to make them more intelligent and adaptive. However, they cannot be adjusted in real time according to system status and external load changes, making it difficult to respond to emergencies and rapidly changing load demands.

Summary of the invention

In view of the problems existing in the existing evaluation method of the excitation system of thermal power units in response to emergencies and rapidly changing load demands, the present invention is proposed.

Therefore, the problem to be solved by the present invention is how to make timely responses to emergencies and rapidly changing load demands.

In order to solve the above technical problems, the present invention provides the following technical solutions:

In a first aspect, an embodiment of the present invention provides an online evaluation method for an excitation system for a thermal power unit, which includes collecting data of the excitation system for the thermal power unit and other relevant data, integrating them, and preprocessing them; extracting features from the preprocessed integrated data, establishing a performance evaluation model for the excitation system of the thermal power unit, and performing performance analysis; using historical data to train and verify the evaluation model, optimize model parameters, and improve the accuracy and reliability of the evaluation model; inputting real-time collected data into the trained evaluation model, performing online evaluation and prediction of excitation system faults, and outputting evaluation results; analyzing the evaluation results to determine whether the performance of the excitation system has declined, achieve early detection of faults, and ensure stable and reliable operation of the excitation system.

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit of the present invention, the preprocessing comprises the following steps: performing linear interpolation processing on the missing data values;

When all data with the same timestamp are lost, linear interpolation is used to process the missing data;

When a data is missing in the data P or D with the same timestamp, the data is supplemented by analyzing other data with the same timestamp; for each data sequence, the rate of change between consecutive time points is calculated; the rate of change of the missing data is estimated based on the correlation of the rate of change, and the data sequences P _i and D _j with the highest correlation with the rate of change of the missing data are selected; if the data sequence P _i is missing, the estimated rate of change of the missing data is ΔP _i ^(t)＝α+β·ΔD _j (t); if the data sequence D _j is missing, the estimated rate of change of the missing data is ΔD _j ^(t)＝α+β·ΔP _i (t), where α and β are coefficients calculated based on historical data; the estimated rate of change is used to supplement the missing data, if P _i (t-1) is known, then P _i (t)＝P _i (t-1)×(1+ΔP _i ^(t));

If D _j (t-1) is known, then D _j (t) = D _j (t-1) × (1 + ΔD _j ^ (t)).

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit of the present invention, the formula for calculating the change rate between consecutive time points for each data sequence is as follows:

ΔP _i (t) = P _i (t-1)P _i (t) - _{P i} (t-1)

ΔD _i (t) = D _j (t-1) D _j (t) - D _j (t-1)

Among them, _Pi and _Dj are data sequences, and t is the timestamp.

The relevant formula for linear interpolation of missing data values is as follows:

Among them, t is the timestamp, and _t1 and _t2 are adjacent time points.

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit of the present invention, the performance analysis includes the following steps: capturing dynamic patterns and shapes in time series data; using a dynamic time warping (DTW) algorithm to compare time series and extract similar patterns, and the relevant formula is as follows:

Among them, S and T are time series.

Estimate the complexity and irregularity of time series data and calculate multiscale entropy. The relevant formula is as follows:

MSE(k)＝-∑p(x)logp(x)

Among them, p(x) is the probability distribution at scale k.

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit of the present invention, the use of the dynamic time warping (DTW) algorithm to compare the time series includes the following steps: when D _DTW ≤θ _deg , it is judged that the performance is normal; when D _DTW >θ _deg , it is judged that the performance is degraded; if the performance is degraded, the value of the mutual information is adjusted, and the relevant formula is as follows:

Among them, B is the number of steps calculated according to the longest step length each time the decay rate is reached, R _decay is the decay rate, and MI is the value of the entire information. The specific formula is as follows:

MI(X;Y)＝∑p(x,y)logp(y)p(x,y)x∈X,y∈Y

Among them, X is the quantitative feature and Y is the fault label.

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit described in the present invention, the method comprises the following steps: when the data collected in real time is input into the evaluation model in real time, real-time anomaly detection is performed; when the input data is normal, the evaluation model is executed according to a predetermined process and a corresponding evaluation result is output; when the input data is abnormal, a corresponding alarm is output and abnormal information is recorded, and different abnormal situations are set with abnormal levels to flexibly handle the abnormal situations; if the abnormal level is level one, the abnormal information is recorded, and the evaluation model is continued to be executed, and the abnormal level is indicated in subsequent reports; if the abnormal level is level two, an alarm is output and the current evaluation task is suspended; if the abnormal level is level three, the current task is immediately stopped, an emergency alarm is output, and a corresponding system self-rescue mechanism is triggered.

As a preferred solution of the online evaluation method for the excitation system of a thermal power unit described in the present invention, the determination of whether the performance of the excitation system has declined includes the following steps: establishing a fault mode by analyzing the evaluation results; when the real-time evaluation results do not match the known fault mode, recording the current real-time evaluation data and model output; when the real-time evaluation results highly match the known fault mode, triggering fault detection to remind more in-depth inspection and maintenance; when the real-time evaluation results poorly match the known fault mode, selecting light processing and recording the current real-time evaluation data and model output.

In a second aspect, an embodiment of the present invention provides an online evaluation system for an excitation system of a thermal power unit, comprising: an acquisition module, for collecting data of the excitation system of the thermal power unit and other relevant data for integration and preprocessing; an extraction module, for extracting features from the preprocessed integrated data, establishing a performance evaluation model for the excitation system of the thermal power unit, and performing performance analysis; an optimization module, for training and verifying the evaluation model using historical data, optimizing model parameters, and improving the accuracy and reliability of the evaluation model; an evaluation module, for inputting real-time collected data into the trained evaluation model, performing online evaluation and prediction of excitation system faults, and outputting evaluation results; an analysis module, for analyzing the evaluation results, determining whether the performance of the excitation system is degraded, achieving early detection of faults, and ensuring stable and reliable operation of the excitation system.

In a third aspect, an embodiment of 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 instructions are executed by the processor, the steps of the online evaluation method for the excitation system of a thermal power unit as described in the first aspect of the present invention are implemented.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the online evaluation method for an excitation system for a thermal power unit as described in the first aspect of the present invention are implemented.

The beneficial effects of the present invention are as follows: the present invention improves the accuracy of fault prediction by combining dynamic time warping DTW and mean square error MSE for feature extraction; uses DTW analysis and performance degradation evaluation to timely identify small changes in the system and potential performance degradation, thereby realizing early detection of faults; by adjusting the mutual information value by considering the performance decay rate, the decision tree model dynamically adapts to changes in the system state, thereby improving the adaptability to new fault modes; and flexibly adjusts and optimizes according to different characteristics and operating conditions of the excitation system of thermal power units, and has good scalability, thereby improving operating efficiency, reducing risks, and ensuring stable and reliable operation of the power system.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

FIG1 is a flow chart of an online evaluation method for an excitation system of a thermal power unit according to Example 1.

FIG. 2 is a system result diagram of the online evaluation method for the excitation system of a thermal power unit in Example 1.

Detailed ways

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

Secondly, the term "one 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 term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

Example 1

1 and 2, which are the first embodiment of the present invention, provide an online evaluation method for an excitation system of a thermal power plant, comprising:

S1: Collect data from the excitation system of thermal power units and other related data, integrate them, and perform preprocessing.

Specifically, parameters P = {p1, p2, ..., _pn } are collected from the excitation system, where _pn represents the i-th parameter, such as current, voltage, etc.; relevant data D = {d1, d2, ..., _dj } are collected from the thermal power units, where _dj represents the j-th parameter, such as load and efficiency.

Furthermore, data integration includes synchronizing the timestamps of the collected data to ensure that the data points of P and D are aligned in time. The specific formula for associating and merging P and D is as follows:

N = merge(P,D)

Among them, N is the data set that is associated and merged with P and D, P is the parameters collected from the excitation system, and D is the relevant data collected from the thermal power unit.

Furthermore, linear interpolation is performed on the missing values of the data. When all the data with the same timestamp are lost, linear interpolation is used to process the missing data. When one data is missing in the data P or D with the same timestamp, the data is supplemented by analyzing other data with the same timestamp. The specific formula for linear interpolation is as follows:

Where t is a timestamp, and _t1 and _t2 are adjacent time points.

Specifically, for each data sequence, the rate of change between consecutive time points is calculated, and the specific formula is as follows:

ΔP _i (t) = P _i (t-1)P _i (t) - _{P i} (t-1)

ΔD _i (t) = D _j (t-1) D _j (t) - D _j (t-1)

Among them, _Pi and _Dj are data sequences, and t is the timestamp.

Furthermore, the correlation between the change rates is analyzed, and the change rate of the missing data is estimated based on the correlation of the change rates; the data sequences P _i and D _j with the highest correlation with the change rate of the missing data are selected. If the data sequence P _i is missing, the estimated change rate of the missing data is ΔP _i ^(t)=α+β·ΔD _j (t); if the data sequence D _j is missing, the estimated change rate of the missing data is ΔD _j ^(t)=α+β·ΔP _i (t), where α and β are coefficients calculated based on historical data.

Furthermore, the estimated rate of change is used to fill in the missing data. If _Pi (t-1) is known, then _Pi (t) = _Pi (t-1) × (1 + _ΔPi ^ (t)); if D _j (t-1) is known, then D _j (t) = D _j (t-1) × (1 + ΔD _j ^ (t)).

S2: Extract features from the preprocessed integrated data, establish a performance evaluation model for the excitation system of the thermal power unit, and perform performance analysis.

Specifically, the dynamic patterns and shapes in time series data are captured, and the dynamic time warping (DTW) algorithm is used to compare time series and extract similar patterns. The specific formula is as follows:

Among them, S and T are time series.

Furthermore, the complexity and irregularity of time series data are estimated, and multi-scale entropy is calculated to quantify the complexity of data at different time scales. The specific formula is as follows:

MSE(k)＝-∑p(x)logp(x)

Among them, p(x) is the probability distribution at scale k.

Furthermore, a performance evaluation model for the excitation system of a thermal power unit is established by dividing the time series data into a training set and a test set.

Specifically, as shown in Table 1, in terms of error, the training set error indicates how well the model fits the known data, the validation set error is an indicator of the model's generalization ability on unseen data, and the test set error provides the performance of the model in actual application scenarios.

Table 1 Performance analysis table

Model Training set error Validation set error Test set error Accuracy Accuracy Recall F1 score Model A 0.012 0.015 0.016 0.88 0.85 0.92 0.88 Model B 0.008 0.011 0.013 0.92 0.89 0.94 0.91 Model C 0.010 0.013 0.014 0.89 0.87 0.91 0.89

Furthermore, accuracy is the proportion of samples correctly classified by the model on the entire test set, while precision focuses on the accuracy of the model in predicting the positive category, and recall focuses on the model's ability to capture the positive category. The F1 score combines precision and recall, providing a metric that balances the two. Through these indicators, the performance of the model can be more comprehensively evaluated to understand its classification capabilities in positive and negative categories.

S3: Use historical data to train and validate the evaluation model, optimize model parameters, and improve the accuracy and reliability of the evaluation model.

Specifically, the model is trained using the training set, and the model parameters are updated by using the decision tree algorithm to adapt it to the pattern of historical data; the performance of the model on unseen data is evaluated using the validation set to monitor whether the model is overfitting or underfitting; the model's hyperparameters are tuned using a random search method to find the best parameter combination; cross-validation technology is used to further verify the model's stability and generalization performance; the model's errors on the training set and validation set are monitored in real time to ensure that the model is not overfitting.

Furthermore, during the training process, save the model parameters with the best performance on the validation set for subsequent use; set a rollback mechanism, and if the model performance does not improve significantly within a certain period, roll back to the previous parameter state to prevent the model from overfitting. Record the model parameters and performance indicators of each experiment for subsequent analysis and tracing.

Specifically, when using historical data to train and verify the model, an error threshold is set. When the model error is greater than the threshold, the model parameters are adjusted and optimized in the hope of reducing the model error. When the model error is equal to the threshold, the model parameters are checked and fine-tuned to improve the performance of the model. When the model error is less than the threshold, the model's performance on other data sets is further verified to ensure the generalization performance of the model.

Furthermore, the error threshold needs to be weighed according to the sensitivity of the specific task, the level of data noise, the application scenario requirements, and the computing resources and efficiency. For tasks with low error tolerance, a smaller error threshold is selected to ensure high accuracy, while for some scenarios that place more emphasis on model robustness and generalization ability, the threshold can be moderately relaxed to improve efficiency.

S4: Input the real-time collected data into the trained evaluation model to perform online evaluation and prediction of excitation system faults, and output the evaluation results.

Specifically, when real-time input data is input into the evaluation model, real-time anomaly detection is performed; when the input data is normal, the evaluation model is executed according to the predetermined process and the corresponding evaluation results are output; when the input data is abnormal, the corresponding alarm is output and the abnormal information is recorded, and different abnormal situations are set with abnormal levels to flexibly handle abnormal situations.

Furthermore, if the abnormality level is level one, the abnormal information is recorded, the evaluation model continues to be executed, and the abnormality level is indicated in the subsequent report; if the abnormality level is level two, an alarm is output and the current evaluation task is suspended; if the abnormality level is level three, the current task is stopped immediately, an emergency alarm is output, and the corresponding system self-rescue mechanism is triggered.

Furthermore, as shown in Table 2, for the model real-time evaluation process record, each row represents an evaluation event, where the "Is the input data normal" column indicates the status of the real-time input data, which may be normal or abnormal. Abnormal situations are divided into different levels, represented by the "Abnormal Level" column, which are divided into no abnormality (0), level 1 abnormality, level 2 abnormality, level 3 abnormality, etc. When the system detects an abnormality, it will output the corresponding "alarm information" to describe the nature and extent of the abnormality. Finally, the "Model Evaluation Result" column records the model's evaluation results for the current data, which may be normal, not evaluated (evaluation interrupted due to abnormality) or other relevant evaluation information.

Table 2 Evaluation results

Is the input data normal? Abnormal Level Alert Message Model evaluation results yes none none normal no 2 Abnormal data, large fluctuations Not evaluated yes 1 Slight fluctuation, continue to evaluate normal no 3 Serious abnormality, stop evaluation Not evaluated

S5: Analyze the evaluation results to determine whether the performance of the excitation system has declined, achieve early detection of faults, and ensure stable and reliable operation of the excitation system.

Specifically, by analyzing the evaluation results, the interdependence between features and faults is quantified. The mutual information value between each feature and the fault label is calculated. According to the DTW analysis results, it is determined whether the system is degraded, and the degradation threshold θ _deg is defined.

Furthermore, when D _DTW ≤θ _deg , the performance is judged to be normal; when D _DTW >θ _deg , the performance is judged to be degraded; if the performance is degraded, the value of the mutual information is adjusted, and the relevant formula is as follows:

Among them, B is the number of steps calculated according to the longest step length each time the decay rate is reached, R _decay is the decay rate, and M is the value of the entire information. The specific formula is as follows:

MI(X;Y)＝∑p(x,y)logp(y)p(x,y)x∈X,y∈Y

Among them, X is the quantitative feature and Y is the fault label.

Furthermore, the fault mode is established by analyzing the evaluation results; when the real-time evaluation results do not match the known fault mode, the current real-time evaluation data and model output are recorded; when the real-time evaluation results highly match the known fault mode, fault detection is triggered to remind more in-depth inspection and maintenance; when the real-time evaluation results poorly match the known fault mode, light processing is selected to record the current real-time evaluation data and model output.

Furthermore, the present embodiment also provides an online evaluation system for an excitation system of a thermal power unit, comprising: an acquisition module, for collecting data of the excitation system of the thermal power unit and other relevant data, integrating them, and performing preprocessing; an extraction module, for performing feature extraction on the preprocessed integrated data, establishing a performance evaluation model for the excitation system of the thermal power unit, and performing performance analysis; an optimization module, for training and verifying the evaluation model using historical data, optimizing model parameters, and improving the accuracy and reliability of the evaluation model; an evaluation module, for inputting real-time collected data into the trained evaluation model, performing online evaluation and prediction of excitation system faults, and outputting evaluation results; an analysis module, for analyzing the evaluation results, determining whether the performance of the excitation system has declined, achieving early detection of faults, and ensuring stable and reliable operation of the excitation system.

This embodiment also provides a computer device, which is applicable to the case of an online evaluation method for an excitation system for a thermal power unit, and includes 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 online evaluation method for an excitation system for a thermal power unit proposed in the above embodiment.

The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the following steps are implemented: collecting data of the excitation system of the thermal power unit and other relevant data, integrating them, and preprocessing them; extracting features from the preprocessed integrated data, establishing a performance evaluation model for the excitation system of the thermal power unit, and performing performance analysis; using historical data to train and verify the evaluation model, optimize model parameters, and improve the accuracy and reliability of the evaluation model; inputting real-time collected data into the trained evaluation model, performing online evaluation and prediction of excitation system faults, and outputting evaluation results; analyzing the evaluation results to determine whether the performance of the excitation system is normal, achieve early detection of faults, and ensure stable and reliable operation of the excitation system.

In summary, the present invention improves the accuracy of fault prediction by combining dynamic time warping DTW and mean square error MSE for feature extraction; uses DTW analysis and performance degradation evaluation to timely identify small changes in the system and potential performance degradation, thereby achieving early detection of faults; by adjusting the mutual information value by considering the performance decay rate, the decision tree model dynamically adapts to changes in the system state and improves the adaptability to new fault modes; it flexibly adjusts and optimizes according to the characteristics and operating conditions of the excitation system of different thermal power units, and has good scalability, thereby improving operating efficiency, reducing risks, and ensuring stable and reliable operation of the power system.

Example 2

Referring to Table 3, which is a second embodiment of the present invention, this embodiment provides an online evaluation method for an excitation system for a thermal power unit. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

Specifically, the exciter generates a DC current that excites the electromagnetic field of the generator. This is usually a small generator, also called an excitation generator; the voltage regulator is used to adjust the current of the exciter to ensure that the voltage output by the generator is maintained at a stable level. The voltage regulator can be mechanical, electromagnetic or digital based on modern electronic components; the excitation power supply provides the DC power required by the exciter, which can be a system consisting of a DC generator, rectifier and capacitors; the excitation control system monitors the voltage, current and other related parameters, and adjusts the output of the voltage regulator and excitation power supply as needed to maintain the normal operation of the generator.

Furthermore, the operating principle of the excitation system is that when the thermal power unit is started, the DC current generated by the exciter is used to excite the electromagnetic field of the generator; usually in the starting phase, the excitation power supply can be provided by an external DC power supply to ensure that the generator can establish an electromagnetic field during the starting process; in the operating phase, if the generator has established a sufficient electromagnetic field, the excitation power supply starts to provide excitation current, the voltage regulator monitors the output voltage of the generator, and adjusts the current of the exciter according to the set value to keep the output voltage at an appropriate level.

Furthermore, when the load changes, the excitation control system adjusts the excitation current by monitoring the generator output to ensure the stability of the output voltage. This is achieved by adjusting the excitation current to change the strength of the electromagnetic field; the excitation system also includes fault protection functions such as overcurrent protection, overvoltage protection and undervoltage protection to ensure that the system can be safely shut down or switched to standby mode under abnormal conditions.

Specifically, as shown in Table 1, in terms of fault prediction accuracy, the accuracy of the traditional method is 75%, while the accuracy of the present invention reaches 92%, which is significantly improved by 17 percentage points. This means that the present invention can more accurately predict the fault of the excitation system of the thermal power unit and help the operator take timely maintenance measures.

Table 3 Comparison between traditional method and the present invention

Furthermore, in terms of early fault detection capability, the traditional method can only detect 50% of early faults, while the present invention increases this to 81%, an increase of 31 percentage points. This means that the present invention has a stronger fault early warning capability and can capture problems at the early stage of a fault, helping to avoid more serious consequences.

Furthermore, the adaptability of the conventional method to new fault modes is generally low, while the present invention has high adaptability. This means that the present invention can effectively deal with new fault modes, and can provide accurate fault diagnosis and prediction even for situations that have not been encountered before.

Specifically, in terms of maintenance cost, the maintenance cost of the traditional method is 30,000, while the maintenance cost of the present invention is 40,000. Although the maintenance cost of the present invention is slightly higher, its improved fault prediction accuracy and early fault detection capabilities, as well as its adaptability to new fault modes, can help reduce the higher cost accident rate and maintenance costs.

Furthermore, the present invention is also superior to traditional methods in terms of system reliability and user-friendliness. By improving the accuracy of fault prediction and early fault detection capabilities, the present invention can improve the reliability of the system and avoid accidents. At the same time, the present invention has also improved user-friendliness, providing an interface and result display that is easier to operate and understand.

Furthermore, the present invention is superior to traditional methods in terms of data processing and analysis time. Due to the use of more advanced algorithms and technologies, the present invention can process and analyze large amounts of data more quickly and provide real-time feedback and results.

In summary, compared with the traditional method, the present invention performs better in fault prediction accuracy, early fault detection capability, adaptability to new fault modes, system reliability, user friendliness, and data processing and analysis time. These improvements can help improve the performance evaluation and fault prediction capabilities of the excitation system of thermal power units and ensure the stable and reliable operation of the excitation system.

It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims (10)

1. An online evaluation method for an excitation system of a thermal power unit, characterized in that: it comprises: Collect data from the excitation system of thermal power units and other related data, integrate them, and perform preprocessing; Extract features from pre-processed integrated data, establish a performance evaluation model for the excitation system of thermal power units, and perform performance analysis; Use historical data to train and verify the evaluation model, optimize model parameters, and improve the accuracy and reliability of the evaluation model; Input the real-time collected data into the trained evaluation model to conduct online evaluation and prediction of excitation system faults and output the evaluation results; Analyze the evaluation results to determine whether the performance of the excitation system has declined, achieve early detection of faults, and ensure stable and reliable operation of the excitation system. 2. The online evaluation method for the excitation system of a thermal power plant according to claim 1, characterized in that the preprocessing comprises the following steps: Linear interpolation is performed on missing data values; When all data with the same timestamp are lost, linear interpolation is used to process the missing data; When one data is missing in data P or D with the same timestamp, the data is supplemented by analyzing other data with the same timestamp; For each data series, the rate of change between consecutive time points is calculated; Estimate the change rate of missing data according to the correlation of the change rate, and select the data series P _i and D _j with the highest correlation with the change rate of missing data; If the data sequence _Pi is missing, the estimated rate of change of the missing data is _ΔPi ^(t) = α + β· _ΔDj (t); If the data sequence D _j is missing, the estimated rate of change of the missing data is ΔD _j ^(t) = α + β·ΔP _i (t), where α and β are coefficients calculated based on historical data; Use the estimated rate of change to fill in the missing data. If _Pi (t-1) is known, then _Pi (t) = _Pi (t-1) × (1 + _ΔPi ^ (t)); if D _j (t-1) is known, then D _j (t) = D _j (t-1) × (1 + ΔD _j ^ (t)). 3. The online evaluation method for the excitation system of a thermal power plant according to claim 2, characterized in that: the relevant formula for calculating the change rate between consecutive time points for each data sequence is as follows: ΔP _i (t) = P _i (t-1)P _i (t) - _{P i} (t-1) ΔD _i (t) = D _j (t-1) D _j (t) - D _j (t-1) Among them, _Pi and _Dj are data sequences, and t is the timestamp; The relevant formula for linear interpolation of missing data values is as follows: Among them, t is the timestamp, and _t1 and _t2 are adjacent time points. 4. The online evaluation method for the excitation system of a thermal power plant according to claim 1, characterized in that the performance analysis comprises the following steps: Capture dynamic patterns and shapes in time series data; The dynamic time warping (DTW) algorithm is used to compare time series and extract similar patterns. The relevant formula is as follows: Among them, S and T are time series; Estimate the complexity and irregularity of time series data and calculate multiscale entropy. The relevant formula is as follows: MSE(k)＝-∑p(x)logp(x) Among them, p(x) is the probability distribution at scale k. 5. The online evaluation method for the excitation system of a thermal power plant according to claim 4, characterized in that: the comparison of time series using a dynamic time warping (DTW) algorithm comprises the following steps: When D _DTW ≤θ _deg , the performance is judged to be normal; When D _DTW >θ _deg , it is judged that the performance is degraded; if the performance is degraded, the value of the mutual information is adjusted, and the relevant formula is as follows: Among them, B is the number of steps calculated according to the longest step length each time the decay rate is reached, R _decay is the decay rate, and MI is the value of the entire information. The specific formula is as follows: MI(X;Y)＝∑p(x,y)logp(y)p(x,y)x∈X,y∈Y Among them, X is the quantitative feature and Y is the fault label. 6. The method for online evaluation of the excitation system for a thermal power plant according to claim 1, characterized in that: the step of inputting the real-time collected data into the trained evaluation model comprises the following steps: Perform real-time anomaly detection when inputting data into the evaluation model in real time; When the input data is normal, the evaluation model is executed according to the predetermined process and the corresponding evaluation results are output; When the input data is abnormal, the corresponding alarm is output and the abnormal information is recorded. Different abnormal situations are set with abnormal levels to handle abnormal situations flexibly. If the abnormality level is level 1, the abnormality information is recorded, the evaluation model continues to be executed, and the abnormality level is indicated in the subsequent report; If the abnormality level is level 2, an alarm is output and the current evaluation task is suspended; If the abnormality level is level three, the current task will be stopped immediately, an emergency alarm will be output, and the corresponding system self-rescue mechanism will be triggered. 7. The method for online evaluation of the excitation system for a thermal power plant according to claim 1, wherein the step of judging whether the performance of the excitation system is degraded comprises the following steps: Establish failure modes by analyzing the evaluation results; When the real-time evaluation result does not match the known fault mode, the current real-time evaluation data and model output are recorded; When the real-time assessment results are highly matched with known fault modes, fault detection is triggered, prompting more in-depth inspection and maintenance; When the real-time evaluation results have a low match with the known fault mode, light processing is selected to record the current real-time evaluation data and model output. 8. An online evaluation system for an excitation system of a thermal power unit, based on the online evaluation method for an excitation system of a thermal power unit according to any one of claims 1 to 7, characterized in that it comprises: The acquisition module is used to collect data from the excitation system of the thermal power unit and other related data for integration and preprocessing; The extraction module extracts features from the preprocessed integrated data, establishes a performance evaluation model for the excitation system of a thermal power unit, and performs performance analysis; The optimization module is used to train and verify the evaluation model using historical data, optimize model parameters, and improve the accuracy and reliability of the evaluation model; The evaluation module is used to input the real-time collected data into the trained evaluation model, perform online evaluation and prediction of excitation system faults, and output the evaluation results; The analysis module is used to analyze the evaluation results, determine whether the performance of the excitation system has declined, realize early detection of faults, and ensure the stable and reliable operation of the excitation system. 9. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein the processor implements the steps of the online evaluation method for an excitation system for a thermal power unit according to any one of claims 1 to 7 when executing the computer program. 10. A computer-readable storage medium having a computer program stored thereon, characterized in that: when the computer program is executed by a processor, the steps of the online evaluation method for an excitation system for a thermal power unit according to any one of claims 1 to 7 are implemented.