Converter valve internal temperature monitoring and early warning method and system

Abstract

The invention discloses a converter valve internal temperature monitoring and early warning method and system, and particularly relates to the technical field of converter valve temperature monitoring. Temperature sensors are arranged on a plurality of key heating units, local temperature data are collected in real time, a time sequence is built in combination with cooling water inlet and outlet temperature, a temperature trend prediction model is built based on temperature rise rate and temperature gradient change, local thermal runaway risks are accurately recognized, and a prediction result is compared with a grading threshold value, so that the local thermal runaway risk is determined. Multi-stage early warning signal output is realized, and a corresponding processing strategy is linked; under the extreme conditions that the local temperature is continuously and abnormally increased and the response of the cooling system is lagged, the system can trigger emergency shutdown protection in an override mode, device failure is avoided, the sensing capacity and response speed of local thermal faults are effectively improved, and the operation safety and the intelligent level of the converter valve system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of converter valve temperature monitoring, and particularly to a method and system for internal temperature monitoring and early warning of a converter valve. Background Art

[0002] The converter valve is a key device in a high-voltage direct current (HVDC) transmission system, and its operating state directly affects the safety and stability of the entire transmission system. Since a large number of power devices (such as thyristors or IGBTs), capacitors, and other electronic components are contained inside the converter valve, these devices generate a large amount of heat during operation and need to be dissipated through a forced cooling system. Currently, the commonly used method is to monitor the inlet and outlet temperatures of the cooling water and trigger an early warning signal after a certain time delay when the temperature exceeds a set threshold.

[0003] The existing technology has the following deficiencies:

[0004] In the existing technology, the time-delay alarm mechanism based on the inlet valve cooling water temperature can work effectively in most operating scenarios. However, in special cases such as sudden high load of the converter valve or local failure of the cooling system, it may cause problems of delayed alarm of local thermal runaway. The specific manifestations are as follows: Some internal power devices or contact areas may experience rapid temperature rise due to blockage, scaling of the cooling pipeline, or local poor ventilation. However, since the overall temperature of the cooling water rises slowly, the system only detects a slight overlimit and then enters a time-delay waiting state without immediately alarming, resulting in local accumulation of heat and ultimately causing device burnout or insulation breakdown. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for internal temperature monitoring and early warning of a converter valve to solve the deficiencies in the background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A method for internal temperature monitoring and early warning of a converter valve, comprising:

[0007] Install multi-point temperature sensors at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time;

[0008] Construct a time series from the local temperature data and the cooling water inlet and outlet temperature data, and based on the local temperature rise rate and temperature gradient change, construct a temperature trend prediction model to determine whether there is a risk of local thermal runaway;

[0009] Compare the judgment result with a gradient risk threshold, generate early warning signals of different levels, and perform corresponding processing;

[0010] If the local temperature continues to increase abnormally, exceeds the safety temperature limit, and the overall temperature of the cooling system does not synchronously reflect the abnormal situation, trigger an emergency shutdown protection instruction.

[0011] Preferably, the key heating unit includes: the thyristor or IGBT module body and its heat dissipation base, the connection terminals of the power capacitor and resistor, the power device bus interface or the copper bar joint, and the surface of the insulation structure in the strong electric field concentration area.

[0012] Preferably, the local temperature data T i (t) collected by the temperature sensors arranged on each key heating unit inside the converter valve and the cooling water inlet temperature T in (t), the outlet water temperature T out (t) are synchronously recorded at a fixed sampling period to form a multi-dimensional temperature time series T(t) = {T1(t), T2(t),..., T m (t), T in (t), T out (t)} under a unified time reference, where m is the total number of temperature data.

[0013] Preferably, for each sensor point, calculate the temperature change rate per unit time. Set the temperature sampling period of the system as Δt. For any temperature sensor point i arranged at the converter valve part, at two consecutive sampling times t k and t k+1 obtain its temperature value and calculate the temperature rise rate per unit time: In the formula, T i (t k ) is the temperature value at the current moment, T i (t k+1 ) is the temperature value at the next moment, and S i (t k ) is the temperature change rate.

[0014] Preferably, for multiple physically adjacent sensor points, calculate the spatial temperature distribution gradient: construct a graph structure G = (V, E, W); V represents the set of all temperature sensor points in the graph, denoted as V = {v1, v2,..., v n}; E is the edge set, indicating the spatial relationship between adjacent sensors; W is the edge weight set; construct a temperature vector T, T = [T1, T2,..., T n S ; S is the vector transpose, calculate the graph Laplacian matrix L: L = D - W; where: D is the degree matrix; for each node v i , define its spatial temperature distribution gradient as: is the set of neighbor nodes of node i, and G is the spatial temperature distribution gradient.

[0015] ​Preferably, the temperature change rate and the spatial temperature distribution gradient are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model takes predicting the local thermal runaway risk value label for each group of comprehensive feature vectors as the prediction target, and minimizing the sum of the prediction errors for all local thermal runaway risk value labels as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence and then the model training is stopped. The local thermal runaway risk value is determined according to the model output result, where the machine learning model is a polynomial regression model.

[0016] Preferably, the obtained local thermal runaway risk value is compared with the gradient risk threshold. The gradient risk threshold includes a first risk threshold and a second risk threshold, and the first risk threshold is less than the second risk threshold. The local thermal runaway risk value is respectively compared with the first risk threshold and the second risk threshold;

[0017] If the risk value is greater than the second risk threshold, it is determined as a severe thermal runaway risk, and a secondary emergency warning or protection action is immediately triggered; if the risk value is between the first and second risk thresholds, it is determined as a moderate risk state, and a primary warning is triggered to prompt operation and maintenance intervention; if the risk value is less than the first risk threshold, it is determined that the current state is safe and stable, and no warning is triggered.

[0018] Preferably, when it is monitored that the temperature of a certain local sensor point or adjacent sensor area shows a continuous increasing trend and the increasing trend meets the following set of combined conditions, the system will consider that potential thermal runaway has occurred:

[0019] The first judgment condition: The temperature of a certain sensor shows a monotonous increase in multiple consecutive sampling periods; and its temperature change rate continuously exceeds the set threshold, and at this time it is considered that the local area has entered an abnormal heat accumulation state;

[0020] The second judgment condition: The current temperature value exceeds the safe temperature limit value allowed by the device, and the temperature difference change amount between the inlet and outlet temperatures T in (t), T out (t) of the surrounding cooling water system is less than the set response threshold ΔT min , that is: |T out (t) - T in (t)| < ΔT min ; indicating that the overall temperature of the cooling system has not reflected the local temperature abnormality, that is, the cooling system may not be able to respond effectively or there is a local failure;

[0021] The third judgment condition: The flow rate or pressure monitoring value of the cooling channel corresponding to the current area is normal, further confirming that it is a local thermal abnormality rather than a system false alarm.

[0022] The present invention also provides a temperature monitoring and early warning system for the internal of a converter valve, including a temperature data acquisition module, a thermal runaway prediction module, an early warning module, and an abnormal handling module;

[0023] Temperature data acquisition module: A multi-point temperature sensor is set at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time;

[0024] Thermal runaway prediction module: Construct a time series with the local temperature data and the temperature data of the inlet and outlet of the cooling water, and based on the local temperature rise rate and the change of the temperature gradient, construct a temperature trend prediction model to judge whether there is a risk of local thermal runaway;

[0025] Early warning module: Compare the judgment result with the gradient risk threshold, generate early warning signals of different levels, and perform corresponding processing;

[0026] Abnormal handling module: If the local temperature continues to increase abnormally, exceeds the safe temperature limit, and the overall temperature of the cooling system does not synchronously reflect the abnormal situation, then trigger an emergency shutdown protection instruction.

[0027] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0028] 1. By arranging multi-point temperature sensors at multiple key heat-generating units inside the converter valve, the present invention collects local temperature data in real time, constructs a temperature time series under a unified time reference in combination with the temperature of the inlet and outlet of the cooling water, further extracts comprehensive feature vectors based on the temperature change rate and the spatial temperature distribution gradient, and uses a polynomial regression model to intelligently predict the risk of local thermal runaway, effectively breaking through the response lag problem brought by the traditional reliance on the average temperature of the cooling water and fixed-delay alarm, being able to identify the local overheating trend earlier, and realizing the transformation from temperature overlimit alarm to thermal risk trend prediction.

[0029] 2. By setting a hierarchical risk threshold, the present invention realizes a hierarchical control mechanism for risk response, and triggers an emergency protection instruction when it is detected that the thermal runaway trend is significant but the cooling system has no obvious response, thereby greatly improving the safety and robustness of the converter valve system under complex working conditions such as sudden high load and local cooling failure. This method has strong real-time performance, predictability and adaptability, can be widely applied to high-voltage direct current transmission systems, and significantly improves the intelligent level and reliability of equipment operation. Description of the Drawings

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0031] Figure 1 This is the flowchart of the warning method of the present invention.

[0032] Figure 2 This is the system module diagram of the present invention. Detailed implementation manners

[0033] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0034] Embodiment 1. Please refer to Figure 1 As shown, a method for monitoring and warning the internal temperature of a converter valve in this embodiment includes:

[0035] Install multi-point temperature sensors at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time;

[0036] Construct a time series from the local temperature data and the temperature data at the inlet and outlet of the cooling water, and based on the local temperature rise rate and the change in temperature gradient, construct a temperature trend prediction model to determine whether there is a risk of local thermal runaway;

[0037] Compare the judgment result with the gradient risk threshold to generate warning signals of different levels and perform corresponding processing;

[0038] If the local temperature continues to increase abnormally, exceeds the safety temperature limit, and the overall temperature of the cooling system does not reflect the abnormal situation synchronously, an emergency shutdown protection instruction is triggered.

[0039] In the present invention, in order to achieve refined monitoring of the internal temperature state of the converter valve, highly sensitive temperature sensors are deployed on multiple key heat-generating units inside the converter valve structure to form a multi-point temperature acquisition network. The specific implementation manners are as follows:

[0040] Object of sensor layout: Select key components inside the converter valve that are prone to generate hot spots or have a risk of failure as the layout points, including: the body of the thyristor or IGBT module and its heat dissipation base, the connection ends of power capacitors and resistors, the bus interfaces or copper bar joints of power devices, and the surface of the insulation structure in the strong electric field concentration area.

[0041] Sensor type selection: Thermistors (NTC / PTC) can be used in medium and low temperature regions; thermocouples or fiber optic temperature sensors can be used in high voltage or strong electromagnetic interference environments; for areas with high accuracy requirements, digital micro temperature chips can be used, which support fast response and remote data transmission.

[0042] Layout strategy: Avoid single-point judgment by setting sensors only at the inlet or outlet of the cooling water. Adopt a layout method of "combining points and surfaces", that is, set multiple measurement points on a single component, and conduct comparative analysis by combining the horizontal and vertical temperature distributions. Add "thermal distribution reference points" for comparing the thermal diffusion rates of different parts.

[0043] Data acquisition and processing: All sensors send real-time temperature data to the monitoring main control unit through a high-speed data bus (such as CAN, RS485 or fiber optic communication). The main control unit is embedded with a temperature analysis module, which conducts data comparison and trend recognition according to the sampling frequency (such as 1Hz - 10Hz), and forms a thermal distribution map. Dynamically calculate the temperature rise rate (ΔT / Δt) and temperature gradient (ΔT / Δx) to implement the abnormal judgment logic.

[0044] The advantages of the present invention are reflected in: achieving rapid positioning of local abnormal temperature rise inside the converter valve; reducing the early warning lag caused by traditional average temperature judgment; enhancing the safety response ability of the equipment in case of sudden load or local cooling failure; providing data support for subsequent big data diagnosis and thermal failure prediction based on AI.

[0045] The local temperature data T i (t) collected by the temperature sensors arranged on each key heat-generating unit inside the converter valve, the cooling water inlet temperature T in (t), and the outlet temperature T out (t) are synchronously recorded at a fixed sampling period (such as 1 second) to form a multi-dimensional temperature time series under a unified time reference. This data structure can be expressed as: T(t) = {T1(t), T2(t),..., T m (t), T in (t), T out (t)}; m is the total number of temperature data.

[0046] Calculate the temperature change rate per unit time for each sensor point: Set the temperature sampling period of the system as Δt (unit: second), for example: Δt = 1s. All temperature sensors perform synchronous sampling according to this period to ensure data time sequence consistency.

[0047] For any temperature sensor point i arranged at a key part of the converter valve, at two consecutive sampling times t k and t k+1 obtain its temperature value and calculate the temperature rise rate per unit time: In the formula, T i (t k ) is the temperature value at the current moment, and T i (t k+1 ) is the temperature value at the next moment. S i (t k ) is the temperature change rate.

[0048] For multiple physically adjacent sensor points, calculate the spatial temperature distribution gradient: construct a graph structure G = (V, E, W); V represents the set of all temperature sensor points in the graph, denoted as V = {v1, v2,..., v n}; E is the edge set, indicating the spatial relationship between adjacent sensors (for example, the distance is less than a threshold); W is the edge weight set, indicating the heat conduction relationship or distance attenuation between nodes, and can be defined as: where d ij is the physical distance between sensors i and j, and σ is the scale parameter for adjusting the heat influence range.

[0049] Construct a temperature vector T, T = [T1, T2,..., T n S ; T is the column vector of temperatures collected by each sensor at a certain moment, with the unit of degree Celsius, and S is the vector transpose. Calculate the graph Laplacian matrix L: L = D - W; where: D is the degree matrix, and the diagonal elements are the node degrees: D ii = ∑ j w ij .

[0050] Calculate the heat gradient energy term (temperature change intensity) E(T): E(T) = ∑ i,j w ij (T i - T j ) 2 ; which represents the sum of the squares of the temperature differences between all adjacent points in the graph, reflecting the spatial change energy of the temperature field, and can be used to determine whether there is a mutation region.

[0051] For each node v i , define its spatial temperature distribution gradient as: is the set of neighbor nodes of node i, and G is the spatial temperature distribution gradient.

[0052] When G is greater than a certain preset threshold G th , it is determined that there is an obvious local temperature difference near this point, which may be a heat runaway source.

[0053] ​The temperature change rate and the spatial temperature distribution gradient are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the local thermal runaway risk value label for each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors for all local thermal runaway risk value labels as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The local thermal runaway risk value is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0054] The obtained local thermal runaway risk value is compared with a gradient risk threshold, the gradient risk threshold includes a first risk threshold and a second risk threshold, and the first risk threshold is less than the second risk threshold, and the local thermal runaway risk value is compared with the first risk threshold and the second risk threshold respectively;

[0055] If the risk value is greater than the second risk threshold, it is judged as a serious thermal runaway risk and a second-level emergency alarm or protection action is immediately triggered;

[0056] If the risk value is between the first and second risk thresholds, it is judged as a moderate risk state, triggering a first-level warning and prompting operation and maintenance intervention;

[0057] If the risk value is less than the first risk threshold, the current state is judged to be safe and stable, and no warning is triggered.

[0058] The system determines that the monitoring point or area has entered a serious overheating state, which is very likely to cause device failure or insulation breakdown. The processing measures include: immediately triggering the secondary alarm signal and pushing high-priority alarms to the operation and maintenance system and control center; initiating emergency protection actions, including: forced switching of cooling modes (such as boosting and increasing flow); load reduction operation; triggering the corresponding converter valve module to automatically exit service (de-parallelization); linkage start-up of the bypass protection system; recording event data, including current feature vectors, heat maps, response measure timestamps, etc., for later analysis and model retraining.

[0059] The system determines that there is a trend or hidden danger of thermal runaway in the current area, but it has not yet reached an emergency state. The treatment measures include: triggering a first-level warning signal and sending a prompt alarm to the operation and maintenance platform; dynamically adjusting the operating parameters of the cooling system, such as: increasing the cooling power of the area; starting the backup cooling channel; starting the heat distribution trend tracking mechanism to continuously monitor the temperature changes in the area at a high frequency in the next few minutes; marking the area as a thermal abnormality area and giving priority to it in the scheduling and inspection plan;

[0060] The system determines that the temperature at the current monitoring point is stable and the operation is normal. The processing methods include: maintaining the regular monitoring frequency and not triggering any alarms; optionally enabling the dynamic threshold adaptive mechanism: if in a low-risk state for a long time, the risk threshold can be appropriately tightened to enhance sensitivity; recording the stable operation data for continuous optimization training of the model.

[0061] If the local temperature continues to increase abnormally, exceeding the safety temperature limit, and the overall temperature of the cooling system does not synchronously reflect the abnormal situation, an emergency shutdown protection instruction is triggered.

[0062] Specifically, when the system monitors that the temperature of a certain local sensor point or adjacent sensor area shows a continuous increasing trend, and this increasing trend meets the following set of combined conditions, the system will consider that potential thermal runaway has occurred:

[0063] The first judgment condition: The temperature of a certain sensor shows a monotonic increase in multiple consecutive sampling periods (for example, N consecutive samplings); and the rate of change of its temperature continuously exceeds the set threshold, and at this time it is determined that the local area has entered an abnormal thermal accumulation state.

[0064] The second judgment condition: The current temperature value exceeds the safety temperature limit value allowed by the device, and the difference change amount of the inlet and outlet temperatures T in (t), T out (t) of the surrounding cooling water system is less than the set response threshold ΔT min , that is: |T out (t) - T in (t)| < ΔT min ; indicating that the overall temperature of the cooling system has not reflected the local temperature abnormality, that is, the cooling system may not be able to respond effectively or there is local failure.

[0065] The third judgment condition: The monitored value of the flow rate or pressure of the cooling channel corresponding to the current area is normal (no cooling interruption), further confirming that it is a local thermal abnormality rather than a system false alarm.

[0066] After meeting the above multiple conditions, the system will comprehensively judge that there is a typical local rapid thermal runaway and ineffective cooling response situation in this area. Since this type of abnormality is often masked by the delay mechanism or average temperature judgment in the existing technology, it may lead to serious device failure or fire risk. In this scenario, the present invention sets a priority trigger path for the emergency shutdown protection instruction: skipping the regular delay and level alarm processes; directly triggering the control system to enter the emergency response state and execute: the faulty converter valve module is taken out of service; the whole machine is protected and shut down or switched to the bypass channel; a red warning signal is sent to the monitoring center and the fault data is locked.

[0067] Example 2, please refer to Figure 2As shown in the figure, a temperature monitoring and early warning system for the internal of a converter valve in this embodiment includes a temperature data acquisition module, a thermal runaway prediction module, an early warning module, and an abnormal handling module;

[0068] Temperature data acquisition module: A multi-point temperature sensor is set at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time;

[0069] Thermal runaway prediction module: Construct a time series from the local temperature data and the temperature data at the inlet and outlet of the cooling water, and based on the local temperature rise rate and the change of temperature gradient, construct a temperature trend prediction model to judge whether there is a risk of local thermal runaway;

[0070] Early warning module: Compare the judgment result with the gradient risk threshold, generate early warning signals of different levels, and perform corresponding processing;

[0071] Abnormal handling module: If the local temperature continues to increase abnormally, exceeds the safe temperature limit, and the overall temperature of the cooling system does not reflect the abnormal situation synchronously, then trigger an emergency shutdown protection instruction.

[0072] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0073] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. These three situations, where A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0074] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0075] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application.

Claims

1. A method for monitoring and warning the internal temperature of a converter valve, characterized in that: Including: Install multi-point temperature sensors at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time; Construct a time series of the local temperature data and the cooling water inlet and outlet temperature data, and based on the local temperature rise rate and temperature gradient change, construct a temperature trend prediction model to determine whether there is a risk of local thermal runaway; Compare the judgment result with the gradient risk threshold, generate warning signals of different levels, and perform corresponding processing; If the local temperature continues to increase abnormally, exceeds the safety temperature limit, and the overall temperature of the cooling system does not synchronously reflect the abnormal situation, trigger an emergency shutdown protection instruction.

2. The internal temperature monitoring and early warning method for a commutation valve according to claim 1, characterized in that: The key heat-generating units include: the thyristor or IGBT module body and its heat dissipation base, the connection ends of power capacitors and resistors, the bus interfaces or copper bar joints of power devices, and the surface of the insulation structure in the strong electric field concentration area.

3. A method for monitoring and warning the internal temperature of a converter valve according to claim 1, characterized in that: The local temperature data T collected by the temperature sensors arranged on each key heating unit inside the converter valve i (t) and the inlet cooling water temperature T in (t), the outlet water temperature T out (t) are synchronously recorded at a fixed sampling period to form a multi-dimensional temperature time series T(t) = {T1(t), T2(t),..., T m (t), T in (t), T out (t)} under a unified time reference, where m is the total number of temperature data.

4. A method for monitoring and warning the internal temperature of a converter valve according to claim 3, characterized in that: Calculate the temperature change rate per unit time for each sensor point. Set the temperature sampling period of the system as Δt. For any temperature sensor point i deployed at the converter valve location, at two consecutive sampling times t k and t k+1 obtain its temperature value and calculate the temperature rise rate per unit time: In the formula, T i (t k ) is the temperature value at the current time, T i (t k+1 ) is the temperature value at the next time, and S i (t k ) is the temperature change rate.

5. A method for monitoring and warning the internal temperature of a converter valve according to claim 4, characterized in that: For multiple physically adjacent sensor points, calculate the spatial temperature distribution gradient: construct a graph structure G = (V, E, W); V represents the set of all temperature sensor points in the graph, denoted as V = {v1, v2,..., v n}; E is the set of edges, indicating the spatial relationship between adjacent sensors; W is the set of edge weights; construct a temperature vector T, T = [T1, T2,..., T n S ;​ Let \(S\) be the vector transpose, and calculate the graph Laplacian matrix \(L\): \(L = D - W\); where: \(D\) is the degree matrix; for each node \(v\) i , define its spatial temperature distribution gradient as: is the set of neighbor nodes of node \(i\), and \(G\) is the spatial temperature distribution gradient.

6. A method for monitoring and warning the internal temperature of a converter valve according to claim 5, characterized in that: Convert the temperature change rate and the spatial temperature distribution gradient into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, take predicting the local thermal runaway risk value label for each group of comprehensive feature vectors as the prediction target, and take minimizing the sum of the prediction errors for all local thermal runaway risk value labels as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and stop the model training, and determine the local thermal runaway risk value according to the model output result, where the machine learning model is a polynomial regression model.

7. A method for monitoring and warning the internal temperature of a converter valve according to claim 6, characterized in that: Compare the obtained local thermal runaway risk value with the gradient risk threshold. The gradient risk threshold includes a first risk threshold and a second risk threshold, and the first risk threshold is less than the second risk threshold. Compare the local thermal runaway risk value with the first risk threshold and the second risk threshold respectively; If the risk value is greater than the second risk threshold, it is judged as a serious thermal runaway risk, and immediately trigger a secondary emergency alarm or protection action; If the risk value is between the first and second risk thresholds, it is judged as a moderate risk state, trigger a first-level warning, and prompt operation and maintenance intervention; If the risk value is less than the first risk threshold, it is judged that the current state is safe and stable, and no warning is triggered.

8. A method for monitoring and warning the internal temperature of a converter valve according to claim 1, characterized in that: When it is monitored that the temperature of a certain local sensor point or adjacent sensor area shows a continuous increasing trend, and the increasing trend meets the following set of combined conditions, the system will regard it as a potential thermal runaway: The first judgment condition: the temperature of a certain sensor shows a monotonic increase in multiple consecutive sampling periods; and its temperature change rate continuously exceeds the set threshold, and at this time it is considered that the local area has entered an abnormal heat accumulation state; Second judgment condition: The current temperature value exceeds the allowable safety temperature limit of the device, and the change in the temperature difference between the inlet and outlet of the surrounding cooling water system T in (t), T out (t) is less than the set response threshold ΔT min , that is: |T out (t) - T in (t)| < ΔT min ; indicating that the overall temperature of the cooling system has not reflected local temperature anomalies, that is, the cooling system may not respond effectively or there may be local failures; The third judgment condition: the flow rate or pressure monitoring value of the cooling channel corresponding to the current area is normal, further confirming that it is a local thermal anomaly rather than a system false alarm.

9. A temperature monitoring and early warning system inside a converter valve is used to implement the temperature monitoring and early warning method for the inside of a converter valve according to any one of claims 1-8, and is characterized in that: Including a temperature data acquisition module, a thermal runaway prediction module, a warning module, and an abnormal processing module; Temperature data acquisition module: Install multi-point temperature sensors at multiple key heat-generating units inside the converter valve to collect local temperature data of each area in real time; Thermal runaway prediction module: Construct a time series of the local temperature data and the cooling water inlet and outlet temperature data, and based on the local temperature rise rate and temperature gradient change, construct a temperature trend prediction model to determine whether there is a risk of local thermal runaway; Early warning module: Compare the judgment result with the gradient risk threshold, generate early warning signals of different levels, and perform corresponding processing; Abnormal handling module: If the local temperature continues to increase abnormally, exceeds the safety temperature limit, and the overall temperature of the cooling system does not reflect the abnormal situation synchronously, trigger an emergency shutdown protection instruction.

Images