Method and system for predicting failure of inlet guide vane of small steam turbine based on ANN model

Abstract

The application relates to a kind of based on ANN model small turbine admission valve fault prediction method and system, comprising: constructing prediction model, based on the historical operation data of unit, with the main steam pressure, main steam temperature, main steam humidity, back pressure, load of small turbine as input, with the small turbine admission valve flow, small turbine speed, small turbine admission valve opening as output, and being constructed by artificial neural network;Collect the current time operation data of unit, utilize prediction model to carry out online prediction to the next time small turbine admission valve flow, small turbine speed, small turbine admission valve opening, and output corresponding prediction value;Set determination criterion, according to determination criterion and based on the deviation of prediction value and corresponding real value, judge whether there is hidden danger of failure.The application realizes early stage to small turbine admission valve sticking failure automatic online prediction, improves the safety of unit operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power generation, and in particular to a method and system for predicting failure of a small turbine admission valve based on an ANN model. BACKGROUND

[0002] In a thermal power generating unit, the task of a small turbine is to drive a feed water pump and to meet the water supply requirement of a boiler. Under normal conditions, the small turbine is supplied with steam from the intermediate pressure cylinder of a main turbine. When the load of the main turbine decreases to the extent that the low pressure steam source cannot meet the requirement of the small turbine, a pressure reducing valve on the high pressure steam (i.e. high pressure cylinder exhaust) pipeline of the small turbine is opened, and the high pressure steam enters the turbine after throttling by the valve. At the same time, the check valve of the low pressure pipeline operates, and the small turbine automatically switches from the low pressure steam source to the high pressure steam source. If the admission valve of the small turbine is stuck during this process, normal operation will be affected.

[0003] The reasons for the sticking of the admission valve of the small turbine include mechanical problems, DEH (digital electro-hydraulic control system of the turbine) faults, and steam problems. The flow rate, speed, and opening degree of the valve of the small turbine caused by these reasons will be manifested in a specific form. However, there is no comprehensive and reasonably designed alarm device in the prior art that can determine whether the admission valve of the small turbine is stuck by detecting weak and special working conditions when the speed is abnormal. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a method and system for predicting failure of an admission valve of a small turbine based on an ANN (artificial neural network) model, with the purpose of automatically predicting the sticking of the admission valve of the small turbine at an early stage of failure.

[0005] The technical solution adopted by the present application is as follows:

[0006] The present application provides a method for predicting failure of an admission valve of a small turbine based on an ANN model, comprising:

[0007] Constructing a prediction model: the prediction model is based on historical operation data of the unit, and the main steam pressure, main steam temperature, main steam humidity, back pressure, and load of the small turbine at the previous moment are input, and the flow rate of the admission valve of the small turbine, the speed of the small turbine, and the opening degree of the admission valve of the small turbine at the next moment are output, and the prediction model is constructed by an artificial neural network;

[0008] Online prediction: collecting operation data of the unit at the current moment, and using the prediction model to perform online prediction on the flow rate of the admission valve of the small turbine, the speed of the small turbine, and the opening degree of the admission valve of the small turbine at the next moment, and outputting the prediction value;

[0009] Intake valve jam failure judgment: Set judgment criteria, and judge whether there is a potential failure risk based on the judgment criteria and the deviation between the predicted value and the corresponding true value at the current moment.

[0010] Further technical solutions are:

[0011] The judgment criteria are:

[0012] The deviation between the actual value and the predicted value of the flow rate of the steam inlet regulating valve of the small steam turbine is greater than the first set value;

[0013] The deviation between the actual value and the predicted value of the small steam turbine speed is greater than the second set value;

[0014] The deviation between the actual value and the predicted value of the opening of the small steam turbine inlet regulating valve is greater than the third set value;

[0015] If the above three conditions are met at the same time, it is determined that there is a potential failure risk;

[0016] The deviation is the absolute value of the difference between the true value and the predicted value.

[0017] The prediction method further includes:

[0018] After determining that there is a hidden danger of failure, an execution signal is sent to the unit control system to automatically adjust the flow of the small steam turbine steam inlet regulating valve. The predicted value of the current small steam turbine steam inlet regulating valve opening is obtained again through the prediction model, and the deviation between the predicted value and the actual value is calculated until the deviation is less than the third set value.

[0019] The third set value is: 5% and lasts for more than a certain period of time, or 10%.

[0020] The prediction method further includes:

[0021] After determining that there is a potential failure risk, an alarm signal is sent to the unit control system.

[0022] The present application also provides a small steam turbine inlet regulating valve fault prediction system based on an ANN model, comprising:

[0023] A prediction model building module is used to build a prediction model based on the historical operation data of the unit, with the main steam pressure, main steam temperature, main steam humidity, back pressure, and load of the small steam turbine at the previous moment as input, and the small steam turbine steam inlet regulating valve flow, small steam turbine speed, and small steam turbine steam inlet regulating valve opening at the next moment as output, and is built through an artificial neural network;

[0024] An online prediction module is used to collect the current operating data of the unit, use the prediction model to make online predictions on the flow rate of the small steam turbine inlet regulating valve, the speed of the small steam turbine, and the opening of the small steam turbine inlet regulating valve at the next moment, and output the predicted values;

[0025] a steam inlet regulating valve stuck failure judgment module, configured to set judgment criteria and judge whether there is a potential failure risk based on the judgment criteria and the deviation between the predicted value and the actual value corresponding to the current moment;

[0026] The determination criteria include:

[0027] The deviation between the actual value and the predicted value of the flow rate of the steam inlet regulating valve of the small steam turbine is greater than the first set value;

[0028] The deviation between the actual value and the predicted value of the small steam turbine speed is greater than the second set value;

[0029] The deviation between the actual value and the predicted value of the opening of the small steam turbine inlet regulating valve is greater than the third set value;

[0030] If the above three conditions are met at the same time, it is determined that there is a potential failure risk;

[0031] The deviation is the absolute value of the difference between the true value and the predicted value.

[0032] Further technical solutions are:

[0033] It also includes a protection execution module, which is used to send an execution signal to the unit control system after determining that there is a potential failure risk, automatically adjust the flow of the small steam turbine steam inlet regulating valve, and again obtain the predicted value of the current small steam turbine steam inlet regulating valve opening through the prediction model, and calculate the deviation between the predicted value and the actual value until the deviation is less than the third set value.

[0034] It also includes an alarm module, which is used to send an alarm signal to the unit control system after determining that there is a potential failure.

[0035] The present application also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is loaded and executed by a processor to implement the ANN model-based small steam turbine steam inlet regulating valve fault prediction method.

[0036] The present application also provides a computer device, which includes a processor and a memory, wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the small steam turbine inlet regulating valve fault prediction method based on the ANN model.

[0037] The beneficial effects of the present invention are as follows:

[0038] This application constructs a prediction model based on the historical operating data of the unit through training, learning and testing of an artificial neural network, which can realize online judgment of the early stage of the failure of the steam inlet regulating valve of a small steam turbine to be stuck. It can be used to issue an alarm in advance, facilitate timely execution of protection operations, avoid failure losses, and improve the safety of thermal power unit operation.

[0039] The application sets a judgment criterion based on logical judgment, realizes automatic detection of the sticking failure of the small turbine admission valve, facilitates subsequent execution protection response, and improves the automation degree of safe operation of the thermal power unit.

[0040] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent from the description, or can be learned by practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 A flowchart of a first embodiment of the prediction method of the embodiment of the present application is shown.

[0042] Figure 2 A logical judgment diagram of the judgment criterion of the prediction method of the embodiment of the present application is shown.

[0043] Figure 3 A flowchart of a second embodiment of the prediction method of the embodiment of the present application is shown. DETAILED DESCRIPTION

[0044] The specific embodiments of the present application are described below with reference to the accompanying drawings.

[0045] One embodiment of the present application provides a small turbine admission valve fault prediction method based on an ANN model, which can be used for early warning of the sticking failure of the small turbine admission valve of the thermal power unit, and guarantees the stability and safety of the unit operation. Artificial neural network (ANN) is a nonlinear, self-adaptive information processing system composed of a large number of processing units interconnected. It is proposed on the basis of modern neuroscience research results, and attempts to process information in the way of simulating brain neural network processing and memory information. It adopts a completely different mechanism from traditional artificial intelligence and information processing technology, overcomes the defects of traditional artificial intelligence based on logical symbols in processing intuition and unstructured information, and has the characteristics of self-adaptation, self-organization and real-time learning.

[0046] Referring to Figure 1 The prediction method of the embodiment includes the following steps:

[0047] S1, constructing a prediction model: the prediction model is based on historical operation data of the unit, and the main steam pressure, the main steam temperature, the main steam humidity, the back pressure and the load of the small turbine at the previous moment are input, the small turbine admission valve flow, the small turbine speed and the small turbine admission valve opening at the next moment are output, and the prediction model is constructed based on artificial neural network training, prediction and testing;

[0048] Specifically, the prediction model construction process includes:

[0049] S11, collect historical operation data of the unit under different working conditions, including operation data of the small turbine under various load conditions, take 10,000 groups of data as sample data, divide the 10,000 groups of data into two parts, 8,000 groups of data for training, and 2,000 groups of data for prediction and testing. Randomly shuffle the order of the data, randomly select 8,000 data as output and 2,000 data as input each time for prediction to ensure the accuracy of the prediction results.

[0050] S12, establish a neural network, the number of neurons in the input layer is 5, the corresponding input is the main steam pressure, the main steam temperature, the main steam humidity, the back pressure and the load of the small turbine at the current time in the prediction time domain, the number of neurons in the output layer is 3, the corresponding output is the flow of the small turbine admission valve, the speed of the small turbine and the opening of the small turbine admission valve at the next time in the prediction time domain, and the activation function is relu. The hidden layer is 3 layers, the number of neurons is 4, 5 and 5 respectively, and the activation function is s type.

[0051] S13, set the optimization model and the loss function. Select mean square error (MSE) as the loss function, and use gradient descent method to minimize the loss function.

[0052] S14, model training. Use randomly selected 8,000 groups of data for training.

[0053] S15, prediction and testing. Use randomly selected 2,000 groups of data for prediction, and compare the prediction results with the actual values generated by the model until the error meets the requirements, and finally obtain the constructed prediction model.

[0054] Those skilled in the art can understand that the preprocessing (normalization, dimensionality reduction, etc.) of the sample data set belongs to the conventional technical means and will not be repeated.

[0055] S2, online prediction: collect the current operation data of the unit, including the main steam pressure, the main steam temperature, the main steam humidity, the back pressure, the load, the flow of the small turbine admission valve, the speed of the small turbine and the opening of the small turbine admission valve at the current time, use the prediction model to predict the flow of the small turbine admission valve, the speed of the small turbine and the opening of the small turbine admission valve at the next time, and output the prediction value.

[0056] S3, admission valve sticking failure determination: set the determination criteria, and determine whether there is a failure hazard based on the deviation between the prediction value and the corresponding true value at the current time according to the determination criteria.

[0057] Referring to Figure 2 , the determination criteria are:

[0058] The deviation between the true value and the prediction value of the flow of the small turbine admission valve is greater than a first set value.

[0059] the deviation between the real value and the predicted value of the small turbine speed is greater than a second set value; and the deviation between the real value and the predicted value of the small turbine admission valve opening is greater than a third set value;

[0060] When the above three conditions are met, it is determined that there is a malfunction risk.

[0061] The deviation is the absolute value of the difference between the real value and the predicted value.

[0062] The first set value can be specifically 100 t / h.

[0063] The first set value can be specifically 50 r / min.

[0064] The third set value can be specifically 5% and last for more than a certain time (for example, 10 s), or 10%.

[0065] Each set value can be adaptively set according to the unit capacity, load and the like in combination with operation experience. Figure 2 The admission valve is the admission valve as described above.

[0066] Referring to Figure 3 In order to perform a protection operation after the malfunction is determined, as another embodiment, the prediction method of the embodiment of the application further comprises:

[0067] S4, after it is determined that there is a malfunction risk, a protection signal is sent to the unit control system to automatically perform load reduction or load increase, so as to adjust the small turbine admission valve flow, and then the predicted value of the small turbine admission valve opening is obtained again through the prediction model, the deviation between the predicted value and the real value is calculated, and the process is repeated until the deviation is less than the third set value, thereby improving the safety of the operation of the thermal power unit.

[0068] In order to send an alarm in time after the malfunction is determined, the prediction method of the embodiment of the application further comprises:

[0069] An alarm signal is sent to the unit control system, so as to manually adjust the operation parameters or perform a corresponding protection operation.

[0070] In summary, the small turbine admission valve fault prediction method based on the ANN model of the embodiment of the application can determine the valve fault on line by training and testing the prediction model based on the historical operation data of the unit through the artificial neural network, can realize early-stage small turbine admission valve jamming malfunction alarm, can avoid losses caused by small turbine admission valve jamming malfunction, and improves the safety of the operation of the thermal power unit.

[0071] The failure determination criterion proposed in the embodiments of the present application realizes automatic detection on the sticking failure of the small turbine admission valve, facilitates subsequent execution protection response, and improves the automation degree of safe operation of the thermal power unit.

[0072] One embodiment of the present application provides a small turbine admission valve failure prediction system based on an ANN model, comprising:

[0073] A prediction model construction module is configured to construct a prediction model, wherein the prediction model is based on historical operation data of the unit, takes the main steam pressure, the main steam temperature, the main steam humidity, the back pressure and the load of the small turbine at the previous moment as inputs, takes the flow, the rotating speed and the opening degree of the small turbine admission valve at the next moment as outputs, and is constructed based on an artificial neural network.

[0074] An online prediction module is configured to collect operation data of the unit at the current moment, perform online prediction on the flow of the small turbine admission valve, the rotating speed of the small turbine and the opening degree of the small turbine admission valve at the next moment by using the prediction model, and output prediction values.

[0075] An admission valve sticking failure determination module is configured to set a determination criterion, and determine whether there is a failure according to the deviation between the prediction values and the real values at the current moment based on the determination criterion.

[0076] The determination criterion is:

[0077] The deviation between the real value and the prediction value of the flow of the small turbine admission valve is greater than a first set value.

[0078] The deviation between the real value and the prediction value of the rotating speed of the small turbine is greater than a second set value.

[0079] The deviation between the real value and the prediction value of the opening degree of the small turbine admission valve is greater than a third set value.

[0080] If the above three conditions are met at the same time, it is determined that there is a failure hidden danger.

[0081] The deviation is an absolute value of the difference between the real value and the prediction value.

[0082] The small turbine admission valve failure prediction system based on the ANN model further comprises a protection execution module configured to send an execution signal to the unit control system to automatically adjust the flow of the small turbine admission valve after it is determined that there is a failure hidden danger, obtain the prediction value of the opening degree of the small turbine admission valve at the current moment again by using the prediction model, calculate the deviation between the prediction value and the real value, and continue until the deviation is less than the third set value.

[0083] The small turbine admission valve failure prediction system based on the ANN model further comprises an alarm module configured to send an alarm signal to the unit control system after it is determined that there is a failure hidden danger.

[0084] One embodiment of the present application provides a computer readable storage medium, wherein the storage medium stores at least one instruction, and the at least one instruction is loaded and executed by a processor to implement the ANN model-based fault prediction method of the inlet guide vane of the small steam turbine.

[0085] One embodiment of the present application provides a computer device, comprising a processor and a memory, wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the ANN model-based fault prediction method of the inlet guide vane of the small steam turbine.

[0086] Those skilled in the art can understand that the above description is only preferred embodiments of the present application, and is not used to limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, and those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for predicting the failure of a small steam turbine steam inlet regulating valve based on an ANN model, characterized in that: include: Constructing a prediction model: The prediction model is based on the historical operating data of the unit, with the main steam pressure, main steam temperature, main steam humidity, back pressure, and load of the small steam turbine at the previous moment as input, and the small steam turbine steam inlet regulating valve flow, small steam turbine speed, and small steam turbine steam inlet regulating valve opening at the next moment as output, and is constructed through an artificial neural network; Online prediction: Collect the current operating data of the unit, use the prediction model to make online predictions on the flow rate of the small steam turbine inlet regulating valve, the speed of the small steam turbine, and the opening of the small steam turbine inlet regulating valve at the next moment, and output the predicted values; Inlet valve jam failure determination: setting a determination criterion, and judging whether there is a potential failure risk based on the determination criterion and the deviation between the predicted value and the actual value corresponding to the current moment; The judgment criteria are: The deviation between the actual value and the predicted value of the flow rate of the steam inlet regulating valve of the small steam turbine is greater than the first set value; The deviation between the actual value and the predicted value of the small steam turbine speed is greater than the second set value; The deviation between the actual value and the predicted value of the opening of the small steam turbine inlet regulating valve is greater than the third set value; If the above three conditions are met at the same time, it is determined that there is a potential failure risk; The deviation is the absolute value of the difference between the true value and the predicted value.

2. The method for predicting a small steam turbine inlet regulating valve fault based on an ANN model according to claim 1 is characterized in that: Also includes: After determining that there is a hidden danger of failure, an execution signal is sent to the unit control system to automatically adjust the flow of the small steam turbine steam inlet regulating valve. The predicted value of the small steam turbine steam inlet regulating valve opening is obtained again through the prediction model, and the deviation between the predicted value and the actual value is calculated until the deviation is less than the third set value.

3. The method for predicting a small steam turbine inlet regulating valve fault based on an ANN model according to claim 1, characterized in that: The third set value is: 5% and lasts for more than a certain period of time, or 10%.

4. The method for predicting a small steam turbine inlet regulating valve fault based on an ANN model according to claim 1, characterized in that: Also includes: After determining that there is a potential failure risk, an alarm signal is sent to the unit control system.

5. A small steam turbine inlet regulating valve fault prediction system based on ANN model, characterized by: include: A prediction model building module is used to build a prediction model based on the historical operation data of the unit, with the main steam pressure, main steam temperature, main steam humidity, back pressure, and load of the small steam turbine at the previous moment as input, and the small steam turbine steam inlet regulating valve flow, small steam turbine speed, and small steam turbine steam inlet regulating valve opening at the next moment as output, and is built through an artificial neural network; An online prediction module is used to collect the current operating data of the unit, use the prediction model to make online predictions on the flow rate of the small steam turbine inlet regulating valve, the speed of the small steam turbine, and the opening of the small steam turbine inlet regulating valve at the next moment, and output the predicted values; a steam inlet regulating valve stuck failure judgment module, configured to set judgment criteria and judge whether there is a potential failure risk based on the judgment criteria and the deviation between the predicted value and the actual value corresponding to the current moment; The determination criteria include: The deviation between the actual value and the predicted value of the flow rate of the steam inlet regulating valve of the small steam turbine is greater than the first set value; The deviation between the actual value and the predicted value of the small steam turbine speed is greater than the second set value; The deviation between the actual value and the predicted value of the opening of the small steam turbine inlet regulating valve is greater than the third set value; If the above three conditions are met at the same time, it is determined that there is a potential failure risk; The deviation is the absolute value of the difference between the true value and the predicted value.

6. The small steam turbine inlet regulating valve fault prediction system based on ANN model according to claim 5 is characterized in that: It also includes a protection execution module, which is used to send an execution signal to the unit control system after determining that there is a potential failure risk, automatically adjust the flow of the small steam turbine steam inlet regulating valve, and again obtain the predicted value of the small steam turbine steam inlet regulating valve opening through the prediction model, and calculate the deviation between the predicted value and the actual value until the deviation is less than the third set value.

7. The small steam turbine inlet regulating valve fault prediction system based on ANN model according to claim 5 is characterized in that: It also includes an alarm module, which is used to send an alarm signal to the unit control system after determining that there is a potential failure.

8. A computer-readable storage medium, characterized in that The storage medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the small steam turbine inlet regulating valve fault prediction method based on the ANN model according to any one of claims 1 to 4.

9. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the small steam turbine inlet regulating valve fault prediction method based on the ANN model as described in any one of claims 1 to 4.

Images