An Internet of Things-based power transformer fault detection system and method

Abstract

The present invention discloses an Internet of Things-based power transformer fault detection system and method, which relates to the technical field of fault detection. The present invention includes a data acquisition module, a data analysis module, a winding analysis module, a solution formulation module, a warning terminal and a database. Whether a winding fails is judged by temperature, amplitude and sound pressure level. If a failure occurs, the failure type of the winding is analyzed, and each potential failure type is analyzed. At the same time, the degree of the winding failure is analyzed. If the degree of the failure is high, the matching degree of each spare winding is analyzed, and a spare winding with a high matching degree is randomly selected for replacement. If the degree of the failure is low, the failure type of the winding, each potential failure type and the operating environment are obtained, and a maintenance solution is formulated. It can accurately judge whether the winding fails, and ensures the operation effect of the power transformer after the winding is replaced and the effectiveness of the maintenance solution.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and particularly relates to a fault detection system and method for a power transformer based on the Internet of Things. Background Art

[0002] Power transformers are crucial equipment in the power system. Their winding faults are one of the main reasons for transformer failures and power system outages. The reliability of the operating state of the windings directly affects the stability of power supply. Through Internet of Things technology, real-time monitoring and data collection of the operating state of power transformer windings can be achieved, and data analysis algorithms can be used to accurately and timely determine whether there are faults in the windings of the transformer and the types of faults.

[0003] Traditional power transformer fault detection systems and methods judge whether the windings of a power transformer have failed through a certain piece of data. If the windings have failed, the reasons and types of the winding failures are judged, and at the same time, the degree of the winding faults is analyzed. If the degree of the winding faults is high, the windings are replaced; if the degree of the winding faults is low, a maintenance plan is formulated. Obviously, such power transformer fault detection systems and methods have at least the following deficiencies: 1. Traditional power transformer fault detection systems and methods judge whether the windings of a power transformer have failed through a single piece of data, but winding failures do not necessarily cause changes in this piece of data, and it is impossible to accurately judge whether the windings have failed.

[0004] 2. When traditional power transformer fault detection systems and methods replace windings, they lack an analysis of the winding matching degree, and it is impossible to judge whether the selected spare windings are suitable for the operation of the power transformer at this time, and thus it is impossible to guarantee the operation effect of the power transformer after the windings are replaced.

[0005] 3. When the windings fail, other potential faults may be caused, but traditional power transformer fault detection systems and methods lack an analysis in this regard and do not consider the influence of the operating environment on maintenance, and it is impossible to guarantee the effectiveness of the maintenance plan. Summary of the Invention

[0006] Aiming at the above-mentioned existing technical deficiencies, the purpose of the present invention is to provide a fault detection system and method for a power transformer based on the Internet of Things.

[0007] To solve the above technical problems, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a fault detection system for a power transformer based on the Internet of Things, including the following modules: a data acquisition module, a data analysis module, a winding analysis module, a plan formulation module, a warning terminal, and a database.

[0008] The data acquisition module is used to set each acquisition point at a preset distance interval and collect the temperature, amplitude, and sound pressure level of each acquisition point in real time.

[0009] The data analysis module is used to obtain the temperature, amplitude, and sound pressure level of each acquisition point at each operating moment, analyze whether the windings of the power transformer are faulty. If a fault occurs, the winding analysis module is executed and a warning is issued.

[0010] The winding analysis module includes a correlation analysis unit and a fault analysis unit.

[0011] The correlation analysis unit is used to obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and each secondary fault type. Combine the main fault type and each secondary fault type during each historical fault of the winding pairwise to obtain various fault combinations. Compare the various fault combinations, divide the fault combinations with the same main fault type into one category, and obtain various categories of fault combinations in this way. Compare the various fault combinations in each category of fault combinations, and call the fault combinations with the same combination the same kind of fault combination. Obtain each same kind of fault combination in each category of fault combinations in this way, and count the number of fault combinations in each same kind of fault combination and the number of fault combinations in each category of fault combinations. In each category of fault combinations, calculate the ratio of the number of fault combinations in each same kind of fault combination to the number of fault combinations in the fault combination category where the same kind of fault combination is located, and call it the proportion of each same kind of fault combination. Obtain the proportion of each same kind of fault combination in this way. Compare a certain same kind of fault combination with other same kind of fault combinations, and call the other same kind of fault combinations each compared same kind of fault combination. If the main fault type in this same kind of fault combination is the same as the secondary fault type in a certain compared same kind of fault combination, and the secondary fault type in this same kind of fault combination is the same as the main fault type in this compared same kind of fault combination, then call this same kind of fault combination and this compared same kind of fault combination an associated combination, call this same kind of fault combination the first fault combination, and call this compared same kind of fault combination the second fault combination. Obtain each associated combination in this way, obtain the proportion of the first fault combination and the proportion of the second fault combination in each associated combination. According to the analysis formula: Obtain the correlation value κ of the c-th associated combination c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation values of each associated combination represent the correlation values of each fault type in each associated combination. Analyze the correlation values between each fault type in this way.

[0012] The fault analysis unit is used to collect various parameters of the winding, analyze the fault type of the winding, and analyze other possible fault types caused by this fault type of the winding.

[0013] The solution formulation module is used to obtain the various operating parameters of the winding, analyze the fault degree of the winding, and formulate a maintenance plan according to the fault degree.

[0014] The warning terminal is used to give a warning when a fault occurs in the winding.

[0015] The database is used to store various fault types that occurred during each historical fault of the winding, various data of the winding when no fault occurred, and various maintenance plans.

[0016] In a second aspect, the present invention provides a method for detecting faults in a power transformer based on the Internet of Things, including the following steps: Step 1, data collection: Set each collection point at a preset distance interval, and collect the temperature, amplitude, and sound pressure level of each collection point in real time.

[0017] Step 2, data analysis: Obtain the temperature, amplitude, and sound pressure level of each collection point at each operating moment, analyze whether a fault occurs in the winding of the power transformer. If a fault occurs, execute the winding analysis module and give a warning.

[0018] Step 3. Correlation analysis: Obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and various secondary fault types. Combine the main fault type and various secondary fault types during each historical fault of the winding pairwise to obtain various fault combinations. Compare the various fault combinations, and divide the fault combinations with the same main fault type into one category. In this way, obtain various categories of fault combinations. Compare the fault combinations in each category of fault combinations, and call the fault combinations with the same combination the same-kind fault combinations. In this way, obtain the same-kind fault combinations in each category of fault combinations, and count the number of fault combinations in each same-kind fault combination and the number of fault combinations in each category of fault combinations. In each category of fault combinations, calculate the ratio of the number of fault combinations in each same-kind fault combination to the number of fault combinations in the fault combination category where the same-kind fault combination is located, and call it the proportion of each same-kind fault combination. In this way, obtain the proportion of each same-kind fault combination. Compare a certain same-kind fault combination with other same-kind fault combinations, and call the other same-kind fault combinations the compared same-kind fault combinations. If the main fault type in this same-kind fault combination is the same as the secondary fault type in a certain compared same-kind fault combination, and the secondary fault type in this same-kind fault combination is the same as the main fault type in this compared same-kind fault combination, then call this same-kind fault combination and this compared same-kind fault combination an associated combination, call this same-kind fault combination the first fault combination, and call this compared same-kind fault combination the second fault combination. In this way, obtain each associated combination, obtain the proportion of the first fault combination and the proportion of the second fault combination in each associated combination. According to the analysis formula: Obtain the correlation value κ of the c-th associated combination c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation values of each associated combination represent the correlation values of various fault types in each associated combination. Analyze the correlation values between various fault types in this way.

[0019] Step 4. Fault analysis: Collect various parameters of the winding, analyze the fault type of the winding, and analyze other possible fault types caused by this fault type of the winding.

[0020] Step 5. Formulate a plan: Obtain various operating parameters of the winding, analyze the fault degree of the winding, and at the same time formulate a maintenance plan according to the fault degree.

[0021] The beneficial effects of the present invention are as follows: 1. A power transformer fault detection system and method based on the Internet of Things determines whether a winding fails by temperature, amplitude, and sound pressure level. If a failure occurs, the type of the winding failure is analyzed, and each potential failure type is analyzed. At the same time, the degree of the winding failure is analyzed. If the degree of the failure is high, the matching degree of each spare winding is analyzed, and a spare winding with a high matching degree is randomly selected for replacement. If the degree of the failure is low, the type of the winding failure, each potential failure type, and the operating environment are obtained, and a maintenance plan is formulated, which can accurately determine whether the winding fails and ensure the operating effect of the power transformer after the winding replacement and the effectiveness of the maintenance plan.

[0022] 2. The present invention analyzes the fault indicators of the winding through the temperature, vibration signal, and sound signal of each acquisition point of the winding. If the fault indicator of the winding is 1, it represents that the winding fails. If the fault indicator of the winding is 0, it represents that the winding does not fail, which can accurately determine whether the winding fails.

[0023] 3. When the winding fails, the present invention first analyzes the correlation value between each fault type of the winding through each fault type that occurred during each previous failure of the winding, then analyzes the type of the winding failure, obtains each potential failure type of the winding according to the type of the winding failure, and formulates a maintenance plan according to the type of the winding failure, each potential failure type, and the operating environment, which ensures the effectiveness of the maintenance plan.

[0024] 4. When replacing the winding, the present invention obtains the performance parameters and dimensions of each spare winding, analyzes the matching degree of each spare winding according to the performance parameters and dimensions of each spare winding, obtains each spare winding with a high matching degree, and randomly selects a spare winding from them for replacement, which ensures the operating effect of the power transformer after the winding replacement. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0026] Figure 1 It is a schematic diagram of the system structure connection of the present invention.

[0027] Figure 2 It is a schematic diagram of the implementation steps flow of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0029] Please refer to Figure 1 As shown, the present invention provides a power transformer fault detection system based on the Internet of Things, including the following modules: a data acquisition module, a data analysis module, a winding analysis module, a solution formulation module, an early warning terminal, and a database.

[0030] The data acquisition module is connected to the data analysis module, the data analysis module is connected to the winding analysis module and the early warning terminal, the winding analysis module is connected to the solution formulation module, and the database is connected to the data analysis module, the winding analysis module, and the solution formulation module.

[0031] The data acquisition module is used to set each acquisition point at a preset distance interval and collect the temperature, amplitude, and sound pressure level of each acquisition point in real time.

[0032] It should be noted that an optical fiber grating sensor, a vibration sensor, and a sound level meter are used to collect the temperature, amplitude, and sound pressure level of each acquisition point in real time.

[0033] The data analysis module is used to obtain the temperature, amplitude, and sound pressure level of each acquisition point at each operating moment, analyze whether the windings of the power transformer are faulty, and if a fault occurs, execute the winding analysis module and give an early warning.

[0034] In a specific embodiment, the process of analyzing whether the windings of the power transformer are faulty is as follows: obtain the temperature, amplitude, and sound pressure level of each acquisition point at each operating moment and input them into the sound anomaly model and the vibration anomaly model, and output the sound anomaly value α a of each acquisition point and the vibration anomaly value β a of each acquisition point. When α a =1 or β a =1, it represents that the sound of the a-th acquisition point is abnormal or the vibration of the a-th acquisition point is abnormal. When α a =0 or β a =0, it represents that the sound of the a-th acquisition point is normal or the vibration of the a-th acquisition point is normal, where a represents the number of each acquisition point, a = 1, 2, 3,..., b, b represents the total number of acquisition points, and both a and b are positive integers.

[0035] Among them, the sound anomaly model: Wherein, R(t) represents the sound pressure level at the t-th running moment, R(t + 1) represents the sound pressure level at the (t + 1)-th running moment, R1 represents the preset threshold value of the difference in sound pressure levels, R2 represents the preset threshold value of the difference in sound pressure levels between adjacent running moments, t represents the number of each running moment, t = 1, 2, 3, ..., c, c represents the total number of running moments, both t and c are positive integers, R′ represents the standard sound pressure level, wherein the preset threshold value of the difference in sound pressure levels is determined by the staff according to the difference between the sound pressure level of the winding when the historical power transformer has no fault and the standard sound pressure level, the preset threshold value of the difference in sound pressure levels between adjacent running moments is set by the staff according to the change in the sound pressure level of the winding at each running moment when the historical power transformer has no fault, and both the preset threshold value of the difference in sound pressure levels and the preset threshold value of the difference in sound pressure levels between adjacent running moments are used to judge whether the sound of the winding is normal, and the standard sound pressure level is set by the staff.

[0036] Among the above, the vibration anomaly model: Wherein, S(t) represents the amplitude at the t-th running moment, S(t + 1) represents the amplitude at the (t + 1)-th running moment, S1 represents the preset threshold value of the amplitude difference, S2 represents the preset threshold value of the amplitude difference between adjacent running moments, S′ represents the standard amplitude, wherein the preset threshold value of the amplitude difference is determined by the staff according to the difference between the amplitude of the winding when the historical power transformer has no fault and the standard amplitude, the preset threshold value of the amplitude difference between adjacent running moments is set by the staff according to the change in the amplitude at each running moment when the historical power transformer has no fault, and both the preset threshold value of the amplitude difference and the preset threshold value of the amplitude difference between adjacent running moments are used to judge whether the vibration of the winding is normal, and the standard amplitude is set by the staff.

[0037] It should be explained that when obtaining from the database that the power transformer has no fault, the amplitudes and sound pressure levels of each acquisition point at each running moment are acquired, and their average values are calculated, and the average values are used as the standard amplitude and the standard sound pressure level.

[0038] The temperatures, sound anomaly values, and vibration anomaly values of each acquisition point at each running moment are acquired and input into the winding anomaly model, and the anomaly value of the winding is output. If the anomaly value of the winding is 1, it means that the winding has a fault; if the anomaly value of the winding is 0, it means that the winding has no fault.

[0039] It should be noted that the expression of the winding anomaly model is as follows: Wherein A a (t) represents the temperature of the a-th acquisition point at the t-th running moment, A a (t + 1) represents the temperature of the a-th acquisition point at the (t + 1)-th running moment, represents the average temperature of the a-th acquisition point, represents a preset temperature difference threshold, and χ represents an abnormal value of the winding. The preset temperature difference threshold is set by the staff to determine whether the temperature change at each collection point is normal.

[0040] The winding analysis module includes a correlation analysis unit and a fault analysis unit.

[0041] The correlation analysis unit is used to obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and various secondary fault types. Combine the main fault type and various secondary fault types during each historical fault of the winding in pairs to obtain various fault combinations. Compare the various fault combinations, divide the fault combinations with the same main fault type into one category, and obtain various categories of fault combinations in this way. Compare the various fault combinations in each category of fault combinations, and call the fault combinations with the same combination the same kind of fault combination. Obtain the same kind of fault combinations in each category of fault combinations in this way, and count the number of fault combinations in each same kind of fault combination and the number of fault combinations in each category of fault combinations. In each category of fault combinations, calculate the ratio of the number of fault combinations in each same kind of fault combination to the number of fault combinations in the fault combination class where the same kind of fault combination is located, and call it the proportion of each same kind of fault combination. Obtain the proportion of each same kind of fault combination in this way. Compare a certain same kind of fault combination with other same kind of fault combinations, and call the other same kind of fault combinations the compared same kind of fault combinations. If the main fault type in this same kind of fault combination is the same as the secondary fault type in a certain compared same kind of fault combination, and the secondary fault type in this same kind of fault combination is the same as the main fault type in this compared same kind of fault combination, then call this same kind of fault combination and this compared same kind of fault combination an associated combination, call this same kind of fault combination the first fault combination, and call this compared same kind of fault combination the second fault combination. Obtain various associated combinations in this way, obtain the proportion of the first fault combination and the proportion of the second fault combination in each associated combination. According to the analysis formula: Obtain the correlation value κ of the c-th associated combination c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation values of each associated combination represent the correlation values of various fault types in each associated combination. Analyze the correlation values between various fault types in this way.

[0042] It should be noted that various fault types include short - circuit faults, open - circuit faults, insulation faults, etc.

[0043] It should also be noted that, for each data value in each fault type, calculate the difference between each data value in each fault type and the data values during normal operation. At the same time, calculate the sum of the differences for each data in each fault type, and make a comparison. The fault type with the largest sum of the differences is regarded as the main fault type, and the other fault types are regarded as the secondary fault types.

[0044] The fault analysis unit is used to collect various parameters of the winding, analyze the fault type of the winding, and analyze other possible fault types caused by this fault type of the winding.

[0045] In a specific embodiment, the fault analysis unit has the following specific process: obtain the DC resistance, dielectric loss factor, no-load current, and no-load loss of the winding, analyze the fault type of the winding according to the obtained data, and at the same time, according to the fault type of the winding, obtain various possible fault types of the winding, and call them potential fault types.

[0046] It should be noted that a DC resistance tester is used to measure the DC resistance of the winding, and a dielectric loss tester is used to measure the dielectric loss factor of the winding.

[0047] It should also be noted that the power transformer is put into no-load operation, and an ammeter and a wattmeter are used to measure and collect the no-load current and no-load loss of the winding.

[0048] Among them, the process of obtaining the possible fault types of the winding is as follows: obtain the correlation value between the fault type of the winding and other fault types. If the correlation value between this fault type and a certain fault type among other fault types is greater than 0.6, it means that this fault type among other fault types is a possible fault type of the winding. Use this method to obtain various possible fault types of the winding.

[0049] In the above, the specific process of analyzing the fault type of the winding is as follows: obtain the DC resistance, real-time dielectric loss factor, real-time no-load current, and real-time no-load loss of the winding of the power transformer when no fault occurs in the database, calculate the average values of the real-time DC resistance, real-time no-load current, and real-time no-load loss, and use the average values as the standard values of the DC resistance, no-load current, and no-load loss. Compare the dielectric loss factors at each operating moment to obtain the standard range of the dielectric loss factor.

[0050] Obtain the DC resistance of the winding, the dielectric loss factor of the winding, the no-load current and no-load loss of each phase winding, then: Where C represents the DC resistance of the winding, D represents the dielectric loss factor of the winding, E represents the no-load current of the winding, F represents the no-load loss of the winding, C′ represents the standard value of the DC resistance, D′ represents the standard range of the dielectric loss factor, E′ represents the standard value of the no-load current, F′ represents the standard value of the no-load loss, and δ represents the return value of the winding fault type. represents the preset threshold value of the DC resistance difference. represents the preset threshold value of the no-load current difference. represents the preset threshold value of the no-load loss difference.

[0051] It should be noted that the preset threshold value of the DC resistance difference is set by the staff to judge whether the DC resistance of the winding is normal.

[0052] It should also be noted that both the preset threshold value of the no-load current difference and the preset threshold value of the no-load loss are set by the staff to judge whether the no-load current and no-load loss of the winding are normal.

[0053] When the return value of the winding fault type is 1, it represents that the fault type of the winding is an open circuit fault. When the return value of the winding fault type is 0, it represents that the fault type of the winding is an insulation fault. When the return value of the winding fault type is -1, it represents that the fault type of the winding is a short circuit fault.

[0054] The scheme formulation module is used to obtain the operating parameters of the winding, analyze the fault degree of the winding, and formulate a maintenance plan according to the fault degree.

[0055] In a specific embodiment, the process of formulating the maintenance plan is as follows: Obtain the values of the operating parameters of the winding at each operating moment when the power transformer has no fault from the database, and compare them to obtain the standard numerical range of each operating parameter. At the same time, use each sensor to collect the operating parameters of the winding at this time, and calculate the return value of the fault degree of the winding. When the return value of the fault degree of the winding is 1, it represents that the fault degree of the winding is high. When the return value of the fault degree of the winding is 0, it represents that the fault degree of the winding is low.

[0056] It should be noted that the operating parameters include inductance, phase, voltage, resistance, etc.

[0057] It should also be noted that an inductance is collected using a mutual inductance sensor, a phase is collected using a Hall effect sensor, a voltage is collected using a voltage transformer, and a resistance is collected using a digital multimeter.

[0058] When the fault degree of the winding is high, obtain the performance parameter values and dimensions of each spare winding from the database, analyze the matching degree of each spare winding, and randomly select a spare winding with a high matching degree for replacement.

[0059] It should be noted that the performance parameters of the winding include withstand voltage strength, magnetic permeability, loss tangent, thermal conductivity, etc.

[0060] When the fault degree of the winding is low, obtain each maintenance plan in the database, obtain the operating parameters of the winding after maintenance in each maintenance plan, calculate the maintenance effect score of each maintenance plan, obtain each fault type repaired in each maintenance plan, and compare it with the fault type of the winding and each potential fault type. If each fault type repaired in a certain maintenance plan includes the fault type of the winding and each potential fault type, then regard this maintenance plan as a standby maintenance plan. Obtain each standby maintenance plan in this way, obtain the maintenance effect, maintenance duration, maintenance cost and each operating environment parameter of each standby maintenance plan, as well as the operating environment parameters of the winding, and perform normalization processing. Then: In the formula represents the value of the h-th operating environment parameter in the g-th standby maintenance plan, J h represents the value of the h-th operating environment parameter of the winding, K g represents the maintenance cost of the g-th standby maintenance plan, L g represents the maintenance duration of the g-th standby maintenance plan, represents the practicality coefficient of the g-th standby maintenance plan, G g represents the maintenance effect score of the g-th standby maintenance plan, g represents the number of each standby maintenance plan, g = 1, 2, 3,..., j, j represents the total number of standby maintenance plans, h represents the number of each operating environment parameter, h = 1, 2, 3,..., i, i represents the total number of operating environment parameters, and g, h, j and i are all positive integers.

[0061] It should be noted that each operating environment parameter includes temperature, humidity, air pressure, etc.

[0062] Compare the practicality coefficients of each main and standby maintenance plan, select the standby maintenance plan with the largest practicality coefficient, and regard this standby maintenance plan as the final maintenance plan.

[0063] In the above, the specific process of calculating the return value of the fault degree of the winding is as follows: Obtain the numerical values of each operating parameter of the winding and the standard numerical range of each operating parameter, and perform normalization processing. Then: In the formula M k represents the value of the k-th operating parameter, M k ′ represents the minimum value of the standard numerical range of the k-th operating parameter, M k″ represents the maximum value of the standard numerical range of the k-th operating parameter, ξ represents the preset threshold of the fault degree index, k represents the number of each operating parameter, k = 1, 2, 3,..., l, l represents the total number of operating parameters, both k and are positive integers, e represents the natural constant, and γ represents the return value of the fault degree of the winding.

[0064] It should be noted that the preset threshold of the fault degree index is set by the staff and is used to judge the fault degree of the winding.

[0065] In the above, the specific process of analyzing the matching degree of each spare winding is as follows: Obtain the numerical values of each performance parameter and the size of each spare winding from the database, and at the same time obtain the standard numerical range of each performance parameter of the winding and the required size of the winding, and perform normalization processing, then: In the formula, N m represents the size of the m-th spare winding, represents the numerical value of the n-th performance parameter of the m-th spare winding, Q n ′ represents the minimum value of the standard numerical range of the n-th performance parameter of the winding, Q n ″ represents the maximum value of the standard numerical range of the n-th performance parameter of the winding, N′ represents the required size of the winding, ζ represents the preset performance coefficient threshold, η m represents the matching degree coefficient of the m-th spare winding, m represents the number of the spare winding, m = 1, 2, 3,..., q, q represents the total number of spare windings, n represents the number of each performance parameter of the winding, n = 1, 2, 3,..., p, p represents the total number of each performance parameter of the winding, and m, n, q, and p are all positive integers.

[0066] It should be noted that the standard numerical range of each performance parameter of the winding is set by the staff to ensure the normal operation of the winding, and the required size of the winding is also set by the staff.

[0067] It should also be noted that

[0068] When the matching degree coefficient of a certain spare winding is 1, it means that the matching degree of this spare winding is high. When the matching degree coefficient of a certain spare winding is 0, it means that the matching degree of this spare winding is low. Analyze the matching degree of each spare winding in this way.

[0069] In the above, the specific process of calculating the maintenance effect score of each maintenance plan is as follows: Obtain each operating parameter of the winding after maintenance, then: In the formula represents the numerical value of the k-th operating parameter of the winding after maintenance in the g-th spare maintenance plan, M k ″′ represents the optimal numerical value of the k-th operating parameter, G g represents the maintenance effect score of the g-th spare maintenance plan.

[0070] It should be noted that the operating efficiency at each operating moment when the power transformer has no fault is obtained from the database, compared, the operating moment with the highest operating efficiency is selected, and the numerical values of each operating parameter at this operating moment are used as the optimal values of each operating parameter.

[0071] The warning terminal is used to give a warning when a winding fault occurs.

[0072] The database is used to store various fault types that occurred during each historical winding fault, various data of the winding when no fault occurred, and various maintenance plans.

[0073] It should be noted that the various data of the winding when no fault occurred include each operating parameter, direct current resistance, dielectric loss factor, no-load current, and no-load loss.

[0074] Please refer to Figure 2 As shown, the present invention provides a method for detecting power transformer faults based on the Internet of Things, including the following steps: Step 1, data acquisition: Each acquisition point is set at a preset distance interval, and the temperature, amplitude, and sound pressure level of each acquisition point are collected in real time.

[0075] Step 2, data analysis: Obtain the temperature, amplitude, and sound pressure level of each acquisition point at each operating moment, analyze whether the winding of the power transformer has a fault. If a fault occurs, execute the winding analysis module and give a warning.

[0076] Step 3. Correlation analysis: Obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and various secondary fault types. Combine the main fault type and various secondary fault types during each historical fault of the winding pairwise to obtain various fault combinations. Compare the various fault combinations, and classify the fault combinations with the same main fault type into one category. In this way, obtain various categories of fault combinations. Compare the fault combinations in each category of fault combinations, and call the fault combinations with the same combination the same kind of fault combination. In this way, obtain the same kind of fault combinations in each category of fault combinations, and count the number of fault combinations in each same kind of fault combination and the number of fault combinations in each category of fault combinations. In each category of fault combinations, calculate the ratio of the number of fault combinations in each same kind of fault combination to the number of fault combinations in the fault combination category where the same kind of fault combination is located, and call it the proportion of each same kind of fault combination. In this way, obtain the proportion of each same kind of fault combination. Compare a certain same kind of fault combination with other same kind of fault combinations, and call the other same kind of fault combinations various compared same kind of fault combinations. If the main fault type in this same kind of fault combination is the same as the secondary fault type in a certain compared same kind of fault combination, and the secondary fault type in this same kind of fault combination is the same as the main fault type in this compared same kind of fault combination, then call this same kind of fault combination and this compared same kind of fault combination an associated combination, call this same kind of fault combination the first fault combination, and call this compared same kind of fault combination the second fault combination. In this way, obtain various associated combinations, obtain the proportion of the first fault combination and the proportion of the second fault combination in each associated combination. According to the analysis formula: Obtain the correlation value κ of the c-th associated combination c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation value of each associated combination represents the correlation value of each fault type in each associated combination. In this way, analyze the correlation value between each fault type.

[0077] Step 4. Fault analysis: Collect various parameters of the winding, analyze the fault type of the winding, and analyze other possible fault types caused by this fault type of the winding.

[0078] Step 5. Formulate a plan: Obtain various operating parameters of the winding, analyze the fault degree of the winding, and at the same time formulate a maintenance plan according to the fault degree.

[0079] In the embodiment of the present invention, whether the winding fails is judged by temperature, amplitude and sound pressure level. If a failure occurs, the failure type of the winding is analyzed, and each potential failure type is analyzed. At the same time, the degree of the winding failure is analyzed. If the degree of the failure is high, the matching degree of each standby winding is analyzed, and a standby winding with a high matching degree is randomly selected for replacement. If the degree of the failure is low, the failure type of the winding, each potential failure type and the operating environment are obtained, and a maintenance plan is formulated. It can accurately judge whether the winding fails, and ensures the operation effect of the power transformer after the winding is replaced and the effectiveness of the maintenance plan.

[0080] The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of the present technology can make various modifications or supplements to the described specific embodiments or use similar methods for substitution, as long as they do not deviate from the concept of the invention or exceed the scope defined by this specification, they should all belong to the protection scope of the present invention.

Claims

1. An Internet of Things-based power transformer fault detection system, characterized in that, It includes the following modules: The data acquisition module is used to set each acquisition point at a preset distance interval and collect the temperature, amplitude, and sound pressure level of each acquisition point in real time; The data analysis module is used to obtain the temperature, amplitude, and sound pressure level of each acquisition point at each operating moment, analyze whether the windings of the power transformer are faulty. If a fault occurs, the winding analysis module is executed and a warning is given; The winding analysis module includes a correlation analysis unit and a fault analysis unit: The correlation analysis unit is used to obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and each secondary fault type. The main fault type and each secondary fault type during each historical fault of the winding are combined pairwise to obtain various fault combinations. The various fault combinations are compared, and the fault combinations with the same main fault type are classified into one category. In this way, various fault combinations are obtained. The various fault combinations in each type of fault combination are compared, and the fault combinations with the same combination are called the same kind of fault combination. In this way, each same kind of fault combination in each type of fault combination is obtained, and the number of fault combinations in each same kind of fault combination and the number of fault combinations in each type of fault combination are counted. In each type of fault combination, the ratio of the number of fault combinations in each same kind of fault combination to the number of fault combinations in the fault combination class where the same kind of fault combination is located is calculated, and it is called the proportion of each same kind of fault combination. In this way, the proportion of each same kind of fault combination is obtained. A certain same kind of fault combination is compared with other same kind of fault combinations, and other same kind of fault combinations are called each compared same kind of fault combination. If the main fault type in this same kind of fault combination is the same as the secondary fault type in a certain compared same kind of fault combination, and the secondary fault type in this same kind of fault combination is the same as the main fault type in this compared same kind of fault combination, then this same kind of fault combination and this compared same kind of fault combination are called an associated combination. This same kind of fault combination is called the first fault combination, and this compared same kind of fault combination is called the second fault combination. In this way, each associated combination is obtained. The proportion of the first fault combination and the proportion of the second fault combination in each associated combination are obtained. According to the analysis formula: The correlation value κ of the c-th associated combination is obtained c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation value of each associated combination represents the correlation value of each fault type in each associated combination. In this way, the correlation value between each fault type is analyzed; The fault analysis unit is used to collect various parameters of the winding, analyze the fault type of the winding, and analyze other possible fault types caused by this fault type of the winding; The solution formulation module is used to obtain the operating parameters of the winding, analyze the degree of winding fault, and formulate a maintenance plan according to the degree of fault; The warning terminal is used to give a warning when a fault occurs in the winding; The database is used to store various fault types that occurred during each historical fault of the winding, the data of the winding when no fault occurred, and various maintenance plans.

2. The power transformer fault detection system based on the Internet of Things according to claim 1, characterized in that The specific process of analyzing whether the windings of the power transformer are faulty is as follows: Obtain the temperature, amplitude, and sound pressure level of each collection point at each operating moment, and input them into the sound anomaly model and vibration anomaly model, and output the sound anomaly value α of each collection point a and the vibration anomaly value β of each collection point a , when α a = 1 or β a = 1, it represents that the sound of the a-th collection point is abnormal or the vibration of the a-th collection point is abnormal. When α a = 0 or β a = 0, it represents that the sound of the a-th collection point is normal or the vibration of the a-th collection point is normal, where a represents the number of each collection point, a = 1, 2, 3,..., b, b represents the total number of collection points, and both a and b are positive integers; Obtain the temperature, sound anomaly value, and vibration anomaly value of each acquisition point at each operating moment, input them into the winding anomaly model, and output the anomaly value of the winding. If the anomaly value of the winding is 1, it means that the winding is faulty. If the anomaly value of the winding is 0, it means that the winding is not faulty.

3. An Internet of Things-based power transformer fault detection system according to claim 1, characterized in that, The specific process of the fault analysis unit is as follows: Obtain the DC resistance, dielectric loss factor, no-load current, and no-load loss of the winding, and analyze the fault type of the winding according to the obtained data. At the same time, according to the fault type of the winding, obtain various possible fault types of the winding, and call them various potential fault types.

4. The fault detection system for a power transformer based on the Internet of Things according to claim 3, characterized in that The specific process of analyzing the fault type of the winding is as follows: Obtain the DC resistance, real-time dielectric loss factor, real-time no-load current, and real-time no-load loss of the winding of the power transformer when no fault occurs in the winding from the database, calculate the average values of the real-time DC resistance, real-time no-load current, and real-time no-load loss, and use these average values as the standard values of the DC resistance, no-load current, and no-load loss. Compare the dielectric loss factors at each operating moment to obtain the standard range of the dielectric loss factor; Obtain the DC resistance of the winding, the dielectric loss factor of the winding, the no-load current and no-load loss of each phase winding, then: Where C represents the DC resistance of the winding, D represents the dielectric loss factor of the winding, E represents the no-load current of the winding, F represents the no-load loss of the winding, C′ represents the standard value of the DC resistance, D′ represents the standard range of the dielectric loss factor, E′ represents the standard value of the no-load current, F′ represents the standard value of the no-load loss, and δ represents the return value of the winding fault type, represents the preset threshold value of the DC resistance difference, represents the preset threshold value of the no-load current difference, represents the preset threshold value of the no-load loss difference; When the return value of the fault type of the winding is 1, it means that the fault type of the winding is an open circuit fault. When the return value of the fault type of the winding is 0, it means that the fault type of the winding is an insulation fault. When the return value of the fault type of the winding is -1, it means that the fault type of the winding is a short circuit fault.

5. The fault detection system for a power transformer based on the Internet of Things according to claim 1, characterized in that The specific process of formulating the maintenance plan is as follows: Obtain the values of the operating parameters of the winding at each operating moment when the power transformer has no fault from the database, compare them to obtain the standard numerical range of each operating parameter. At the same time, use each sensor to collect the operating parameters of the winding at this time, and calculate the return value of the degree of winding fault. When the return value of the degree of winding fault is 1, it means that the degree of winding fault is high. When the return value of the degree of winding fault is 0, it means that the degree of winding fault is low; When the fault degree of the winding is high, obtain the performance values and dimensions of each spare winding from the database, analyze the matching degree of each spare winding, and randomly select a spare winding with a high matching degree for replacement; When the fault degree of the winding is low, obtain each maintenance plan in the database, obtain the operating parameters of the winding after maintenance in each maintenance plan, calculate the maintenance effect score of each maintenance plan, obtain each fault type repaired in each maintenance plan, and compare it with the fault type of the winding and each potential fault type. If each fault type repaired in a certain maintenance plan includes the fault type of the winding and each potential fault type, then use this maintenance plan as a spare maintenance plan. Obtain the maintenance effects, maintenance durations, maintenance costs and each operating environment parameter of each spare maintenance plan, as well as each operating environment parameter of the winding, and perform normalization processing. Then: In the formula represents the value of the h-th operating environment parameter in the g-th alternative maintenance plan, J h represents the value of the h-th operating environment parameter of the winding, K g represents the maintenance cost of the g-th alternative maintenance plan, L g represents the maintenance duration of the g-th alternative maintenance plan, represents the practicality coefficient of the g-th alternative maintenance plan, G g represents the maintenance effect score of the g-th alternative maintenance plan. g represents the number of each alternative maintenance plan, g = 1, 2, 3,..., j, j represents the total number of alternative maintenance plans, h represents the number of each operating environment parameter, h = 1, 2, 3,..., i, i represents the total number of operating environment parameters, and g, h, j, and i are all positive integers; Compare the practicality coefficients of each main and spare maintenance plans, select the spare maintenance plan with the largest practicality coefficient, and use this spare maintenance plan as the final maintenance plan.

6. The fault detection system for a power transformer based on the Internet of Things according to claim 5, characterized in that, The specific process of calculating the return value of the fault degree of the winding is as follows: Obtain the numerical values of each operating parameter of the winding and the standard numerical range of each operating parameter, and perform normalization processing. Then: Where M k represents the value of the k-th operating parameter, M k ' represents the minimum value of the standard value range of the k-th operating parameter, M k '' represents the maximum value of the standard value range of the k-th operating parameter, ξ represents the preset fault degree index threshold, k represents the number of each operating parameter, k = 1, 2, 3,..., l, l represents the total number of operating parameters, both k and are positive integers, e represents the natural constant, and γ represents the fault degree return value of the winding.

7. The fault detection system for a power transformer based on the Internet of Things according to claim 5, characterized in that, The specific process of analyzing the matching degree of each spare winding is as follows: Obtain the performance parameter values and dimensions of each spare winding from the database. At the same time, obtain the standard numerical range of each performance parameter of the winding and the required dimensions of the winding, and perform normalization processing. Then: where N m represents the size of the m-th spare winding, represents the value of the n-th performance parameter of the m-th spare winding, Q n ′ represents the minimum value of the standard numerical range of the n-th performance parameter of the winding, Q n ″ represents the maximum value of the standard numerical range of the n-th performance parameter of the winding, N′ represents the required size of the winding, ζ represents the preset performance coefficient threshold, η m represents the matching degree coefficient of the m-th spare winding, m represents the number of the spare winding, m = 1, 2, 3,..., q, q represents the total number of spare windings, n represents the number of each performance parameter of the winding, n = 1, 2, 3,..., p, p represents the total number of each performance parameter of the winding, m, n, q, and p are all positive integers; When the matching degree coefficient of a certain spare winding is 1, it means that the matching degree of this spare winding is high. When the matching degree coefficient of a certain spare winding is 0, it means that the matching degree of this spare winding is low. Analyze the matching degree of each spare winding in this way.

8. An Internet of Things-based power transformer fault detection system according to claim 6, characterized in that, The specific process of calculating the maintenance effect score of each maintenance plan is as follows: Obtain the operating parameters of the winding after maintenance. Then: where represents the value of the k-th operating parameter of the winding after repair in the g-th alternative repair plan, M k ″′ represents the optimal value of the k-th operating parameter, G g represents the repair effect score of the g-th alternative repair plan.

9. A fault detection method for an Internet of Things-based power transformer fault detection system according to any one of claims 1-8, characterized in that, Including: Step 1, data collection: Set each collection point at a preset distance interval, and collect the temperature, amplitude and sound pressure level of each collection point in real time; Step 2, data analysis: Obtain the temperature, amplitude and sound pressure level of each collection point at each operating moment, analyze whether the winding of the power transformer fails. If it fails, execute the winding analysis module and give an early warning; Step 3. Correlation analysis: Obtain various fault types during each historical fault of the winding from the database, and classify them into the main fault type and various secondary fault types. Combine the main fault type and various secondary fault types during each historical fault of the winding in pairs to obtain various fault combinations. Compare the various fault combinations, divide the fault combinations with the same main fault type into one category. In this way, obtain various categories of fault combinations. Compare the fault combinations in each category of fault combinations, and call the fault combinations with the same combination the same-kind fault combinations. In this way, obtain the same-kind fault combinations in each category of fault combinations, and count the number of fault combinations in each same-kind fault combination and the number of fault combinations in each category of fault combinations. In each category of fault combinations, calculate the ratio of the number of fault combinations in each same-kind fault combination to the number of fault combinations in the fault combination category where the same-kind fault combination is located, and call it the proportion of each same-kind fault combination. In this way, obtain the proportion of each same-kind fault combination. Compare a certain same-kind fault combination with other same-kind fault combinations, and call the other same-kind fault combinations the compared same-kind fault combinations. If the main fault type in this same-kind fault combination is the same as the secondary fault type in a certain compared same-kind fault combination, and the secondary fault type in this same-kind fault combination is the same as the main fault type in this compared same-kind fault combination, then call this same-kind fault combination and this compared same-kind fault combination an associated combination, call this same-kind fault combination the first fault combination, and call this compared same-kind fault combination the second fault combination. In this way, obtain each associated combination, obtain the proportion of the first fault combination and the proportion of the second fault combination in each associated combination. According to the analysis formula: Obtain the correlation value κ of the c-th associated combination c , where A1 represents the proportion of the first fault combination, A2 represents the proportion of the second fault combination, c represents the number of each associated combination, c = 1, 2, 3,..., d, d represents the total number of associated combinations, and both c and d are positive integers. Then the correlation values of each associated combination represent the correlation values of various fault types in each associated combination. Analyze the correlation values between various fault types in this way; Step 4, fault analysis: Collect each parameter of the winding, analyze the fault type of the winding, and analyze other fault types that may be caused by this fault type of the winding; Step 5, formulate a plan: Obtain the operating parameters of the winding, analyze the fault degree of the winding, and formulate a maintenance plan according to the fault degree at the same time.

Images