A temperature early warning system and method for a rod control rod position system cabinet
By combining the temperature prediction model of the bar-controlled bar position system cabinet with the POT model, real-time monitoring and dynamic threshold generation of the bar-controlled bar position system cabinet temperature were realized, solving the problem of low accuracy in temperature anomaly judgment in the existing technology and improving the reliability and accuracy of early warning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CNNC FUJIAN FUQING NUCLEAR POWER
- Filing Date
- 2024-12-09
- Publication Date
- 2026-06-09
AI Technical Summary
In the existing bar control and position system cabinet temperature monitoring, manually setting fixed alarm thresholds leads to low accuracy in abnormal judgment and cannot effectively warn of abnormal temperatures.
By combining the temperature prediction model of the rod control and positioning system cabinet with the POT model, data is acquired through sensors for real-time prediction and dynamic threshold generation, and a deep learning network is used to determine temperature anomalies.
It improves the accuracy of temperature anomaly warnings, reduces the number of false alarms, ensures timely warnings before temperature anomalies occur, and protects critical components.
Smart Images
Figure CN119803725B_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of power plant maintenance and operation technology, specifically relating to a temperature early warning system and method for a rod control and rod position system cabinet. Background Technology
[0002] The Rod Control and Positioning System (RGL) comprises the Rod Control System (RCS) and the Rod Positioning System. As a core system of a nuclear power plant, the RGL is responsible for reactor reactivity control and measuring the position of control rods within the reactor core. The weakest link in the reliability of the RGL lies primarily in the MDP (Mechanical Distribution Module) boards within the cabinet. Based on board maintenance experience at Fuqing and other power plants, it has been clearly identified that current sensors and thyristors are key components restricting the normal operation of the MDP boards, with high-temperature aging being the primary failure factor. Currently, cabinet temperature monitoring relies on manual recording during periodic inspections, which cannot fully reflect the temperature status and changes within the RGL cabinet, nor can it provide early warnings when abnormal temperatures occur. Therefore, it is necessary to establish a real-time temperature monitoring and early warning platform for the RGL cabinet to improve the ability to warn of temperature anomalies and enhance system reliability.
[0003] Given the characteristics of server rack structure and internal temperature distribution, temperature monitoring using a combination of various thermal imaging devices and temperature sensors has been widely applied in the field of real-time server rack temperature monitoring. Document CN210242989U discloses a substation server rack temperature monitoring device, which uses multiple thermal imaging devices evenly installed on the inner wall support of the rack to detect the interior of the rack from multiple angles and with comprehensive coverage. An external alarm can be installed for timely alerts to prevent overheating and spontaneous combustion. Document CN109405982A discloses a server rack temperature monitoring device, method, and system. This system uses temperature sensors installed at intervals along the length of connecting components, driven by a device that moves them on a guide rail to collect real-time temperature data. This allows the server to more accurately map the temperature field, thus reflecting the temperature distribution on the air intake side of the rack.
[0004] However, in the actual operation of the rod control and positioning system cabinet, the probability of abnormal temperature data is very small. Over long periods of operation, the data sample accumulates to a large size, and the scarcity of abnormal data makes it difficult to locate anomalies. Against this backdrop, manually setting fixed alarm thresholds and using predicted temperatures for anomaly detection yields very low accuracy. Summary of the Invention
[0005] In view of this, this application aims to provide a temperature early warning system and method for bar-controlled bar position system cabinets. By combining the bar-controlled bar position system cabinet temperature prediction model, POT model algorithm and real-time cabinet temperature monitoring technology, it solves the problem that the accuracy of existing bar-controlled bar position system cabinet temperature early warning systems, which use manually set fixed alarm thresholds and predicted temperatures for anomaly judgment, is very low.
[0006] This application provides a temperature warning system for a rod-controlled rod positioning system cabinet. The system includes sensors, a temperature prediction module, a dynamic threshold module, and a decision module. The sensors are connected to the temperature prediction module via a first serial port and to the dynamic threshold module via a second serial port. The temperature prediction module and the dynamic threshold module are connected to the decision module via a third serial port. The sensors include both temperature and wind speed sensors. The temperature prediction module contains a temperature prediction model for the rod-controlled rod positioning system cabinet. It acquires temperature and wind speed data from the sensors, uses the model to predict the cabinet temperature in real time, and outputs the predicted temperature data to the decision module. The dynamic threshold module contains a POT model. It acquires temperature and wind speed data from the sensors, uses the POT model to generate a real-time threshold, and outputs an alarm threshold temperature value to the decision module. The decision module compares the predicted temperature value with the alarm threshold temperature value and issues an alarm.
[0007] In one specific embodiment of this application, the temperature prediction model for the bar-controlled bar position system cabinet is a feature fusion network model established based on the characteristics of the feature factors related to the temperature of the bar-controlled bar position system cabinet.
[0008] In one specific embodiment of this application, the temperature prediction model of the rod control and positioning system cabinet consists of three network layers: a recurrent network layer, a fully convolutional layer, and a feature fusion layer.
[0009] The second aspect of this application provides a temperature early warning method for a rod-controlled rod-positioning system cabinet. This method includes: using a temperature prediction module in a temperature early warning system for a rod-controlled rod-positioning system cabinet according to the first aspect of this application to acquire data signals from sensors, and inputting these data signals into a temperature prediction model for real-time cabinet temperature prediction, obtaining predicted temperature data, and then outputting the predicted temperature data to a decision module; using a dynamic threshold module in a temperature early warning system for a rod-controlled rod-positioning system cabinet according to the first aspect of this application to acquire data signals from sensors, generating a real-time threshold using a POT model, and outputting an alarm threshold temperature value to the decision module; and using a decision module in a temperature early warning system for a rod-controlled rod-positioning system cabinet according to the first aspect of this application to compare the real-time predicted temperature value with the alarm threshold temperature value, make a decision, and issue an alarm.
[0010] In one specific embodiment of this application, a temperature prediction module in a temperature early warning system for a rod-controlled rod-positioning system cabinet according to the first aspect of this application acquires data signals from sensors and inputs these data signals as input modules into a temperature prediction model for the rod-controlled rod-positioning system cabinet to perform real-time cabinet temperature prediction. After obtaining the predicted temperature data, the predicted temperature data is output to a decision module, including:
[0011] The temperature prediction module in the temperature early warning system of the rod control and rod position system cabinet of the first aspect of this application acquires the data signal of the sensor in real time, and marks the data signal before the current time as historical data and the data signal at the current time as real-time data;
[0012] Historical and real-time data are preprocessed to obtain preprocessed historical and real-time data.
[0013] The preprocessed historical data is fed into the temperature prediction model of the bar-controlled bar position system cabinet as input for model training, and the temperature prediction model of the bar-controlled bar position system cabinet cabinet is updated regularly to obtain the data fitting network model of the temperature of the bar-controlled bar position system cabinet.
[0014] The preprocessed real-time data is used as input to the data fitting network model of the bar control and positioning system cabinet temperature, and then the predicted temperature data is output.
[0015] In one specific embodiment of this application, a dynamic threshold module in a temperature early warning system for a rod control and positioning system cabinet according to the first aspect of this application acquires data signals from sensors, generates real-time thresholds using a POT model, and outputs alarm threshold temperature values to a decision module, including:
[0016] The dynamic threshold module in the temperature early warning system of the rod control and position system cabinet of the first aspect of this application acquires the data signal of the sensor in real time, and calibrates the data signal before the current moment as historical data and the data signal at the current moment as real-time data;
[0017] Historical and real-time data are preprocessed to obtain preprocessed historical and real-time data.
[0018] The preprocessed historical data is traversed to determine the maximum voltage value in the historical bar control system cabinet temperature data. The preprocessed real-time data is then used as the input to the POT model. The POT model is used to calculate the threshold temperature value based on the maximum temperature value, and the alarm threshold temperature value is output to the decision module.
[0019] In one specific embodiment of this application, the decision module in the temperature early warning system of the rod control and positioning system cabinet of the first aspect of this application compares the predicted temperature value with the alarm threshold temperature value in real time, makes a decision, and issues an alarm, including:
[0020] In the temperature warning system of the rod control and rod position system cabinet of the first aspect of this application, the decision module reads the predicted temperature data output by the temperature prediction module and the alarm threshold temperature value output by the dynamic threshold module, checks whether the predicted temperature data and the alarm threshold temperature value are valid, and if they are valid, proceeds to the next step; otherwise, the data is invalid and the task ends.
[0021] The system compares the predicted temperature value with a threshold value. Only when the predicted temperature value is greater than the threshold value is the temperature abnormality confirmed and an alarm triggered.
[0022] The beneficial effects of the technical solution of this application are as follows: On the basis of the existing real-time temperature monitoring device for the cabinet, a dynamic alarm threshold temperature value output by the POT model is added as a reference point, and the predicted temperature data generated by the cabinet temperature prediction model of the rod control and rod position system is used as the basis to determine whether the cabinet has an abnormal temperature. It can generate early warning information before the abnormal temperature causes substantial damage to the MDP card. At the same time, through the dynamic alarm threshold temperature value, the number of false alarms is effectively reduced and the reliability of the alarm is improved. Attached Figure Description
[0023] Figure 1 The diagram shown is a structural schematic of a temperature warning system for a rod control and rod position system cabinet according to an embodiment of this application.
[0024] Figure 2 The diagram shown is a schematic representation of the specific structure of a rack temperature prediction model for a rod control system according to an embodiment of this application.
[0025] Figure 3 The diagram shown is a schematic flowchart of a temperature early warning method for a rod control and positioning system cabinet according to an embodiment of this application. Detailed Implementation
[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0027] At least one embodiment of this application provides a temperature early warning system for a bar control and positioning system cabinet. (Reference) Figure 1The cabinet temperature early warning system includes a sensor 101, a temperature prediction module 102, a dynamic threshold module 103, and a decision module 104. The sensor 101 is connected to the temperature prediction module 102 via a first serial port and to the dynamic threshold module 103 via a second serial port. The temperature prediction module 102 and the dynamic threshold module 103 are connected to the decision module 104 via a third serial port. The sensor 101 includes both temperature and wind speed sensors. The temperature prediction module 102 contains a rod-controlled rack position system cabinet temperature prediction model. The temperature prediction module 102 acquires temperature and wind speed data from the sensor 101, uses the rod-controlled rack position system cabinet temperature prediction model to perform real-time cabinet temperature prediction, and outputs the predicted temperature data to the decision module 104. The dynamic threshold module 103 contains a POT model. The dynamic threshold module 103 acquires temperature and wind speed data from the sensor 101, uses the POT model to generate real-time threshold values, and outputs alarm threshold temperature values to the decision module 104. The decision module 104 is used to compare the predicted temperature value with the alarm threshold temperature value in real time, make a decision, and issue an alarm.
[0028] It should be noted that the rack of the rod control and positioning system can include any suitable combination of logic rack, power rack, processing rack, measurement rack and test rack.
[0029] According to the technical solution provided in the embodiments of this application, based on the existing real-time temperature monitoring device for server racks, a dynamic alarm threshold temperature value output by the POT model is added as a reference point, and the predicted temperature data generated by the rack temperature prediction model of the bar control and bar position system is used as the basis to determine whether the rack has an abnormal temperature. It can generate early warning information before the abnormal temperature causes substantial damage to the MDP card. At the same time, by using the dynamic alarm threshold temperature value, the number of false alarms is effectively reduced and the reliability of the alarm is improved.
[0030] In at least one embodiment of this application, the temperature prediction model for the bar-controlled bar position system cabinet is a feature fusion network model established based on the characteristics of the feature factors related to the temperature of the bar-controlled bar position system cabinet.
[0031] In at least one embodiment of this application, the temperature prediction model for the rod control and positioning system cabinet consists of three network layers: a recurrent network layer, a fully convolutional layer, and a feature fusion layer. This establishes a feature fusion network structure based on PLSTM-FCN.
[0032] For example, the specific structure of the temperature prediction model for the rack of the rod control and positioning system can be as follows: Figure 2 As shown, the three network layers have the following characteristics.
[0033] The recurrent network layer consists of LSTM (Long Short Term Memory Network) neurons and linear layers. It is the part of PLSTM-FCN responsible for learning the long-term dependencies of time series data. The linear layer is also known as the fully connected layer. In the linear layer, each neuron is connected to all neurons in the previous layer and performs linear transformations through linear combinations of neurons in the previous layer.
[0034] Fully Convolutional Networks (FCNs), as the functional module of the PLSTM-FCN network, extract features from sliced temporal data. They consist of convolutional layers, batch normalization, and activation functions. Dropout layers temporarily hide a portion of neurons in the connected network layers during training, stopping training until the next training iteration. Dropout layers can mitigate overfitting during model training by generalizing the model.
[0035] The concatenate layer is used to connect the data features extracted by the LSTM and FCN layers. By concatenating multiple features, it achieves data augmentation and reduces overfitting during the training of deep network models.
[0036] POT model: The maximum value is defined as x, and the threshold is y. The threshold y satisfies: y = c1 * ln(c2 * x + 1) + b; where b is a preset lower limit, and c1 and c2 are preset fitting logarithmic function parameters. The preset fitting logarithmic function parameters are calculated through N reference points on the fitting logarithmic function curve.
[0037] At least one embodiment of this application also provides a temperature early warning method for a bar-controlled bar positioning system cabinet. This temperature early warning method for the bar-controlled bar positioning system cabinet is implemented using the cabinet temperature early warning system provided in the above embodiments of this application. (Reference) Figure 3 The temperature warning method for the rod control and positioning system cabinet includes the following steps.
[0038] S1: The temperature prediction module 102 in the temperature early warning system of the rod control and position system cabinet provided in this application embodiment obtains data signals from the sensor 101, and inputs the data signals as input modules into the temperature prediction model of the rod control and position system cabinet to perform real-time cabinet temperature prediction. After obtaining the predicted temperature data, the predicted temperature data is output to the decision module 104.
[0039] S2: The dynamic threshold module 103 in the temperature warning system of the rod control and position system cabinet of this application obtains data signals from the sensor 101, generates real-time thresholds using the POT model, and outputs alarm threshold temperature values to the decision module 104.
[0040] Specifically, after acquiring the data signal from the sensor 101, the dynamic threshold module 103 performs data preprocessing and obtains the current alarm threshold temperature value based on historical data. The alarm threshold temperature value can also be an abnormal temperature threshold.
[0041] S3: The decision module 104 of the temperature early warning system of the rod control and position system cabinet in an embodiment of this application compares the predicted temperature value with the alarm threshold temperature value in real time, makes a decision, and issues an alarm.
[0042] Specifically, the decision module 104 receives the predicted temperature data output from the temperature prediction module 102 and the alarm threshold temperature value output from the dynamic threshold module 103, and then performs an alarm analysis.
[0043] The technical solution provided in this application, based on existing real-time cabinet temperature monitoring devices, combines real-time prediction data from a bar-controlled rack position system cabinet temperature prediction model with dynamic alarm threshold temperature values generated by a POT model to propose a cabinet temperature early warning method. By combining the bar-controlled rack position system cabinet temperature prediction model, the POT model algorithm, and real-time cabinet temperature monitoring technology, it is possible to simultaneously monitor the cabinet temperature in real time, use a deep learning network to predict the cabinet temperature in real time to achieve early warning conditions, and improve the accuracy of alarms by applying a dynamic threshold using the POT model.
[0044] In at least one embodiment of this application, S1.1 to S1.4 are specific implementations of the above-mentioned S1.
[0045] S1.1: The temperature prediction module 102 in the temperature early warning system of the rod control and position system cabinet of this application uses the data signal of the sensor 101 in real time, and marks the data signal before the current time as historical data and the data signal at the current time as real-time data.
[0046] Specifically, the temperature prediction module 102 acquires temperature and wind speed data through sensors installed in the rod control and position system cabinet, stores the data in the data module in real time, and marks the previous data as historical data.
[0047] S1.2: Preprocess historical and real-time data to obtain preprocessed historical and real-time data.
[0048] Specifically, the temperature prediction module 102 sends real-time and historical data to the data processing module for data preprocessing.
[0049] S1.3: The preprocessed historical data is fed into the temperature prediction model of the bar-controlled bar position system cabinet as input for model training, and the temperature prediction model of the bar-controlled bar position system cabinet cabinet is updated regularly to obtain the data fitting network model of the temperature of the bar-controlled bar position system cabinet.
[0050] Specifically, preprocessed historical data can be used as input to the temperature prediction model of the bar-controlled bar-position system cabinet to train the model and obtain a data fitting network model for the temperature of the bar-controlled bar-position system cabinet.
[0051] It should be noted that the temperature prediction model of the rod control and position system cabinet can be temporarily constructed or pre-constructed and directly called when used. This application embodiment does not make specific limitations on this.
[0052] S1.4: The preprocessed real-time data is used as the input to the data fitting network model of the rod control and positioning system cabinet temperature, and then the predicted temperature data is output.
[0053] In at least one embodiment of this application, S2.1 to S2.3 are specific implementations of the above-mentioned S2.
[0054] S2.1: The dynamic threshold module 103 in the temperature warning system of the rod control and position system cabinet of the present application uses the data signal of the sensor 101 in real time to acquire the data signal of the sensor 101, and the data signal before the current time is marked as historical data, and the data signal at the current time is marked as real-time data.
[0055] Specifically, the temperature prediction module 102 acquires temperature and wind speed data through sensors installed in the rod control and position system cabinet, stores the data in the data module in real time, and marks the previous data as historical data.
[0056] S2.2: Preprocess historical and real-time data to obtain preprocessed historical and real-time data.
[0057] S2.3: Traverse the preprocessed historical data to determine the maximum voltage value in the historical bar control bar position system cabinet temperature data, and use the preprocessed real-time data as the input of the POT model. Calculate the threshold temperature value based on the maximum temperature value using the POT model, and output the alarm threshold temperature value to the decision module 104.
[0058] In at least one embodiment of this application, S3.1 and S3.2 are specific implementations of the above-described S3.
[0059] S3.1: Using the temperature warning system of a rod control and position system cabinet according to an embodiment of this application, the decision module 104 reads the predicted temperature data output by the temperature prediction module 102 and the alarm threshold temperature value output by the dynamic threshold module 103, checks whether the predicted temperature data and the alarm threshold temperature value are valid. If they are valid, proceed to the next step; otherwise, the data is invalid, and the task ends.
[0060] Specifically, the decision module 104 can determine whether the acquired predicted temperature data and alarm threshold temperature value are complete and synchronized. If the above conditions are not met, the decision module 104 will terminate.
[0061] S3.2: The predicted temperature value is compared with the threshold value. Only when the predicted temperature value is greater than the threshold value is the temperature abnormality confirmed and an alarm is triggered.
[0062] For example, if the predicted temperature data T1 is greater than the alarm threshold data T2, a temperature anomaly alarm will be triggered and the module will terminate; if the predicted temperature data T1 is less than the alarm threshold data T2, the module will terminate directly.
[0063] It should be noted that the combination of the technical features in the embodiments of this application is not limited to the combination methods described in the embodiments of this application or the combination methods described in specific embodiments. All technical features described in this application can be freely combined or combined in any way, unless they contradict each other.
[0064] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications or equivalent substitutions made within the spirit and principles of this application should be included within the protection scope of this application.
Claims
1. A temperature early warning system for a rod control and positioning system cabinet, characterized in that, It includes a sensor, a temperature prediction module, a dynamic threshold module, and a decision module. The sensors are connected to the temperature prediction module via a first serial port and to the dynamic threshold module via a second serial port. The temperature prediction module and the dynamic threshold module are connected to the decision module via a third serial port. The sensors include both temperature and wind speed sensors. The temperature prediction module is equipped with a rod control and positioning system cabinet temperature prediction model. The temperature prediction module is used to acquire temperature and wind speed data from sensors, use the rod control and positioning system cabinet temperature prediction model to perform real-time cabinet temperature prediction, and output the predicted temperature data to the decision module. The dynamic threshold module contains a POT model. The dynamic threshold module is used to acquire temperature and wind speed data from the sensor, use the POT model to generate real-time thresholds, and output alarm threshold temperature values to the decision module. The decision module is used to compare the predicted temperature value with the alarm threshold temperature value, make a decision, and issue an alarm.
2. The temperature early warning system for a rod control and rod position system cabinet according to claim 1, characterized in that, The temperature prediction model for the bar-controlled rack system cabinet is a feature fusion network model established based on the characteristics of the feature factors related to the temperature of the bar-controlled rack system cabinet.
3. A temperature early warning system for a rod control and positioning system cabinet according to claim 1 or 2, characterized in that, The temperature prediction model for the bar control and positioning system cabinet consists of three network layers: a recurrent network layer, a fully convolutional layer, and a feature fusion layer.
4. A temperature early warning method for a rod control and positioning system cabinet, characterized in that, include: The temperature prediction module in the temperature early warning system of the rod control and position system cabinet as described in any one of claims 1 to 3 acquires data signals from the sensor, inputs the data signals into the temperature prediction model of the rod control and position system cabinet for real-time cabinet temperature prediction, and outputs the predicted temperature data to the decision module after obtaining the predicted temperature data. The dynamic threshold module in the temperature early warning system of the rod control and position system cabinet acquires data signals from the sensor, uses the POT model to generate real-time thresholds, and outputs the alarm threshold temperature value to the decision module. The temperature warning system of the rod control and positioning system cabinet uses a decision module to compare the real-time predicted temperature value with the alarm threshold temperature value and make a decision and issue an alarm.
5. A temperature early warning method for a rod control and positioning system cabinet according to claim 4, characterized in that, The temperature prediction module acquires data signals from sensors and inputs these signals into the rack temperature prediction model of the rod control and positioning system for real-time rack temperature prediction. After obtaining the predicted temperature data, it outputs the predicted temperature data to the decision module, including: The temperature prediction module acquires data signals from the sensor in real time, and calibrates the data signals before the current moment as historical data, and the data signals at the current moment as real-time data; Historical and real-time data are preprocessed to obtain preprocessed historical and real-time data. The preprocessed historical data is fed into the temperature prediction model of the bar-controlled bar position system cabinet as input for model training, and the temperature prediction model of the bar-controlled bar position system cabinet cabinet is updated regularly to obtain the data fitting network model of the temperature of the bar-controlled bar position system cabinet. The preprocessed real-time data is used as input to the data fitting network model of the bar control and positioning system cabinet temperature, and then the predicted temperature data is output.
6. The temperature early warning method for a rod control and positioning system cabinet according to claim 4, characterized in that, The dynamic threshold module in the temperature warning system of the rod control and positioning system cabinet acquires data signals from sensors, generates real-time thresholds using the POT model, and outputs alarm threshold temperature values to the decision module, including: The dynamic threshold module in the temperature warning system of the rod control and position system cabinet is used to acquire sensor data signals in real time, and the data signals before the current moment are calibrated as historical data, and the data signals at the current moment are calibrated as real-time data; Historical and real-time data are preprocessed to obtain preprocessed historical and real-time data. The preprocessed historical data is traversed to determine the maximum voltage value in the historical bar control system cabinet temperature data. The preprocessed real-time data is then used as the input to the POT model. The POT model is used to calculate the threshold temperature value based on the maximum temperature value, and the alarm threshold temperature value is output to the decision module.
7. A temperature early warning method for a rod control and positioning system cabinet according to claim 4, characterized in that, The temperature warning system of the rod control and positioning system cabinet uses a decision module to compare the real-time predicted temperature value with the alarm threshold temperature value and make a decision and issue an alarm, including: Using the temperature warning system of the rod control and position system cabinet, the decision module reads the predicted temperature data output by the temperature prediction module and the alarm threshold temperature value output by the dynamic threshold module, and checks whether the predicted temperature data and alarm threshold temperature value are valid. If they are valid, proceed to the next step; otherwise, the data is invalid and the task ends. The system compares the predicted temperature value with a threshold value. Only when the predicted temperature value is greater than the threshold value is the temperature abnormality confirmed and an alarm triggered.