Equipment early warning system for thermal control of thermal power plant

Abstract

The invention provides an early warning system for thermal control equipment of a thermal power plant, and belongs to the technical field of industrial management, and the system comprises a model building module which obtains internal environment data of the thermal power plant and operation monitoring data of the thermal control equipment of the thermal power plant, and builds a normal behavior model of the operation of the thermal control equipment of the thermal power plant; the deviation determination module is used for determining the real-time operation deviation condition of the thermal control equipment of the thermal power plant and further determining the real-time abnormal data of the thermal control equipment of the thermal power plant; the score determination module is used for determining a real-time health score of the thermal control equipment of the thermal power plant; the strategy determination module is used for determining a real-time fault reason of the thermal control equipment of the thermal power plant, and further determining a comprehensive early warning strategy based on the real-time fault reason of the thermal control equipment of the thermal power plant; and the early warning execution module is used for determining an early warning level based on the real-time health score of the thermal control equipment of the thermal power plant and sending out an early warning signal. And the fault prediction and early warning capabilities of the thermal control equipment of the thermal power plant are improved.

Description

Technical Field

[0001] The invention relates to the technical field of industrial management, and in particular to an early warning system for thermal control equipment in a thermal power plant. Background Art

[0002] With the expansion of the scale of thermal power plants and the continuous advancement of technology, the role of thermal control equipment in thermal power plants has become increasingly critical. The thermal control equipment of thermal power plants is responsible for ensuring the efficient and stable operation of the power generation process, and their performance directly affects the safety and efficiency of the entire power generation system. In order to avoid reduced power generation efficiency or safety accidents caused by equipment failure, these devices must be continuously monitored and maintained. However, due to the complex internal environment of thermal power plants and the changing operating conditions of thermal control equipment, traditional monitoring and maintenance methods are difficult to fully predict and promptly handle equipment abnormalities. Therefore, it is crucial to develop intelligent systems that can monitor in real time and provide early warnings.

[0003] Currently, the monitoring of thermal power plant equipment mainly relies on the installation of conventional sensors and regular inspections by maintenance personnel. Existing monitoring systems can usually only detect relatively obvious faults, and often the occurrence of faults is inevitable, resulting in abnormal equipment operation before they can react. In addition, although some fault diagnosis systems based on data analysis and modeling have been applied in industrial environments, they usually rely on pre-defined abnormality criteria and fail to fully utilize the historical operating data of the equipment and complex environmental changes to build accurate prediction models.

[0004] Therefore, the present invention provides an early warning system for thermal control equipment in a thermal power plant. Summary of the invention

[0005] The present invention provides an early warning system for thermal control equipment in a thermal power plant, which is used to solve the problems in the prior art of lack of dynamic real-time prediction, insufficient use of historical data and inflexible early warning strategies. By combining real-time monitoring of equipment, historical data analysis and intelligent early warning strategies, the defects of the prior art such as inability to issue early warnings and insufficient use of data are overcome, and the fault prediction and early warning capabilities of thermal control equipment in thermal power plants are effectively improved.

[0006] The present invention provides an early warning system for thermal control equipment in a thermal power plant, comprising: Model building module: Deploy sensors in thermal control equipment and inside thermal power plants to obtain internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants. Based on the internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants, establish a normal behavior model of thermal control equipment of thermal power plants; Deviation determination module: Determine the real-time operation deviation of thermal control equipment in a thermal power plant based on the normal behavior model of the thermal control equipment in a thermal power plant and the real-time operation monitoring data of the thermal control equipment in a thermal power plant, and then analyze the real-time operation deviation of the thermal control equipment in a thermal power plant to determine the real-time abnormal data of the thermal control equipment in a thermal power plant; Scoring determination module: determines the real-time health score of thermal control equipment in a thermal power plant based on the historical abnormal data of thermal control equipment in a thermal power plant and the real-time abnormal data of thermal control equipment in a thermal power plant; Strategy determination module: Analyze the real-time health scores of thermal control equipment in thermal power plants and the historical abnormal data of thermal control equipment in thermal power plants to determine the real-time failure causes of thermal control equipment in thermal power plants, and then determine a comprehensive early warning strategy based on the real-time failure causes of thermal control equipment in thermal power plants; Early warning execution module: Determines the early warning level based on the real-time health score of thermal control equipment in thermal power plants, and issues early warning signals based on the early warning level and comprehensive early warning strategy.

[0007] The present invention provides an early warning system for thermal control equipment in a thermal power plant, a deviation determination module, comprising: Data fusion unit: fuses the internal environment data of the thermal power plant with the operation monitoring data of the thermal control equipment of the thermal power plant based on the preset fusion algorithm; Model building unit: Based on the fused data, normal behavior models of several thermal control equipment in thermal power plants are established; Model combination unit: based on the analysis of the operation monitoring data of the thermal control equipment of the thermal power plant, the non-fault state data of the thermal control equipment of the thermal power plant is obtained, the non-fault state data of the thermal control equipment of the thermal power plant is input into the normal behavior model of the thermal control equipment of each thermal power plant, and the output result of the normal behavior model of the thermal control equipment of each thermal power plant is obtained. Based on the output result of the normal behavior model of the thermal control equipment of each thermal power plant and the preset judgment algorithm, the combination weight of the normal behavior model of the thermal control equipment of each thermal power plant is determined, and then based on the combination weight and the preset combination rule, several normal behavior models of the thermal control equipment of the thermal power plant are combined to obtain a combination model.

[0008] The present invention provides an early warning system for thermal control equipment in a thermal power plant, a model combination unit, comprising: Model screening subunit: constructing a candidate model set based on normal behavior models of thermal control equipment of several thermal power plants, determining the score of each normal behavior model of thermal control equipment of thermal power plants in the candidate model set based on a preset scoring algorithm, and determining the normal behavior models of thermal control equipment of several thermal power plants whose scores are greater than a preset score threshold as the first model; A weight determination subunit: obtaining the output of each first model, and then determining the weight of each first model in combination with a preset algorithm; Model combination subunit: based on the weights of all the first models, all the first models are combined according to preset fusion rules to generate a basic combination model; Model generation subunit: takes the output of the basic combination model as input and generates the final combination model based on the preset learner and the internal environmental data of the thermal power plant.

[0009] The present invention provides an early warning system for thermal control equipment in a thermal power plant, a model building module, comprising: Parameter determination unit: determines the deviation value of the real-time operation prediction of thermal control equipment in thermal power plants based on the preset model set technology and the final combined model; Data acquisition unit: Determine the real-time operation deviation of all thermal control equipment in the thermal power plant based on the deviation value of the real-time operation prediction of the thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant:

[0010]

[0011]

[0012] in, It is the real-time operation deviation of all thermal control equipment in thermal power plants. It is the basic operation deviation of all thermal control equipment in thermal power plants. is the historical operating deviation of all thermal control equipment in thermal power plants, is the weight coefficient of the i-th sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the actual measurement value of the sensor at time t that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the deviation value predicted by the final combined model at time t by the sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant for the ith time, is the historical data impact factor, is the correlation coefficient of the sensor data of the jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant relative to the sensor data of the ith sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, reflecting the correlation between the sensors. The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time The actual value of is the time lag, indicating that at time forward The state of the moment, The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time According to the deviation value predicted by the final combined model, is the environmental impact factor, is the environmental value at time t, is the standard environmental value under the normal operating environment of the equipment; m is the total number of sensors that obtain the operating monitoring data of thermal control equipment in thermal power plants; Situation determination unit: determining the real-time deviation situation of all thermal control equipment of the thermal power plant based on the real-time operation deviation amount of all thermal control equipment of the thermal power plant; Threshold determination unit: determines the abnormality judgment threshold based on the real-time operation deviation of all thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant:

[0013] in, is the abnormality judgment threshold, g is the number of operating parameters of thermal control equipment in thermal power plants, is the mean value of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is the standard deviation of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is a preset adjustable coefficient that determines the sensitivity of the threshold, and , is the sth real-time operating parameter of thermal control equipment in thermal power plants, is the coefficient used to adjust the effect of deviation on the threshold, It is the real-time operation deviation of thermal control equipment in thermal power plants. is the time change rate of the sth operating parameter of thermal control equipment in thermal power plants, The preset weights corresponding to the real-time operation status of thermal control equipment in thermal power plants. The preset weights corresponding to the real-time deviations of thermal control equipment in thermal power plants; Data determination unit: determines the abnormal situation of the equipment based on the real-time deviation situation of the equipment and the abnormality judgment threshold, and then determines the real-time deviation data of the equipment and the corresponding real-time operation parameters as the equipment abnormality data.

[0014] The present invention provides an early warning system for thermal control equipment in a thermal power plant, a scoring determination module, comprising: Establish health scoring rules based on multiple factors, combine historical abnormal data and real-time abnormal data to score the health status of thermal control equipment in thermal power plants, and perform weighted fusion of the scoring results of each factor based on the fuzzy logic scoring model; Obtain the service life, maintenance history, and load changes of thermal control equipment in thermal power plants and adjust the scoring results based on the preset adjustment algorithm; The adjusted scoring results are compared and analyzed with the historical scoring trends, and the real-time health score of the thermal control equipment in the thermal power plant is determined based on the analysis results.

[0015] The present invention provides a thermal power plant thermal control equipment early warning system, a strategy determination module, comprising: Chain formation unit: Analyze historical abnormal data based on the preset reasoning model to determine the failure cause chain of thermal control equipment in thermal power plants; Graph construction unit: constructs an abnormal graph of thermal control equipment in a thermal power plant based on the fault cause chain of thermal control equipment in a thermal power plant; Spectrum analysis unit: Analyze the abnormal spectrum of thermal control equipment in thermal power plants to determine the comprehensive risk value of thermal control equipment failures in thermal power plants:

[0016] in, is the comprehensive risk value of thermal control equipment in thermal power plants, n is the number of failure types of thermal control equipment in thermal power plants, is the weight of the kth fault type, is the basic risk value of the kth fault type, is the time decay factor, is the correlation coefficient between the kth fault type and other fault types except the kth fault type, Divide the basic risk value of the hth fault type by the kth fault type, where is the basic risk value of the kth fault type:

[0017] in, is the occurrence frequency of the kth fault type, is the importance of the kth fault type, is the detection difficulty of the kth fault type, Score the health status of thermal control equipment in thermal power plants; Strategy determination unit: Determine the initial early warning strategy based on the comprehensive risk value of failures of thermal control equipment in thermal power plants, and continuously collect real-time early warning effect data after the execution of the early warning strategy, and optimize the early warning strategy in combination with the preset optimization algorithm, and then determine the comprehensive early warning strategy.

[0018] The present invention provides an early warning system for thermal control equipment in a thermal power plant, and the early warning signal includes: Parameter determination unit: Analyze the historical abnormal data of thermal control equipment in thermal power plants, determine several abnormal states and corresponding associated equipment parameters; Transition determination unit: determines the transition between all abnormal states based on the abnormal state and the corresponding associated device parameters, and then constructs an abnormal state transition diagram; Path determination unit: Analyze the abnormal state transition diagram to determine the critical path on each transfer path of each abnormal state; Cause determination unit: determines the fault based on the abnormal state, and determines the cause of the fault based on the critical path on each transfer path of each abnormal state; History analysis unit: determines the impact of each fault in several preset time periods based on historical abnormal data; Parameter series connection unit: connects the fault, fault cause and corresponding associated equipment parameters in series according to time and impact degree, thus forming a fault cause chain.

[0019] The present invention provides an early warning system for thermal control equipment in a thermal power plant, and the early warning signal includes: early warning level, equipment name and location, fault type, real-time health score of the equipment, fault cause, early warning recommended measures, historical abnormal records corresponding to the fault cause, and predictive maintenance prompts corresponding to the fault cause.

[0020] The present invention provides an early warning system for thermal control equipment in a thermal power plant, which is used to solve the problems in the prior art of lack of dynamic real-time prediction, insufficient use of historical data and inflexible early warning strategy. By combining real-time monitoring of equipment, historical data analysis and intelligent early warning strategy, the defects of the prior art such as inability to issue early warning and insufficient data utilization are overcome, and the fault prediction and early warning capabilities of thermal control equipment in thermal power plants are effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 The present invention is a schematic structural diagram of a thermal power plant thermal control equipment early warning system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Example 1 like Figure 1 As shown, an embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, comprising: Model building module: Deploy sensors in thermal control equipment and inside thermal power plants to obtain internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants. Based on the internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants, establish a normal behavior model of thermal control equipment of thermal power plants; Deviation determination module: Determine the real-time operation deviation of thermal control equipment in a thermal power plant based on the normal behavior model of the thermal control equipment in a thermal power plant and the real-time operation monitoring data of the thermal control equipment in a thermal power plant, and then analyze the real-time operation deviation of the thermal control equipment in a thermal power plant to determine the real-time abnormal data of the thermal control equipment in a thermal power plant; Scoring determination module: determines the real-time health score of thermal control equipment in a thermal power plant based on the historical abnormal data of thermal control equipment in a thermal power plant and the real-time abnormal data of thermal control equipment in a thermal power plant; Strategy determination module: Analyze the real-time health scores of thermal control equipment in thermal power plants and the historical abnormal data of thermal control equipment in thermal power plants to determine the real-time failure causes of thermal control equipment in thermal power plants, and then determine a comprehensive early warning strategy based on the real-time failure causes of thermal control equipment in thermal power plants; Early warning execution module: Determines the early warning level based on the real-time health score of thermal control equipment in thermal power plants, and issues early warning signals based on the early warning level and comprehensive early warning strategy.

[0025] In this embodiment, the thermal control equipment in the thermal power plant refers to various types of equipment used to control and adjust the thermal system in the thermal power plant, which ensures the effective conversion and utilization of thermal energy in the power generation process. They include boilers, steam turbines, condensers, heat exchangers and other equipment, as well as sensors, actuators and control systems that control the operation of these equipment. For example, the boiler control system of a thermal power plant regulates fuel combustion and steam production to ensure that the boiler operates within a safe range. If the boiler overheats, it may cause failures or dangers, so continuous monitoring is required; In this embodiment, the process of establishing a normal behavior model for the operation of thermal control equipment in a thermal power plant is as follows: Data collection: Equipment operation monitoring data (such as temperature, pressure, etc.) and environmental data (such as humidity, room temperature, etc.) are collected through sensors to form time series data. Data preprocessing: The collected data is cleaned, completed, denoised and normalized to eliminate outliers and noise to ensure the accuracy and consistency of the data. Feature extraction: Extract key features, including time series features (sliding average, rate of change), related features (relationships between equipment parameters) and equipment behavior pattern features. Model selection and training: Select a suitable algorithm (such as linear regression, ARIMA, LSTM, etc.), train the model based on historical data, and learn the behavior pattern of the equipment under normal operating conditions. Model verification and optimization: Evaluate the accuracy of the model through cross-validation and performance indicators (such as MSE, R²), and adjust hyperparameters to optimize model performance. Model deployment: Deploy the model to the real-time monitoring system and dynamically update it according to new data to ensure that the model adapts to changes in equipment operation.

[0026] In this embodiment, the internal environmental data of a thermal power plant refers to information related to the operating environment in the thermal power plant, which usually includes temperature, humidity, pressure, airflow, flue gas composition, etc. Since these environmental parameters will affect the operating status of the equipment, real-time monitoring of these data is essential to ensure the normal operation of the equipment. For example, in a boiler room, changes in temperature and pressure may have a direct impact on the operation of the boiler. If the temperature in the room is too high, it may reduce the efficiency of the equipment or even cause equipment failure; In this embodiment, the operation monitoring data of the thermal control equipment of the thermal power plant refers to the operation parameters obtained in real time from the thermal control equipment of the thermal power plant through sensors, including temperature, pressure, flow rate, power, vibration frequency and other data. The monitoring data is used to determine the current working state of the equipment. For example, the speed, steam inlet temperature and pressure of the steam turbine are monitored to ensure that the steam turbine operates under the best working conditions and avoid abnormal situations. In this embodiment, the normal behavior model is a reference model constructed based on the historical data and environmental data of the device, which represents the operating state and parameter range of the device under normal conditions. By comparing the current data with the model, it can be determined whether the device is in a normal operating state. For example, for a boiler system in a thermal power plant, the normal behavior model may define the range of parameters such as pressure and temperature in the boiler under different load conditions. If the current data deviates from this range, the system will determine it as abnormal.

[0027] In this embodiment, the real-time health score is a quantitative result of the health status of the device obtained by analyzing the current operating status of the device (including real-time monitoring data and abnormal conditions). This score reflects the health status of the device and is used to evaluate whether the device needs maintenance or repair. For example, a health score can be from 0 to 100, 100 means that the device is in good operating condition, and a score below 50 may mean that the device has serious hidden dangers and needs immediate maintenance; In this embodiment, the real-time fault cause is the specific cause of the equipment failure obtained based on the analysis of real-time data and historical data. Understanding the cause of the fault helps to take targeted maintenance measures to avoid more serious problems. For example, if data analysis finds that the abnormal temperature of the boiler is due to a blockage in the fuel delivery system, the system will give the specific cause of the fault so that repairs can be carried out; In this embodiment, the comprehensive early warning strategy is a processing strategy generated by the system based on the health score of the equipment, the cause of the fault, etc. The strategy may include multiple measures such as reminding the operator, automatically shutting down, starting the backup system, etc. to ensure the safe operation of the equipment and the thermal power plant. For example, if the system detects that the boiler is about to overheat and may cause damage, the comprehensive early warning strategy may include issuing an alarm, reducing the fuel supply, or automatically shutting down the boiler.

[0028] In this embodiment, the warning level is a different alarm level set according to the health score of the equipment and the degree of abnormality. Different levels represent the severity of the problem, usually divided into three levels: low, medium and high. For example, when the boiler pressure is only slightly beyond the normal range, the system may issue a low-level alarm, but if the pressure rises sharply, the system will issue a high-level alarm, which may be accompanied by an automatic shutdown.

[0029] In this embodiment, the early warning signal is a prompt message issued by the system when a problem is detected. It is usually conveyed to the operator or the remote monitoring center in the form of sound, light, text message, email, etc., prompting them to take measures. For example, when the boiler is found to be overheated, the system emits a red flashing light and sends a text message to the responsible engineer to inform them that immediate measures need to be taken.

[0030] The beneficial effects of the above technical solution are: by combining real-time monitoring of equipment, historical data analysis and intelligent early warning strategies, it overcomes the defects of the existing technology such as the inability to provide early warning and insufficient data utilization, effectively improves the fault prediction and early warning capabilities of thermal control equipment in thermal power plants, and solves the problems of lack of dynamic real-time prediction, insufficient utilization of historical data and inflexible early warning strategies in the existing technology. Example 2 An embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, a deviation determination module, including: Data fusion unit: fuses the internal environment data of the thermal power plant with the operation monitoring data of the thermal control equipment of the thermal power plant based on the preset fusion algorithm; Model building unit: Based on the fused data, normal behavior models of several thermal control equipment in thermal power plants are established; Model combination unit: based on the analysis of the operation monitoring data of the thermal control equipment of the thermal power plant, the non-fault state data of the thermal control equipment of the thermal power plant is obtained, the non-fault state data of the thermal control equipment of the thermal power plant is input into the normal behavior model of the thermal control equipment of each thermal power plant, and the output result of the normal behavior model of the thermal control equipment of each thermal power plant is obtained. Based on the output result of the normal behavior model of the thermal control equipment of each thermal power plant and a preset algorithm, the combination weight of the normal behavior model of the thermal control equipment of each thermal power plant is determined, and then based on the combination weight and the preset combination rule, several normal behavior models of the thermal control equipment of the thermal control equipment of the thermal power plant are combined to obtain a combination model.

[0031] In this embodiment, the preset fusion algorithm refers to a pre-designed data processing method for integrating data from different sources (such as environmental data and equipment operation data). Through data fusion, more complete and accurate system status information can be obtained, including weighted average method, Kalman filtering, etc. Assume that the operation monitoring data of the boiler includes temperature, pressure, etc., and the environmental data includes indoor temperature and humidity. The preset fusion algorithm may be a weighted average method, in which the operation monitoring data accounts for 70% of the weight and the environmental data accounts for 30%, so as to obtain a comprehensive equipment status value; In this embodiment, data fusion refers to combining data from multiple different sources (such as equipment monitoring data and environmental data) through a certain algorithm to obtain more reliable comprehensive information. For example, the data collected by the boiler's temperature sensor and the room's ambient temperature sensor are fused to generate a comprehensive temperature value to more accurately judge the boiler's actual working status.

[0032] In this embodiment, establishing normal behavior models of the operation of several thermal control equipment of thermal power plants based on the fused data means that after completing the data fusion, normal behavior models of multiple devices are established based on these fused data. Each model may focus on different equipment or operating parameters, such as temperature, pressure, flow rate, etc. For example, for a boiler system, multiple models can be established to monitor different parameters, such as one model specifically monitoring temperature and another model monitoring pressure. These models together reflect the behavior of the boiler during normal operation.

[0033] In this embodiment, the non-fault state data refers to the operation data collected when the device is working normally. By analyzing these data, a normal behavior model of the device can be established to determine whether the current device is in a normal operating state. For example, in a boiler system, the non-fault state data may include that the temperature of the device under normal conditions is 400°C and the pressure is 5MPa. Any data that deviates from these values ​​may indicate that the device is in an abnormal state.

[0034] In this embodiment, the combined weight of the normal behavior model of each thermal control equipment of each thermal power plant is determined based on the output results of the normal behavior model of each thermal control equipment of each thermal power plant and the preset algorithm, which means that each normal behavior model will produce an output result, which is used to judge the state of the equipment. The importance of each model in the comprehensive judgment, that is, the combined weight, is determined through the preset judgment algorithm. Different models may have different effects on the overall prediction, so it is necessary to assign appropriate weights to them. For example, for the operation monitoring of the boiler, the output results of the pressure model and the temperature model may be different. In order to accurately reflect the health status of the boiler, the system will set an algorithm based on historical data, give the pressure model a weight of 70%, and the temperature model a weight of 30%, because the pressure has a greater impact on the system.

[0035] In this embodiment, the normal behavior models of the thermal control equipment of several thermal power plants are combined based on the combination weights and preset combination rules. Based on the combination weights of each model, multiple normal behavior models are combined together through preset rules to obtain a more accurate comprehensive judgment model. The combined model can comprehensively consider multiple parameters and conditions to conduct a more comprehensive analysis of the equipment operation. It is assumed that there are three models in the operation monitoring of the boiler, which monitor the temperature, pressure and flow rate respectively. By setting weights and rules, the results of these models are combined to form a more comprehensive model for judging the overall health status of the boiler.

[0036] In this embodiment, the combined model is a final model that integrates multiple single normal behavior models (such as temperature models, pressure models, etc.) according to the set combination weights and rules. The model can provide a comprehensive health status assessment and operation prediction of the equipment. For example, the final combined model can combine data from different dimensions such as temperature, pressure and flow rate of the boiler to provide a unified health score or status analysis. This model can not only detect whether the equipment is in an abnormal state, but also warn of potential fault problems earlier.

[0037] The beneficial effects of the above technical solution are: by integrating the internal environmental data of the thermal power plant and the equipment operation monitoring data, a normal behavior model is established, and the model combination is optimized according to the non-fault state data. By combining different data sources and model weights, a more accurate equipment early warning model is generated, which improves the accuracy and efficiency of equipment fault detection from an unconventional perspective, effectively preventing faults from occurring and reducing maintenance costs.

[0038] Example 3 An embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, a model combination unit, comprising: Model screening subunit: constructing a candidate model set based on normal behavior models of thermal control equipment of several thermal power plants, determining the score of each normal behavior model of thermal control equipment of thermal power plants in the candidate model set based on a preset scoring algorithm, and determining the normal behavior models of thermal control equipment of several thermal power plants whose scores are greater than a preset score threshold as the first model; A weight determination subunit: obtaining the output of each first model, and then determining the weight of each first model in combination with a preset algorithm; Model combination subunit: based on the weights of all the first models, all the first models are combined according to preset fusion rules to generate a basic combination model; Model generation subunit: takes the output of the basic combination model as input and generates the final combination model based on the preset learner and the internal environmental data of the thermal power plant.

[0039] In this embodiment, the candidate model set is a set of normal behavior models of several thermal control equipment in thermal power plants. These models are all potential available models obtained through data fusion and construction. The model that best reflects the current state of the equipment is selected from this set. Assume that for boiler operation monitoring, 10 different normal behavior models have been established, and these models focus on parameters such as temperature, pressure, flow rate, fuel input, etc. These 10 models are put together to form a candidate model set, waiting for further screening of better models; In this embodiment, the preset scoring algorithm is an algorithm for evaluating the performance and accuracy of each model in the candidate model set. The scoring algorithm will score each model based on the performance of the model, such as prediction accuracy, error rate, stability, etc., to determine whether they should be selected. For example, when scoring, the algorithm may consider the prediction accuracy of the model in the past 30 days. Assuming that the accuracy of one model is 95% and the other model is 85%, the scoring algorithm will give a higher score to the first model because it performs better; in this embodiment, the score of the normal behavior model refers to the score of each model calculated according to the preset scoring algorithm. The score reflects the accuracy and stability of the model in predicting the normal behavior of the equipment. A high score indicates that the model performs better. For example, the monitoring model for boiler temperature is scored 92 points according to the preset scoring algorithm, while another model for pressure is scored 88 points. These scores are used to determine which models are more suitable for subsequent combinations; In this embodiment, the preset score threshold is a pre-set score limit used to screen the models in the candidate model set. Only models with scores higher than this threshold can be selected as the "first model" for subsequent combination. If the preset score threshold is set to 90 points, then in the candidate model set, only the boiler temperature model with a score of 92 points is selected as the "first model", and the pressure model with a score of 88 points is eliminated; Get the output of each first model, and then combine the preset algorithm and the weight of the first model. The first model refers to the model that has been screened and has a score higher than the preset threshold and is selected for subsequent combination. They represent the optimal normal behavior model and are used to build more complex combination models. For example, the boiler temperature model scored 92 points, which exceeded the threshold of 90 points, so it was selected as the first model. This model will be used in the subsequent combination process.

[0040] Obtain the output of each first model, and then combine the preset algorithm and the weight of the first model. Obtain the output of each first model, and then combine the preset algorithm and the weight of the first model to calculate the importance of each first model. According to the output of each first model and the preset algorithm, determine the weight of these models in the combined model. The higher the weight, the greater the influence of the model in the combination. Suppose there are two first models in the system, one is the temperature model (score 92) and the other is the flow rate model (score 95). The preset algorithm may assign weights according to the score ratio, and may eventually assign 45% weight to the temperature model and 55% weight to the flow rate model.

[0041] In this embodiment, the preset fusion rules refer to the rules used to integrate the outputs of the various first models in the model combination process. These rules determine how to combine the outputs of different models to generate a comprehensive judgment result. For example, the weighted average method performs weighted average on the outputs of the temperature model and the flow rate model according to their weights (45% and 55%) to obtain a comprehensive state value as the preliminary output of the combined model. In this embodiment, the basic combination model is a model formed by integrating the outputs of all the first models through preset fusion rules, which integrates the information of each first model and is used to describe the overall operating status of the equipment. For example, the comprehensive status value obtained by weighted averaging of the temperature model and the flow rate model constitutes the basic combination model, which can reflect the overall health status of the boiler.

[0042] In this embodiment, the output of the basic combination model is used as input, and the final combination model is generated based on the preset learner and the internal environmental data of the thermal power plant. This means that the result of the basic combination model is input into a preset learner, and combined with the internal environmental data of the thermal power plant, further training and optimization are performed to generate a more accurate final combination model. The preset learner can be various machine learning algorithms, such as neural networks, decision trees, etc. For example, the basic combination model outputs a comprehensive health score (such as 70 points), and this score is input into a neural network learner together with environmental data (such as indoor temperature and humidity). Through further training, the learner can optimize the predictive ability of the model, and finally generate a more accurate combination model to evaluate the operating status of thermal power plant equipment in real time.

[0043] The beneficial effects of the above technical solution are: by screening, weighting and combining multiple normal behavior models, a basic combination model is generated, and then the final combination model is generated based on the preset learner and environmental data. The model screening and weight optimization steps are introduced, and combined with environmental data, a more accurate equipment early warning model is realized, which effectively improves the intelligence and prediction accuracy of the early warning system and significantly reduces the possibility of equipment failure.

[0044] Example 4 An embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, a model building module, including: Parameter determination unit: determines the deviation value of the real-time operation prediction of thermal control equipment in thermal power plants based on the preset model set technology and the final combined model; Data acquisition unit: Determine the real-time operation deviation of all thermal control equipment in the thermal power plant based on the deviation value of the real-time operation prediction of the thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant:

[0045]

[0046]

[0047] in, It is the real-time operation deviation of all thermal control equipment in thermal power plants. It is the basic operation deviation of all thermal control equipment in thermal power plants. is the historical operating deviation of all thermal control equipment in thermal power plants, is the weight coefficient of the i-th sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the actual measurement value of the sensor at time t that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the deviation value predicted by the final combined model at time t by the sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant for the ith time, is the historical data impact factor, is the correlation coefficient of the sensor data of the jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant relative to the sensor data of the ith sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, reflecting the correlation between the sensors. The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time The actual value of is the time lag, indicating that at time forward The state of the moment, The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time According to the deviation value predicted by the final combined model, is the environmental impact factor, is the environmental value at time t, is the standard environmental value under the normal operating environment of the equipment; m is the total number of sensors that obtain the operating monitoring data of thermal control equipment in thermal power plants; Situation determination unit: determining the real-time deviation situation of all thermal control equipment of the thermal power plant based on the real-time operation deviation amount of all thermal control equipment of the thermal power plant; Threshold determination unit: determines the abnormality judgment threshold based on the real-time operation deviation of all thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant:

[0048] in, is the abnormality judgment threshold, g is the number of operating parameters of thermal control equipment in thermal power plants, is the mean value of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is the standard deviation of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is a preset adjustable coefficient that determines the sensitivity of the threshold, and , is the sth real-time operating parameter of thermal control equipment in thermal power plants, is the coefficient used to adjust the effect of deviation on the threshold, It is the real-time operation deviation of thermal control equipment in thermal power plants. is the time change rate of the sth operating parameter of thermal control equipment in thermal power plants, The preset weights corresponding to the real-time operation status of thermal control equipment in thermal power plants. The preset weights corresponding to the real-time deviations of thermal control equipment in thermal power plants; Data determination unit: determines the abnormal situation of the equipment based on the real-time deviation situation of the equipment and the abnormality judgment threshold, and then determines the real-time deviation data of the equipment and the corresponding real-time operation parameters as the equipment abnormality data.

[0049] In this embodiment, the preset model set technology uses multiple trained model sets to predict the operating status of the equipment, and selects the optimal model for early warning by comparing the prediction results of different models. For example, in a thermal power plant, the preset model set technology may include multiple algorithm models, such as linear regression models, neural network models, etc., to predict the future status of the equipment; In this embodiment, the deviation value of the real-time operation prediction is the difference between the predicted value and the actual operation data of the equipment. The larger the deviation value, the greater the possibility that the equipment is abnormal. For example, if the predicted boiler temperature is 500°C, but the actual temperature is 520°C, then the deviation value is 20°C.

[0050] In this embodiment, the real-time operation deviation is the overall deviation between the real-time operation status of all equipment and the standard operation status. It is an important parameter for judging the health status of equipment. For example, by combining data from multiple sensors, the operation deviation of equipment such as boilers and turbines is obtained, and it is found that it is generally higher or lower than the normal value.

[0051] In this embodiment, the historical data impact factor is used to measure the impact of historical data on the current deviation calculation. Historical data may provide important trends and patterns. If the device temperature has been rising over the past few months, this trend will affect the current prediction, and the historical data impact factor will be larger; In this embodiment, the environmental impact factor reflects the degree of influence of external environmental factors on the operating state of the equipment. External environment such as temperature and humidity may affect the performance of the equipment. If the ambient temperature rises abnormally, this may affect the rise of the boiler temperature and the environmental impact factor will increase.

[0052] In this embodiment, the real-time deviation is the operating state of the current device obtained based on the real-time operation deviation, indicating the current health status of the device. For example, if the real-time deviation of a device exceeds a preset value, it means that the device is operating abnormally, and the real-time deviation is "abnormal"; In this embodiment, the abnormality judgment threshold is a critical value used to judge whether the device enters an abnormal state. When the real-time operation deviation exceeds this value, the device is considered abnormal. For example, the temperature threshold of a device is set to 550°C. When the temperature exceeds this value, the system will trigger an abnormal alarm.

[0053] In this embodiment, the preset adjustable coefficient is a coefficient used to adjust the abnormality judgment threshold, and can be flexibly adjusted according to different devices or scenarios to improve the sensitivity of early warning. For example, for key equipment, such as the main steam pipeline, the preset adjustable coefficient can be increased to detect abnormalities earlier; In this embodiment, the time rate of change is the speed at which the equipment operating parameters change over time. This indicator can help identify sudden changes or gradual deterioration in the equipment operating status. If the boiler temperature rises sharply in a short period of time, the time rate of change will be high, which may indicate that the equipment is malfunctioning.

[0054] In this embodiment, the abnormal condition of the equipment is determined based on the real-time deviation of the equipment and the abnormal judgment threshold. The real-time deviation of the equipment is compared with the preset abnormal judgment threshold to determine whether the equipment is in an abnormal state and record the abnormal data. If the sensor data shows that the real-time deviation of the boiler pressure exceeds the abnormal judgment threshold, the system will identify that the boiler is in an abnormal state and record the relevant data.

[0055] The beneficial effects of the above technical solution are: Example 5 An early warning system for thermal control equipment in a thermal power plant provided by an embodiment of the present invention includes a scoring determination module, comprising: Establish health scoring rules based on multiple factors, combine historical abnormal data and real-time abnormal data to score the health status of thermal control equipment in thermal power plants, and perform weighted fusion of the scoring results of each factor based on the fuzzy logic scoring model; Obtain the service life, maintenance history, and load changes of thermal control equipment in thermal power plants and adjust the scoring results based on the preset adjustment algorithm; The adjusted scoring results are compared and analyzed with the historical scoring trends, and the real-time health score of the thermal control equipment in the thermal power plant is determined based on the analysis results.

[0056] In this embodiment, the health scoring rules of multiple factors are rules formulated based on multiple factors that affect the health of the equipment, and each factor represents a part of the equipment operation, such as temperature, pressure, load, etc. The scoring rules evaluate the health status of the equipment by monitoring and analyzing these factors. For example, the health scoring rules of thermal power plant equipment may include factors such as temperature monitored by sensors, pressure changes, and equipment operation time. Each factor is assigned different weights for scoring according to its importance.

[0057] In this embodiment, the health status of thermal control equipment in a thermal power plant is scored by combining historical abnormal data and real-time abnormal data. The health status of the equipment is evaluated by scoring the historical faults and current monitoring data of the equipment. Historical data helps to understand long-term trends, while real-time data reflects the current status. If a steam turbine has experienced multiple abnormal shutdowns in the past year, and the recent monitoring data also shows that it deviates from the normal operating status, these data will jointly affect the health score of the equipment.

[0058] In this embodiment, the fuzzy logic scoring model is a scoring method based on fuzzy logic, which is used to process uncertain and ambiguous data. The scoring results of different factors will be processed according to the fuzzy logic rules, making the scoring more flexible and adaptable. For example, the equipment in a thermal power plant may have multiple factors affecting the health status (such as load fluctuations, ambient temperature, etc.). The fuzzy logic model can handle the uncertainty between these factors and give a comprehensive health score.

[0059] In this embodiment, the preset adjustment algorithm is an algorithm for adjusting the health score result. It is usually corrected based on information such as the equipment's age, maintenance history, load changes, etc. to reflect the equipment's actual health status. If a device has been in use for 10 years and its maintenance history shows that it is not maintained in a timely manner, the preset adjustment algorithm will lower its score to reflect the device's greater risk of failure.

[0060] In this embodiment, the comparative analysis is to compare the adjusted score result with the historical score trend, analyze the change trend of the score, and thus predict the future state or potential failure risk of the device. If the health score of a device has been declining in the past few months, the comparative analysis can help predict the possibility of its impending failure and take preventive measures; In this embodiment, the real-time health score is an instant health assessment value obtained by integrating the current state of the device, historical data and the adjusted score result, reflecting the current health status of the device.

[0061] The beneficial effects of the above technical solution are: by introducing health scoring rules and fuzzy logic scoring models based on multiple factors, and combining historical and real-time data, dynamic scoring of equipment health status is achieved, and the score is adjusted based on service life, maintenance history, etc. to make it more accurate, and the predictability and flexibility of real-time health scores are improved through trend comparison, which significantly improves the accuracy and reliability of equipment fault warnings and optimizes equipment maintenance and management.

[0062] Example 6 An embodiment of the present invention provides a thermal power plant thermal control equipment early warning system, a strategy determination module, including: Chain formation unit: Analyze historical abnormal data based on the preset reasoning model to determine the failure cause chain of thermal control equipment in thermal power plants; Graph construction unit: constructs an abnormal graph of thermal control equipment in a thermal power plant based on the fault cause chain of thermal control equipment in a thermal power plant; Spectrum analysis unit: Analyze the abnormal spectrum of thermal control equipment in thermal power plants to determine the comprehensive risk value of thermal control equipment failures in thermal power plants:

[0063] in, is the comprehensive risk value of thermal control equipment in thermal power plants, n is the number of failure types of thermal control equipment in thermal power plants, is the weight of the kth fault type, is the basic risk value of the kth fault type, is the time decay factor, is the correlation coefficient between the kth fault type and other fault types except the kth fault type, Divide the basic risk value of the hth fault type by the kth fault type, where is the basic risk value of the kth fault type:

[0064] in, is the occurrence frequency of the kth fault type, is the importance of the kth fault type, is the detection difficulty of the kth fault type, Score the health status of thermal control equipment in thermal power plants; Strategy determination unit: Determine the initial early warning strategy based on the comprehensive risk value of failures of thermal control equipment in thermal power plants, and continuously collect real-time early warning effect data after the execution of the early warning strategy, and optimize the early warning strategy in combination with the preset optimization algorithm, and then determine the comprehensive early warning strategy.

[0065] In this embodiment, the fault cause chain refers to identifying and arranging the various cause nodes that cause equipment failure based on historical abnormal data and fault modes to form a cause-and-effect chain. For example, in a thermal power plant, if a sensor fails and causes temperature abnormality, and the temperature abnormality causes the turbine to overheat and shut down, then this process can form a fault cause chain, namely "sensor failure → temperature abnormality → turbine overheating".

[0066] In this embodiment, the equipment abnormality map is a graphical representation based on the chain of equipment failure causes, which is used to display various possible abnormal states and their associations. The map helps understand the propagation path of the fault and the relationship between various factors. For example, the equipment abnormality map of the thermal control system of a thermal power plant may display abnormal signals of multiple sensors and the relationship between these signals and system overheating, load abnormality, etc., providing a global view to help analyze the problem.

[0067] In this embodiment, the detection difficulty reflects the complexity and technical requirements of identifying a certain type of fault, which is usually determined by factors such as the characteristic signal of the fault, the accuracy of the detection method, the influence of the equipment and the environment, and is quantified into numerical values ​​in the following ways: Signal significance analysis: By statistically analyzing the characteristic signal strength of the fault type in the historical monitoring data (such as fluctuation amplitude, signal-to-noise ratio, etc.), the weaker the signal, the higher the detection difficulty; Detection technical requirements: Based on the current technical means used to detect the fault type, evaluate the accuracy of the required sensors, the complexity of the algorithm, etc. The higher the complexity, the greater the difficulty; Environmental interference factors: Combined with the complexity of the operating environment of the thermal power plant (such as high temperature, high humidity, etc.), the more factors that interfere with the detection process, the higher the difficulty.

[0068] In this embodiment, the comprehensive risk value of equipment failure is a quantitative value calculated based on multiple factors such as failure type, frequency, severity, and detection difficulty, and is used to assess the failure risk faced by the equipment. For example, a high comprehensive risk value of a certain equipment may be because the equipment failure has a high frequency, high severity, and difficult to detect. For example, the comprehensive risk value of a boiler may be higher than that of a low-risk auxiliary equipment.

[0069] In this embodiment, the time decay factor is used to describe the process in which the impact of historical faults or anomalies on the current risk assessment gradually weakens over time. The time decay factor can reduce the impact of old data on the current judgment, indicating the change of the risk of the fault type over time. The older the fault, the larger the time decay factor, which reduces its impact. For example, the fault event of a certain equipment in a thermal power plant in the past year has a small impact in the recent risk assessment. Therefore, the time decay factor is set to gradually reduce the impact of the earlier fault event.

[0070] In this embodiment, the real-time warning effect refers to the actual effect of the warning system monitored in real time after the early warning strategy is implemented, including whether it can accurately predict failures or reduce the accident rate. If, after implementing a certain early warning strategy, the system successfully warns and avoids the overheating shutdown of the turbine, the real-time warning effect of the strategy is successful, and the feedback can be used to optimize future early warning strategies.

[0071] The beneficial effects of the above technical solution are: by constructing a fault cause chain and anomaly map, comprehensively considering factors such as fault type, correlation and time decay, and quantitatively evaluating the comprehensive risk value of the equipment. Innovatively introducing real-time early warning effect feedback and optimization algorithms, the early warning strategy can be dynamically adjusted, improving the accuracy and adaptability of the early warning system, thereby effectively reducing the risk of failure and maintenance costs, and significantly improving the safety and reliability of thermal control equipment in thermal power plants.

[0072] Example 7 An embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, and an early warning signal includes: Parameter determination unit: Analyze the historical abnormal data of thermal control equipment in thermal power plants, determine several abnormal states and corresponding associated equipment parameters; Transition determination unit: determines the transition between all abnormal states based on the abnormal state and the corresponding associated device parameters, and then constructs an abnormal state transition diagram; Path determination unit: Analyze the abnormal state transition diagram to determine the critical path on each transfer path of each abnormal state; Cause determination unit: determines the fault based on the abnormal state, and determines the cause of the fault based on the critical path on each transfer path of each abnormal state; History analysis unit: determines the impact of each fault in several preset time periods based on historical abnormal data; Parameter series connection unit: connects the fault, fault cause and corresponding associated equipment parameters in series according to time and impact degree, thus forming a fault cause chain.

[0073] In this embodiment, the analysis of historical abnormal data refers to the analysis of abnormal data recorded during the operation of the equipment to identify potential failure modes and associated factors. For example, in a thermal power plant, the abnormal operation data of a certain equipment in the past five years was analyzed, and it was found that multiple parameters had obvious fluctuations before the equipment failure, such as abnormal increases in temperature and pressure; In this embodiment, an abnormal state refers to a state in which the equipment is different from normal operation, and the corresponding associated equipment parameters are equipment operation data related to these abnormal states. For example, the abnormal state may be that the boiler temperature is too high, and the associated equipment parameters include temperature sensor readings, steam pressure, fuel input, etc.

[0074] In this embodiment, the transition situation between abnormal states describes the process of how one abnormal state develops or transforms into another abnormal state, which is usually based on the change trend of equipment parameters. For example, an excessively high boiler temperature may cause an abnormal increase in steam pressure, which then causes the turbine to shut down due to overload. This is the transition situation between abnormal states. In this embodiment, the abnormal state transition diagram is a graphical representation of the abnormal states of the equipment and their mutual transfer paths, which helps to understand the evolution process of equipment failure. For example, states such as abnormal boiler temperature, abnormal steam pressure, and turbine shutdown can be represented as a coherent state transfer process through the abnormal state transition diagram.

[0075] In this embodiment, the critical path on each transfer path of each abnormal state refers to the critical path that has the greatest impact on the fault evolution during the abnormal state transfer process. For example, in the path where the boiler temperature is too high and the turbine shuts down, the critical path may be the failure of the temperature sensor to adjust the fuel input in time.

[0076] In this embodiment, the fault is determined based on the abnormal state, and the cause of the fault is determined based on the key path on each transfer path of each abnormal state. The specific fault of the equipment is judged by analyzing the abnormal state and its transfer path, and the cause of the fault is determined based on the key factors on the transfer path. For example, the boiler shutdown fault is confirmed as excessive temperature through the abnormal state chain, and the cause is determined to be the failure of the temperature sensor to feedback in time.

[0077] In this embodiment, determining the impact of each fault in several preset time periods based on historical abnormal data is to analyze historical data and determine the impact of the fault on the equipment or system in different time periods to evaluate the severity and frequency of the fault. For example, by analyzing the historical data of the past three months, it is found that boiler failures are more frequent in summer than in winter, and the impact range is larger.

[0078] In this embodiment, the fault cause chain is obtained based on the analysis of the historical data of the equipment, showing the causal chain of various cause nodes and their mutual relationships from the initial abnormal state to the final fault. Each node in the chain represents a fault or abnormal state, and each connection indicates that the fault leads to the occurrence of another fault. For example, the cause chain of a boiler failure may include "temperature sensor failure → abnormal temperature rise → steam pressure increase → turbine overload shutdown".

[0079] The beneficial effects of the above technical solution are: by analyzing historical abnormal data, constructing abnormal state transition diagrams and determining critical paths, the faults and their causes can be accurately located. The faults are dynamically analyzed in combination with time and impact, realizing full-chain tracking of equipment failures. The system breaks the limitations of traditional fault warnings, enhances the early detection and diagnosis capabilities of complex and variable faults, and effectively improves the operating stability and maintenance efficiency of thermal control equipment in thermal power plants.

[0080] Example 8 An embodiment of the present invention provides an early warning system for thermal control equipment in a thermal power plant, wherein the early warning signal includes: early warning level, equipment name and location, fault type, real-time health score of the equipment, fault cause, early warning recommended measures, historical abnormal records corresponding to the fault cause, and predictive maintenance prompts corresponding to the fault cause.

[0081] In this embodiment, the historical abnormal record corresponding to the fault cause is a historical record of the fault cause retained in the system, which helps analyze its frequency and pattern. For example, in the past six months, the boiler temperature sensor has had three failure records; In this embodiment, the predictive maintenance prompt corresponding to the fault cause is that the system provides maintenance suggestions on preventing such faults by analyzing historical data and the fault cause. For example, it is recommended to regularly check and calibrate the temperature sensor to prevent future faults.

[0082] In this embodiment, the early warning signal, assuming that the early warning signal of a thermal power plant is as follows: Warning level: Intermediate warning Equipment name: Unit 4 cooling system, Equipment location: 3rd floor, Area A, Generator room, Abnormal type: Coolant temperature is too high, Health score: 72 / 100, Fault cause analysis: Coolant circulation efficiency decreases, causing the temperature to exceed the safety threshold, Recommended measures: Reduce equipment load, check coolant level and circulation system, Historical abnormality record: 3 similar abnormalities have occurred in the past week, Warning time: 2024-05-13 11:30, Abnormal duration: 15 minutes.

[0083] The beneficial effects of the above technical solution are: by providing detailed early warning signal information, such as warning level, equipment health score, fault cause and predictive maintenance prompts, it can achieve accurate fault positioning, early warning and proactive maintenance suggestions, integrate historical data with real-time analysis, make equipment failure prevention and maintenance more intelligent, improve the safety and operation and maintenance efficiency of thermal power plant equipment, and reduce unplanned downtime.

[0084] Finally, 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 aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these 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.

Claims

1. A thermal power plant thermal control equipment early warning system, characterized in that: include: Model building module: Deploy sensors in thermal control equipment and inside thermal power plants to obtain internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants. Based on the internal environmental data of thermal power plants and operation monitoring data of thermal control equipment of thermal power plants, establish a normal behavior model of thermal control equipment of thermal power plants; Deviation determination module: Determine the real-time operation deviation of thermal control equipment in a thermal power plant based on the normal behavior model of the thermal control equipment in a thermal power plant and the real-time operation monitoring data of the thermal control equipment in a thermal power plant, and then analyze the real-time operation deviation of the thermal control equipment in a thermal power plant to determine the real-time abnormal data of the thermal control equipment in a thermal power plant; Scoring determination module: determines the real-time health score of thermal control equipment in a thermal power plant based on the historical abnormal data of thermal control equipment in a thermal power plant and the real-time abnormal data of thermal control equipment in a thermal power plant; Strategy determination module: Analyze the real-time health scores of thermal control equipment in thermal power plants and the historical abnormal data of thermal control equipment in thermal power plants to determine the real-time failure causes of thermal control equipment in thermal power plants, and then determine a comprehensive early warning strategy based on the real-time failure causes of thermal control equipment in thermal power plants; Early warning execution module: Determines the early warning level based on the real-time health score of thermal control equipment in thermal power plants, and issues early warning signals based on the early warning level and comprehensive early warning strategy.

2. The thermal control equipment early warning system for a thermal power plant according to claim 1, characterized in that: Deviation determination module, including: Data fusion unit: fuses the internal environment data of the thermal power plant with the operation monitoring data of the thermal control equipment of the thermal power plant based on the preset fusion algorithm; Model building unit: Based on the fused data, normal behavior models of several thermal control equipment in thermal power plants are established; Model combination unit: based on the analysis of the operation monitoring data of the thermal control equipment of the thermal power plant, the non-fault state data of the thermal control equipment of the thermal power plant is obtained, the non-fault state data of the thermal control equipment of the thermal power plant is input into the normal behavior model of the thermal control equipment of each thermal power plant, and the output result of the normal behavior model of the thermal control equipment of each thermal power plant is obtained. Based on the output result of the normal behavior model of the thermal control equipment of each thermal power plant and the preset judgment algorithm, the combination weight of the normal behavior model of the thermal control equipment of each thermal power plant is determined, and then based on the combination weight and the preset combination rule, several normal behavior models of the thermal control equipment of the thermal power plant are combined to obtain a combination model.

3. The thermal control equipment early warning system for a thermal power plant according to claim 2, characterized in that: Model combination unit, including: Model screening subunit: constructing a candidate model set based on normal behavior models of thermal control equipment of several thermal power plants, determining the score of each normal behavior model of thermal control equipment of thermal power plants in the candidate model set based on a preset scoring algorithm, and determining the normal behavior models of thermal control equipment of several thermal power plants whose scores are greater than a preset score threshold as the first model; A weight determination subunit: obtaining the output of each first model, and then determining the weight of each first model in combination with a preset algorithm; Model combination subunit: based on the weights of all the first models, all the first models are combined according to preset fusion rules to generate a basic combination model; Model generation subunit: takes the output of the basic combination model as input and generates the final combination model based on the preset learner and the internal environmental data of the thermal power plant.

4. The thermal control equipment early warning system for a thermal power plant according to claim 3, characterized in that: Model building modules, including: Parameter determination unit: determines the deviation value of the real-time operation prediction of thermal control equipment in thermal power plants based on the preset model set technology and the final combined model; Data acquisition unit: Determine the real-time operation deviation of all thermal control equipment in the thermal power plant based on the deviation value of the real-time operation prediction of the thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant: in, It is the real-time operation deviation of all thermal control equipment in thermal power plants. It is the basic operation deviation of all thermal control equipment in thermal power plants. is the historical operating deviation of all thermal control equipment in thermal power plants, is the weight coefficient of the i-th sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the actual measurement value of the sensor at time t that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, is the deviation value predicted by the final combined model at time t by the sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant for the ith time, is the historical data impact factor, is the correlation coefficient of the sensor data of the jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant relative to the sensor data of the ith sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant, reflecting the correlation between the sensors. The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time The actual value of is the time lag, indicating that at time forward The state of the moment, The jth sensor that obtains the operation monitoring data of the thermal control equipment of the thermal power plant at time According to the deviation value predicted by the final combined model, is the environmental impact factor, is the environmental value at time t, is the standard environmental value under the normal operating environment of the equipment; m is the total number of sensors that obtain the operating monitoring data of thermal control equipment in thermal power plants; Situation determination unit: determining the real-time deviation situation of all thermal control equipment of the thermal power plant based on the real-time operation deviation amount of all thermal control equipment of the thermal power plant; Threshold determination unit: determines the abnormality judgment threshold based on the real-time operation deviation of all thermal control equipment in the thermal power plant and the operation monitoring data of the thermal control equipment in the thermal power plant: in, is the abnormality judgment threshold, g is the number of operating parameters of thermal control equipment in thermal power plants, is the mean value of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is the standard deviation of the sth operating parameter of the thermal control equipment in the thermal power plant under healthy conditions, is a preset adjustable coefficient that determines the sensitivity of the threshold, and , is the sth real-time operating parameter of thermal control equipment in thermal power plants, is the coefficient used to adjust the effect of deviation on the threshold, It is the real-time operation deviation of thermal control equipment in thermal power plants. is the time change rate of the sth operating parameter of thermal control equipment in thermal power plants, The preset weights corresponding to the real-time operation status of thermal control equipment in thermal power plants. Preset weights corresponding to the real-time deviations of thermal control equipment in thermal power plants; Data determination unit: determines the abnormal situation of the equipment based on the real-time deviation situation of the equipment and the abnormality judgment threshold, and then determines the real-time deviation data of the equipment and the corresponding real-time operation parameters as the equipment abnormality data.

5. The thermal control equipment early warning system for a thermal power plant according to claim 1, characterized in that: Rating determination module, including: Establish health scoring rules based on multiple factors, combine historical abnormal data and real-time abnormal data to score the health status of thermal control equipment in thermal power plants, and perform weighted fusion of the scoring results of each factor based on the fuzzy logic scoring model; Obtain the service life, maintenance history, and load changes of thermal control equipment in thermal power plants and adjust the scoring results based on the preset adjustment algorithm; The adjusted scoring results are compared and analyzed with the historical scoring trends, and the real-time health score of the thermal control equipment in the thermal power plant is determined based on the analysis results.

6. The thermal control equipment early warning system for a thermal power plant according to claim 1, characterized in that: The strategy determination module includes: Chain formation unit: Analyze historical abnormal data based on the preset reasoning model to determine the failure cause chain of thermal control equipment in thermal power plants; Graph construction unit: constructs an abnormal graph of thermal control equipment in a thermal power plant based on the fault cause chain of thermal control equipment in a thermal power plant; Spectrum analysis unit: Analyze the abnormal spectrum of thermal control equipment in thermal power plants to determine the comprehensive risk value of thermal control equipment failures in thermal power plants: in, is the comprehensive risk value of thermal control equipment in thermal power plants, n is the number of failure types of thermal control equipment in thermal power plants, is the weight of the kth fault type, is the basic risk value of the kth fault type, is the time decay factor, is the correlation coefficient between the kth fault type and other fault types except the kth fault type, Divide the basic risk value of the hth fault type by the kth fault type, where is the basic risk value of the kth fault type: in, is the occurrence frequency of the kth fault type, is the importance of the kth fault type, is the detection difficulty of the kth fault type, Score the health status of thermal control equipment in thermal power plants; Strategy determination unit: Determine the initial early warning strategy based on the comprehensive risk value of failures of thermal control equipment in thermal power plants, and continuously collect real-time early warning effect data after the execution of the early warning strategy, and optimize the early warning strategy in combination with the preset optimization algorithm, and then determine the comprehensive early warning strategy.

7. The thermal control equipment early warning system for a thermal power plant according to claim 6, characterized in that: Warning signs include: Parameter determination unit: Analyze the historical abnormal data of thermal control equipment in thermal power plants, determine several abnormal states and corresponding associated equipment parameters; Transition determination unit: determines the transition between all abnormal states based on the abnormal state and the corresponding associated device parameters, and then constructs an abnormal state transition diagram; Path determination unit: Analyze the abnormal state transition diagram to determine the critical path on each transfer path of each abnormal state; Cause determination unit: determines the fault based on the abnormal state, and determines the cause of the fault based on the critical path on each transfer path of each abnormal state; History analysis unit: determines the impact of each fault in several preset time periods based on historical abnormal data; Parameter series connection unit: connects the fault, fault cause and corresponding associated equipment parameters in series according to time and impact degree, thus forming a fault cause chain.

8. The thermal control equipment early warning system for a thermal power plant according to claim 1, characterized in that: Early warning signals include: warning level, equipment name and location, fault type, real-time health score of the equipment, fault cause, early warning recommended measures, historical abnormal records corresponding to the fault cause, and predictive maintenance tips corresponding to the fault cause.

Images