Degradation trend early warning method and system for power plant equipment and medium

Abstract

The invention provides a degradation trend early warning method and system for power plant equipment and a medium. The method comprises the following steps: acquiring data of a plurality of real-time operation parameters of equipment under different working conditions; performing degradation prediction based on the data of the plurality of real-time operation parameters to obtain degradation prediction data; according to the various operation condition data of the equipment, training a preset normal operation threshold value of the equipment to obtain an alarm threshold value under the corresponding operation condition; judging whether the equipment has a degradation trend or not according to the degradation prediction data and the alarm threshold value; and carrying out early warning prompt on the equipment with the degradation trend. Degradation trend early warning is carried out by combining the warning threshold values of the equipment under different working conditions and the degradation prediction data of the multiple pieces of real-time operation parameter data, the early warning sensitivity can be effectively improved, the risk of false alarm and missing alarm is reduced, and therefore safe and stable operation of the equipment is guaranteed.

Description

Technical Field

[0001] The present application relates to the technical field of power plant equipment detection, and in particular to a method, system and medium for early warning of degradation trend of power plant equipment. Background Art

[0002] During long-term operation, the loss of power plant equipment gradually accumulates, and the performance of the equipment gradually decreases or deteriorates. This will not cause failures in the short term, but will cause the equipment to gradually lose efficiency, increase energy consumption, produce wear and tear, shorten its lifespan, and affect the long-term stability of the equipment.

[0003] The existing equipment status degradation trend warning is a fault warning based on a single parameter and a single operating condition. It can only monitor the equipment operating status of a certain operating parameter under the same warning condition. The warning sensitivity is not enough, and the risk of false alarm and missed alarm is high. For example, when the fan is running at full load, the vibration may increase, while when it is running at low load, the vibration will decrease. When running at full load, the natural increase in vibration may be misjudged as a fault, resulting in frequent system alarms, which is actually the normal operating state of the fan. When running at low load, even if there is a slight vibration abnormality, the system may think that the vibration value is still within the normal range, ignoring potential hidden dangers. Therefore, when the system fails to distinguish between normal vibration caused by load changes and abnormal vibration caused by faults, based on the same vibration alarm value, the possibility of missed alarm and false alarm increases; for example, when the fan is in a high temperature environment, the lubricating oil of the bearing may become thinner, resulting in an increase in the bearing temperature, which is not a bearing fault. However, since the system does not consider external temperature factors, the alarm threshold is still based on the standard temperature range or the standard vibration range, which may cause the system to frequently trigger a high temperature alarm. In addition, a high temperature environment may cause a decrease in lubrication performance, thereby causing a slight increase in vibration. This change may also be misjudged as an equipment abnormality, resulting in an increase in the false alarm rate. Summary of the invention

[0004] In response to the above problems, the present application provides a method, system and medium for early warning of degradation trends of power plant equipment, which performs early warning of degradation trends based on early warning conditions under different operating conditions through multiple operating data of the equipment, thereby improving the early warning sensitivity and reducing the risk of false alarms and missed alarms.

[0005] To achieve the purpose of this application, this application provides the following technical solutions:

[0006] In a first aspect of an embodiment of the present application, a method for early warning of degradation trend of power plant equipment is provided, the method comprising:

[0007] Obtain data on multiple real-time operating parameters of the equipment under different working conditions;

[0008] Based on the data of the plurality of real-time operating parameters, performing degradation prediction to obtain corresponding degradation prediction data;

[0009] According to various operating condition data of the equipment, the preset normal operating threshold is trained to obtain the alarm threshold under the corresponding operating condition;

[0010] Comparing the degradation prediction data with the alarm threshold, when at least one degradation prediction data value among multiple degradation prediction data of the device is not less than the alarm threshold, using the device information corresponding to the degradation prediction data value as alarm information;

[0011] An alarm is issued based on the alarm information.

[0012] In a possible embodiment, performing degradation prediction based on the data of the multiple real-time operating parameters to obtain corresponding degradation prediction data includes:

[0013] Extracting data features of each real-time operating parameter of the device under different working conditions;

[0014] The data features are input into a pre-trained degradation prediction model to obtain degradation prediction data values ​​under corresponding working conditions of the equipment.

[0015] In a possible embodiment, the degradation prediction model is pre-trained in the following manner:

[0016] Obtaining a sample degradation prediction model with operating data above a preset normal operating threshold;

[0017] Collecting data of multiple historical operating parameters of the equipment under multiple operating conditions and extracting historical fault data from the collected data of the historical operating parameters;

[0018] Extracting the data of the historical operating parameters and the data features of the historical fault data, and obtaining the data features corresponding to the historical operating data and the historical fault data;

[0019] Based on the data features of the equipment under each operating condition, the sample degradation prediction model is trained to obtain a degradation prediction model.

[0020] In a possible embodiment, the training of the preset normal operating threshold value according to the various operating condition data of the device to obtain the alarm threshold value under the corresponding operating condition includes:

[0021] Get the preset normal operation threshold of the device;

[0022] Collect historical operation data of the equipment under multiple operating conditions and historical fault data extracted from the historical operation data;

[0023] The normal operation threshold is trained based on the historical operation data and the historical fault data to obtain the alarm threshold under the corresponding operating condition.

[0024] In a possible embodiment, it further includes: data preprocessing;

[0025] A normal variation range of the data of the equipment operating parameters under multiple operating states is set. When the data of the equipment operating parameters exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed data of the operating parameters under different working conditions.

[0026] In a second aspect of an embodiment of the present application, a degradation trend early warning system for power plant equipment is provided, comprising:

[0027] A data acquisition module is used to obtain data of multiple real-time operating parameters of the equipment under different working conditions, wherein the data of the real-time operating parameters include: temperature, vibration, and pressure;

[0028] A degradation prediction module, wherein the degradation prediction module is provided with a degradation prediction model, and the degradation prediction model is used to predict the degradation data of the data of different operating parameters of the equipment based on the data of each real-time operating parameter collected;

[0029] A comparison module, configured to determine, based on the degradation prediction data and the preset alarm threshold, whether at least one of the multiple degradation prediction data values ​​of the device is not less than its preset alarm threshold, and if so, send device information that the predicted degradation data is not less than the preset threshold to the degradation signal receiving module;

[0030] A degradation signal receiving module, used to receive device information that the predicted degradation data is not less than a preset alarm threshold, and send an alarm signal that the received predicted degradation data is not less than the preset threshold to an alarm module;

[0031] Alarm module: used to generate alarm based on received alarm information.

[0032] In a possible embodiment, the degradation prediction model includes:

[0033] A data collection module, used to collect data of multiple historical operating parameters of the equipment under multiple operating conditions and historical fault data extracted from the collected data of historical operating parameters; the operating conditions include load, ambient temperature, and equipment service life;

[0034] A training module is used to train a degradation prediction model based on the historical data operating parameter data and the historical fault data.

[0035] In a possible embodiment, it also includes a pre-processing module;

[0036] The preprocessing module is used to filter the data collected by the data acquisition module, wherein the filtering process includes setting a normal variation range of data of multiple operating parameters of the equipment. When the data of the operating parameters of the equipment exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed operating data.

[0037] In a possible embodiment, it further includes: a control system;

[0038] The control system controls the data acquisition module, the degradation prediction module, the comparison module, the degradation signal receiving module, the alarm module, and the communication module.

[0039] In a third aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, any method step described in the first aspect is implemented.

[0040] Beneficial effects of the embodiments of the present application:

[0041] The degradation trend warning method, system and medium of power plant equipment provided in the embodiments of the present application. First, degradation prediction can be performed based on multiple real-time operating data of the equipment under different operating conditions to obtain degradation prediction data, and then the preset normal operating threshold is trained according to the various operating condition data of the equipment to obtain the alarm threshold under the corresponding operating condition, and then based on the degradation prediction data and the alarm threshold, it is determined whether the equipment has a degradation trend, so as to give an early warning to the equipment with a degradation trend. By combining the alarm threshold of the equipment under different operating conditions with the degradation prediction data of multiple real-time operating parameter data to perform equipment degradation trend warning, it is possible to effectively improve the warning sensitivity and reduce the risk of false alarms and missed alarms, thereby ensuring the safe and stable operation of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0043] Figure 1 A schematic diagram of a process flow of a method for early warning of a degradation trend of power plant equipment provided in an embodiment of the present application; DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features; in the description of this application, unless otherwise specified, "plurality" means two or more.

[0046] Embodiment 1:

[0047] In order to more clearly illustrate the degradation trend warning method for power plant equipment provided in the embodiment of the present application, a possible application scenario of the degradation trend warning method provided in the embodiment of the present application will be exemplified below. It can be understood that the following example is only a possible application scenario of the degradation trend warning method provided in the embodiment of the present application. In other possible embodiments, the degradation trend warning method provided in the embodiment of the present application can also be applied to other possible application scenarios, and the following example does not impose any limitations on this.

[0048] When the operating environment changes greatly, the operating data fluctuates greatly, and multi-dimensional factors are not considered, the sensitivity of the equipment degradation trend warning based on the alarm status under a single operating state and the data of a single operating parameter is not enough, and the risk of false alarms and missed alarms is high. In order to reduce the risk of false alarms and missed alarms, equipment degradation trend warnings can be performed based on different thresholds under multiple operating conditions and degradation prediction data of multiple real-time operating data.

[0049] It is understandable that the degradation of equipment is not limited to the influence of a single operating data, but the changes in multiple operating data will have an impact on the degradation of equipment, and the operating data under different working conditions are different. For example, when the fan is running at full load, the vibration may increase, while when it is running at low load, the vibration will decrease. When running at full load, the natural increase in vibration may be misjudged as a fault, causing the system to frequently alarm, but in fact this is the normal operation of the fan. When running at low load, even if there is a slight vibration abnormality, the system may think that the vibration value is still within the normal range, ignoring the potential hidden dangers. Therefore, when the system fails to distinguish between normal vibration caused by load changes and abnormal vibration caused by faults, based on the same vibration alarm value, the possibility of missed alarms and false alarms increases; for example, when the fan is in a high temperature environment, the lubricating oil of the bearing may become thinner, causing the bearing temperature to rise, which is not a bearing fault, but because the system does not consider external temperature factors, the alarm threshold is still based on the standard temperature range or the standard vibration range, which may cause the system to frequently trigger a high temperature alarm. In addition, a high temperature environment may cause a decrease in lubrication performance, thereby causing a slight increase in vibration. This change may also be misjudged as an equipment abnormality, resulting in an increase in the false alarm rate.

[0050] Based on this, the embodiment of the present application provides a method for early warning of degradation trend of power plant equipment, which can be applied to any system and storage medium with early warning capability of degradation trend. The method can be found in Figure 1 , Figure 1 The figure shows a flow chart of a method for early warning of degradation trend of power plant equipment provided in an embodiment of the present application, including:

[0051] S1, obtain data of multiple real-time operating parameters of the equipment under different working conditions.

[0052] Specifically, the data of real-time operating parameters may be data of different parameter types according to different embodiments. For example, the equipment may be a fan, a water pump, a steam turbine, etc. The operating condition is the working environment and performance at a certain moment or within a certain period of time, which may be load, operating environment, service life, etc. The data is a parameter used to represent the performance of the equipment, which may be temperature, vibration, pressure, etc.

[0053] The PLC control system may be based on sensors to collect operating data corresponding to parameters such as temperature, vibration, pressure, etc. collected during the operation of fans, water pumps, steam turbines and other equipment. Since how to obtain data on real-time operating parameters of the equipment is not the main application point of this application, it will not be elaborated here.

[0054] S2, performing degradation prediction based on the data of the plurality of real-time operating parameters to obtain corresponding degradation prediction data.

[0055] Specifically, the steps include:

[0056] S21, extracting data features of each real-time operating parameter of the device under different working conditions.

[0057] Specifically, since the objects to be studied in many machine learning problems may be categorical variables, texts, or even images, we need to first perform feature extraction processing on the data of the real-time operating parameters to obtain the data features corresponding to the data of the real-time operating parameters.

[0058] S22, inputting the data features into a pre-trained degradation prediction model to obtain a degradation prediction data value under a corresponding working condition of the equipment.

[0059] The degradation prediction data value can be the key performance indicator of each real-time operating parameter of the extracted equipment under different working conditions (such as different loads, ambient temperatures, service life, etc.), which can be regarded as temperature fluctuations under different loads and ambient temperatures, and power consumption characteristics extracted based on load changes and equipment service life, such as power change rate and stability.

[0060] Specifically, the degradation prediction model may be pre-trained, including the following steps:

[0061] S221, obtaining a sample degradation prediction model whose operating data is higher than a preset normal operating threshold.

[0062] Specifically, the sample degradation prediction model in this solution is a self-encoding neural network model. The activation functions of the input and output layers of the self-encoding neural network model select the sigmoid activation function, and the activation function of the middle layer selects the tanh activation function.

[0063] It should be noted that the autoencoding neural network model is only a preferred implementation of the present solution. Those skilled in the art can select different neural network models for training according to actual needs. The difference in the neural network models does not affect the protection scope of the present application.

[0064] S222, collecting data of multiple historical operating parameters of the equipment under multiple operating conditions and historical fault data extracted from the collected data of the historical operating parameters.

[0065] Specifically, in the preferred embodiment of the present solution, the operating conditions include equipment performance parameters: load, ambient temperature, operating time, efficiency, etc. The data of historical operating parameters can be various operating data (such as temperature, pressure, vibration, power, speed, etc.) of the equipment collected in real time by various sensors (such as temperature sensors, vibration sensors, pressure sensors, current and voltage sensors, etc.) of the equipment and stored in the database, and historical fault data (such as fault type, occurrence time, fault impact, etc.).

[0066] It should be noted that the historical operation data includes but is not limited to the above data, and those skilled in the art can flexibly select different data parameters that can represent the operating status of the device according to the actual device or operating environment.

[0067] In order to accurately obtain historical operation data, in a preferred embodiment of the present solution, before obtaining the historical operation data of the device during normal operation, the historical operation data needs to be preprocessed, which will be discussed in detail below and will not be repeated here.

[0068] S223, extracting the data of the historical operating parameters and the data features of the historical fault data, and obtaining the data features corresponding to the historical operating data;

[0069] Specifically, since the objects to be studied in many machine learning problems may be categorical variables, text, or even images, we need to first perform feature extraction on the historical operation data to obtain the data features corresponding to the historical operation data. After obtaining the data features, the sample degradation prediction model is trained according to the data features to obtain the degradation prediction model of the equipment.

[0070] In order to obtain the data features corresponding to the historical operation data, in the preferred implementation of the present application, feature extraction processing is performed on the historical operation data, and the data features corresponding to the historical operation data are obtained, specifically:

[0071] Perform data cleaning on the historical operation data:

[0072] Normalize the cleaned historical operation data, and perform dimension reduction on the normalized historical operation data through principal component analysis;

[0073] The data features are obtained through the correlation coefficient method and the historical operation data after dimensionality reduction processing.

[0074] Specifically, the historical operation data needs to be preprocessed first, which will be discussed in detail below and will not be repeated here.

[0075] S223, training the sample degradation prediction model based on the data features corresponding to the historical operation data and the historical fault data of the equipment under each operating condition to obtain a degradation prediction model.

[0076] Specifically, when training the sample degradation prediction model, the historical operation data and historical fault data can be trained by machine learning methods to obtain the degradation prediction model, or the historical operation data and historical fault data can be trained by deep learning methods to obtain the degradation prediction model.

[0077] Specifically, according to the operating data of the equipment under multiple working conditions (such as load, temperature, vibration, etc.) and historical fault data (such as fault type, occurrence time, duration, etc.), the data is preprocessed to construct a training data set; based on deep learning models such as LSTM (long short-term memory network) or GRU (gated recurrent unit), the mean size of each parameter data of the equipment under different working conditions in the training data set when failure occurs and the parameter data of each parameter during normal operation are extracted, and the difference between the two is calculated. The difference is used as the prediction error, and the training goal is to minimize the sum of squares of the prediction errors; the degradation prediction model is trained until the sum of squares of the prediction errors converges and the training is stopped.

[0078] It is understandable that the vibration characteristics of fans under different loads vary significantly, with increased vibration at full load and reduced vibration at low load. If the prediction model is not trained based on these differences in operating conditions, it may lead to misinterpretation of vibration data, resulting in erroneous degradation prediction data. For example, when operating at full load, the model may not correctly distinguish between normal load-induced vibration and vibration caused by a fault.

[0079] It should be noted that the autoencoding neural network model is only a preferred implementation of the present solution. Those skilled in the art can select different neural network models for training according to actual needs. The difference in the neural network models does not affect the protection scope of the present application.

[0080] In S3, the dynamic adjustment of the alarm threshold specifically includes the following steps:

[0081] S31, obtaining a preset normal operation threshold of the device;

[0082] The preset normal operation thresholds of the equipment can be set by the equipment manufacturer or operation and maintenance personnel based on the equipment's design parameters, technical specifications, and historical operation experience. These thresholds are used to determine whether the equipment is in a normal operating state. The normal operation thresholds can also be dynamically adjusted based on the operation experience, failure mode, and other data accumulated during the use of the equipment.

[0083] S32, collecting historical operation data of the equipment under multiple operation conditions and historical fault data extracted from the historical operation data;

[0084] Specifically, historical operation data should include time series data of various operating parameters (such as temperature, vibration, pressure, current, etc.) of the equipment under different working conditions. The data should cover both normal operation and failure situations. Historical failure data: includes parameter changes when the equipment fails, which may include equipment type, failure type, various parameter values ​​when the failure occurs (such as abnormal vibration, temperature, etc.) and corresponding fault diagnosis labels (such as "bearing failure", "overheating", etc.).

[0085] S33, training the normal operation threshold based on the historical operation data and the historical fault data to obtain the alarm threshold under the corresponding operating condition.

[0086] Specifically, select features that are highly correlated with equipment failures from historical data, such as temperature, vibration, pressure, oil level, load, etc. For fault data, label them according to the fault type and indicate the parameter status when the fault occurs; for normal data, set the label to "normal".

[0087] Furthermore, for labeled data (i.e., marked with normal and fault conditions), supervised learning algorithms are used to train the model. The model will learn the alarm thresholds of various parameters under different working conditions (such as load, ambient temperature, equipment service life, etc.); for unlabeled data, unsupervised learning algorithms can be used for anomaly detection, and the model automatically finds data outside the parameter range and uses it as a potential alarm threshold; under different working conditions (such as load, temperature, humidity, etc.), the model can learn which parameter fluctuations under specific working conditions indicate equipment abnormalities by comparing the changing trends of normal and faulty states. For example, the standard deviation of a parameter is larger under high load conditions and relatively smaller under low loads, and the model will learn the alarm threshold under such conditions.

[0088] The relationship between the operating parameters and fault data of the equipment can be trained through regression models (such as linear regression, support vector regression (SVR), etc.) to obtain the alarm thresholds of the equipment under different working conditions. For example, the alarm thresholds under different working conditions can be trained based on parameters such as temperature and vibration. Machine learning algorithms (such as random forests, XGBoost, etc.) can also be used to train alarm thresholds based on historical operating data and fault data. The model will output alarm thresholds based on the input data of different working conditions (such as load, ambient temperature, etc.).

[0089] It can be understood that one way to obtain the alarm threshold is to use load changes as input variables, establish a relationship model between load and vibration, identify the impact of load changes on vibration, and adjust the alarm threshold based on a joint analysis of load and vibration.

[0090] It is further necessary to explain that when performing degradation prediction and dynamic adjustment of alarm thresholds for all data, the collected data needs to be preprocessed, specifically, including the following steps:

[0091] A normal variation range of the data of the equipment operating parameters is set. When the data of the equipment operating parameters exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed data of the operating parameters under different working conditions.

[0092] Specifically, we first clean the historical operation data. When we collect data manually, it is inevitable that there will be omissions or miscollections, and null values ​​(NAN) and spaces may often appear in the data. Before data analysis and processing, data cleaning is a necessary process and is also the most important part of the entire data analysis process.

[0093] The cleaned historical operation data is normalized. The specific role of normalization is to summarize the statistical distribution of the unified samples. Normalization means the same, unified and unified. Whether for modeling or calculation, the basic measurement unit must be the same. The neural network is trained and predicted based on the statistical probability of the sample in the event, and the value of the sigmoid function mentioned above is between 0 and 1, as is the output of the last node of the network, so the output of the sample is often normalized. In addition, there are often singular sample data in the data. The existence of singular sample data increases the network training time and may cause the network to fail to converge. In order to avoid this situation and facilitate the subsequent data processing and speed up the network learning speed, the input signal can be normalized so that the mean of the input signal of all samples is close to 0 or very small compared to its mean square error.

[0094] After completing the normalization processing, the normalized historical operation data is subjected to dimensionality reduction processing through principal component analysis, linearly uncorrelated historical operation data are selected, and then the correlation between the historical operation data and the training results is evaluated through the correlation coefficient method, so as to obtain the data features corresponding to the historical operation data.

[0095] Secondly, define the normal range of variation. By analyzing the historical data of the equipment under different working conditions, determine the normal range of variation of each operating parameter. This can be defined by statistical methods such as mean, standard deviation, maximum value, and minimum value. For example, for temperature parameters, you can set "the temperature is between the historical maximum and minimum values, and the fluctuation is within ±5%". Among them, under different working conditions (such as load, ambient temperature, equipment operating life, etc.), the data is grouped and the normal range of variation for specific working conditions is established. For example, the data can be classified according to load (low, medium, and high load) and the normal range under each load category is calculated.

[0096] Parameters that are outside the normal range are then flagged in the dataset and removed from the analysis or replaced with reasonable default values ​​(such as mean, median, etc.).

[0097] In S4, each degradation prediction data of the equipment under different working conditions is compared with its corresponding alarm threshold. If each predicted degradation prediction data is less than its corresponding preset probability threshold, return to S3; if at least one degradation prediction data among the degradation prediction data under different working conditions is greater than its corresponding alarm threshold, the information that the equipment may fail is sent to the alarm module through the communication module.

[0098] Different alarm signals are output according to different degradation prediction data. For example, a mild warning: the equipment is still within the normal working range, but close to the threshold, reminding personnel to pay attention; a moderate warning: the equipment parameters obviously deviate from the normal range and need to be inspected; a serious alarm: the equipment is at risk of failure or seriously degraded, and the equipment must be shut down immediately for inspection.

[0099] In S5, an alarm is issued based on the received alarm information, which can be performed by triggering a sound alarm or a visual alarm through a light sign or LED display screen, such as a green light: the equipment is operating normally; a yellow light: the equipment is approaching a fault state or a warning state; a red light: the equipment has a serious fault or is about to fail and needs to be shut down for inspection immediately.

[0100] Furthermore, the alarm information includes device information, such as device number or name, device model or type, and device location; it also includes alarm level and alarm cause, such as minor (warning), moderate (caution), severe (fault), and causes such as excessive temperature, excessive vibration, etc.

[0101] In order to monitor the degradation trend of the equipment in real time, in a preferred embodiment of the present application, the method may further include: storing the real-time operation data as new historical operation data in a database, and updating the degradation warning model in real time. Specifically, the real-time operation data is stored in a database as new historical operation data to achieve real-time updating of the degradation warning model, thereby improving the accuracy and scope of application of the degradation warning model.

[0102] In actual applications, sensors installed on the fan bearings collect data on parameters such as bearing temperature, vibration amplitude, rotation speed, and ambient temperature in real time. The real-time characteristics of each operating parameter are extracted through the data acquisition module, and the real-time collected operating parameters are input into a pre-trained degradation prediction model to predict the degradation trend of the bearing under the current operating conditions and provide degradation prediction data. Based on the historical operating data and fault data of the bearings, dynamic alarm thresholds suitable for different operating conditions (such as low wind speed, high wind speed, and high load) have been established. If the temperature, vibration, or other degradation prediction data exceeds the threshold, the system generates an alarm message, and the operation and maintenance personnel perform corresponding processing (such as lubrication replenishment, deceleration operation, or shutdown for maintenance) based on the alarm message.

[0103] For example, for different operating conditions, the alarm threshold setting of the fan bearing can be dynamically adjusted according to the differences in operating conditions. For example, when the fan is running at full load, the vibration level rises to 12mm / s, but because the alarm threshold after dynamic adjustment is 15mm / s, the system does not trigger an early warning. If the predicted vibration exceeds 15mm / s, the system triggers an early warning to indicate that there may be an abnormality. When the fan is running at low load, the normal vibration level is 6mm / s. If insufficient lubrication causes the predicted vibration to rise to 8mm / s, the system will trigger an early warning in time to indicate that there may be a fault.

[0104] For example, under low load conditions (load <50%), the bearing temperature alarm threshold is: >55℃ (minor warning), >70℃ (serious warning), and the bearing vibration alarm threshold is: >0.3mm / s (minor warning), >0.5mm / s (serious warning); under normal load conditions (load 50%-80%), the bearing temperature alarm threshold is: >70℃ (minor warning), >85℃ (serious warning), and the bearing vibration alarm threshold is: >0.5mm / s (minor warning), >0.8mm / s (serious warning); under high load conditions (load >80%), the bearing temperature alarm threshold is: >80℃ (minor warning), >95℃ (serious warning), and the bearing vibration alarm threshold is: >0.8mm / s (minor warning), >1.2mm / s (serious warning).

[0105] Secondly, during the operation of the fan, the controller controls the temperature sensor and vibration sensor to collect the working status data of the fan bearing. For example, the bearing temperature of fan B is 67°C and the vibration is 0.75mm / s. After inputting the degradation prediction model, the bearing temperature of fan B is 70°C and the vibration is 0.8mm / s.

[0106] After comparison, the bearing temperature of fan B is not less than the alarm threshold of normal load conditions, and the vibration is not less than the alarm threshold of high load conditions. It can be seen that the fan is at risk of aging and requires maintenance.

[0107] By adopting a degradation trend early warning method for power plant equipment provided in this embodiment, when the equipment is working, it is only necessary to collect the working data of the equipment and input it into the degradation prediction model to obtain degradation prediction data. It is only necessary to compare the degradation data with multiple alarm thresholds corresponding to different working conditions. The system can automatically determine whether the equipment is in a degradation trend, adjust its working status or issue an alarm in time, and ensure the maximum utilization efficiency of the equipment.

[0108] Embodiment 2:

[0109] This embodiment provides a degradation trend early warning system for power plant equipment, including:

[0110] The data acquisition module is used to obtain data of multiple real-time operating parameters of the equipment under different working conditions, and the real-time operating parameter data includes: temperature, vibration, and pressure.

[0111] The degradation prediction module is provided with a degradation prediction model, and the degradation prediction model is used to predict the degradation data of different operating parameters of the equipment based on the collected data of each real-time operating parameter.

[0112] Among them, the degradation prediction model includes:

[0113] A data collection module, used to collect data of multiple historical operating parameters of the equipment under multiple operating conditions and historical fault data extracted from the collected data of historical operating parameters; the operating conditions include load, ambient temperature, and equipment service life;

[0114] A training module is used to train a degradation prediction model based on the historical data operating parameter data and the historical fault data.

[0115] The comparison module is used to determine whether at least one of the multiple degradation prediction data values ​​of the device is not less than its preset alarm threshold based on the degradation prediction data and the preset alarm threshold. If so, the device information that the predicted degradation data is not less than the preset threshold is sent to the degradation signal receiving module.

[0116] The degradation signal receiving module is used to receive device information that the predicted degradation data is not less than a preset alarm threshold, and send an alarm signal that the received predicted degradation data is not less than the preset threshold to the alarm module.

[0117] Alarm module: used to generate alarm based on received alarm information.

[0118] Exemplarily, a degradation prediction model based on deep learning or traditional machine learning has analyzed the historical data of the equipment under multiple operating conditions and given the following degradation prediction data: Equipment A: vibration prediction value: 0.15 mm / s / s, temperature prediction value: 75°C, lubricating oil prediction value: 30%; Equipment B: vibration prediction value: 0.25 mm / s, temperature prediction value: 82°C, lubricating oil prediction value: 50%; Assume that the following are the alarm thresholds of the equipment: vibration threshold: 0.20 mm / s, temperature threshold: 80°C, lubricating oil quantity threshold: 40% (lubricating oil needs to be replaced if it is lower than this value: lubricating oil cooling medium flow needs to be adjusted in time if it is higher than this value); Equipment A: lubricating oil prediction value 30%, which is lower than the alarm threshold of 40%, and the information of Equipment A is sent; Equipment B: vibration prediction value 0.25 mm / s and temperature prediction value 82°C, both exceed the alarm threshold, and the information of Equipment B is sent.

[0119] It should be noted that if at least one of the disadvantageous prediction data is greater than its preset alarm threshold, the control module will send an alarm message to the device.

[0120] In a possible embodiment, it also includes a pre-processing module;

[0121] The preprocessing module is used to filter the data collected by the data acquisition module, wherein the filtering process includes setting a normal variation range of data of multiple operating parameters of the equipment. When the data of the operating parameters of the equipment exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed operating data.

[0122] The control system controls the data acquisition module, the degradation prediction module, the comparison module, the degradation signal receiving module, the alarm module and the communication module.

[0123] Embodiment 3:

[0124] The embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned degradation trend early warning method of any power plant equipment are implemented. . For the specific steps of the method, reference can be made to the corresponding contents disclosed in the above-mentioned embodiment, and no further description will be given here.

[0125] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0126] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0127] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0128] The above embodiments are only used to illustrate the technical solution of the present application, but not to limit it. The present application is not limited to the exact structure described above and illustrated in the accompanying drawings, and it cannot be determined that the specific implementation of the present application is limited to these descriptions. For ordinary technicians in the technical field to which the present application belongs, various changes and modifications made without departing from the concept of the present application should be deemed to belong to the protection scope of the present application.

Claims

1. A method for early warning of degradation trend of power plant equipment, characterized in that: The method comprises: Obtain data on multiple real-time operating parameters of the equipment under different working conditions; Based on the data of the plurality of real-time operating parameters, performing degradation prediction to obtain corresponding degradation prediction data; According to various operating condition data of the equipment, the preset normal operating threshold is trained to obtain the alarm threshold under the corresponding operating condition; Comparing the degradation prediction data with the alarm threshold, when at least one degradation prediction data value among multiple degradation prediction data of the device is not less than the alarm threshold, using the device information corresponding to the degradation prediction data value as alarm information; An alarm is issued based on the alarm information.

2. The method for early warning of degradation trend of power plant equipment according to claim 1, characterized in that: Based on the data of the multiple real-time operating parameters, degradation prediction is performed to obtain corresponding degradation prediction data, including: Extracting data features of each real-time operating parameter of the device under different working conditions; The data features are input into a pre-trained degradation prediction model to obtain degradation prediction data values ​​under corresponding working conditions of the equipment.

3. The method for early warning of degradation trend of power plant equipment according to claim 2, characterized in that: The degradation prediction model is pre-trained in the following way: Obtaining a sample degradation prediction model with operating data above a preset normal operating threshold; Collecting data of multiple historical operating parameters of the equipment under multiple operating conditions and extracting historical fault data from the collected data of the historical operating parameters; Extracting data features of the historical operating parameters and the historical fault data, and obtaining data features corresponding to the historical operating parameters and the historical fault data; Based on the data features of the equipment under each operating condition, the sample prediction model is trained to obtain a degradation prediction model.

4. The method for early warning of degradation trend of power plant equipment according to claim 1, characterized in that: The method of training the preset normal operation threshold value of the device according to the various operating condition data of the device to obtain the alarm threshold value under the corresponding operating condition includes: Get the preset normal operation threshold of the device; Collect historical operation data of the equipment under multiple operating conditions and historical fault data extracted from the historical operation data; The normal operation threshold is trained based on the historical operation data and the historical fault data to obtain the alarm threshold under the corresponding operating condition.

5. The method for early warning of degradation trend of power plant equipment according to claim 1, characterized in that: Also includes: Data preprocessing; A normal variation range of the data of the equipment operating parameters is set. When the data of the equipment operating parameters exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed data of the operating parameters under different working conditions.

6. A degradation trend early warning system for power plant equipment, characterized in that: include: A data acquisition module is used to obtain data of multiple real-time operating parameters of the equipment under different working conditions, wherein the data of the real-time operating parameters include: temperature, vibration, and pressure; A degradation prediction module, wherein the degradation prediction module is provided with a degradation prediction model, and the degradation prediction model is used to predict the degradation data of the data of different operating parameters of the equipment based on the data of each real-time operating parameter collected; A comparison module, configured to determine, based on the degradation prediction data and the preset alarm threshold, whether at least one of the multiple degradation prediction data values ​​of the device is not less than its preset alarm threshold, and if so, send device information that the predicted degradation data is not less than the preset threshold to the degradation signal receiving module; A degradation signal receiving module, used to receive device information that the predicted degradation data is not less than a preset alarm threshold, and send an alarm signal that the received predicted degradation data is not less than the preset threshold to an alarm module; Alarm module: used to generate alarm based on received alarm information.

7. The degradation trend early warning system for power plant equipment according to claim 6, characterized in that: The degradation prediction model comprises: A data collection module, used to collect data of multiple historical operating parameters of the equipment under multiple operating conditions and historical fault data extracted from the collected data of historical operating parameters; the operating conditions include load, ambient temperature, and equipment service life; A training module is used to train a degradation prediction model based on the historical data operating parameter data and the historical fault data.

8. The degradation trend early warning system for power plant equipment according to claim 6, characterized in that: It also includes a pre-processing module; The preprocessing module is used to filter the data collected by the data acquisition module, wherein the filtering process includes setting a normal variation range of data of multiple operating parameters of the equipment. When the data of the operating parameters of the equipment exceeds the normal variation range, the data of the operating parameters are eliminated to obtain several groups of processed operating data.

9. The degradation trend early warning system for power plant equipment according to claim 6, characterized in that: Also includes: Control systems; The control system controls the data acquisition module, the degradation prediction module, the comparison module, the degradation signal receiving module, the alarm module, and the communication module.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps according to any one of claims 1 to 5 are implemented.

Images