A new energy transformer leak detection method, system, terminal and medium

Abstract

The present application belongs to the technical field of transformer leak detection, and specifically discloses a new energy transformer leak detection method, system, terminal and medium. The method comprises the following steps: obtaining the initial measurement pressure, initial ambient temperature and box surface temperature after the transformer is inflated and stabilized; synchronously collecting real-time measurement pressure and corresponding real-time temperature data in the pressure maintaining monitoring stage; establishing a dynamic heat transfer model representing the internal and external heat conduction boundary and thermal inertia delay characteristics, and calculating the dynamic equivalent gas temperature inside the transformer through time series iteration; fusing the transformer shell thermal elastic deformation compensation rate and the real gas nonlinear state equation to calculate the theoretically predicted pressure after environmental compensation; comparing and analyzing the deviation between the real-time measurement pressure and the theoretically predicted pressure, and judging whether there is a leak and intelligently diagnosing the leak point type according to the dynamic evolution characteristics of the deviation. The present application overcomes the defect of false alarm of the traditional fixed threshold, and realizes high sensitivity capture and accurate qualitative diagnosis of early micro leakage.

Description

Technical Field

[0001] This invention belongs to the field of transformer leak testing, specifically relating to a method, system, terminal, and medium for leak testing of new energy transformers. Background Technology

[0002] Oil-immersed new energy transformers are the core hub of the power system, and their tank sealing performance directly affects the stability of the insulation medium and the safety of power grid operation. Sealing tests are a routine preventative test that must be performed on transformers before and after installation. Traditional leak testing methods typically involve injecting dry, compressed gas above the transformer oil level and monitoring the pressure changes of the gas during a specified long-term pressure holding period to check for potential leaks. With the advancement of digital factory construction, online monitoring systems are increasingly replacing manual inspections.

[0003] Existing transformer leak monitoring systems typically install digital pressure sensors on-site to collect pressure data in real time during the pressure holding period and upload it to a remote server or monitoring terminal via a wireless network. In terms of leak detection logic, most existing systems rely on preset fixed empirical thresholds. For example, when the real-time monitored pressure value drops by more than a fixed percentage of the initial pressure, the system determines that the transformer has a leak and triggers an alarm.

[0004] However, in actual leak testing, the above-mentioned technology has significant drawbacks. First, the leak testing cycle is long and the temperature difference in the workshop is large. The existing system ignores the strong thermal inertia and heat transfer hysteresis effect of large transformers, and is prone to misjudging normal physical pressure drop caused by ambient cooling as a real leak. Second, the existing calculations do not compensate for the dynamic volume change caused by the thermoelastic deformation of the metal shell and the non-ideal compressibility characteristics of real gases, resulting in a distorted pressure reference. Finally, a single fixed threshold judgment cannot detect early trace leaks, and after an alarm, it cannot identify whether it is a weld pinhole or a gasket leak, which greatly restricts the efficiency of fault diagnosis and repair. Summary of the Invention

[0005] This invention addresses the problems in existing technologies by providing a method, system, terminal, and medium for leak testing of new energy transformers. It solves the problem in the background technology that neglects the strong thermal inertia and heat transfer hysteresis effect of large transformers, easily misjudging normal physical pressure drops caused by ambient cooling as actual leaks. It also solves the problem of pressure benchmark distortion caused by the uncompensated dynamic volume changes due to the thermoelastic deformation of the metal shell and the non-ideal compressibility characteristics of real gases. Furthermore, it solves the problem that a single fixed threshold cannot capture early, minute leaks, and that after an alarm, it cannot identify whether the leak is from weld pinholes or gaskets, greatly restricting the efficiency of fault diagnosis and repair.

[0006] The technical solution adopted in this invention is as follows: Firstly, this application provides a method for leak testing of a new energy transformer, which includes the following steps: Step S1: Fill the transformer with gas to the preset pressure. After the pressure stabilizes, obtain and record the initial measured pressure, initial ambient temperature and initial transformer tank surface temperature of the transformer. Step S2: During the pressure holding monitoring stage, the real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature of the transformer are collected synchronously according to the preset collection cycle. Step S3: Establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. Use the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, obtain the dynamic equivalent gas temperature inside the transformer under the current acquisition period. Step S4: Based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature, and equivalent gas temperature, combined with the gas thermodynamic state characteristics, calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle. Step S5: Compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure. Based on the dynamic change characteristics of the deviation, determine whether there is a leak in the transformer and output the corresponding leak test diagnosis results or alarm signals.

[0007] Furthermore, the specific process of establishing the dynamic heat transfer model and performing calculations in step S3 includes: The transformer casing is considered as a lumped parameter heat absorber with thermal resistance and heat capacity. A heat transfer differential equation is established from the external environment to the transformer casing and then from the transformer casing to the internal gas. By introducing a comprehensive thermal inertia time constant to characterize thermal inertia delay, the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are discretized, and a time-series iterative calculation formula in discrete state is constructed to calculate the dynamic equivalent gas temperature.

[0008] Furthermore, the specific formula for calculating the time-series iteration under discrete states is as follows:

[0009] in, For the current number Dynamic equivalent gas temperature for each acquisition cycle This represents the dynamic equivalent gas temperature from the previous data collection cycle. The time interval of the data collection cycle. The overall thermal inertia time constant of the transformer system, This is the surface weighting coefficient for heat transfer from the transformer casing to the internal gas. This represents the surface temperature of the transformer tank during the current data collection period. This represents the ambient temperature during the current data collection period.

[0010] Furthermore, the specific process of calculating the theoretically predicted pressure after environmental compensation in step S4 includes: A nonlinear pressure prediction coupled model is established that integrates transformer shell thermoelastic deformation compensation with the real gas equation of state. The volume expansion coefficient of the transformer tank material is obtained. Based on the temperature difference between the initial transformer tank surface temperature and the real-time transformer tank surface temperature in the current acquisition period, the deformation compensation rate that causes the change in the dynamic effective volume inside the transformer is calculated. The second virial coefficient, which characterizes the intermolecular interaction force of the test gas, is extracted. Combined with the real-time measured pressure of the previous acquisition cycle and the equivalent gas temperature in the current acquisition cycle, the nonlinear gas compression compensation term is calculated. The relative change ratio of the initial measured pressure, the equivalent gas temperature, the deformation compensation rate, and the nonlinear gas compression compensation term are calculated in a multidimensional coupling manner through a nonlinear pressure prediction coupling model to obtain the theoretical predicted pressure after environmental compensation under the current acquisition cycle.

[0011] Furthermore, the specific formula for multidimensional coupling calculation is as follows:

[0012] Among them, the nonlinear gas compression compensation term The computational expansion is as follows:

[0013] In the formula, The theoretically predicted stress after environmental compensation under the current collection cycle; Initial measurement pressure; The equivalent gas temperature for the current acquisition cycle; The equivalent gas temperature under the initial steady-state condition; The equivalent volumetric expansion coefficient of the transformer housing material; This represents the surface temperature of the transformer enclosure during the current data collection period. This is the initial surface temperature of the transformer enclosure; and These are the temperature-dependent second virial coefficients of the test gas in the current and initial states, respectively. The real-time measured pressure from the previous data acquisition cycle; This is the universal gas constant.

[0014] Furthermore, in step S5, the real-time measured pressure is compared and analyzed with the theoretically predicted pressure. Based on the dynamic changes in the deviation, it is determined whether there is a leak in the transformer. Specifically, this includes: Calculate the real-time measured pressure under the current acquisition period. Compared with theoretically predicted pressure The transient pressure residual between the transformer and the gas flow rate is mapped inversely to a dynamic equivalent leakage cross-sectional area characterizing the actual internal gas mass loss, based on the transformer's initial volume and gas flow state equation. The mapping formula is:

[0015] A time-series sliding observation window is constructed based on the dynamic equivalent leakage cross-sectional area, and the leakage energy drift characteristic value within the sliding observation window is calculated using the cumulative sum algorithm. :

[0016] Extract the system background fluctuation noise of the transformer in the uncharged state, and set a dynamic adaptive decision threshold related to the ambient temperature. When the leakage energy drift characteristic value It shows a monotonically increasing trend, and When the transformer is found to have a real physical leak, a leak test alarm signal is output. in, This refers to the rated initial cavity volume inside the transformer. The set gas equivalent flow coefficient, The time interval of the data collection cycle. This represents the leakage energy drift characteristic value from the previous acquisition cycle. This represents the average background noise level of the system. This is the set anti-interference drift tolerance constant.

[0017] Furthermore, after confirming the existence of a real leak in the transformer, the method also includes a step of intelligently diagnosing the leak type based on dynamic change characteristics, specifically including: Extracting the dynamic equivalent leakage cross-sectional area The evolutionary trajectory over time is calculated, and the slope of the linear regression evolution of this trajectory is determined. Discrete step derivative with adjacent periods :

[0018]

[0019] If the evolution slope The discrete step derivative is less than the set threshold for small variables and occurs throughout the pressure holding monitoring phase. All values ​​are less than the set mutation threshold, indicating that the area of ​​the leakage channel remains a non-zero constant and exhibits rigid geometric characteristics. The system diagnoses the current leakage type as micropore leakage in the transformer tank weld. If the evolution slope The value exceeds the set threshold for small variables, exhibiting a nonlinear expansion trend, or the discrete step derivative at any moment during pressure holding monitoring. If the value exceeds the mutation threshold, it indicates that the leakage channel has deformed due to the relaxation of compressive stress, and the system diagnoses the current leakage type as gasket leakage at a flexible interface. in, This represents the total number of samples collected within the evolutionary trajectory analysis window. For the first The collection time for each sample To analyze the average acquisition time within the window, This is to analyze the average value of the dynamic equivalent leakage cross-sectional area within the window.

[0020] Secondly, this application provides a new energy transformer leak testing system for implementing the new energy transformer leak testing method as described in the first aspect. The system includes: The initial state acquisition module is used to fill the transformer with gas to a preset pressure. After the pressure stabilizes, it acquires and records the transformer's initial measured pressure, initial ambient temperature, and initial transformer tank surface temperature. The real-time data acquisition module is used to synchronously acquire the transformer's real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature according to a preset acquisition cycle during the pressure holding monitoring stage. The equivalent temperature calculation module is used to establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. The real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are used as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, the dynamic equivalent gas temperature inside the transformer under the current acquisition period is obtained. The predicted pressure compensation module is used to calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature and equivalent gas temperature, combined with the gas thermodynamic state characteristics. The deviation analysis and diagnosis module is used to compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure, determine whether there is a leak in the transformer based on the dynamic change characteristics of the deviation, and output the corresponding leak test diagnosis results or alarm signals.

[0021] Thirdly, this application provides a terminal, including: The memory is used to store the leak test program for new energy transformers; The processor is used to implement the steps of the new energy transformer leak testing method as described in the first aspect when executing the new energy transformer leak testing program.

[0022] Fourthly, this application provides a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the new energy transformer leak testing method as described in the first aspect.

[0023] As can be seen from the above technical solutions, the advantages of the present invention are: By simultaneously acquiring temperature and pressure data, combining dynamic heat transfer and thermodynamic models to calculate and predict pressure theoretically, and then comparing the predicted pressure with the actual pressure for deviation analysis, this scheme eliminates the physical pressure fluctuations caused by temperature variations and removes the judgment errors introduced by the fixed pressure threshold method. It provides an objective calculation benchmark for the transformer's sealing status, improving the identification and accuracy of leaks under varying operating conditions.

[0024] A differential equation for heat transfer from the external environment to the internal gas is established, and a comprehensive thermal inertia time constant is introduced to quantify the heat transfer hysteresis characteristics during the heating and cooling process of the metal container into a mathematical model. This scheme corrects the timing misalignment error caused by directly using the external temperature to replace the internal gas temperature, restores the true thermodynamic response process of the internal gas, and provides a time-dimensional parameter compensation basis for pressure prediction calculations.

[0025] This paper provides a discretized time-series iterative algebraic formula for dynamic equivalent gas temperature, transforming the continuous heat transfer differential equation into computational logic adapted to a fixed period of the monitoring terminal. This scheme clarifies the recursive calculation rules and solution path for the microprocessor, reduces the resource consumption of continuous integration operations on system hardware computing power, and ensures that the heat transfer compensation model can be stably executed in industrial application environments.

[0026] By incorporating the volumetric expansion coefficient of the transformer casing material and the second virial coefficient of the gas into the model, the internal volumetric deformation caused by temperature difference and the non-ideal gas compressibility characteristics are calculated. This method couples the solid mechanical deformation parameters with the nonlinear variables of the gas, correcting the inherent calculation biases of treating the equipment as an absolutely rigid container and the gas as an ideal gas, thus improving the underlying physical analysis model for transformer leak testing.

[0027] A specific algebraic expression for multidimensional coupled calculation is given, clarifying the mathematical hierarchy of the nonlinear gas compression compensation term. This calculation formula uses the measurement feedback value of the previous cycle and the virial coefficient of the current state to form a correlation equation, establishing a calculation path for converting various initial parameters and environmental parameters into theoretically predicted pressures, and transforming physical change characteristics into quantitative numerical mapping rules that can be recognized by computers.

[0028] The pressure residual is inversely mapped to an equivalent leakage cross-sectional area, and a cumulative sum statistical model is introduced to accumulate the minute numerical drifts within the observation window over time. This scheme, combined with an adaptive background noise decision threshold, amplifies the weak leakage parameters on the time axis and filters out background interference, solving the problem that early minute leaks are difficult to identify by conventional single-measurement comparison methods.

[0029] The linear regression slope and discrete step derivative of the equivalent leakage cross-sectional area evolution trajectory are calculated and used as classification criteria. Based on the physical properties of different materials, this method maps a constant area to rigid weld leakage and an expanding area to flexible gasket leakage, transforming the fault diagnosis process into a quantitative logic for automatic system calculation, providing objective results to distinguish the causes of leakage. Attached Figure Description

[0030] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a flowchart of the new energy transformer leak testing method of the present invention; Figure 2 This is an architectural diagram of the new energy transformer leak testing system of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0033] Please see Figure 1 As shown, this application provides a method for leak testing of a new energy transformer, including the following steps: Step S1: Fill the transformer with gas to the preset pressure. After the pressure stabilizes, obtain and record the initial measured pressure, initial ambient temperature and initial transformer tank surface temperature of the transformer. In practice, operators use specialized inflation equipment to inject dry compressed air or high-purity nitrogen into the transformer tank. Because gas compression releases heat, after reaching the preset pressure, the tank must be left to stand for a period to allow the internal gas to reach a preliminary thermal equilibrium with the external environment. After this period, the system's main control unit acquires the initial absolute pressure inside the transformer using a high-precision digital pressure sensor. Simultaneously, it records the initial baseline temperature data using ambient temperature sensors deployed around the transformer and surface-mount temperature probes attached to the tank. For example, during a factory leak test of a certain type of new energy photovoltaic step-up transformer, dry air is injected to the set test pressure. After two hours of static pressure stabilization, the system automatically records the initial measured pressure as 35.05 kPa, the initial ambient temperature in the workshop as 20°C, and the average initial temperature of the tank surface as 20.5°C. These baseline data will serve as the sole starting point for all subsequent environmental compensation calculations.

[0034] Step S2: During the pressure holding monitoring stage, the real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature of the transformer are collected synchronously according to the preset collection cycle. In practice, pressure monitoring typically needs to continue for several hours or even more than 24 hours. To ensure the timeliness and accuracy of the data, the system performs high-frequency data sampling at pre-set time intervals. To overcome the uneven temperature distribution caused by sunlight or local heat sources affecting large transformer tanks, surface temperature acquisition usually employs multi-point distributed control, and extreme outliers are removed using data cleaning algorithms before averaging. All sensor data is synchronized via an industrial IoT module to ensure that pressure and temperature data within the same sampling period are aligned on the time axis. For example, with a data acquisition cycle set at 15 minutes, during a 24-hour pressure holding test, the system automatically wakes up every 15 minutes to simultaneously capture the current pressure value, workshop ambient temperature, and multiple surface temperatures on the sun-facing and shaded sides of the tank. At a specific sampling point, the system recorded a slight drop in real-time measured pressure to 34.98 kPa, while the real-time ambient temperature dropped to 18°C ​​due to local ventilation in the workshop, and the real-time transformer tank surface temperature slowly dropped to 19.2°C.

[0035] Step S3: Establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. Use the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, obtain the dynamic equivalent gas temperature inside the transformer under the current acquisition period. In practical implementation, due to the large metal casing of the transformer, its thermal resistance and heat capacity are extremely high, resulting in a significant time delay in the conduction of changes in external ambient temperature to the internal gas. This step does not use the traditional method of directly averaging; instead, it treats the transformer as a dynamic system with thermal inertia. The system presets a comprehensive thermal inertia time constant based on the material thickness and heat dissipation area of ​​the transformer casing. Combining this with historical temperature data from the previous period, it uses a discretized time-series recursive algorithm to gradually approximate the true physical temperature of the internal gas. For example, during nighttime, the workshop ambient temperature drops sharply by 5°C within one hour. If the external temperature is directly used for estimation, it would be mistakenly assumed that the internal gas also cooled instantly. However, the system's dynamic heat transfer model calculations show that, due to the insulation effect of the transformer's tens of tons of metal, the internal equivalent gas temperature only slowly decreases by 1.5°C. This dynamic equivalent gas temperature output by the system restores the delayed evolution of the thermodynamic state inside the sealed cavity, eliminating interference parameters caused by sudden temperature changes for subsequent calculations.

[0036] Step S4: Based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature, and equivalent gas temperature, combined with the gas thermodynamic state characteristics, calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle. In its implementation, the system not only considers the conventional thermal expansion and contraction effects but also incorporates the thermoelastic deformation of the transformer's metal casing and the non-ideal state of the high-voltage test gas into multi-dimensional coupled calculations. With temperature fluctuations, the effective volume of the transformer undergoes slight physical scaling, while the intermolecular forces of the gas change with the state. The system calculates the volume change rate based on a preset volume expansion coefficient and extracts the virial coefficient of the gas at the corresponding temperature to compensate for the gas's nonlinear compression characteristics, ultimately deriving the standard theoretical pressure that should be present under absolutely sealed conditions and at the current temperature. For example, in the aforementioned nighttime cooling scenario, the internal equivalent gas temperature drops by 1.5°C, and the casing undergoes microscopic contraction due to cooling. The system rigorously extrapolates various thermodynamic characteristics, concluding that, without any physical leakage, the theoretically predicted pressure at the current moment should naturally drop to 34.99 kPa due to the temperature decrease. This theoretical value serves as a dynamic benchmark for subsequent judgment of whether a real leak has occurred, avoiding the limitations of traditional fixed pressure thresholds.

[0037] Step S5: Compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure, determine whether there is a leak in the transformer based on the dynamic change characteristics of the deviation, and output the corresponding leak test diagnosis results or alarm signals. In practical implementation, the system calculates the residual between real-time measured values ​​and theoretical predicted values, transforming fuzzy pressure fluctuations into an equivalent leakage cross-sectional area characterizing the actual mass loss of the transformer. To extract weak leakage signals from the natural fluctuation noise of the sensors, the system introduces an accumulation and statistical algorithm to accumulate energy from small deviations over multiple consecutive cycles. When the accumulated drift characteristic value continuously shows a monotonically increasing trend and exceeds the adaptive anti-interference threshold, it is determined to be a real leak. Furthermore, the system also observes the evolution trajectory of the leakage area. If it exhibits a stationary constant, it is qualitatively characterized as rigid weld leakage; if it shows a nonlinear expansion, it is qualitatively characterized as flexible gasket leakage. For example, in the comparison, the system found that although the real-time measured pressure of 34.98 kPa differed from the theoretical predicted pressure of 34.99 kPa by only 0.01 kPa, the single deviation was extremely small, but in the subsequent consecutive acquisition cycles, this deviation persisted and steadily accumulated upwards, while the leakage cross-sectional area trajectory showed a horizontal straight-line characteristic. The system then filters out normal physical background noise, accurately determines that there is a micropore leak in the transformer's tank weld, and pushes a red leak alarm signal with a specific fault type label to the monitoring screen of the digital factory, guiding on-site personnel to carry out targeted repair welding and troubleshooting.

[0038] In some embodiments, the specific process of establishing the dynamic heat transfer model and performing calculations in step S3 includes: The transformer casing is considered as a lumped parameter heat absorber with thermal resistance and heat capacity. A heat transfer differential equation is established from the external environment to the transformer casing and then from the transformer casing to the internal gas. The specific continuous-state differential formula for the heat transfer differential equation is as follows:

[0039] in, The transient rate of change of the equivalent gas temperature inside the transformer over time; The overall thermal inertia time constant of the transformer system; The surface weighting coefficient of the transformer casing for heat transfer to the internal gas represents the contribution of the casing metal to the heat conduction of the internal gas. The surface temperature of the transformer enclosure over a continuous period of time; The external ambient temperature over a continuous period of time; The value represents the dynamic equivalent gas temperature inside the transformer over a continuous period of time.

[0040] By introducing a comprehensive thermal inertia time constant to characterize thermal inertia delay, the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are discretized, and a time-series iterative calculation formula in discrete state is constructed to calculate the dynamic equivalent gas temperature.

[0041] In practical implementation, since transformer shells are typically welded from thick steel plates, possessing significant mass and specific heat capacity, fluctuations in external ambient temperature cannot instantly penetrate the shell to alter the physical state of the internal gas. The system, based on the thermodynamic lumped parameter method, equates the complex three-dimensional transformer shell to a heat-absorbing node with uniform thermal resistance and heat capacity characteristics. Based on physical parameters such as the shell's material density, specific heat capacity, convective heat transfer coefficient, and effective heat dissipation surface area, the system derives a comprehensive thermal inertia time constant reflecting the heat transfer hysteresis effect of this node. To adapt to the periodic data sampling characteristics of the on-site digital processor, the system transforms the originally continuous partial differential heat transfer equation into discretized iterative algebraic logic based on a fixed time step. In each fixed acquisition cycle, the system calls the internal equivalent gas temperature calculated in the previous cycle as a baseline prior value, superimposing the temperature change increment driven by the current sampling cycle's tank surface temperature and ambient temperature, thereby achieving time-series recursive tracking of the actual thermal response process of the internal gas.

[0042] For example, when performing a sealing test on a 3150kVA new energy transformer, the system calculates and sets its comprehensive thermal inertia time constant to 45 minutes based on the structural parameters of this type of equipment. When a thunderstorm causes the workshop ambient temperature to drop sharply by 8 degrees Celsius in a short period of time, because the comprehensive thermal inertia time constant is much larger than the set 15-minute data acquisition cycle, the system uses a time-series iterative calculation formula to deduce that in the first 15-minute acquisition cycle, the equivalent gas temperature inside the transformer is only slightly affected by cold conduction, and its temperature drop is only 0.2 degrees Celsius. This calculation method based on discretized time-series iteration quantifies the damping and buffering effect of the metal shell on drastic temperature fluctuations, objectively restores the tailing phenomenon of the slow decay of the internal gas temperature over time, and eliminates the system calculation deviation caused by directly substituting the sudden change value of the external ambient temperature.

[0043] In some embodiments, the time-series iterative calculation formula for discrete states is specifically as follows:

[0044] in, For the current number Dynamic equivalent gas temperature for each acquisition cycle This represents the dynamic equivalent gas temperature from the previous data collection cycle. The time interval of the data collection cycle. The overall thermal inertia time constant of the transformer system, This is the surface weighting coefficient for heat transfer from the transformer casing to the internal gas. This represents the surface temperature of the transformer tank during the current data collection period. This represents the ambient temperature during the current data collection period.

[0045] The comprehensive thermal inertia time constant characterizing the thermal inertia delay properties in step S3 Instead of being set as fixed empirical values, these values ​​are obtained through calculation using the thermodynamic lumped parameter method, based on the transformer's three-dimensional structural parameters and material thermal properties. The specific acquisition process is as follows: First, the system interfaces with the Manufacturing Execution System (MES) or Bill of Materials (BOM) database of the digital factory to obtain the equivalent total mass of the metal casing of the currently tested transformer. and the specific heat capacity at constant pressure of the shell material Therefore, the equivalent total heat capacity of the transformer casing can be calculated. Secondly, the system extracts the total effective convective heat transfer area of ​​the transformer casing, including the corrugated pipes or heat sinks. In conjunction with the natural airflow conditions at the leak test site, a corresponding comprehensive surface convective heat transfer coefficient is assigned. Calculate the equivalent thermal admittance of the transformer relative to the external environment. : Finally, the system will have an equivalent total heat capacity. Divide by equivalent thermal conductivity The comprehensive thermal inertia time constant characterizing the heat penetration hysteresis effect was calculated. : ; In a specific engineering embodiment, regarding the mass of a certain model of empty shell... for The new energy transformer system retrieves the specific heat capacity of its carbon steel material. for The equivalent total heat capacity was calculated. for Simultaneously extract its effective heat dissipation area on its outer surface. for Combined with the heat transfer coefficient of micro-wind convection in the workshop (Typical value) The equivalent thermal admittance was calculated. The system substitutes the above parameters into the formula to calculate the overall thermal inertia time constant of the transformer. (about ).

[0046] In some embodiments, the specific process of calculating the theoretically predicted pressure after environmental compensation in step S4 includes: A nonlinear pressure prediction coupled model is established that integrates transformer shell thermoelastic deformation compensation with the real gas equation of state. The volume expansion coefficient of the transformer tank material is obtained. Based on the temperature difference between the initial transformer tank surface temperature and the real-time transformer tank surface temperature in the current acquisition period, the deformation compensation rate that causes the change in the dynamic effective volume inside the transformer is calculated. The specific formula for calculating the deformation compensation rate is as follows: ; in, This represents the dynamic effective volumetric deformation compensation rate under the current acquisition cycle. The equivalent volumetric expansion coefficient of the transformer enclosure metal material obtained by connecting with the digital factory database; This represents the real-time transformer tank surface temperature during the current data acquisition period. This is the reference transformer tank surface temperature under the initial steady-state condition.

[0047] The second virial coefficient, which characterizes the intermolecular interaction force of the test gas, is extracted. Combined with the real-time measured pressure of the previous acquisition cycle and the equivalent gas temperature in the current acquisition cycle, the nonlinear gas compression compensation term is calculated. The relative change ratio of the initial measured pressure, the equivalent gas temperature, the deformation compensation rate, and the nonlinear gas compression compensation term are coupled in a multidimensional calculation to obtain the theoretical predicted pressure after environmental compensation under the current acquisition period.

[0048] In practical implementation, the transformer is treated as a metal pressure vessel. When performing predictive calculations, the system first retrieves the equivalent volumetric expansion coefficient of the steel used in the transformer's casing from its built-in material property database. During long-term leak testing, significant fluctuations in ambient temperature cause non-negligible thermal expansion and contraction in the casing metal, leading to dynamic changes in the actual effective volume of the sealed cavity. Based on the temperature difference between the casing surface at the initial reference time and the current sampling period, the system calculates the microscopic deformation ratio of the volume and generates the corresponding deformation compensation rate. Simultaneously, the high-pressure test gas injected into the transformer experiences non-negligible collisions and volume changes at the molecular level under high pressure and variable temperature conditions. The system incorporates a high-precision physical property state table for the corresponding test gas and extracts the second virial coefficient, dependent on the current equivalent gas temperature, through data mapping. This coefficient characterizes the actual intermolecular interaction forces. Subsequently, the system substitutes the actual feedback pressure from the previous sampling period and the current equivalent gas temperature into the calculation, dynamically correcting for the gas's non-ideal compressibility and deriving a nonlinear gas compression compensation term. Finally, the initial reference pressure, the ratio of temperature change reflecting thermodynamic work, the volumetric deformation compensation rate representing solid mechanics, and the gas compression compensation term reflecting fluid mechanics are multi-dimensionally coupled to correct the theoretical drift of conventional linear calculations near temperature extremes.

[0049] In some embodiments, the specific formula for multidimensional coupling calculation is as follows:

[0050] Among them, the nonlinear gas compression compensation term The computational expansion is as follows:

[0051] In the formula, The theoretically predicted stress after environmental compensation under the current collection cycle; Initial measurement pressure; The equivalent gas temperature for the current acquisition cycle; The equivalent gas temperature under the initial steady-state condition; The equivalent volumetric expansion coefficient of the transformer housing material; This represents the surface temperature of the transformer enclosure during the current data collection period. This is the initial surface temperature of the transformer enclosure; and These are the temperature-dependent second virial coefficients of the test gas in the current and initial states, respectively. The real-time measured pressure from the previous data acquisition cycle; This is the universal gas constant.

[0052] In some embodiments, step S5 involves comparing the real-time measured pressure with the theoretically predicted pressure to determine whether the transformer is leaking based on the dynamic characteristics of the deviation. Specifically, this includes: Calculate the real-time measured pressure under the current acquisition period. Compared with theoretically predicted pressure The transient pressure residual between the transformer and the gas flow rate is mapped inversely to a dynamic equivalent leakage cross-sectional area characterizing the actual internal gas mass loss, based on the transformer's initial volume and gas flow state equation. The mapping formula is:

[0053] A time-series sliding observation window is constructed based on the dynamic equivalent leakage cross-sectional area, and the leakage energy drift characteristic value within the sliding observation window is calculated using the cumulative sum algorithm. :

[0054] Extract the system background fluctuation noise of the transformer in the uncharged state, and set a dynamic adaptive decision threshold related to the ambient temperature. When the leakage energy drift characteristic value It shows a monotonically increasing trend, and When the transformer is found to have a real physical leak, a leak test alarm signal is output. in, This refers to the rated initial cavity volume inside the transformer. The set gas equivalent flow coefficient, The time interval of the data collection cycle. This represents the leakage energy drift characteristic value from the previous acquisition cycle. This represents the average background noise level of the system. This is the set anti-interference drift tolerance constant.

[0055] In some embodiments, after confirming the existence of a real leak in the transformer, the method further includes a step of intelligently diagnosing the leak type based on dynamic change characteristics, specifically including: Extracting the dynamic equivalent leakage cross-sectional area The evolutionary trajectory over time is calculated, and the slope of the linear regression evolution of this trajectory is determined. Discrete step derivative with adjacent periods :

[0056]

[0057] If the evolution slope The discrete step derivative is less than the set threshold for small variables and occurs throughout the pressure holding monitoring phase. All values ​​are less than the set mutation threshold, indicating that the area of ​​the leakage channel remains a non-zero constant and exhibits rigid geometric characteristics. The system diagnoses the current leakage type as micropore leakage in the transformer tank weld. If the evolution slope The value exceeds the set threshold for small variables, exhibiting a nonlinear expansion trend, or the discrete step derivative at any moment during pressure holding monitoring. If the value exceeds the mutation threshold, it indicates that the leakage channel has deformed due to the relaxation of compressive stress, and the system diagnoses the current leakage type as gasket leakage at a flexible interface. in, This represents the total number of samples collected within the evolutionary trajectory analysis window. For the first The collection time for each sample To analyze the average acquisition time within the window, This is to analyze the average value of the dynamic equivalent leakage cross-sectional area within the window.

[0058] Please see Figure 2 As shown, in some embodiments, this application provides a new energy transformer leak testing system for implementing a new energy transformer leak testing method. The system includes: The initial state acquisition module is used to fill the transformer with gas to a preset pressure. After the pressure stabilizes, it acquires and records the transformer's initial measured pressure, initial ambient temperature, and initial transformer tank surface temperature. The real-time data acquisition module is used to synchronously acquire the transformer's real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature according to a preset acquisition cycle during the pressure holding monitoring stage. The equivalent temperature calculation module is used to establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. The real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are used as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, the dynamic equivalent gas temperature inside the transformer under the current acquisition period is obtained. The predicted pressure compensation module is used to calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature and equivalent gas temperature, combined with the gas thermodynamic state characteristics. The deviation analysis and diagnosis module is used to compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure, determine whether there is a leak in the transformer based on the dynamic change characteristics of the deviation, and output the corresponding leak test diagnosis results or alarm signals.

[0059] In some embodiments, this application provides a terminal, including: The memory is used to store the leak test program for new energy transformers; The processor is used to implement the steps of the new energy transformer leak testing method when executing the new energy transformer leak testing program.

[0060] In some embodiments, this application provides a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the new energy transformer leak testing method.

[0061] The above description is merely a preferred embodiment of one or more embodiments of this specification and is not intended to limit the scope of one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of this specification should be included within the protection scope of one or more embodiments of this specification.

Claims

1. A method for leak testing of a new energy transformer, characterized in that, Includes the following steps: Step S1: Fill the transformer with gas to the preset pressure. After the pressure stabilizes, obtain and record the initial measured pressure, initial ambient temperature and initial transformer tank surface temperature of the transformer. Step S2: During the pressure holding monitoring stage, the real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature of the transformer are collected synchronously according to the preset collection cycle. Step S3: Establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. Use the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, obtain the dynamic equivalent gas temperature inside the transformer under the current acquisition period. Step S4: Based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature, and equivalent gas temperature, combined with the gas thermodynamic state characteristics, calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle. Step S5: Compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure. Based on the dynamic change characteristics of the deviation, determine whether there is a leak in the transformer and output the corresponding leak test diagnosis results or alarm signals.

2. The method for leak testing of new energy transformers according to claim 1, characterized in that, The specific process of establishing the dynamic heat transfer model and performing calculations in step S3 includes: The transformer casing is considered as a lumped parameter heat absorber with thermal resistance and heat capacity. A heat transfer differential equation is established from the external environment to the transformer casing and then from the transformer casing to the internal gas. By introducing a comprehensive thermal inertia time constant to characterize thermal inertia delay, the acquired real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are discretized, and a time-series iterative calculation formula in discrete state is constructed to calculate the dynamic equivalent gas temperature.

3. The method for leak testing of new energy transformers according to claim 2, characterized in that, The specific formula for time-series iterative calculation in discrete states is as follows: in, For the current number Dynamic equivalent gas temperature for each acquisition cycle This represents the dynamic equivalent gas temperature from the previous data collection cycle. The time interval of the data collection cycle. The overall thermal inertia time constant of the transformer system, This is the surface weighting coefficient for heat transfer from the transformer casing to the internal gas. This represents the surface temperature of the transformer tank during the current data collection period. This represents the ambient temperature during the current data collection period.

4. The method for leak testing of new energy transformers according to any one of claims 1-3, characterized in that, The specific process for calculating the theoretically predicted pressure after environmental compensation in step S4 includes: A nonlinear pressure prediction coupled model is established that integrates transformer shell thermoelastic deformation compensation with the real gas equation of state. The volume expansion coefficient of the transformer tank material is obtained. Based on the temperature difference between the initial transformer tank surface temperature and the real-time transformer tank surface temperature in the current acquisition period, the deformation compensation rate that causes the change in the dynamic effective volume inside the transformer is calculated. The second virial coefficient, which characterizes the intermolecular interaction force of the test gas, is extracted. Combined with the real-time measured pressure of the previous acquisition cycle and the equivalent gas temperature in the current acquisition cycle, the nonlinear gas compression compensation term is calculated. The relative change ratio of the initial measured pressure, the equivalent gas temperature, the deformation compensation rate, and the nonlinear gas compression compensation term are calculated in a multidimensional coupling manner through a nonlinear pressure prediction coupling model to obtain the theoretical predicted pressure after environmental compensation under the current acquisition cycle.

5. The method for leak testing of new energy transformers according to claim 4, characterized in that, The specific formula for multidimensional coupling calculation is as follows: Among them, the nonlinear gas compression compensation term The computational expansion is as follows: In the formula, The theoretically predicted stress after environmental compensation under the current collection cycle; Initial measurement pressure; The equivalent gas temperature for the current acquisition cycle; The equivalent gas temperature under the initial steady-state condition; The equivalent volumetric expansion coefficient of the transformer housing material; This represents the surface temperature of the transformer enclosure during the current data collection period. This is the initial surface temperature of the transformer enclosure; and These are the temperature-dependent second virial coefficients of the test gas in the current and initial states, respectively. The real-time measured pressure from the previous data acquisition cycle; This is the universal gas constant.

6. The method for leak testing of new energy transformers according to claim 5, characterized in that, Step S5 involves comparing the real-time measured pressure with the theoretically predicted pressure to analyze the deviation. Based on the dynamic changes in the deviation, it is determined whether the transformer is leaking. Specifically, this includes: Calculate the real-time measured pressure under the current acquisition period. Compared with theoretically predicted pressure The transient pressure residual between the transformer and the gas flow rate is mapped inversely to a dynamic equivalent leakage cross-sectional area characterizing the actual internal gas mass loss, based on the transformer's initial volume and gas flow state equation. The mapping formula is: A time-series sliding observation window is constructed based on the dynamic equivalent leakage cross-sectional area, and the leakage energy drift characteristic value within the sliding observation window is calculated using the cumulative sum algorithm. : Extract the system background fluctuation noise of the transformer in the uncharged state, and set a dynamic adaptive decision threshold related to the ambient temperature. When the leakage energy drift characteristic value It shows a monotonically increasing trend, and When the transformer is found to have a real physical leak, a leak test alarm signal is output. in, This refers to the rated initial cavity volume inside the transformer. The set gas equivalent flow coefficient, The time interval of the data collection cycle. This represents the leakage energy drift characteristic value from the previous acquisition cycle. This represents the average background noise level of the system. This is the set anti-interference drift tolerance constant.

7. The method for leak testing of new energy transformers according to claim 6, characterized in that, After confirming the existence of a real leak in the transformer, the method also includes a step of intelligently diagnosing the leak type based on dynamic change characteristics, specifically including: Extracting the dynamic equivalent leakage cross-sectional area The evolutionary trajectory over time is calculated, and the slope of the linear regression evolution of this trajectory is determined. Discrete step derivative with adjacent periods : If the evolution slope The discrete step derivative is less than the set threshold for small variables and occurs throughout the pressure holding monitoring phase. All values ​​are less than the set mutation threshold, indicating that the area of ​​the leakage channel remains a non-zero constant and exhibits rigid geometric characteristics. The system diagnoses the current leakage type as micropore leakage in the transformer tank weld. If the evolution slope The value exceeds the set threshold for small variables, exhibiting a nonlinear expansion trend, or the discrete step derivative at any moment during pressure holding monitoring. If the value exceeds the mutation threshold, it indicates that the leakage channel has deformed due to the relaxation of compressive stress, and the system diagnoses the current leakage type as gasket leakage at a flexible interface. in, This represents the total number of samples collected within the evolutionary trajectory analysis window. For the first The collection time for each sample To analyze the average acquisition time within the window, This is to analyze the average value of the dynamic equivalent leakage cross-sectional area within the window.

8. A leak testing system for new energy transformers, used to implement the leak testing method for new energy transformers as described in claim 1, characterized in that, The system includes: The initial state acquisition module is used to fill the transformer with gas to a preset pressure. After the pressure stabilizes, it acquires and records the transformer's initial measured pressure, initial ambient temperature, and initial transformer tank surface temperature. The real-time data acquisition module is used to synchronously acquire the transformer's real-time measured pressure, real-time ambient temperature, and real-time transformer tank surface temperature according to a preset acquisition cycle during the pressure holding monitoring stage. The equivalent temperature calculation module is used to establish a dynamic heat transfer model that characterizes the thermal conduction boundary and thermal inertia delay characteristics inside and outside the transformer. The real-time ambient temperature sequence and the real-time transformer tank surface temperature sequence are used as the observation input variables of the dynamic heat transfer model. Through time lag compensation and time-series iterative calculation, the dynamic equivalent gas temperature inside the transformer under the current acquisition period is obtained. The predicted pressure compensation module is used to calculate the theoretical predicted pressure after environmental compensation under the current acquisition cycle based on the initial measured pressure, initial ambient temperature, initial transformer tank surface temperature and equivalent gas temperature, combined with the gas thermodynamic state characteristics. The deviation analysis and diagnosis module is used to compare and analyze the deviation between the real-time measured pressure and the theoretical predicted pressure, determine whether there is a leak in the transformer based on the dynamic change characteristics of the deviation, and output the corresponding leak test diagnosis results or alarm signals.

9. A terminal, characterized in that, include: The memory is used to store the leak test program for new energy transformers; The processor is used to implement the steps of the new energy transformer leak testing method as described in claim 1 when executing the new energy transformer leak testing program.

10. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions. When the computer reads the computer instructions from the storage medium, the computer executes the new energy transformer leak testing method as described in claim 1.

Images