Thermal signal analysis early warning method and system based on knowledge graph

Abstract

The invention relates to the technical field of thermal signal early warning, and discloses a thermal signal analysis early warning method and system based on a knowledge graph, and the method comprises the following steps: S101, collecting thermal signal data; s102, acquiring an enthalpy value and a mass flow rate, and generating an energy conservation residual error; s103, evaluating a temperature change trend, generating a thermal inertia prediction residual error, evaluating a channel coupling degree to obtain a cross-channel coefficient, and determining a cross-channel coefficient residual error; s104, normalizing the energy conservation residual error, the thermal inertia prediction residual error and the cross-channel coefficient residual error respectively, and fusing to generate an abnormal score; and S105, mounting the energy conservation residual error, the thermal inertia prediction residual error, the cross-channel coefficient residual error and the abnormal score to a preset knowledge graph to generate a fault early warning triple. According to the invention, high-precision signal processing, multi-dimensional fault detection, intelligent reasoning and accurate early warning of the thermal system are realized.

Description

Technical Field

[0001] The present invention belongs to the field of thermal signal early warning technology, and specifically relates to a thermal signal analysis and early warning method and system based on knowledge graph. Background Art

[0002] In the operation of thermal systems (such as boilers, steam turbines, and heat exchangers) in the energy, chemical, and power sectors, real-time monitoring and fault warning are core requirements for ensuring safe, stable, and efficient system operation. Traditional thermal signal analysis methods primarily rely on single-parameter threshold alarms, which suffer from the following significant flaws: 1. Lack of multi-parameter coupling: Thermal system parameters such as temperature, pressure, flow, and power exhibit strong coupling. Traditional methods analyze single parameters in isolation, failing to capture anomalies in the correlations between parameters and prone to missing early-stage complex faults. 2. Inadequate dynamic characteristic modeling: The system's dynamic behaviors, such as thermal inertia and energy conservation, lack quantitative modeling. This makes it impossible to effectively predict temperature trends or assess the extent of energy imbalances, making it difficult to provide early warnings of progressive faults. 3. Highly subjective threshold setting: Traditional fixed thresholds are not dynamically adjusted based on historical system operation data and cannot adapt to slow-changing characteristics such as equipment aging and changing operating conditions. They often trigger false alarms due to environmental interference or sensor drift, resulting in low warning reliability. 4. Single fault diagnosis dimension: Lack of multi-dimensional integrated analysis of the system's global status makes it difficult to distinguish between apparent anomalies and actual faults. Operations and maintenance personnel must manually check multi-source data, resulting in low response efficiency. Summary of the Invention

[0003] The present invention provides a thermal signal analysis and early warning method and system based on knowledge graph, which solves the technical problem in related technologies that thermal signal early warning relies only on a single physical quantity or empirical threshold and is difficult to simultaneously capture energy imbalance, thermal inertia deviation and channel coupling anomaly.

[0004] The present invention provides a thermal signal analysis and early warning method based on knowledge graph, comprising the following steps: S101, collecting thermal signal data from a thermal system within a first preset time period and performing preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; S102, obtaining enthalpy and mass flow rate based on the preprocessed thermal signal data, thereby determining inlet energy flow and outlet energy flow, and simultaneously evaluating the deviation of energy conservation to form an energy conservation residual; S103: Evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data, and generate a thermal inertia prediction residual. Evaluate the channel coupling degree based on the preprocessed temperature and pressure to obtain a cross-channel coefficient, and then determine the cross-channel coefficient residual. S104, normalizing the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual respectively, and fusing them according to preset weights to generate an anomaly score; S105, the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score are mounted as attributes to the corresponding device node in the preset knowledge graph, and knowledge reasoning is performed on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, fault type; and based on the fault warning triplet, a warning signal is output according to the preset warning strategy.

[0005] Furthermore, the denoising adopts sliding window moving average filtering; In the baseline drift correction, the baseline mean of each parameter is obtained by calculating the average value of temperature, pressure, flow and power within a preset historical healthy period, and the difference between each parameter of the thermal signal data and the baseline mean of the corresponding parameter is calculated to obtain the corrected parameter.

[0006] Furthermore, the enthalpy value and mass flow rate are obtained based on the pre-processed thermal signal data, specifically including: The enthalpy value is obtained by multiplying the difference between the pre-processed temperature and the preset saturation temperature by the preset pressure-specific heat capacity, and then adding the product to the saturation enthalpy corresponding to the preset saturation temperature; The fluid density is obtained by using the pre-processed temperature and pressure as query conditions in the preset fluid state table, and the mass flow rate is obtained by multiplying the fluid density by the pre-processed flow rate.

[0007] Furthermore, in the plurality of inlet channels, the enthalpy value corresponding to each inlet channel and the corresponding mass flow rate are sequentially multiplied and accumulated to obtain the inlet energy flow; In multiple outlet channels, the enthalpy value corresponding to each outlet channel and the corresponding mass flow rate are multiplied and added in sequence to obtain the outlet energy flow.

[0008] Furthermore, the system energy storage change is determined based on the preset system heat capacity and the preprocessed temperature difference between adjacent sampling moments, the difference between the inlet energy flow and the outlet energy flow is compared with the system energy storage change, and the absolute difference between the two is used as the energy conservation residual at the current sampling moment.

[0009] Furthermore, within adjacent sampling moments, based on the preprocessed temperature at the previous sampling moment, the inlet energy flow at the previous sampling moment, the outlet energy flow at the previous sampling moment, and the preset system heat capacity, the predicted temperature at the current sampling moment is calculated through the thermal inertia model, and then the predicted temperature is compared with the preprocessed temperature at the current sampling moment, and the absolute difference between the two is used as the thermal inertia prediction residual at the current sampling moment.

[0010] Furthermore, the determination of the cross-channel coefficient residual includes: S201, within a preset sliding window period, based on the pre-processed temperature and pressure, calculate the Pearson correlation coefficient between the two within the sliding window and use it as the cross-channel coefficient; S202, within a preset historical health period, according to a preset sliding window time period, the Pearson correlation coefficients of multiple sliding windows are obtained using the same method as S201, and the average is calculated to obtain a reference cross-channel coefficient; S203 : Compare the cross-channel coefficient with the reference cross-channel coefficient, and use the absolute difference between the two as the cross-channel coefficient residual of the sliding window corresponding to the current sampling moment.

[0011] Furthermore, the specific steps of S104 include: S301, within a preset historical health period, composing energy conservation residuals, thermal inertia prediction residuals, and cross-channel coefficient residuals at multiple sampling moments into an energy conservation residual sequence, a thermal inertia prediction residual sequence, and a cross-channel coefficient residual sequence, respectively, in chronological order; S302, statistically processing the energy conservation residual sequence, thermal inertia prediction residual sequence, and cross-channel coefficient residual sequence respectively to obtain the 99th percentile of each sequence, and use it as the healthy period threshold corresponding to the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual; S303: performing a ratio operation on the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual at the sampling time within the first preset time period and the corresponding healthy period threshold value to obtain normalized energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual; S304 , performing weighted accumulation on the normalized energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual according to pre-set weights to obtain an anomaly score.

[0012] Furthermore, knowledge reasoning is performed on the mount attributes to generate fault warning triples, including: When the energy conservation residual of a device node is greater than or equal to the preset energy imbalance threshold, and the anomaly scores of the three most recent consecutive sampling moments are all greater than or equal to 1, a fault warning triplet with the fault type being energy imbalance fault is generated; When the thermal inertia prediction residual of the device node is greater than or equal to the preset thermal inertia threshold, and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet is generated with the fault type being temperature trend deviation fault; When the cross-channel coefficient residual of the device node is greater than or equal to the preset coupling threshold and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet with the fault type being coupling failure is generated; When a fault warning triplet with the fault type of energy imbalance fault already exists in the preset knowledge graph and the device node has a high-voltage state attribute, a fault warning triplet with the fault type of overheating high-voltage fault is generated, where the high-voltage state means that the pressure after preprocessing is greater than or equal to the preset safety pressure threshold.

[0013] The present invention provides a thermal signal analysis and early warning system based on a knowledge graph, comprising: a data acquisition and preprocessing module, configured to acquire thermal signal data from the thermal system within a first preset time period and perform preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and wherein the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; The energy conservation deviation evaluation module is used to obtain the enthalpy value and mass flow rate based on the pre-processed thermal signal data, and then determine the inlet energy flow and outlet energy flow, while evaluating the deviation of energy conservation to form the energy conservation residual; A multi-dimensional residual calculation module is used to evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data and generate thermal inertia prediction residuals. It also evaluates the channel coupling degree based on the preprocessed temperature and pressure to obtain the cross-channel coefficient and further determine the cross-channel coefficient residual; The residual fusion module is used to normalize the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual, and fuse them according to the preset weights to generate anomaly scores; The reasoning and warning module is used to mount the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score as attributes to the corresponding device nodes in the preset knowledge graph, and perform knowledge reasoning on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, and fault type; and based on the fault warning triplet, output a warning signal according to the preset warning strategy.

[0014] The beneficial effects of the present invention are as follows: the present invention improves signal accuracy by collecting thermal signals and performing denoising and baseline drift correction; generates energy conservation residuals by calculating enthalpy, mass flow, and energy flow, combined with system heat capacity and temperature differences, and comprehensively evaluates energy balance; generates multi-dimensional residuals by using thermal inertia models and Pearson correlation coefficients, normalizes and weightedly fuses them into anomaly scores based on historical data; mounts residual attributes on a knowledge graph and performs multi-conditional reasoning to generate fault warning triples, and outputs responses based on a hierarchical warning strategy. This achieves high-precision preprocessing of thermal system signals, multi-dimensional fault feature extraction, knowledge-driven intelligent reasoning, and precise hierarchical warnings, improving the accuracy, comprehensiveness, and timeliness of fault detection, enhancing the safety and maintenance efficiency of system operation, and reducing false alarm rates and potential risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the thermal signal analysis and early warning method based on knowledge graph of the present invention. DETAILED DESCRIPTION

[0016] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0017] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprising" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, but do not exclude other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0018] like Figure 1 As shown in FIG, the thermal signal analysis and early warning method based on the knowledge graph includes the following steps: S101, collecting thermal signal data from a thermal system within a first preset time period and performing preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; S102, obtaining enthalpy and mass flow rate based on the preprocessed thermal signal data, thereby determining inlet energy flow and outlet energy flow, and simultaneously evaluating the deviation of energy conservation to form an energy conservation residual; S103: Evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data, and generate a thermal inertia prediction residual. Evaluate the channel coupling degree based on the preprocessed temperature and pressure to obtain a cross-channel coefficient, and then determine the cross-channel coefficient residual. S104, normalizing the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual respectively, and fusing them according to preset weights to generate an anomaly score; S105, the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score are mounted as attributes to the corresponding device node in the preset knowledge graph, and knowledge reasoning is performed on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, fault type; and based on the fault warning triplet, a warning signal is output according to the preset warning strategy.

[0019] In one embodiment of the present invention, a thermal resistor sensor is used to obtain temperature, a diffused silicon pressure transmitter is used to obtain pressure, a vortex flowmeter is used to obtain flow, and a power transmitter is used to collect voltage and current to calculate power.

[0020] In one embodiment of the present invention, the denoising adopts sliding window moving average filtering. Specifically, for each parameter of temperature, pressure, flow and power, a sliding window of fixed length is set on the time series, and the arithmetic mean of the signal value in the window is calculated as the denoised value of the window center point. Each time the sliding window moves forward one sampling point, the average value in the window is recalculated, and the entire time series is iteratively processed. This method can effectively filter out high-frequency random noise and retain the true trend characteristics of the signal.

[0021] In the baseline drift correction described, the baseline mean of each parameter is obtained by calculating the average value of temperature, pressure, flow, and power during a preset historical healthy period. The difference between each parameter in the thermal signal data and the corresponding baseline mean is then calculated to obtain the corrected parameter. Specifically, after the initial commissioning of the thermal system, a continuous period of time known to be operating normally is selected as the historical healthy period, and sample data for each parameter during this period is collected. For each parameter, its arithmetic mean during the historical healthy period is calculated as the baseline mean. During real-time monitoring, the corresponding baseline mean is subtracted from the current parameter value to obtain the corrected parameter value. This method can effectively eliminate sensor zero drift and baseline offset caused by long-term system operation, thereby improving signal accuracy.

[0022] In one embodiment of the present invention, obtaining enthalpy and mass flow rate based on preprocessed thermal signal data specifically includes: The enthalpy value is obtained by multiplying the difference between the pre-processed temperature and the preset saturation temperature by the preset pressure-specific heat capacity, and then adding the product to the saturation enthalpy corresponding to the preset saturation temperature. According to the pre-processed pressure, the corresponding saturation temperature is obtained by querying the preset pressure-saturation temperature correspondence table. The calculation formula of the enthalpy value is: , where h represents the enthalpy value, Indicates the preset saturation enthalpy, represents the preset pressure specific heat capacity, T represents the temperature after pretreatment, Indicates the preset saturation temperature; In the preset fluid state table, the fluid density is obtained using the pre-processed temperature and pressure as query conditions, and the fluid density is multiplied by the pre-processed flow rate to obtain the mass flow rate; wherein, the preset fluid state table is pre-constructed based on the thermodynamic properties of the working fluid and can be determined through experimental measurement to ensure that the correspondence between the density value and the temperature and pressure is accurate and reliable.

[0023] The enthalpy and mass flow calculated using this method accurately reflect the energy state and actual flow rate of the working fluid in the thermal system. Enthalpy calculation combines the working fluid's saturation characteristics with its superheat contribution, avoiding the limitations of single-parameter measurement. Mass flow is corrected for density through temperature and pressure compensation, eliminating the impact of changes in the medium's state on flow measurement. This process provides critical foundational data for subsequent calculations of inlet and outlet energy flows and energy conservation residuals, ensuring the accuracy of system energy balance analysis and, in turn, improving the reliability and timeliness of fault warnings.

[0024] In one embodiment of the present invention, in a plurality of inlet channels, the enthalpy value corresponding to each inlet channel and the corresponding mass flow rate are sequentially multiplied and accumulated to obtain the inlet energy flow, wherein the inlet channel refers to a pipeline in a thermal system that transports fluid into the interior of the device; In multiple outlet channels, the enthalpy value corresponding to each outlet channel and the corresponding mass flow rate are multiplied and accumulated in sequence to obtain the outlet energy flow. The outlet channel refers to the pipeline in the thermal system used to transport fluid from the inside of the equipment to the outside.

[0025] The inlet energy flow and the outlet energy flow can accurately reflect the energy input and output balance relationship of the thermal system; this process enables the thermal system to more sensitively detect energy imbalance faults such as leakage and efficiency reduction, and significantly enhances the safety and reliability of the thermal system.

[0026] In one embodiment of the present invention, the system energy storage change is determined based on the preset system heat capacity and the pre-processed temperature difference between adjacent sampling moments, the difference between the inlet energy flow and the outlet energy flow is compared with the system energy storage change, and the absolute difference between the two is used as the energy conservation residual at the current sampling moment; The calculation formula for the system energy storage change is: , Indicates the change of system energy storage, C indicates the preset system heat capacity, represents the pre-processed temperature difference between adjacent sampling moments, and Represent the kth and k-1th sampling moments respectively, k represents the sampling moment index, Indicates the sampling interval.

[0027] The formula for calculating the energy conservation residual is: ,in, represents the energy conservation residual, and They represent the inlet energy flow and outlet energy flow, respectively. The energy conservation residual is used to measure the degree of deviation between the actual energy balance of the thermal system and the theoretical value. When there is leakage in the system or the energy conversion efficiency decreases, the difference between the inlet energy flow and the outlet energy flow will significantly deviate from the energy storage change, resulting in an increase in the residual value. This can effectively reflect the global energy balance state of the thermal system. Compared with traditional monitoring methods that only rely on a single parameter, this process comprehensively considers the dynamic relationship between energy input, output and energy storage, thereby improving the safety and operational stability of the thermal system.

[0028] In one embodiment of the present invention, within adjacent sampling moments, a predicted temperature at the current sampling moment is calculated using a thermal inertia model based on the preprocessed temperature at the previous sampling moment, the inlet energy flow at the previous sampling moment, the outlet energy flow at the previous sampling moment, and a preset system heat capacity. The predicted temperature is then compared with the preprocessed temperature at the current sampling moment, and the absolute difference between the two is used as the thermal inertia prediction residual at the current sampling moment. The calculation formula for the predicted temperature is: , represents the predicted temperature at the kth sampling moment, represents the pre-processed temperature at the k-1th sampling moment, and The predicted temperature at the kth sampling moment is compared with the preprocessed temperature, and the absolute value of the difference between the two is calculated as the thermal inertia prediction residual at the kth sampling moment. This residual reflects the degree of deviation between the actual temperature and the temperature predicted by the thermal inertia model. When there are anomalies in the thermal system, such as changes in thermal resistance, leakage, or decreased energy conversion efficiency, the actual temperature change will deviate from the model prediction value, causing this residual to increase.

[0029] Predicting temperature changes and calculating residuals using a thermal inertia model can effectively detect abnormal changes in the system's thermal dynamic characteristics. This method leverages the physical relationship between the system's heat capacity and energy flow to dynamically predict temperature trends, improving the thermal system's fault warning capabilities and operational reliability.

[0030] In one embodiment of the present invention, determining the cross-channel coefficient residual includes: S201: Within a preset sliding window period, based on the preprocessed temperature and pressure, calculate the Pearson correlation coefficient between the two within the sliding window and use it as the cross-channel coefficient. The Pearson correlation coefficient measures the correlation between temperature and pressure. The closer the Pearson correlation coefficient is to 1, the stronger the correlation between temperature and pressure. Using a sliding window can adapt to the slow time-varying characteristics of thermal systems and avoid interference from single-moment data fluctuations. S202: Within a preset historical health period, based on a preset sliding window time period, the Pearson correlation coefficients of multiple sliding windows are obtained using the same method as S201, and the average is calculated to obtain a reference cross-channel coefficient. The reference cross-channel coefficient is used to represent the coupling relationship between temperature and pressure during normal operation of the thermal system, providing a benchmark for real-time monitoring. S203 : Compare the cross-channel coefficient with the reference cross-channel coefficient, and use the absolute difference between the two as the cross-channel coefficient residual of the sliding window corresponding to the current sampling moment.

[0031] By quantifying the degree of cross-channel coupling between temperature and pressure through the Pearson correlation coefficient and establishing a reference benchmark based on historical health period data, it is possible to effectively capture subtle changes in the operating status of the thermal system and improve the accuracy of early warning.

[0032] In one embodiment of the present invention, the specific steps of S104 include: S301, within a preset historical health period, composing energy conservation residuals, thermal inertia prediction residuals, and cross-channel coefficient residuals at multiple sampling moments into an energy conservation residual sequence, a thermal inertia prediction residual sequence, and a cross-channel coefficient residual sequence, respectively, in chronological order; S302: Statistically process the energy conservation residual sequence, the thermal inertia prediction residual sequence, and the cross-channel coefficient residual sequence, respectively, to obtain the 99th percentile of each sequence, and use it as the healthy period threshold corresponding to the energy conservation residual, the thermal inertia prediction residual, and the cross-channel coefficient residual. Specifically, sort the values ​​in each sequence from small to large, and take the value at the 99th percentile as the healthy period threshold of the corresponding residual. S303: Ratio calculations are performed on the energy conservation residuals, thermal inertia prediction residuals, and cross-channel coefficient residuals at the sampling moments within the first preset time period with the corresponding healthy period thresholds to obtain normalized energy conservation residuals, thermal inertia prediction residuals, and cross-channel coefficient residuals. The effects of dimensional differences on the fusion calculation are eliminated by uniformly mapping residuals of different dimensions to a range of 0 to 1. S304 , performing weighted accumulation of the normalized energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual according to pre-set weights to obtain an anomaly score; wherein the energy conservation residual is given a higher weight.

[0033] By dynamically determining the residual threshold based on historical health period data, the subjectivity of the traditional fixed threshold is avoided, making the threshold more in line with the actual operating characteristics of the system. Normalization processing eliminates dimensional differences and ensures the comparability of multi-dimensional residuals. The weighted fusion algorithm can highlight the role of key monitoring indicators, such as energy conservation residuals, so that the anomaly score can comprehensively reflect the overall status.

[0034] In one embodiment of the present invention, the device nodes in the pre-defined knowledge graph correspond to individual devices or components in an actual thermal system and are the primary objects where faults occur. Each device node is associated with various attributes, such as operating parameters like pressure and temperature, as well as device status information. These attributes serve as crucial indicators for determining whether the device is operating properly.

[0035] Fault types are categorized descriptions of possible equipment faults within the knowledge graph, such as energy imbalance, temperature deviation, coupling failure, and overheating and high-voltage faults. The knowledge graph clearly defines the relationship between different fault types and device nodes and their associated attributes.

[0036] Based on the device node mounting attributes in the preset knowledge graph, different types of fault warning triples are generated through multi-conditional logical judgment. Each triple consists of the device node, the warning, and the fault type, which is used to accurately describe the fault subject, whether a warning was issued, and the nature of the fault. The specific reasoning rules are as follows: When the energy conservation residual of a device node is greater than or equal to the preset energy imbalance threshold, and the anomaly scores of the three most recent consecutive sampling moments are all greater than or equal to 1, a fault warning triplet with the fault type being energy imbalance fault is generated; When the thermal inertia prediction residual of the device node is greater than or equal to the preset thermal inertia threshold, and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet is generated with the fault type being temperature trend deviation fault; When the cross-channel coefficient residual of the device node is greater than or equal to the preset coupling threshold and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet with the fault type being coupling failure is generated; When a fault warning triplet with the fault type of energy imbalance fault already exists in the preset knowledge graph and the device node has a high-voltage state attribute, a fault warning triplet with the fault type of overheating high-voltage fault is generated, where the high-voltage state means that the pressure after preprocessing is greater than or equal to the preset safety pressure threshold.

[0037] In one embodiment of the present invention, the preset warning strategy includes: When a single fault warning triplet is generated, a yellow prompt will flash at the corresponding position on the local monitoring interface; When a fault warning triplet of the same fault type is repeatedly generated within three consecutive sampling moments, the operation and maintenance personnel will be notified via SMS or email; When a fault warning triplet of an overheating and high-voltage fault type is generated, a red emergency shutdown alarm is triggered and on-site personnel are notified through an audible and visual alarm.

[0038] The early warning strategy achieves accurate push and efficient processing of alarm information through a hierarchical response mechanism, which can improve operation and maintenance efficiency and system reliability, while reducing labor costs and potential economic losses.

[0039] An embodiment of the present invention also provides a thermal signal analysis and early warning system based on a knowledge graph, including: a data acquisition and preprocessing module, configured to acquire thermal signal data from the thermal system within a first preset time period and perform preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and wherein the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; The energy conservation deviation evaluation module is used to obtain the enthalpy value and mass flow rate based on the pre-processed thermal signal data, and then determine the inlet energy flow and outlet energy flow, while evaluating the deviation of energy conservation to form the energy conservation residual; A multi-dimensional residual calculation module is used to evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data and generate thermal inertia prediction residuals. It also evaluates the channel coupling degree based on the preprocessed temperature and pressure to obtain the cross-channel coefficient and further determine the cross-channel coefficient residual; The residual fusion module is used to normalize the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual, and fuse them according to the preset weights to generate anomaly scores; The reasoning and warning module is used to mount the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score as attributes to the corresponding device nodes in the preset knowledge graph, and perform knowledge reasoning on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, and fault type; and based on the fault warning triplet, output a warning signal according to the preset warning strategy.

[0040] It should be noted that the intervals and thresholds are set for ease of comparison. The threshold size depends on the amount of sample data and the cardinality set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless numerical calculations. These formulas are derived from software simulations of the most recent real-world conditions using large amounts of data. The preset parameters in these formulas are set by those skilled in the art based on actual conditions.

[0041] The above describes the embodiments of the present invention, but the present invention is not limited to the above specific implementation methods. The above specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A thermal signal analysis and early warning method based on knowledge graph, characterized in that: The following steps are involved: S101, collecting thermal signal data from a thermal system within a first preset time period and performing preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; S102, obtaining enthalpy and mass flow rate based on the preprocessed thermal signal data, thereby determining inlet energy flow and outlet energy flow, and simultaneously evaluating the deviation of energy conservation to form an energy conservation residual; S103: Evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data, and generate a thermal inertia prediction residual. Evaluate the channel coupling degree based on the preprocessed temperature and pressure to obtain a cross-channel coefficient, and then determine the cross-channel coefficient residual. S104, normalizing the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual respectively, and fusing them according to preset weights to generate an anomaly score; S105, the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score are mounted as attributes to the corresponding device node in the preset knowledge graph, and knowledge reasoning is performed on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, fault type; and based on the fault warning triplet, a warning signal is output according to the preset warning strategy.

2. The thermal signal analysis and early warning method based on knowledge graph according to claim 1 is characterized in that: The denoising adopts sliding window moving average filtering; In the baseline drift correction, the baseline mean of each parameter is obtained by calculating the average value of temperature, pressure, flow and power within a preset historical healthy period, and the difference between each parameter of the thermal signal data and the baseline mean of the corresponding parameter is calculated to obtain the corrected parameter.

3. The thermal signal analysis and early warning method based on knowledge graph according to claim 2 is characterized in that: Obtain enthalpy and mass flow rate based on pre-processed thermal signal data, including: The enthalpy value is obtained by multiplying the difference between the pre-processed temperature and the preset saturation temperature by the preset pressure-specific heat capacity, and then adding the product to the saturation enthalpy corresponding to the preset saturation temperature; The fluid density is obtained by using the pre-processed temperature and pressure as query conditions in the preset fluid state table, and the mass flow rate is obtained by multiplying the fluid density by the pre-processed flow rate.

4. The thermal signal analysis and early warning method based on knowledge graph according to claim 1 is characterized in that: In multiple inlet channels, the enthalpy value corresponding to each inlet channel and the corresponding mass flow rate are multiplied and accumulated in sequence to obtain the inlet energy flow; In multiple outlet channels, the enthalpy value corresponding to each outlet channel and the corresponding mass flow rate are multiplied and added in sequence to obtain the outlet energy flow.

5. The thermal signal analysis and early warning method based on knowledge graph according to claim 4 is characterized in that: The system energy storage change is determined based on the preset system heat capacity and the preprocessed temperature difference between adjacent sampling moments. The difference between the inlet energy flow and the outlet energy flow is compared with the system energy storage change, and the absolute difference between the two is used as the energy conservation residual at the current sampling moment.

6. The thermal signal analysis and early warning method based on knowledge graph according to claim 1 is characterized in that: At adjacent sampling moments, the predicted temperature at the current sampling moment is calculated through the thermal inertia model based on the preprocessed temperature at the previous sampling moment, the inlet energy flow at the previous sampling moment, the outlet energy flow at the previous sampling moment, and the preset system heat capacity. The predicted temperature is then compared with the preprocessed temperature at the current sampling moment, and the absolute difference between the two is taken as the thermal inertia prediction residual at the current sampling moment.

7. The thermal signal analysis and early warning method based on knowledge graph according to claim 2 is characterized in that: The determination of the cross-channel coefficient residual includes: S201, within a preset sliding window period, based on the pre-processed temperature and pressure, calculate the Pearson correlation coefficient between the two within the sliding window and use it as the cross-channel coefficient; S202, within a preset historical health period, according to a preset sliding window time period, the Pearson correlation coefficients of multiple sliding windows are obtained using the same method as S201, and the average is calculated to obtain a reference cross-channel coefficient; S203 : Compare the cross-channel coefficient with the reference cross-channel coefficient, and use the absolute difference between the two as the cross-channel coefficient residual of the sliding window corresponding to the current sampling moment.

8. The thermal signal analysis and early warning method based on knowledge graph according to claim 2 is characterized in that: The specific steps of S104 include: S301, within a preset historical health period, composing energy conservation residuals, thermal inertia prediction residuals, and cross-channel coefficient residuals at multiple sampling moments into an energy conservation residual sequence, a thermal inertia prediction residual sequence, and a cross-channel coefficient residual sequence, respectively, in chronological order; S302, statistically processing the energy conservation residual sequence, thermal inertia prediction residual sequence, and cross-channel coefficient residual sequence respectively to obtain the 99th percentile of each sequence, and use it as the healthy period threshold corresponding to the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual; S303: performing a ratio operation on the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual at the sampling time within the first preset time period and the corresponding healthy period threshold value to obtain normalized energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual; S304 , performing weighted accumulation on the normalized energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual according to pre-set weights to obtain an anomaly score.

9. The thermal signal analysis and early warning method based on knowledge graph according to claim 1 is characterized in that: Perform knowledge reasoning on the mount attributes to generate fault warning triples, including: When the energy conservation residual of a device node is greater than or equal to the preset energy imbalance threshold, and the anomaly scores of the three most recent consecutive sampling moments are all greater than or equal to 1, a fault warning triplet with the fault type being energy imbalance fault is generated; When the thermal inertia prediction residual of the device node is greater than or equal to the preset thermal inertia threshold, and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet is generated with the fault type being temperature trend deviation fault; When the cross-channel coefficient residual of the device node is greater than or equal to the preset coupling threshold and the anomaly score at the current sampling moment is greater than or equal to 1, a fault warning triplet with the fault type being coupling failure is generated; When a fault warning triplet with the fault type of energy imbalance fault already exists in the preset knowledge graph and the device node has a high-voltage state attribute, a fault warning triplet with the fault type of overheating high-voltage fault is generated, where the high-voltage state means that the pressure after preprocessing is greater than or equal to the preset safety pressure threshold.

10. The thermal signal analysis and early warning system based on knowledge graph is characterized by: The thermal signal analysis and early warning method based on the knowledge graph according to any one of claims 1 to 9 comprises: a data acquisition and preprocessing module, configured to acquire thermal signal data from the thermal system within a first preset time period and perform preprocessing, wherein the thermal signal data includes temperature, pressure, flow rate, and power, and wherein the preprocessing includes performing denoising and baseline drift correction on the thermal signal data to obtain preprocessed thermal signal data; The energy conservation deviation evaluation module is used to obtain the enthalpy value and mass flow rate based on the pre-processed thermal signal data, and then determine the inlet energy flow and outlet energy flow, while evaluating the deviation of energy conservation to form the energy conservation residual; A multi-dimensional residual calculation module is used to evaluate the temperature change trend according to the thermal inertia model based on the preprocessed thermal signal data and generate thermal inertia prediction residuals. It also evaluates the channel coupling degree based on the preprocessed temperature and pressure to obtain the cross-channel coefficient and further determine the cross-channel coefficient residual; The residual fusion module is used to normalize the energy conservation residual, thermal inertia prediction residual, and cross-channel coefficient residual, and fuse them according to the preset weights to generate anomaly scores; The reasoning and warning module is used to mount the energy conservation residual, thermal inertia prediction residual, cross-channel coefficient residual and anomaly score as attributes to the corresponding device nodes in the preset knowledge graph, and perform knowledge reasoning on the mounted attributes to generate a fault warning triplet, wherein the fault warning triplet includes: device node, occurrence warning, and fault type; and based on the fault warning triplet, output a warning signal according to the preset warning strategy.

Images