A method for fault alarm of heat tracing cable based on comparison of ambient temperature

Abstract

This invention discloses a fault alarm method for heat tracing cables based on ambient temperature comparison, belonging to the field of cable fault early warning technology. It includes continuously and synchronously acquiring the ambient temperature sequence output by an ambient temperature sensor and the cable temperature sequence output by a heat tracing cable surface temperature sensor; identifying natural variation segments in the ambient temperature sequence in real time, and distinguishing these segments into heating and cooling segments; the innovation of this invention lies in combining temperature-dependent piecewise linear normalization processing to unify the response evaluation standard across the entire temperature range, overcoming detection bias caused by material temperature characteristics. It uses the overall asymmetry after multi-point fusion, temperature change patterns, long-cycle attenuation trends, fault feature matching results, and extreme operating condition safety strategies as multiple control criteria, changing the previous detection mode of instantaneous comparison of a single parameter, and comprehensively covering explicit faults, latent aging, gradual degradation, sensor deterioration, and extreme operating risks.

Description

Technical Field

[0001] This invention belongs to the field of cable fault early warning technology, specifically a fault alarm method for heat tracing cables based on ambient temperature comparison. Background Technology

[0002] As a core component for freeze protection and insulation of industrial pipelines and equipment, the timeliness and stability of fault alarms of heat tracing cables directly determine the operational safety of industrial pipelines. Currently, there are two main detection methods: electrical parameter detection and single temperature difference comparison. Electrical detection can only identify hard faults such as short circuits and open circuits, making it difficult to detect progressive, latent defects such as insulation aging and deterioration of heat transfer performance. Conventional temperature comparison methods rely solely on instantaneous temperature difference thresholds for judgment, failing to consider dynamic response differences during temperature rises and falls and parameter shifts caused by temperature. This makes them susceptible to interference from diurnal environmental fluctuations and seasonal temperature changes, resulting in significant issues of misjudgment and missed detection.

[0003] In actual industrial sites, the operating environment of heat tracing cables varies greatly, and extremely low temperature freezing scenarios are common. Conventional active detection logic does not take into account the safety constraints of low temperature and high-risk operating conditions. Blindly applying power disturbances can easily exacerbate the risk of pipeline freezing and cracking or cause insufficient heating.

[0004] Meanwhile, existing technologies lack a safety decision-making mechanism based on operating conditions, failing to distinguish between conventional constant-temperature scenarios and extremely low-temperature high-risk scenarios. Detection actions and safety protection actions cannot be linked, resulting in significant shortcomings in overall operational protection and adaptability to operating conditions. Furthermore, traditional detection modes mostly rely on fixed thresholds for passive fault determination, lacking the ability to accumulate fault data, autonomously learn features, and transfer rules across scenarios. A single threshold is insufficient to adapt to the operational differences of equipment in different environments and with varying degrees of aging, making it impossible to achieve early fault prediction and trend-based risk warnings. Global adaptability and intelligence levels are limited. Summary of the Invention

[0005] The purpose of this invention is to provide a method for alarming faults in heat tracing cables based on ambient temperature comparison, so as to solve the problems mentioned in the background art.

[0006] A method for fault alarm of heat tracing cable based on ambient temperature comparison, comprising: A1. Continuously and synchronously acquire the ambient temperature sequence output by the ambient temperature sensor and the cable temperature sequence output by the surface temperature sensor of the heating cable; A2. Real-time identification of natural variation segments in the ambient temperature sequence, and differentiation of the natural variation segments into warming segments and cooling segments; A3. Calculate the dynamic response characteristics of the cable temperature sequence relative to the ambient temperature sequence in the heating and cooling sections, respectively. The calculation includes temperature-dependent normalization of the response delay time and the response decay coefficient. The temperature-dependent normalization adopts a piecewise linear mapping based on the rated operating temperature range of the heat tracing cable to map the response delay time and response decay coefficient calculated under different ambient temperatures to a unified dimensionless scale. The dynamic response characteristics include the normalized response delay time and response decay coefficient. A4. Compare the asymmetry between the dynamic response characteristics of the heating section and the cooling section. If the asymmetry is less than a preset asymmetry threshold, determine that the heat tracing cable has failed and output an alarm signal. A5. If a complete heating and cooling segment is not identified within a preset time (e.g., 24 hours, which can be adjusted according to the average frequency of occurrence of historical natural variation segments), an electrothermal disturbance of a predetermined amplitude is actively applied to the heating cable. This disturbance is used as an excitation signal to form a simulated heating and cooling segment, and the process returns to step A3 to calculate the dynamic response characteristics of the simulated heating and cooling segments.

[0007] A complete fault alarm method for heat tracing cables is provided, comprising the following basic steps: Simultaneously acquiring ambient and cable temperature sequences, automatically identifying natural temperature change segments and distinguishing between heating / cooling segments, calculating the dynamic response characteristics (response delay time and response attenuation coefficient) after temperature dependence normalization, comparing the asymmetry of heating / cooling segments, and determining a fault when the temperature falls below a threshold; furthermore, when there is no complete natural temperature change segment for an extended period, actively applying electrothermal disturbances to simulate the heating / cooling process, ensuring uninterrupted detection. This scheme overcomes the problem of traditional threshold methods ignoring dynamic response differences and temperature influences, achieving all-weather fault monitoring that combines passive and active methods.

[0008] In some possible implementations, when distinguishing between the heating and cooling phases in step A2, the transition points within the natural temperature change phase are further identified. These transition points refer to the moments when the ambient temperature changes from rising to falling or from falling to rising. Step A2 further includes: The temperature difference between adjacent moments in an ambient temperature sequence is detected, and the location of the transition point is determined based on the positive or negative change of the difference. Using the transition point as the boundary, the temperature sequence before the transition point is assigned to the end of the previous natural change segment, and the temperature sequence after the transition point is assigned to the beginning of the next natural change segment, so that the heating segment and the cooling segment share the same transition point. Discard natural variation segments whose duration is less than a preset minimum duration.

[0009] The transition point is determined by detecting the positive or negative change in the temperature difference between adjacent time points, and the heating and cooling segments are seamlessly connected by sharing the transition point, while discarding natural change segments with too short a duration. Its advantages include: eliminating boundary gaps or overlaps caused by segmentation, avoiding contamination of diagnostics by short-term invalid temperature fluctuations (such as light winds or instantaneous sunlight interference), and making the division of heating / cooling intervals used for feature comparison more consistent with the natural evolution of ambient temperature, thereby improving the reliability of subsequent asymmetry calculations.

[0010] In some possible implementations, when calculating the response delay time in step A3, the transition point is used as the time origin, and step A3 further includes: Taking the conversion point as the origin, backtracking for a first preset time period, within the first preset time period, calculate the first response delay of the cable temperature sequence relative to the ambient temperature sequence, as the response delay of the heating segment; Taking the conversion point as the origin, backtracking for a second preset time period, within the second preset time period, calculate the second response delay of the cable temperature sequence relative to the ambient temperature sequence, as the response delay of the cooling segment; If the calculated result of the first response delay or the second response delay exceeds a preset reasonable range, the natural change segment is marked as an invalid segment and the system waits for the next transition point.

[0011] Using the transition point as the time origin, the response delays for the heating and cooling phases are calculated by backtracking for fixed durations before and after the transition point, and a reasonable response delay range is set for validity verification. The benefits are: it provides a unified time reference for calculating the heating and cooling response delays, ensuring the comparability of parameters under the two operating conditions; simultaneously, the validity verification automatically eliminates abnormal natural variation segments (such as out-of-range delays caused by sensor jitter or local faults), preventing invalid data from entering the fault judgment process, thus improving the system's anti-interference capability and judgment accuracy.

[0012] In some possible implementations, the asymmetry in step A4 is obtained by comparing the dynamic response characteristics of the heating and cooling segments point by point within the same natural variation segment. Step A4 further includes: Arrange the ambient temperature values ​​in the heating and cooling sections in ascending order to form a temperature axis; Multiple discrete temperature sampling points are selected on the temperature axis, and the response delay time and response decay coefficient corresponding to each temperature sampling point in the heating segment and the cooling segment are extracted respectively. For each temperature sampling point, calculate the ratio of the absolute value of the difference between the response delay times of the heating and cooling segments to the sum of the two, and the ratio of the absolute value of the difference between the response attenuation coefficients to the sum of the two. Add the two ratios to obtain the local asymmetry at that temperature sampling point. The average value of the local asymmetry at all temperature sampling points is taken as the overall asymmetry of the natural variation segment.

[0013] Within the same natural temperature range, the ambient temperature values ​​of the heating and cooling phases are sorted to construct a temperature axis. Discrete temperature sampling points are selected and paired point-by-point with response delay time and response decay coefficient. The local asymmetry at each point is calculated, and the average value is taken as the overall asymmetry. Its advantages are: it achieves refined quantification of bidirectional response differences across the entire temperature range, overcomes the shortcomings of traditional overall mean comparison which may mask fault characteristics in local temperature ranges, and makes the asymmetry index more representative and sensitive.

[0014] In some possible implementations, after the fourth step of step A4, the temperature dependence of the asymmetry is further determined, including the following steps: Analyze the trend of local asymmetry at each temperature sampling point as the temperature decreases; If the local asymmetry increases monotonically as the temperature decreases, the heat tracing cable is determined to be in normal working condition. If the local asymmetry tends to zero or remains unchanged as the temperature decreases, the heat tracing cable is determined to be in a failure state. If the local asymmetry fluctuates non-monotonicly as the temperature decreases, it is determined that the ambient temperature sensor or the cable temperature sensor is faulty, and a sensor self-test alarm signal is output.

[0015] By utilizing the physical law that the thermal resistance of heat tracing cable materials naturally increases at low temperatures, and using the temperature dependence of asymmetry as an additional criterion, it is possible to effectively distinguish between cable failure, normal aging temperature-sensitive characteristics, and temperature sensor hardware failure, thus realizing fault classification diagnosis and online self-testing of sensors.

[0016] In some possible implementations, the following steps are also included between step A4 and step A5: Detect the asymmetry sequence calculated from multiple consecutive natural variation segments preceding the current natural variation segment; Determine whether the asymmetry sequence exhibits a gradually decreasing trend; If the cable exhibits a gradual decay trend and the asymmetry of the current natural change segment is lower than a preset absolute threshold, then the heating cable is determined to be in a state of gradual degradation. Based on the decay rate of the asymmetry sequence, calculate the number of remaining natural variation segments required for the asymmetry to decrease to the asymmetry threshold; Based on the average occurrence frequency of historical natural change segments, the remaining number of natural change segments is converted into remaining running time, and a maintenance warning signal containing the remaining running time is output.

[0017] In some possible implementations, when the electrothermal disturbance is actively applied in step A5, the specific method of applying the electrothermal disturbance includes the following steps: The system detects the rate of change of the current ambient temperature. If the absolute value of the rate of change is less than a preset stability threshold, the environment is determined to be in a constant temperature state. Under constant temperature conditions, a first power pulse and a second power pulse are applied to the heat tracing cable in sequence. The first power pulse is a positive heating pulse with 50% to 80% of the rated power, and the second power pulse is a cooling pulse with the output power reduced to below 10% of the rated power (including zero power). During the second pulse, the cable temperature gradually decreases due to natural heat dissipation. A preset interval time is set between the two pulses, and the interval time is configured to be greater than half of the thermal time constant of the heat tracing cable and less than the thermal time constant. Record the cable temperature response curves to the first power pulse and the second power pulse, and use these response curves as data for the simulated heating and cooling phases.

[0018] In some possible implementations, after applying the double-pulse sequence in step A5, the method further includes a step of directly determining the asymmetry using the simulated data: Calculate the asymmetry between the cable temperature response delay time to the first power pulse and the response delay time to the second power pulse; The calculated asymmetry is compared with the asymmetry threshold in step A4, so that the active perturbation mode and the natural change mode share the same set of judgment logic. If an effective asymmetry cannot be obtained after applying a double pulse sequence multiple times, it is determined that the ambient temperature sensor or cable temperature sensor is not sensitive enough, and a sensor maintenance alarm signal is output.

[0019] In some possible implementations, if the complete heating and cooling phases are not identified within a preset time in step A5, a safety decision-making step is also included before actively applying electrothermal disturbance: Detect whether the current ambient temperature is lower than a preset extremely low temperature threshold; If the current ambient temperature is lower than the extremely low temperature threshold, then it is further detected whether the rate of change of the current ambient temperature is positive and exceeds a preset safe heating rate. If the ambient temperature is below the extremely low threshold and is not in a rapid heating state, the current working condition is recorded as a high-risk freezing state, the predetermined amplitude of the electrothermal disturbance is limited to zero and the preset time is shortened (for example, the preset time is shortened to half of the original value, i.e., 12 hours), a forced alarm signal is directly output, and the heat tracing cable is kept continuously powered. If the ambient temperature is not lower than the extremely low threshold, or if it is lower than the extremely low threshold but is in a state of rapid heating, then an electrothermal disturbance is allowed to be applied, and the predetermined amplitude is limited to a preset low power amplitude (e.g., 30% of the rated power), and the disturbance duration is limited to a preset short duration (e.g., 15 minutes).

[0020] It should be noted that the low power amplitude is less than 50% of the rated power.

[0021] In some possible implementations, a fault feature learning and transfer step is also included: Once a fault is determined in the heat tracing cable, the ambient temperature sequence and cable temperature sequence in the last complete natural change segment before the fault occurred are extracted as fault feature samples. The fault feature samples are compared differentially with multiple normal feature samples collected during the historical normal operation of the same heat tracing cable to extract the differential feature curves. The differential feature curves are classified and stored according to fault type to form a fault feature library; In subsequent operation, the dynamic response characteristics of the current natural change segment are calculated in real time, and the current characteristics are matched with the characteristics in the fault feature library; If the matching degree exceeds a preset similarity threshold, an early warning signal for this type of fault will be output in advance. The successfully matched fault characteristics and corresponding environmental conditions are uploaded to the cloud for updating the global fault judgment rules.

[0022] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention establishes a high-quality data foundation by continuously and synchronously acquiring ambient temperature and cable temperature sequences, combined with automatic identification of temperature rise / fall during natural temperature variations, precise location of transition points, and elimination of invalid fluctuations. It introduces piecewise linear normalization processing based on temperature dependence to overcome detection biases caused by shifts in the thermal properties of the heating cable material across different temperature ranges, ensuring uniform comparability of response delay time and response attenuation coefficient across the entire temperature range. Using the overall asymmetry of the dynamic response characteristics of temperature rise and fall within the same natural temperature variation segment as the core criterion, and combining point-by-point temperature pairing, local asymmetry fusion, and temperature dependence trend analysis, it achieves accurate differentiation between overt faults, sensor anomalies, and normal material temperature-sensitive characteristics.

[0023] Furthermore, by analyzing the trend of long-period asymmetry sequences, the gradual degradation process can be captured and the remaining operating time can be quantified, supporting preventive maintenance. In constant temperature or extreme low temperature scenarios without natural temperature fluctuations, the hierarchical active disturbance strategy based on safety decisions (including dual-pulse excitation, power duration limitation and high-risk interlocking) takes into account both detection reliability and operational safety. Finally, by leveraging fault feature self-learning and cloud migration mechanisms, fault judgment rules are continuously optimized, enhancing intelligent early warning capabilities across all scenarios and the entire lifecycle. In summary, this invention changes the previous detection mode of instantaneous comparison of a single parameter, comprehensively covering explicit faults, latent aging, gradual degradation, sensor deterioration, and extreme operational risks, thereby improving the accuracy of fault alarms for heat tracing cables. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0025] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 This application provides a method for fault alarm of heat tracing cable based on ambient temperature comparison, including: A1. Continuously and synchronously acquire the ambient temperature sequence output by the ambient temperature sensor and the cable temperature sequence output by the surface temperature sensor of the heating cable. It can be understood that the ambient temperature sensor is used to collect temperature data of the external environment where the heating cable is located, and the cable temperature sensor is used to collect temperature data of the surface of the heating cable. The sampling frequency is 1Hz, and the ambient and cable temperature sensors are acquired synchronously, strictly ensuring that the timestamps of the two data streams are completely aligned.

[0027] The ambient temperature sequence refers to the set of ambient temperature values ​​arranged sequentially according to the acquisition time, while the cable temperature sequence refers to the set of cable surface temperature values ​​arranged sequentially on the same time axis. The two sets of sequence data are completely synchronized in length and acquisition time.

[0028] To ensure data acquisition accuracy, the ambient temperature sensor should be installed 10 to 15 centimeters away from the heating cable and in a location unobstructed by heat sources. The cable temperature sensor should be snap-fitted to the cable surface, ensuring a tight fit with the insulation layer. The measurement range of all sensors is set to -40 degrees Celsius to 120 degrees Celsius, with a measurement error not exceeding ±0.5 degrees Celsius. During data acquisition, the data acquisition module receives analog signals from both types of sensors in real time. After analog-to-digital conversion, these signals are stored in a local database, with the storage interval consistent with the sampling frequency.

[0029] In industrial settings, there is often a misalignment in the timing of data acquisition between ambient temperature and cable surface temperature. Traditional detection methods typically employ independent asynchronous sampling, resulting in a lack of a unified time reference for both types of temperature data. However, the heat conduction of heat tracing cables inherently possesses a time-dependent physical characteristic, and this timing misalignment directly disrupts the correspondence in the thermal response process. Based on this fundamental principle of heat transfer, this solution adopts a dual-end continuous synchronous sequence acquisition mode, establishing a time-by-time correspondence between the ambient temperature change process and the cable temperature follow-up process.

[0030] By using a synchronous time-series data acquisition method, the temporal correlation characteristics of heat transfer can be fully preserved, avoiding the feature fragmentation problem caused by asynchronous sampling. This provides complete and physical conduction logic-aligned raw data for the subsequent refined analysis of dynamic response delay and attenuation patterns. This time-series collaborative acquisition approach differs from the industry-standard random single-point temperature measurement method and has a non-obvious monitoring foundation construction logic.

[0031] A2. Identify natural temperature variation segments in the ambient temperature sequence in real time, and distinguish between warming and cooling segments.

[0032] It should be noted that the natural change segment refers to a temperature data segment in which the ambient temperature continuously exhibits a single changing trend without human intervention. Its determination requires consideration of both the magnitude and duration of the temperature change. The criteria for determining the warming segment are an ambient temperature rise rate ≥ 0.1 degrees Celsius per minute and a duration ≥ 10 minutes; the criteria for determining the cooling segment are an ambient temperature drop rate ≤ -0.1 degrees Celsius per minute and a duration ≥ 10 minutes.

[0033] In one implementation of this step, a sliding window method is used for identification. The window length is set to 10 data points. The ambient temperature sequence is traversed sequentially, and the average temperature change rate within each window is calculated. If the average rate meets the aforementioned threshold condition, the time period corresponding to that window is marked as the corresponding natural change segment. Furthermore, to avoid interference from short-term fluctuations, the ambient temperature sequence is processed by moving average filtering before calculating the temperature change rate. The smoothed temperature data is then used for change segment identification, which can effectively improve the stability of the identification results.

[0034] In step A2, when distinguishing between the heating and cooling phases, the transition points within the natural temperature change phases are further identified. A transition point refers to the moment when the ambient temperature changes from rising to falling or from falling to rising. Step A2 further includes: The temperature difference between adjacent moments in the ambient temperature sequence is detected, and the location of the transition point is determined based on the positive or negative change of the difference. Using the transition point as the boundary, the temperature sequence before the transition point is assigned to the end of the previous natural change segment, and the temperature sequence after the transition point is assigned to the beginning of the next natural change segment, so that the heating segment and the cooling segment share the same transition point. Natural change segments with a duration less than a preset minimum duration are discarded.

[0035] It should be understood that diurnal fluctuations in ambient temperature and climate changes can create trend reversal points, which are the critical moments when temperature rises and falls switch between each other. These critical points are easily ignored by conventional segmentation algorithms, resulting in blurred boundaries between the two change intervals.

[0036] In this step, the temperature difference between adjacent time points refers to the value obtained by calculating the difference between the ambient temperatures of two sampling time points. When the difference changes from a positive value to a negative value, it means that the overall temperature trend has switched from rising to falling. When the difference changes from a negative value to a positive value, it means that the overall temperature trend has switched from falling to rising. By relying on the inverse change of the difference, the time point corresponding to the trend switch can be accurately located, thereby determining the location of the transition point.

[0037] Understandably, using a unique transition point as the common boundary between two change intervals can eliminate the blank intervals caused by segmentation, prevent data at trend switching points from being repeatedly removed or counted, and ensure that the time intervals of the warming and cooling phases are closely connected, with no data omissions or interval overlaps throughout the process.

[0038] In this embodiment, the minimum duration is uniformly set to 5 minutes. For the identified natural change segments, as long as the overall continuous duration is less than this fixed duration, they are directly determined as invalid fluctuation segments and discarded, thereby filtering out invalid temperature fluctuation interference caused by short-term breezes and instantaneous changes in sunlight.

[0039] Conventional segmentation methods mostly employ rigid time truncation or fixed boundary division, ignoring the objective laws of natural temperature transitions. This can easily lead to problems such as misalignment of temperature rise and fall intervals and the mixing of short-period invalid fluctuations into valid data, directly interfering with the accuracy of subsequent bidirectional response feature comparisons.

[0040] Based on the objective physical law of alternating changes in ambient temperature trends, this invention adds a transition point identification and interval linkage division mechanism. It accurately captures the moment of trend turning point by relying on the positive and negative changes in adjacent temperature differences, and uses shared transition points to achieve seamless connection between two operating conditions. At the same time, it is combined with a strategy to eliminate short-term invalid segments.

[0041] Through a multi-layered design that includes precise location of turning points, collaborative division of interval boundaries, and active filtering of invalid fluctuations, the boundaries of temperature rise and fall segments can be standardized, short-term temperature disturbances with no diagnostic value can be eliminated, and the divided change intervals can better reflect the real evolution of natural ambient temperature. This breaks through the conventional technical limitations of extensive segmentation identification and provides regular and reliable interval data support for the subsequent accurate comparison of bidirectional asymmetric features.

[0042] Under normal operating conditions, the heat conduction path and thermal resistance of heating cables naturally differ during ambient heating and cooling processes. This is an inherent physical property of heat exchange in solid materials. Most existing conventional testing methods directly use mixed temperature data from all time periods for unified analysis, ignoring the differences in thermal characteristics under the two operating conditions of heating and cooling. Forcibly mixing the calculations will smooth out the differentiated response signals caused by cable faults.

[0043] Based on the aforementioned objective laws of heat exchange, this invention separately identifies two independent change intervals: the heating segment and the cooling segment, enabling partitioned analysis of these two operating conditions. By segmenting the natural temperature change trend, it is possible to isolate the mutual interference between different heat exchange conditions, accurately pinpoint the independent response patterns under two opposing temperature changes, and further enhance the rigor of interval division by combining conversion point correction and short segment elimination optimization. This breaks the conventional inertia of industry-wide unified analysis, providing rigorous thermal theoretical support for the subsequent fault judgment logic based on bidirectional feature comparison.

[0044] A3. Calculate the dynamic response characteristics of the cable temperature sequence relative to the ambient temperature sequence in the heating and cooling stages respectively. The calculation includes temperature-dependent normalization of the response delay time and response decay coefficient. The temperature-dependent normalization adopts a piecewise linear mapping based on the rated operating temperature range of the heat tracing cable, mapping the response delay time and response decay coefficient calculated under different ambient temperatures to a unified dimensionless scale. The dynamic response characteristics include the normalized response delay time and response decay coefficient.

[0045] Temperature dependence normalization employs a piecewise linear mapping based on the rated operating temperature range of the heat tracing cable. This maps the response delay time and response decay coefficient calculated under different ambient temperatures to a unified dimensionless scale. The dynamic response characteristics include the normalized response delay time and response decay coefficient. As an implementable example, for a PTC self-regulating heat tracing cable with a rated power of 30W / m, the rated operating temperature range of -20°C to 80°C is divided into three segments: -20°C to 20°C, 20°C to 50°C, and 50°C to 80°C. Each segment uses the following linear mapping formula: The first segment is defined as y = 0.5x + 0.2, the second segment as y = 0.8x + 0.1, and the third segment as y = 1.2x - 0.3, where x is the original response delay time (minutes) or the original response attenuation coefficient (dimensionless), and y is the normalized dimensionless scale value. In practical applications, the mapping coefficients of each segment can be calibrated using the following method: at at least three typical temperature points (e.g., -10°C, 25°C, 65°C), measure the response delay time and response attenuation coefficient of the normal heat tracing cable, and fit the linear function coefficients of each segment using the least squares method. Different cable models should be recalibrated.

[0046] It should be understood that response delay time refers to the time difference between when the ambient temperature changes and when the cable temperature begins to change accordingly. The method for determining this is as follows: within the natural change period, when the cumulative change in ambient temperature reaches 2 degrees Celsius, record this time point as the start time of the ambient temperature change. Subsequently, monitor the cable temperature, and when the cumulative change in cable temperature reaches 30% of the cumulative change in ambient temperature, record this time point as the start time of the cable temperature response. The difference between the two times is the response delay time.

[0047] The response attenuation coefficient refers to the proportion of the change in cable temperature with the change in ambient temperature. It is quantified by calculating the ratio of the total change in cable temperature to the total change in ambient temperature within the natural change range. This ratio is the response attenuation coefficient, and its value ranges from 0 to 1. The closer the ratio is to 1, the more sensitive the cable temperature response is.

[0048] In this step, the core of temperature dependence normalization is to eliminate the influence of differences in ambient temperature baselines on response parameters. The rated operating temperature range of the heat tracing cable is set to -20°C to 80°C, and this range is divided into three segments: -20°C to 20°C, 20°C to 50°C, and 50°C to 80°C. Each segment corresponds to a linear mapping rule. In this embodiment, the thermal conductivity of the cable material and the thermal resistance of the insulation layer will naturally shift under different temperature ranges. Piecewise linear mapping can specifically correct the parameter deviations caused by these differences in physical properties, uniformly converting the physical parameters of the dispersed ranges into dimensionless values ​​that can be compared laterally.

[0049] In step A3, when calculating the response delay time, the transition point is taken as the time origin. Step A3 further includes: Taking the transition point as the origin, backtracking backward for a first preset time period, the first response delay of the cable temperature sequence relative to the ambient temperature sequence is calculated within the first preset time period, which is used as the response delay of the heating stage; taking the transition point as the origin, backtracking backward for a second preset time period, the second response delay of the cable temperature sequence relative to the ambient temperature sequence is calculated within the second preset time period, which is used as the response delay of the cooling stage. If the calculated result of the first response delay or the second response delay exceeds the preset reasonable range (1 minute to 15 minutes), the natural change segment is marked as invalid and the system waits for the next transition point. The reasonable range is obtained by conducting no less than 50 natural temperature rise and fall cycles on the same type of heating cable under ambient temperature conditions of -20°C to 40°C and wind speed of 0 to 2 m / s, and statistically analyzing the 95% confidence interval of the response delay time. This range can be recalibrated for different cable models based on their actual thermal response characteristics.

[0050] It should be noted that the transition point, as the critical node for the switching of temperature trends, corresponds to the complete heating and cooling processes before and after it. Using this as the time origin can establish a unified time series reference benchmark, avoiding the response delay calculation deviation caused by the traditional arbitrary selection of intervals.

[0051] In this embodiment, both the first preset duration and the second preset duration are set according to the thermal inertia characteristics of the heat tracing cable and are uniformly fixed at 20 minutes. This duration can cover the complete response process of the cable temperature following the change of ambient temperature, and can also avoid introducing additional temperature interference factors due to excessive duration.

[0052] Specifically, 20 minutes are traced back from the transition point as the origin. This time period corresponds to the complete data of the heating segment before the transition point. The first response delay obtained within this interval is directly used as the core response parameter of the heating segment by using the aforementioned response delay calculation method. 20 minutes are traced back from the transition point as the origin. The second response delay obtained by tracing back to the complete data of the cooling segment after the transition point is used as the core response parameter of the cooling segment.

[0053] This symmetrical interval selection method based on the transition point allows the calculation of response parameters for the heating and cooling cycles to be based on datasets of equal duration and time series, ensuring the comparability of the two parameters.

[0054] Furthermore, the preset reasonable range is set based on statistical data from a large amount of normal heat tracing cable operation. In this embodiment, the reasonable range for response delay is fixed at 1 minute to 15 minutes. When the first response delay is less than 1 minute or greater than 15 minutes, it indicates that the cable temperature response in the heating section is abnormally fast or slow, which may be due to data acquisition interference or a local cable fault. When the second response delay exceeds this range, the cooling segment response data is similarly deemed invalid. Once any response delay exceeds a reasonable range, the system directly marks that group of natural change segments as invalid, discards the calculated response parameters, re-enters the transition point identification process, and waits for the next complete trend switching node.

[0055] Traditional response delay calculations often randomly extract arbitrary intervals within the changing range, lacking a unified benchmark for interval duration and starting position. This results in a lack of a basis for cross-sectional comparison of response parameters across temperature rise and fall periods. Furthermore, the absence of a parameter validity verification mechanism allows abnormal data to be directly used in subsequent judgments, easily leading to false alarms. Based on the objective laws governing the correlation of thermal response time series, this invention uses the transition point as a unified time origin and constructs a symmetrical interval extraction and parameter calculation logic. This ensures that the calculation scenarios and duration standards for the two response delay parameters are completely consistent, thereby guaranteeing the fairness of parameter comparison.

[0056] At the same time, a response delay rationality verification mechanism is added to filter abnormal data by fixing a reasonable range, so as to avoid the interference of invalid intervals on the fault judgment results.

[0057] The insulation and sheathing materials of heat tracing cables exhibit significant temperature sensitivity. Their thermal properties spontaneously shift in high and low temperature environments, resulting in completely different response parameters for the same fault condition under different ambient temperatures. This has long been an inherent technical challenge hindering the improvement of detection accuracy in this field. Traditional detection methods generally use raw parameters for direct comparison, assuming that response parameters across different temperature ranges are directly comparable, which reveals a clear bias in technical understanding.

[0058] Based on the objective mechanism of material thermal properties changing with temperature, this invention introduces a piecewise linear mapping temperature dependence normalization strategy to perform a unified scaling transformation on response delay time and response decay coefficient. Combined with symmetric calculation and validity verification under the conversion point benchmark, the consistency and reliability of parameters are further enhanced.

[0059] By using a differentiated correction design with temperature segmentation normalization, the parameter offset problem caused by material temperature drift can be offset, allowing the bidirectional response characteristics across temperature ranges to have equal comparison conditions. It is not a simple parameter value correction, but rather a optimization of the detection logic from both the underlying material properties and the time series benchmark.

[0060] A4. Compare the asymmetry between the dynamic response characteristics of the heating and cooling sections. If the asymmetry is less than the preset asymmetry threshold, the heat tracing cable is determined to be faulty and an alarm signal is output.

[0061] In one implementation of this step, the asymmetry is calculated by comparing the difference between the response delay and attenuation coefficient of the heating and cooling sections. Let the normalized response delay time of the heating section be t1 and the normalized response attenuation coefficient be k1, and the normalized response delay time of the cooling section be t2 and the normalized response attenuation coefficient be k2. The asymmetry is calculated by separately calculating the absolute difference between the two parameters, then combining the mean to convert it into a relative difference, and finally integrating the results of the two parameters to obtain the final asymmetry value. A preset asymmetry threshold is set to 0.15. When the asymmetry is less than 0.15, a fault is identified and an alarm is triggered. This threshold is derived from statistical analysis of a large amount of test data from normally operating heat tracing cables and meets the judgment requirements of actual industrial application scenarios.

[0062] The preset asymmetry threshold is set to 0.15. The calibration method for this threshold is as follows: Select no fewer than 30 normally operating heat tracing cables. Under standard environmental conditions (ambient temperature 25℃±5℃, windless or light wind conditions), record the asymmetry of no fewer than 50 naturally varying segments for each cable. Calculate the average value μ and standard deviation σ of all asymmetry data, and take μ-3σ as the threshold. Based on the above method, in this embodiment, μ=0.32, σ=0.057, and μ-3σ≈0.15. In practical applications, if the user cannot calibrate it themselves, 0.15 can be directly used as the default threshold.

[0063] The asymmetry in step A4 is obtained by comparing the dynamic response characteristics of the heating and cooling segments point by point within the same natural variation segment. Step A4 further includes: Arrange the ambient temperature values ​​in the heating and cooling sections in ascending order to form a temperature axis; select multiple discrete temperature sampling points on the temperature axis, and extract the response delay time and response decay coefficient corresponding to each temperature sampling point in the heating and cooling sections respectively (using linear interpolation, the response delay time and decay coefficient of the entire heating section calculated in step A3 are mapped to each sampling point according to the temperature change ratio; the same applies to the cooling section). For each temperature sampling point, calculate the ratio of the absolute value of the difference between the response delay times of the heating and cooling segments to their sum, and the ratio of the absolute value of the difference between the response attenuation coefficients to their sum. Add the two ratios to obtain the local asymmetry at that temperature sampling point. Take the average of the local asymmetry at all temperature sampling points as the overall asymmetry of the natural variation segment.

[0064] It should be noted that the thermal conductivity difference of the heat tracing cable will continuously change with the ambient temperature. The bidirectional response difference at different temperature points is not uniform. A single global average comparison will mask the subtle fault characteristics of local temperature ranges and cannot achieve fine identification.

[0065] This invention limits the asymmetry calculation to be carried out within the same complete natural change segment, ensuring that the heating and cooling data come from the same round of temperature fluctuation process, avoiding background interference from different time periods and different environmental conditions, and making the comparison samples have homogeneity and reference value.

[0066] In practice, all effective ambient temperature values ​​for the heating and cooling sections within the same segment are first collected and then arranged in an orderly manner from smallest to largest to construct a continuous and uniform standardized temperature axis, eliminating the comparison deviation caused by the disordered timing of the original data collection.

[0067] Then, multiple discrete temperature sampling points are selected at equal intervals on the complete temperature axis. The number of sampling points is set according to the on-site detection accuracy requirements. Under normal working conditions, selecting eight to twelve sampling points can balance the computational load and recognition accuracy.

[0068] For each fixed temperature sampling point, the normalized response delay time and normalized response decay coefficient under heating and cooling conditions are matched and retrieved respectively, so as to achieve one-to-one pairing of features under the same temperature but different trends.

[0069] For a single temperature sampling point, a relative difference ratio is used for calculation. The sum of parameters is used as the denominator, and the absolute value of the parameter difference is used as the numerator. This eliminates calculation bias caused by the size of the parameter base, providing a unified standard for measuring delay and attenuation parameters of different orders of magnitude. The relative ratios of the two types of parameters at a single point are superimposed to synthesize the local asymmetry at that temperature location, accurately reflecting the degree of difference in temperature rise and fall responses within a local temperature range. Finally, the local asymmetry values ​​corresponding to all discrete sampling points are arithmetically averaged, and the differential information across the entire temperature range is integrated to generate an overall asymmetry value that characterizes the entire temperature fluctuation process. This overall asymmetry value serves as the core quantitative indicator for final fault determination.

[0070] After step A4, the temperature dependence of asymmetry is further determined, including the following steps: analyzing the trend of local asymmetry at each temperature sampling point as the temperature decreases. If the local asymmetry increases monotonically as the temperature decreases, the heat tracing cable is considered to be in normal working condition. If the local asymmetry increases monotonically as the temperature decreases (i.e., the percentage of positive local asymmetry differences between adjacent temperature sampling points exceeds 90%), the heating cable is determined to be in normal working condition. If the local asymmetry tends to zero or remains unchanged as the temperature decreases (i.e., the difference between the maximum and minimum local asymmetry values ​​at each sampling point is less than 0.05), the heating cable is determined to be in a failed state. If the local asymmetry exhibits non-monotonic fluctuations as the temperature decreases (i.e., the number of positive and negative alternations of the difference exceeds half the number of sampling points), the ambient temperature sensor or cable temperature sensor is determined to be faulty.

[0071] The physical principle upon which this judgment is based is as follows: the thermal conductivity of commonly used insulation materials (such as cross-linked polyolefins and fluoroplastics) and PTC heating core layers in heat tracing cables decreases as the temperature decreases, resulting in an increase in the thermal response delay time and a decrease in the response attenuation coefficient of the cable in low-temperature environments; at the same time, the heat exchange during the heating process is inhibited by the external low-temperature environment, while the cooling process is dominated by the cable's own heat capacity. The difference between the two is amplified at low temperatures, manifested as a monotonically increasing local asymmetry as the temperature decreases.

[0072] It should be understood that under normal operating conditions, the insulation material and heat-conducting structure of the heat tracing cable will exhibit a regular change in thermal resistance as the ambient temperature decreases. The bidirectional difference in heat conduction in low-temperature environments will be gradually amplified, naturally forming a monotonic change pattern where local asymmetry gradually increases with decreasing temperature. This pattern is an inherent physical property of the cable material and heat transfer structure. After solving for the overall asymmetry, relying on multiple sets of discrete temperature sampling points established in the early stages, the continuous change trend of local asymmetry under different temperature gradients can be analyzed longitudinally. This allows for the discovery of deep temperature correlation characteristics, overcoming the limitations of relying solely on a single mean value for judgment.

[0073] When the local asymmetry at each sampling point throughout the entire process increases monotonically as the ambient temperature decreases, it indicates that the difference in bidirectional heat conduction during cable temperature rise and fall conforms to the inherent law of material temperature change characteristics. The insulation structure and heat-conducting layer are intact and without deterioration, and it can be directly determined that the overall working condition of the cable is normal.

[0074] When the local asymmetry gradually approaches zero as the temperature decreases or remains at a fixed value for a long period of time, it indicates that the difference in thermal response between the hot and cold sides of the cable has completely disappeared, the heat transfer structure tends to be homogenized, and the insulation aging, moisture corrosion or heat tracing layer deterioration problems have occurred, and the cable as a whole has entered a failure operation state.

[0075] Furthermore, if the local asymmetry decreases monotonically as the temperature decreases (i.e., gradually decreases from a higher value), its physical essence is also that the heat transfer difference tends to disappear, which should be classified as a failure state and can be handled with reference to the criterion of "tending to zero".

[0076] When the local asymmetry fluctuates randomly and non-monotonicly during the temperature reduction process, without a stable trend, it indicates that there is distortion, abnormal delay, or acquisition deviation in the two temperature acquisition data. After ruling out cable body faults, the hardware abnormality of the ambient temperature sensor or cable temperature sensor can be directly located, and the dedicated sensor self-test alarm signal can be triggered simultaneously to achieve fault classification and differentiation.

[0077] Most existing detection methods only focus on the single numerical result of asymmetry, completely ignoring the inherent correlation between asymmetry and temperature gradient changes. This makes it impossible to distinguish between three different types of defects: cable body faults, material degradation, and sensor acquisition anomalies, easily leading to misjudgments of fault types and insufficient targeted maintenance. Based on the temperature dependence characteristics of the thermal parameters of heat tracing cables, this invention adds a temperature-dependent trend analysis mechanism for asymmetry to the overall asymmetry calculation. It utilizes the trend differences of continuous temperature samples from multiple points to construct a hierarchical fault determination logic.

[0078] By differentiating the three types of working conditions—normal trend matching, failure trend identification, and abnormal sensor fluctuation identification—it is possible to upgrade from single numerical judgment to regular feature judgment, effectively distinguishing between equipment failure and data acquisition hardware failure.

[0079] In another implementation, the asymmetry can be achieved by spatial distance conversion, which adapts to the operating requirements of low-computing-power field controllers and ensures the compatibility of the algorithm implementation.

[0080] Under normal operating conditions, the thermal response of a heat tracing cable is inevitably affected by the difference in thermal inertia and thermal resistance, resulting in a reasonable asymmetric characteristic during heating and cooling. However, when the cable experiences latent faults such as insulation aging, localized moisture absorption, or deterioration in heat transfer, the overall heat conduction structure of the cable tends to homogenize. The naturally existing difference in bidirectional response will continue to shrink, and the asymmetric characteristics will gradually weaken. This is an irreversible physical change law in the fault evolution process. Existing technologies generally focus on single indicators such as temperature difference and single resistance parameters for fault diagnosis, completely ignoring the homogenization of bidirectional response caused by the fault, thus forming a fixed, single-dimensional detection mindset.

[0081] Based on the underlying physical changes in the above-mentioned fault evolution, this invention uses the asymmetry of dynamic response to temperature rise and fall as the core criterion, combined with a refined calculation mode of temperature point-by-point pairing and temperature dependence trend analysis, which can accurately capture the global and local differential feature attenuation trend caused by the fault, while realizing online self-testing of the sensor.

[0082] By using a multi-level quantitative comparison method that integrates local feature fusion at multiple points with temperature trend analysis, early hidden faults that cannot be captured by conventional one-way detection and overall mean comparison can be identified. This breaks away from the traditional limitations of single-parameter threshold comparison in this field, thereby effectively avoiding the defects of missed detection, false alarms and ambiguous fault classification in single-index detection.

[0083] It also includes the following steps between steps A4 and A5: detecting the asymmetry sequence calculated in multiple consecutive natural variation segments preceding the current natural variation segment; Determine whether the asymmetry sequence exhibits a gradually decreasing trend (using least squares linear fitting; if the fitting slope is less than -0.005 and the goodness of fit R0 is good, the result is positive).2 If the value is ≥0.6, it is determined to be gradually decaying; if it shows a gradually decaying trend and the asymmetry of the current natural change segment is lower than a preset absolute threshold, it is determined that the heating cable is in a state of gradual degradation. Based on the decay rate of the asymmetry sequence, calculate the number of remaining natural change segments required for the asymmetry to drop to the asymmetry threshold ((remaining number of segments = (current asymmetry - asymmetry threshold) / decay rate)); based on the average occurrence frequency of historical natural change segments, convert the number of remaining natural change segments into the remaining running time ((remaining days = remaining number of segments / average occurrence frequency)), and output a maintenance warning signal containing the remaining running time.

[0084] It should be noted that the insulation aging and material degradation of heat tracing cables are mostly gradual processes that evolve slowly over a long period. The asymmetry value of a single natural change segment can only reflect the instantaneous operating state and is difficult to reflect the long-term performance evolution pattern. This step retrieves historical data from multiple consecutive sets of valid natural change segments before the current time period, arranges them in chronological order to form a complete asymmetry sequence, and relies on long-term continuous data to achieve long-term tracking of degradation trends, making up for the shortcomings of instantaneous detection that cannot predict.

[0085] In practice, the overall asymmetry values ​​corresponding to historical natural change segments are continuously collected and stored. The sequence length is set according to the operation and maintenance requirements. Typically, the most recent eight to twelve consecutive effective natural change segments are selected as analysis samples to ensure the stability and accuracy of trend judgment.

[0086] By comparing the changes in values ​​within a sequence group by group, it is determined whether the overall data exhibits a gradual decline, thus distinguishing between short-term fluctuations caused by instantaneous environmental disturbances and long-term degradation caused by material deterioration. Based on the stable degradation trend of the sequence, the dual condition of the current asymmetry being below a preset absolute threshold is added to avoid misjudgment based on a single trend and accurately pinpoint the progressive degradation state of the cable.

[0087] Furthermore, by statistically analyzing the average decrease in the asymmetry sequence within a unit period, a standardized decay rate is obtained. Combined with the difference between the current value and the limit asymmetry threshold, the number of remaining natural change segments that the parameter needs to undergo to decay to the critical threshold is calculated in reverse.

[0088] Simultaneously, the average daily frequency of natural change segments during long-term operation is statistically analyzed, and a conversion relationship between the number of change segments and the actual running time is established. The abstract remaining cycle number is transformed into an intuitive remaining running time, and finally, a maintenance early warning signal with time reference is generated, providing a quantitative basis for on-site planned maintenance and advance allocation of spare parts.

[0089] Traditional monitoring methods generally adopt a passive protection mode of over-limit alarm, which can only trigger a prompt after the cable performance has completely failed. It completely lacks the ability to track the degradation process, quantify the attenuation rate, and predict the lifespan in advance, and cannot meet the control needs of preventive maintenance of industrial equipment.

[0090] Based on the objective evolution law of slow degradation of cable performance, this invention adds a long-period asymmetry sequence trend analysis mechanism on the basis of instantaneous fault judgment and sensor self-test, and constructs a full-chain predictive logic of trend identification, state classification, life prediction and early warning output.

[0091] By analyzing historical data across multiple periods, numerical fluctuations caused by accidental environmental disturbances can be filtered out, gradual hidden degradation problems can be accurately identified, and the remaining running time can be visualized and predicted by quantitative conversion of decay rate.

[0092] It also includes fault feature learning and transfer steps: Once a fault is determined in the heat tracing cable, the ambient temperature sequence and cable temperature sequence in the last complete natural change segment before the fault occurred are extracted as fault feature samples. The fault feature samples are compared differentially with multiple normal feature samples collected during the historical normal operation of the same heat tracing cable to extract the differential feature curves. Differential feature curves are categorized and stored according to fault type to form a fault feature library. During subsequent operation, the dynamic response characteristics of the current natural variation segment (including the response delay time versus temperature curve and the response attenuation coefficient versus temperature curve) are calculated in real time, and the current characteristics are matched with features in the fault feature library. The matching algorithm employs dynamic time warping. Align the response delay time series of the current natural variation segment with the response delay time series of each fault sample in the fault feature library, calculate the DTW distance, and then use the sigmoid function to map the distance to the [0,1] interval to obtain the similarity. Similarly, DTW matching is performed on the response decay coefficient sequence, and the arithmetic mean of the similarity between the two is taken as the final matching degree; if the matching degree exceeds the preset similarity threshold (default 75%), an early warning signal for this fault type is output in advance; The successfully matched fault characteristics and corresponding environmental conditions are uploaded to the cloud for updating the global fault judgment rules.

[0093] It should be noted that the steady-state operating data before the equipment failure contains implicit features that can reflect material degradation and structural abnormalities. Unlike the temperature response pattern under normal operating conditions, the time series data of the last complete natural change segment before the failure is selected as the sample. This can ensure the timeliness and relevance of the features, avoid the interference of abnormal data after the failure, and accurately pinpoint the original operating pattern in the early stage of the failure.

[0094] By selecting multiple sets of characteristic samples from stable and normal operation within the same equipment's historical period for horizontal differential comparison, common interferences caused by environmental temperature fluctuations, seasonal operating condition changes, and inherent differences in equipment parameters can be offset, while amplifying unique response deviation information under fault conditions. Continuous differential feature curves are generated through point-by-point numerical differential operations, transforming discrete parameter differences into visualized continuous change patterns, enhancing the identification of fault characteristics, and providing a standardized data carrier for subsequent classification, storage, and rapid matching.

[0095] Based on specific fault categories such as insulation aging, abnormal heat dissipation, sensor misalignment, and insufficient heating power, differential characteristic curves are categorized and stored locally, gradually accumulating a dedicated fault feature library adapted to the operating environment of each individual device. This categorized storage model enables accurate differentiation of fault types, avoiding confusion between different defect features. Simultaneously, relying on long-term data accumulation, the number of samples is continuously enriched, gradually improving the coverage and recognition accuracy of the feature library, and building a data foundation for localized self-learning.

[0096] During normal operation, the system continuously collects dynamic response characteristics of each naturally changing segment in real time, simultaneously performs normalization processing and feature formatting adaptation, and conducts similarity matching calculations with various tag features within the fault feature library. By relying on existing historical fault data for real-time comparison and screening, it eliminates the need for rigid judgment based on fixed thresholds, enabling early detection of slowly evolving trend anomalies and compensating for the shortcomings of fixed thresholds in identifying progressive faults.

[0097] The preset similarity threshold is calibrated based on the on-site identification accuracy requirements, and is typically set at 75%, serving as the critical judgment condition for prediction and early warning. When the matching amplitude between real-time features and inventory fault features exceeds the threshold range, the corresponding fault type is directly bound and an early warning signal is issued. This changes the traditional passive mode of alarming only after a fault occurs, enabling defect prediction and early intervention, and reducing the impact of sudden faults on the pipeline system.

[0098] The system synchronously uploads locally validated fault characteristic data, corresponding ambient temperature ranges, load conditions, and other related information to the cloud platform, completing data aggregation and rule iteration across devices and scenarios. Through cloud-based big data aggregation and analysis, it extracts general fault judgment logic and feature matching rules, which are then distributed back to each local control terminal. This enables cross-scenario migration and optimization of the fault identification model, continuously improving the system's general detection capabilities under different operating conditions and in different regions.

[0099] Traditional heat tracing cable detection technologies generally employ a static, fixed threshold judgment mode, lacking data accumulation and self-learning capabilities. This makes them unable to adapt to operational characteristic shifts caused by equipment aging and environmental changes, resulting in a high rate of missed detections for progressive and trend-based faults. This invention achieves a paradigm shift from passive alarm to proactive prediction through a closed-loop learning architecture encompassing local fault sample extraction, differential feature enhancement, classification and database construction, real-time feature matching, and cloud-based rule iteration. Utilizing the equipment's own historical data for personalized feature training allows for adaptation to individual operational differences in single devices. Simultaneously, cloud-based data migration updates global judgment rules, continuously optimizing multi-scenario adaptability. This intelligent design, combining localized learning and global migration optimization, overcomes the technical limitations of traditional mechanical judgment, effectively improving the identification accuracy and predictive capability of progressive, latent faults.

[0100] A5. If the complete heating and cooling segments are not identified within the preset time, a safety decision step is performed before actively applying the electrothermal disturbance. Then, based on the safety decision result, the electrothermal disturbance is selectively applied to excite the formation of the simulated heating and cooling segments, and the process returns to step A3 to complete the feature calculation and state determination.

[0101] If the complete heating and cooling phases are not identified within a preset time in step A5, a safety decision-making step is also included before actively applying electrothermal disturbance: detecting whether the current ambient temperature is lower than a preset extremely low temperature threshold. If the current ambient temperature is below the extremely low temperature threshold, then further detect whether the rate of change of the current ambient temperature is positive and exceeds a preset safe heating rate; if the ambient temperature is below the extremely low threshold and is not in a rapid heating state, then record the current working condition as a high-risk freezing state, limit the predetermined amplitude of the electric heating disturbance to zero and shorten the preset time, directly output a forced alarm signal, and keep the heating cable continuously energized. If the ambient temperature is not lower than the extremely low threshold, or if it is lower than the extremely low threshold but is in a state of rapid temperature rise, then an electrothermal disturbance is allowed to be applied, and the predetermined amplitude is limited to a preset low power amplitude, and the disturbance duration is limited to a preset short duration.

[0102] It should be noted that in industrial cryogenic pipeline scenarios, static operating conditions without natural temperature fluctuations are often accompanied by a high risk of freezing. Conventional detection methods that directly start power disturbances are prone to intermittent power outages and power reduction, which weakens the heat tracing and insulation capabilities and exacerbates the risk of pipeline freezing and blockage.

[0103] This step adds a pre-emptive safety decision-making logic before the artificial disturbance is initiated, using the absolute ambient temperature and temperature change rate as dual judgment criteria to achieve graded control of high-risk working conditions and normal constant temperature working conditions, taking into account both testing needs and equipment operation safety.

[0104] The extremely low temperature threshold is calibrated according to the on-site pipeline antifreeze requirements, and is conventionally set to -10 degrees Celsius. This serves as a dividing line between ordinary low temperatures and high-risk freezing temperatures, and is used to quickly identify extreme operating conditions that require priority protection. By directly comparing the real-time ambient temperature values ​​with the threshold, the first-level operating condition determination can be completed quickly. The logic is simple and the response is rapid, making it suitable for fast calculation and execution by on-site edge controllers.

[0105] The safe heating rate is the critical warming index in low-temperature environments, and the preset safe heating rate is set to 0.2 degrees Celsius per minute. When the ambient temperature has fallen into the extremely low temperature range, the sign and magnitude of the temperature change rate are used to determine whether the external environment has the conditions for natural warming. This distinguishes between continuous extreme cold solidification conditions and short-term low temperature fluctuation conditions, avoiding misjudgments of operating conditions caused by a single temperature threshold.

[0106] When the temperature is in an extremely low range and there is no rapid warming trend, it indicates that the site will maintain an extremely cold and static environment for a long time, and the risk of pipeline medium freezing is at the highest level. At this time, any form of power disturbance output is completely prohibited, and the predetermined disturbance amplitude is directly limited to zero to prevent power adjustment fluctuations caused by detection actions. Simultaneously, the original preset detection time is shortened to reduce the waiting period for the next round of operating condition identification. At the same time, a mandatory alarm signal is immediately issued to remind maintenance personnel to intervene and handle the situation. The heating cable is kept continuously powered throughout the process to ensure pipeline operation safety with maximum insulation power and prioritize the avoidance of freezing accidents.

[0107] When the ambient temperature is above the extremely low threshold, or when the outside temperature is low but the environment is rapidly warming up, the risk of medium freezing is controllable, allowing the active detection process to start normally. To balance energy consumption and equipment safety, the output power of the electrothermal disturbance is uniformly limited to a preset low power range, typically 30% of the rated power. Simultaneously, the disturbance duration is compressed to a preset short duration range, with the duration of a single disturbance controlled within 15 minutes. This limitation method can reduce power fluctuation impacts while meeting the data acquisition requirements for simulated heating and cooling, preventing frequent power adjustments in low-temperature environments from accelerating cable aging, and achieving a balance between detection functionality and equipment protection.

[0108] Existing active detection technologies generally only focus on data acquisition needs and lack safety interlocking logic for extreme low-temperature scenarios. Detection actions and antifreeze protection actions are prone to conflict, which can easily lead to secondary operational risks.

[0109] This invention constructs a dual safety judgment system based on temperature threshold and rate of change to achieve accurate identification and active interlocking of high-risk freezing conditions, and establishes a hierarchical control mechanism with graded detection permissions, power levels, and alarm levels. By prohibiting disturbances, locking continuous power supply, and triggering forced alarms in high-risk conditions, it can avoid safety hazards in extremely cold scenarios from the control logic level. At the same time, in ordinary low-temperature scenarios, it adopts a low-power, short-duration restricted disturbance mode to reduce the impact of detection on the equipment.

[0110] When actively applying the electrothermal disturbance in step A5, the specific method of applying the electrothermal disturbance includes the following steps: The system detects the rate of change of the current ambient temperature. If the absolute value of the rate of change is less than a preset stability threshold, the environment is determined to be in a constant temperature state. Under constant temperature conditions, a first power pulse and a second power pulse are applied to the heat tracing cable in sequence. The first power pulse is a positive heating pulse with 50% to 80% of the rated power, and the second power pulse is a cooling pulse with the output power reduced to below 10% of the rated power (including zero power). During the second pulse, the cable temperature gradually decreases due to natural heat dissipation. A preset interval time is set between the two pulses. The interval time is configured to be greater than half of the thermal time constant of the heat tracing cable and less than the thermal time constant. The response curve of the cable temperature to the first power pulse (heating segment) and the response curve to the second power pulse (cooling segment) are recorded, and the response curves are used as simulated heating segment and cooling segment data.

[0111] The rate of change in ambient temperature is calculated using a 10-minute statistical period. It calculates the temperature change amplitude per unit time based on continuously collected ambient temperature data, with a preset stability threshold of 0.05 degrees Celsius per minute. When the absolute value of the rate of change is below this threshold, it indicates that the ambient temperature has remained stable for a long period without natural temperature fluctuations. In this case, data on natural temperature changes cannot be obtained, and an active perturbation mechanism needs to be activated. This judgment logic can accurately distinguish between natural temperature fluctuations and a constant temperature state, avoiding blindly applying perturbations during natural temperature changes, thus balancing detection continuity and energy efficiency.

[0112] The positive power pulse refers to a positive power supply load of 50% to 80% of the rated power input to the heating cable, causing the cable temperature to rise steadily in a short period of time, fully simulating the heating process under natural operating conditions. The second power pulse (cooling pulse) is implemented based on the load regulation mode. By reducing the output power to below 10% of the rated power (including complete power cut-off), the temperature gradually drops due to natural heat dissipation from the cable and the environment, simulating the cooling process in the natural environment. The two sets of pulses maintain opposite power output logics in terms of thermal effect: the first power pulse is the positive heating excitation, and the second power pulse is the natural heat dissipation excitation after power reduction / power cut-off. This ensures that the mechanisms of the heating excitation and the cooling excitation correspond to each other, providing a symmetrical basis for comparison of the two sets of simulation data and reducing detection deviations caused by artificial disturbances.

[0113] The thermal time constant of the heat tracing cable is a fixed physical parameter characterizing the thermal inertia of the equipment. It can be pre-entered into the system through factory calibration or on-site measurement. Limiting the pulse interval time to the range of half to the thermal time constant avoids incomplete response data caused by excessively short intervals, while preventing temperature steady-state solidification caused by excessively long intervals. This ensures that the cable temperature forms continuous and characteristic change curves under the action of two consecutive pulses, guaranteeing the validity and correlation of the simulation data.

[0114] Throughout the continuous application of the dual-pulse sequence, the two temperature sensors maintain synchronous high-frequency sampling to fully capture the temperature rise and fall curves of the cable as a result of the pulse excitation. The temperature rise curve corresponding to the first positive power pulse is directly defined as the simulated heating segment data, and the temperature fall curve corresponding to the second cooling pulse (i.e., the second power pulse) is directly defined as the simulated cooling segment data. Both are uniformly incorporated into the system's data processing queue to ensure uninterrupted operation of the detection process under constant temperature conditions.

[0115] After applying the dual-pulse sequence in step A5, the method further includes a step of directly judging the asymmetry using simulated data: calculating the asymmetry between the response delay time of the cable temperature to the first power pulse and the response delay time to the second power pulse. The calculated asymmetry is compared with the asymmetry threshold in step A4, so that the active perturbation mode and the natural change mode share the same set of judgment logic. If an effective asymmetry cannot be obtained after applying a double pulse sequence multiple times, it is determined that the ambient temperature sensor or cable temperature sensor is not sensitive enough, and a sensor maintenance alarm signal is output.

[0116] It should be noted that the simulated heating and cooling data generated by dual-pulse excitation possess complete dynamic response characteristics, requiring no additional data format correction; the parameter extraction algorithm for the naturally varying segment can be directly used. This step separately extracts the cable temperature response delay time corresponding to the two sets of pulses, using the difference in bidirectional response delay as the core calculation basis to quickly solve the specific asymmetry under artificial disturbance conditions, shortening the detection and calculation cycle in constant temperature scenarios and improving the efficiency of state determination.

[0117] The asymmetry value calculated under active disturbance mode is directly reused with the fixed asymmetry threshold set in step A4 for comparison and judgment, unifying the evaluation criteria and judgment logic for the two types of working conditions. This integrated design can avoid the judgment confusion caused by multiple scenarios and multiple thresholds, simplify the system control logic, make the results of natural environment detection and artificial stimulation detection horizontally comparable, and ensure the consistency and rigor of fault judgment in all scenarios.

[0118] If repeated application of the dual-pulse sequence and continuous acquisition of temperature response data fails to output a valid asymmetry value after multiple calculations, it indicates hardware-level performance degradation in the temperature acquisition link. After ruling out abnormal cable thermal performance and control module failures, the problem can be directly traced to sensitivity degradation or response lag in the ambient temperature sensor or cable temperature sensor. Simultaneously, a dedicated sensor maintenance alarm signal is issued, enabling early screening of sensor performance degradation and expanding the scope of equipment self-testing.

[0119] Traditional active detection methods in constant temperature scenarios often employ independent judgment criteria and computational logic, which increases the difficulty of program development and subsequent maintenance costs. Furthermore, they mostly focus only on identifying cable faults, lacking long-term monitoring of the performance of the acquisition sensors. This invention achieves algorithm sharing between natural and artificially disturbed operating conditions by unifying the asymmetry calculation method and threshold standard, thus improving the consistency of detection results.

[0120] In practice, the power control module, combined with the results of the prior safety decision, controls the output power, duration and start / stop permissions of the electrothermal disturbance in a hierarchical manner. Under normal restricted operating conditions, the dual-pulse sequence is output in a low-power, short-duration mode, while under high-risk freezing conditions, the disturbance output is completely blocked.

[0121] After data acquisition, dynamic response feature calculation, asymmetry comparison, and sensor performance verification are performed synchronously. Combined with the fault feature library, the prediction results are matched in real time to complete the cable status classification judgment, trend abnormality early warning, and output of multiple types of alarm signals, ensuring the closed-loop safe operation of the detection process in all scenarios such as extreme low temperature, normal constant temperature, and natural fluctuation.

[0122] Let's consider another scenario: in complex industrial settings such as temperature-controlled workshops, underground utility tunnels, and extremely cold exposed pipelines, the natural ambient temperature is consistently stable or extremely cold. A single detection logic cannot meet the protection needs of various operating conditions. Furthermore, the failure patterns of equipment vary across different regions and with different ages, making it difficult for fixed rules to be universally applicable. This invention safeguards safety under extreme conditions through proactive safety decision-making and tiered control. It autonomously learns and accumulates personalized data based on local fault characteristics, and continuously optimizes the algorithm through cloud-based rule migration, comprehensively improving the system's adaptability to complex environments and its intelligent predictive capabilities.

[0123] The periodic fluctuations in ambient temperature are a necessary prerequisite for passive temperature detection. Continuous power supply for freeze protection in extremely cold and high-risk scenarios is the safety baseline for industrial pipeline operation. Shifts in response characteristics caused by long-term equipment aging are a major cause of progressive failures. These various operational needs are intertwined and constrain each other. Conventional technologies cannot simultaneously meet the multiple requirements of safety protection, real-time detection, fault prediction, and global adaptation. They suffer from rigid operating modes, limited identification methods, and weak scalability.

[0124] Based on the design requirements of full-cycle, full-scenario, and intelligent operation, this invention incorporates a multi-layer security protection architecture and an autonomous learning and transfer mechanism, and connects the complete link of on-site data collection, feature extraction, classification and storage, real-time matching, and cloud iteration to achieve deep integration of security protection, fault diagnosis, trend prediction, and rule optimization.

[0125] The innovation of this invention lies in its continuous and synchronous acquisition of ambient temperature and cable temperature sequences, combined with automatic identification of temperature rise / fall during natural temperature variations, precise location of transition points, and elimination of invalid fluctuations, laying a high-quality data foundation. The introduction of piecewise linear normalization processing based on temperature dependence overcomes detection biases caused by shifts in the thermal properties of the heating cable material across different temperature ranges, ensuring uniform comparability of response delay time and response decay coefficient across the entire temperature range. Using the overall asymmetry of the dynamic response characteristics of temperature rise and fall within the same natural temperature variation segment as the core criterion, and in conjunction with point-by-point temperature pairing, local asymmetry fusion, and temperature dependence trend analysis, it achieves accurate differentiation between overt faults, sensor anomalies, and normal temperature-sensitive characteristics of the material. Furthermore, through trend analysis of long-period asymmetry sequences, it can capture progressive degradation processes and quantify remaining operating time, supporting preventative maintenance. In constant temperature or extreme low temperature scenarios without natural temperature fluctuations, the hierarchical active disturbance strategy based on safety decisions (including dual-pulse excitation, power duration limitation, and high-risk interlocking) balances detection reliability and operational safety. Finally, by leveraging fault feature autonomous learning and cloud migration mechanisms, fault judgment rules are continuously optimized, enhancing intelligent early warning capabilities across all scenarios and the entire lifecycle. In summary, this invention changes the previous detection mode of instantaneous comparison of a single parameter, comprehensively covering explicit faults, latent aging, gradual degradation, sensor deterioration, and extreme operational risks, thereby improving the accuracy, robustness, and adaptability of heat tracing cable fault alarms.

[0126] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of fault alarm for heat tracing cable based on ambient temperature comparison, characterized in that, include: A1. Continuously and synchronously acquire the ambient temperature sequence output by the ambient temperature sensor and the cable temperature sequence output by the surface temperature sensor of the heating cable; A2. Real-time identification of natural variation segments in the ambient temperature sequence, and differentiation of the natural variation segments into warming segments and cooling segments; A3. Calculate the dynamic response characteristics of the cable temperature sequence relative to the ambient temperature sequence in the heating and cooling sections, respectively. The calculation includes temperature-dependent normalization of the response delay time and the response decay coefficient. The temperature-dependent normalization adopts a piecewise linear mapping based on the rated operating temperature range of the heat tracing cable to map the response delay time and response decay coefficient calculated under different ambient temperatures to a unified dimensionless scale. The dynamic response characteristics include the normalized response delay time and response decay coefficient. A4. Compare the asymmetry between the dynamic response characteristics of the heating section and the cooling section. If the asymmetry is less than a preset asymmetry threshold, determine that the heat tracing cable has failed and output an alarm signal. A5. If a complete heating and cooling segment is not identified within a preset time, an electrothermal disturbance of a predetermined amplitude is actively applied to the heating cable. This disturbance is used as an excitation signal to form a simulated heating and cooling segment, and the process returns to step A3 to calculate the dynamic response characteristics of the simulated heating and cooling segments.

2. A method of fault alarm for heat tracing cable based on comparison of ambient temperature as claimed in claim 1, wherein, In step A2, when distinguishing between the heating and cooling phases, the transition points within the natural temperature change phases are further identified. These transition points refer to the moments when the ambient temperature changes from rising to falling or from falling to rising. Step A2 further includes: The temperature difference between adjacent moments in the ambient temperature sequence is detected, and the location of the transition point is determined based on the positive or negative change of the difference. Using the transition point as the boundary, the temperature sequence before the transition point is assigned to the end of the previous natural change segment, and the temperature sequence after the transition point is assigned to the beginning of the next natural change segment, so that the heating segment and the cooling segment share the same transition point. Discard natural variation segments whose duration is less than a preset minimum duration.

3. The method of claim 1, wherein the method further comprises: When calculating the response delay time in step A3, the transition point is used as the time origin. Step A3 further includes: Taking the conversion point as the origin, backtracking for a first preset time period, within the first preset time period, calculate the first response delay of the cable temperature sequence relative to the ambient temperature sequence, as the response delay of the heating segment; Taking the conversion point as the origin, backtracking for a second preset time period, within the second preset time period, calculate the second response delay of the cable temperature sequence relative to the ambient temperature sequence, as the response delay of the cooling segment; If the calculated result of the first response delay or the second response delay exceeds a preset reasonable range, the natural change segment is marked as an invalid segment and the system waits for the next transition point.

4. The method of claim 1, wherein the method further comprises: The asymmetry in step A4 is obtained by comparing the dynamic response characteristics of the heating and cooling segments point by point within the same natural variation segment. Step A4 further includes: Arrange the ambient temperature values ​​in the heating and cooling sections in ascending order to form a temperature axis; Multiple discrete temperature sampling points are selected on the temperature axis, and the response delay time and response decay coefficient corresponding to each temperature sampling point in the heating segment and the cooling segment are extracted respectively. For each temperature sampling point, calculate the ratio of the absolute value of the difference between the response delay times of the heating and cooling segments to the sum of the two, and the ratio of the absolute value of the difference between the response attenuation coefficients to the sum of the two. Add the two ratios to obtain the local asymmetry at that temperature sampling point. The average value of the local asymmetry at all temperature sampling points is taken as the overall asymmetry of the natural variation segment.

5. An ambient temperature comparison based heat tracing cable fault alarm method according to claim 4, characterized in that, Following the fourth step of step A4, the temperature dependence of the asymmetry is further determined, including the following steps: Analyze the trend of local asymmetry at each temperature sampling point as the temperature decreases; If the local asymmetry increases monotonically as the temperature decreases, the heat tracing cable is determined to be in normal working condition. If the local asymmetry tends to zero or remains unchanged as the temperature decreases, the heat tracing cable is determined to be in a failure state. If the local asymmetry fluctuates non-monotonicly as the temperature decreases, it is determined that the ambient temperature sensor or the cable temperature sensor is faulty, and a sensor self-test alarm signal is output.

6. The method of claim 1, wherein the method further comprises: It also includes the following steps between steps A4 and A5: Detect the asymmetry sequence calculated from multiple consecutive natural variation segments preceding the current natural variation segment; Determine whether the asymmetry sequence exhibits a gradually decreasing trend; If the cable exhibits a gradual decay trend and the asymmetry of the current natural change segment is lower than a preset absolute threshold, then the heating cable is determined to be in a state of gradual degradation. Based on the decay rate of the asymmetry sequence, calculate the number of remaining natural variation segments required for the asymmetry to decrease to the asymmetry threshold; Based on the average occurrence frequency of historical natural change segments, the remaining number of natural change segments is converted into remaining running time, and a maintenance warning signal containing the remaining running time is output.

7. The method of claim 1, wherein the method further comprises: When actively applying the electrothermal disturbance in step A5, the specific method of applying the electrothermal disturbance includes the following steps: The system detects the rate of change of the current ambient temperature. If the absolute value of the rate of change is less than a preset stability threshold, the environment is determined to be in a constant temperature state. Under constant temperature conditions, a first power pulse and a second power pulse are applied to the heat tracing cable in sequence. The first power pulse is a positive heating pulse with 50% to 80% of the rated power, and the second power pulse is a cooling pulse with the output power reduced to less than 10% of the rated power, so that the cable temperature decreases slowly during the second pulse period by relying on natural heat dissipation. A preset interval time is set between the two pulses, and the interval time is configured to be greater than half of the thermal time constant of the heat tracing cable and less than the thermal time constant. Record the cable temperature response curves to the first power pulse and the second power pulse, and use these response curves as data for the simulated heating and cooling phases.

8. An environmental temperature comparison based heat tracing cable fault alarm method according to claim 7, characterized in that, After applying the double-pulse sequence in step A5, the method further includes a step of directly determining the asymmetry using the simulated data: Calculate the asymmetry between the cable temperature response delay time to the first power pulse and the response delay time to the second power pulse; The calculated asymmetry is compared with the asymmetry threshold in step A4, so that the active perturbation mode and the natural change mode share the same set of judgment logic. If an effective asymmetry cannot be obtained after applying a double pulse sequence multiple times, it is determined that the ambient temperature sensor or cable temperature sensor is not sensitive enough, and a sensor maintenance alarm signal is output.

9. The method of claim 1, wherein the method further comprises: If a complete heating and cooling phase is not identified within a preset time in step A5, a safety decision-making step is also included before actively applying electrothermal disturbance: Detect whether the current ambient temperature is lower than a preset extremely low temperature threshold; If the current ambient temperature is lower than the extremely low temperature threshold, then it is further detected whether the rate of change of the current ambient temperature is positive and exceeds a preset safe heating rate. If the ambient temperature is below the extremely low threshold and is not in a rapid heating state, the current working condition is recorded as a high-risk freezing state, the predetermined amplitude of the electrothermal disturbance is limited to zero and the preset time is shortened, a forced alarm signal is directly output, and the heat tracing cable is kept continuously energized. If the ambient temperature is not lower than the extremely low threshold, or if it is lower than the extremely low threshold but is in a state of rapid temperature rise, then an electrothermal disturbance is allowed to be applied, and the predetermined amplitude is limited to a preset low power amplitude, and the disturbance duration is limited to a preset short duration.

10. The method of claim 1, wherein the method further comprises: It also includes fault feature learning and transfer steps: Once a fault is determined in the heat tracing cable, the ambient temperature sequence and cable temperature sequence in the last complete natural change segment before the fault occurred are extracted as fault feature samples. The fault feature samples are compared differentially with multiple normal feature samples collected during the historical normal operation of the same heat tracing cable to extract the differential feature curves. The differential feature curves are classified and stored according to fault type to form a fault feature library; In subsequent operation, the dynamic response characteristics of the current natural change segment are calculated in real time, and the current characteristics are matched with the characteristics in the fault feature library; If the matching degree exceeds a preset similarity threshold, an early warning signal for this type of fault will be output in advance. The successfully matched fault characteristics and corresponding environmental conditions are uploaded to the cloud for updating the global fault judgment rules.

Images