Life estimation of electrolytic capacitors

Abstract

A device for estimating the life consumption of one or more electrolytic capacitors, comprising a temperature sensor thermally insulated from the surrounding environment by being arranged between two equal electrolytic capacitors connected in parallel with each other or on the housing of an electrolytic capacitor covered with a layer of thermally insulating material. The device further comprises a controller adapted to estimate the life consumption based on the measurement data of the temperature sensor.

Description

[0001] This application is a divisional application. The original application has the application number 202210719344.2 and the original application date is June 23, 2022. The entire contents of the original application are incorporated herein by reference. Technical Field

[0002] The present invention relates to an apparatus and method for estimating the lifespan of an electrolytic capacitor, and to an apparatus for using an electrolytic capacitor and including an apparatus for estimating its lifespan according to the present invention. Background Technology

[0003] It is well known that electrolytic capacitors age, and this aging depends on the temperature of the electrolyte. Electrolytic capacitor manufacturers typically provide their customers with formulas or noctilinear graphs to estimate lifetime at a given temperature. Similar to the Arrhenius equation, which describes the rate of a chemical reaction as a function of temperature, lifetime is estimated using an exponential function.

[0004] Documents DE 102012105198 B4 (ebm-papst), DE 112018003260 T5 (OMRON), DE102004035723 (Siemens), and JP H 0627175 (Toyoda) teach the use of temperature sensors located at different locations on the device to measure temperature values ​​that can be used to estimate lifespan consumption.

[0005] DE 10 2012 105 198 B4 (ebm-papst) teaches that electrolyte temperature is important for lifespan and notes that this temperature is affected by ambient temperature and ripple current. However, the document does not address methods for determining this ambient temperature.

[0006] DE 11 2018 003 260 T5 (OMRON) focuses primarily on determining the ambient temperature from temperature readings of a sensor located inside the device. In one embodiment, the document teaches that the temperature sensor is bonded to one side of one of the electrolytic capacitors, and that this measurement, in addition to the ambient temperature, is used to estimate the remaining lifespan of the device.

[0007] DE 10 2004 035 723 (Siemens) considers the case of an electric motor. The power stored in the capacitor and the ambient temperature are the input parameters for a thermal model that determines the capacitor core temperature. The capacitor core temperature is then used to estimate the remaining lifespan. The ambient temperature is assumed to be the temperature of the circulating coolant used to cool the motor.

[0008] JP H 0 627 175 (Toyoda) ultimately proposed two different methods: In the first example, a temperature sensor is placed inside the capacitor to directly measure the core temperature. In the second example, a detector measures the ripple current, and a temperature sensor measures the temperature near the capacitor's periphery; these two measurements are used to estimate the core temperature.

[0009] Lifetime consumption is exponentially related to core temperature. Therefore, uncertainty in estimating core temperature leads to significant uncertainty in estimating lifetime. Since lifetime consumption integrates over time to produce total lifetime consumption, the uncertainty can reach values ​​that render the results essentially useless. Placing sensors inside the capacitor is technically challenging and carries the risk that the electrolyte evaporates faster than without sensors due to additional openings in the seal, limiting the capacitor's lifespan. Using thermal models of the device's interior is difficult because, generally, different electrical components generate different amounts of heat under different conditions. Furthermore, in most cases, how a given device will be positioned in its final usage space is unknown, so the cooling efficiency of different parts of the device casing is often unknown to the manufacturer. Therefore, thermal models that correlate a temperature measured at some location inside or outside the device with the core temperature of the electrolytic capacitor located at a specific location inside the device have significant uncertainty. Due to the exponential relationship, the large uncertainty in electrolyte temperature leads to even greater uncertainty in estimating lifetime consumption. Since remaining lifetime is estimated using total consumed lifetime, which is an integral of many lifetime consumptions, the uncertainty increases even more, limiting the practical value of the estimate. If the relative uncertainty of the preliminary lifespan prediction is greater than + or -25%, then the remaining lifespan estimate preferably has limited practical value. Summary of the Invention

[0010] Therefore, the object of the present invention is to create a device for estimating the life of one or more electrolytic capacitors, which allows for the estimation of the remaining life with small uncertainty and can be readily used in various devices.

[0011] In a preferred embodiment, the device according to the invention is designed such that, if the device is used in a regular, repetitive cycle, the relative uncertainty of the preliminary lifespan prediction is between -10% and 10%, particularly preferably between -5% and 5%.

[0012] The solution of the present invention is specified by the features of the appended claims. According to the present invention, an apparatus for estimating the lifespan of one or more electrolytic capacitors (ELCOs) includes a temperature sensor and a controller adapted to estimate the lifespan consumption of the electrolytic capacitors based on measurement data from the temperature sensor.

[0013] The temperature sensor is insulated from its surroundings. This insulation is achieved by placing the temperature sensor between two equal electrolytic capacitors connected in parallel. Alternatively, the temperature sensor can be placed on the casing of the electrolytic capacitor and insulated by covering it with a layer of insulating material.

[0014] Both temperature sensor configurations ensure that the measured temperature is the temperature of the electrolytic capacitor's casing. This avoids localized heating caused by, for example, another nearby heat source. In one embodiment, a layer of insulating material prevents localized heating. In another embodiment, the temperature sensor is positioned between two equal heat sources that protect the sensor from localized heating or cooling by other heat sources, thus providing insulation against the surrounding environment.

[0015] Furthermore, the sensor setup is independent of the size of the device used, the environment, and the configuration of other components. The casing of an electrolytic capacitor is typically made of metals such as aluminum, which have excellent thermal conductivity. Therefore, temperature variations with the casing are small, and measurements taken at insulated locations are representative of the casing temperature. Consequently, the thermal model used as the basis for estimating remaining life depends solely on the characteristics of the electrolytic capacitor. Therefore, this device can be used in many different devices and situations.

[0016] In the preferred embodiment, where a temperature sensor is positioned between two identical electrolytic capacitors connected in parallel, accurate measurement is allowed without reducing the surface area available for cooling the electrolytic capacitors. Because the electrolytic capacitors are identical and because they are connected in parallel, they may have substantially the same case temperature after an initial time interval in which one electrolytic capacitor ages faster than the other. Therefore, the sensor positioned between them is insulated from other heat sources or heat sinks.

[0017] In a preferred embodiment where the temperature sensor is positioned between two equal electrolytic capacitors, the capacitors are approximately cylindrical with radius R and height h. The capacitors are arranged so that their longitudinal axes are parallel and they are spaced 2D apart. In this case, the temperature sensor, in a direction perpendicular to the line connecting the two longitudinal axes, is less than the square root of 8 and the distance 2D between the two capacitors. Further, in a direction parallel to and perpendicular to the line connecting the two longitudinal axes, the temperature sensor is less than the difference between the height h and the distance 2D minus twice the radius R. The temperature sensor is preferably placed within a cuboid volume, the width of which (B...)... Q ) equals the square root of 8 multiplied by the difference between the radius and the distance 2D (B) Q = R-2D), the height of the cuboid's volume (H) Q The height of the capacitor is equal to the difference between its distance 2D and twice the radius R (H). Q= h-(2D-2R)), and the depth of the cuboid's volume (T) Q The difference is equal to twice the square root of the difference between the distance 2D between the two capacitors and the square of the radius R and the square of half the width of the volume (T). Q = 2D-2 The volume is located between the two capacitors such that its height is parallel to the longitudinal axis of the two capacitors and its depth is parallel to the line connecting the two capacitors. The center of the volume is located at the middle of the line connecting the two capacitors. The capacitors are partially located within the volume, allowing the temperature sensor according to this embodiment to be mounted on the housing of one of the capacitors.

[0018] When the capacitor shape is different, the volume into which the temperature sensor should be placed is preferably estimated to be a set of points with the following characteristics: the aperture of each of these points is not blocked by the current capacitor, and there is at least one straight line perpendicular to the observation direction in each observation direction, limited to at least 90°.

[0019] The results showed that by selecting the location of the temperature sensor within the volume, the desired thermal insulation effect of the capacitor was particularly significant.

[0020] By placing the temperature sensor on the casing of the electrolytic capacitor and insulating it by covering it with a layer of insulating material, the device can include only one electrolytic capacitor. To minimize the loss of cooling capacity, the temperature sensor is preferably placed in an area that is difficult to cool even without insulating material, such as between the electrolytic capacitor and the mounting plate or component of the device's casing.

[0021] In the following description, the lifespan of a capacitor is described by different parameters:

[0022] "Lifetime" is an estimated lifespan of a capacitor under current conditions, covering its entire service life. Lifetime is expressed in units of time, such as hours. As conditions change over time, lifetime will vary and may increase or decrease.

[0023] "Total lifespan" is the time a capacitor has operated under conditions that have already affected it or will affect it. A clock that starts when the capacitor is manufactured and stops when the capacitor fails will display the "Total lifespan." Total lifespan is expressed in units of time.

[0024] "Basic life" L0 is the current life of the capacitor at its maximum rated category temperature T0 and subjected to its reference ripple current I. S Lifespan under influence. Basic lifespan is expressed in units of time, such as hours. Basic lifespan is a parameter in models used to estimate lifespan. Therefore, in many cases, unless there is an intentional change in units, the unit of basic lifespan is lifespan.

[0025] The "life" of a capacitor refers to the time from its manufacture to the present. Life is expressed in units of time.

[0026] "Maximum service life" refers to the duration a capacitor can function properly when unused and stored at room temperature without specific protection or exposure to extreme conditions. Maximum service life is expressed in time units, typically years. The typical maximum service life of a standard electrolytic capacitor is approximately 15 years.

[0027] "Lifespan loss" is the fraction of lifespan lost per unit of time. It can be expressed as a percentage or a fraction of lifespan per unit of time. Lifespan loss depends on current conditions. Lifespan loss is inversely proportional to lifespan.

[0028] "Total lifespan depletion" is the lifespan of the capacitor since it was first put into use. It is the integral of lifespan depletion over time, from the start of the capacitor's use until the present moment. Typically, the value of total lifespan depletion is between 0 and 1, or 0% and 100%, depending on the unit chosen. If the total lifespan depletion is greater than 1 or 100%, the risk of capacitor failure or non-nominal behavior is high. The time required for total lifespan depletion to rise from 0 to 1, or from 0% to 100%, is the total lifespan if this time is shorter than the maximum lifespan.

[0029] "Usage Cycle Lifetime Consumption" refers to the lifetime consumption within a single usage cycle.

[0030] "Preliminary lifespan forecast" is the predicted duration for total lifespan depletion from 0 to 1, or from 0% to 100%. This forecast is based on an inference of total lifespan depletion. Preliminary lifespan forecast is expressed in units of time. As the inference of total lifespan depletion may be revised, the preliminary lifespan forecast may change with the lifespan of the capacitor. If the lifespan of the capacitor is not limited by its maximum lifespan, the preliminary lifespan forecast is an estimate of the total lifespan.

[0031] "Remaining lifetime" is the smaller difference between the initial projected lifetime and the maximum lifespan, and the lifespan of the capacitor. It is expressed in units of time.

[0032] In this document, a “failed” electrolytic capacitor is defined as a capacitor whose characteristics, such as capacitance or equivalent series resistance, have undergone an unacceptable change relative to its nominal value. While acceptable deviations from the nominal value depend on the application, these deviations typically increase rapidly once the total lifespan is reached. Therefore, the exact choice of the threshold for declaring a capacitor as failed is irrelevant here. Preferably, the basic lifespan L0, the maximum lifespan, and the lifespan estimate use the same threshold to define “failure.”

[0033] In a preferred embodiment, the device includes instruments capable of measuring output or input data. The output or input data includes voltage, power, or current affected by or influencing one or more electrolytic capacitors. The controller in this embodiment is adapted to include the measured output or input data in its estimate of the lifespan of the electrolytic capacitors.

[0034] In particular, aluminum electrolytic capacitors generate internal heat when ripple current flows through them. Measuring the output or input data, along with the known characteristics of the device using the electrolytic capacitor, and therefore the implicit information about how the measured data relates to the ripple current flowing through the capacitor, allows for the estimation of this internally generated heat, thus enabling the use of a more accurate thermal model of the electrolytic capacitor.

[0035] Preferably, if changes in voltage, power, or current alter the ripple current through the capacitor, then the voltage, power, or current will affect or be affected by the capacitor. In addition to the magnitude of the voltage, power, or current, the measured characteristics can be frequency, mode, waveform, and / or number of phases.

[0036] The heat generated by ripple current depends on its frequency. A given device, with its input and output specified, will react in a similar manner over its lifetime. Therefore, if the input and output of the device are the same, the ripple current flowing through the electrolytic capacitor will be substantially the same.

[0037] In many cases, since the device is typically built for specific input conditions that are known, such as single-phase or three-phase voltages with a given RMS voltage and frequency, and known line voltage waveforms or patterns, etc., and the number of phases and patterns of the output are usually given by the circuit and device specifications, in such embodiments, measurements of the specified output voltage, power, or current are sufficient to determine the condition of the device and derive the effective ripple current therefrom.

[0038] In many cases, the output conditions are known because devices are typically built to provide specific output conditions, such as single-phase or three-phase voltages with a given RMS voltage, mode, frequency, and known output voltage waveform. Furthermore, the number of phases and mode of the input are usually given by the device specifications. Therefore, in such embodiments, measurements of the specified input voltage, power, or current are sufficient to determine the device's condition and derive the effective ripple current.

[0039] The input or output waveform can be, for example, a sine wave, a square wave, a triangle wave, or a flat wave. The input or output mode can be, for example, DC or AC voltage. Preferably, the controller includes or has access to a memory that provides a correlation between measured characteristics such as output voltage, current, or power and effective ripple current. This correlation can be formed, for example, in the form of a lookup table or a formula.

[0040] Preferably, this correlation is derived by the device manufacturer and read into the controller when the device is installed in a given device. This correlation can be determined by measuring the ripple current in a laboratory at a sufficiently high time resolution for a representative set of possible input and output conditions for the current device. The measured ripple current is analyzed, for example by Fourier transform, to obtain the frequency and RMS value for each relevant frequency. The equivalent series resistance (ESR) of the electrolytic capacitor depends on the ripple current frequency. This dependence is measurable. The frequency multiplier for each frequency can be obtained by calculating the square root of the fraction of the ESR at the standard frequency with the ESR at the frequency in question. The effective ripple current is defined as the square root of the sum of the squares of the quotients of the ripple current at a given frequency and the frequency multiplier at the frequency in question.

[0041] This embodiment is computationally efficient by directly storing the effective ripple current: no detailed frequency analysis is required, nor is it necessary to calculate the effective ripple current during nominal operation.

[0042] In other embodiments, the device's outputs and inputs are always the same or there are only a finite set of different inputs and outputs during the use of the current device. In such embodiments, the lifetime consumption estimate may be based purely on the housing temperature measurement, and in some embodiments, on information applied based on one of the input and output conditions from a given set. In other embodiments, the current input and output conditions are determined based on the development of the housing temperature. In all these embodiments, the effective ripple current value may be explicitly obtained by the controller, or the influence of one or a few of these effective ripple current values ​​may be implicitly incorporated into the way the controller estimates the lifetime consumption.

[0043] In a device where it is known that only a very small amount of power is dissipated inside the capacitor, the effect of internal heating can be neglected in the estimation and can be considered as part of the uncertainty.

[0044] The device according to the invention comprises at least one electrolytic capacitor and the apparatus according to the invention.

[0045] In many cases, the lifespan of electrolytic capacitors limits the lifespan of devices that include such capacitors. Equipping a device with the device according to the invention allows for the estimation of when the device or its electrolytic capacitors need to be replaced.

[0046] In one embodiment, the device includes first and second detectors or ammeters. The controller is adapted to include measurements from the first and second detectors and the ammeter in the estimation of lifetime consumption.

[0047] The first detector is suitable for measuring the input voltage, input current, and / or input power of the device. The second detector is suitable for measuring the output voltage, output current, and / or output power of the device. The ammeter is suitable for measuring the current through one or more electrolytic capacitors.

[0048] The use of first and second detectors allows the device to be used in devices where the characteristics of the input and output may change and are therefore not previously known to the device manufacturer. In these cases, the first and second detectors can detect which of the possible input and output conditions is given. For example, measurements can be used to detect the RMS value and number of phases of the input voltage or current, and to determine whether the device output is AC or DC voltage, whether the number of phases is applicable, and its voltage and / or current. In one embodiment, the measurements of the first detector are also used to detect the input frequency and line voltage waveform. In one embodiment, the measurements of the second detector are also used to detect the output frequency and output voltage waveform. The line voltage waveform can be, for example, sinusoidal, rectangular, or triangular in shape, or it can be a flat wave representing a DC voltage.

[0049] Preferably, the controller includes or has access to a memory that provides a correlation between the detected input and output characteristics and the effective ripple current. This correlation can be formed, for example, in the form of a set of lookup tables or one or more formulas. Preferably, if applicable, separate lookup tables or formulas exist for different possible combinations of input and output phases for AC and DC outputs, while the tables and / or formulas connect the RMS voltage, current, or power detected by the first and second detectors to the effective ripple current for the current device. Preferably, this correlation is derived by the device manufacturer and read into the controller when the device is installed in a given device.

[0050] Measuring the current through one or more electrolytic capacitors using an ammeter allows for direct measurement of ripple current. In embodiments where the device's response to input and output conduction is unknown, or where the device only has access to information directly around the electrolytic capacitors, directly measuring the ripple current and calculating the effective ripple current in the controller is preferred.

[0051] Ultimately, in some embodiments, either due to device specifications or due to the detection of the first and second detectors, the controller may possess information about the number of phases at the device input and characteristics at the output. In these cases, the controller can access a frequency multiplier that lists and sets of frequencies expected to appear in the ripple current. Only the contribution of these frequencies to the measured current is used to estimate the effective ripple current. This embodiment reduces the requirements on the first and second detectors, as they only need to detect the number of phases or distinguish between AC and DC voltages. The computational power required to analyze the ripple current is less than in the case where there are no prior assumptions about the occurrence frequencies.

[0052] In a preferred embodiment of the device according to the invention, the controller estimates the effective ripple current I by considering the output mode of the device and the number of phases and / or modes and / or waveforms on the device inputs, as well as the measured or given power or current or input voltage values. a .

[0053] The results showed that the frequency of the ripple current is strongly influenced by the number of phases, the mode, and sometimes the waveform of the device's input and output modes. In the preferred embodiment where the device is a power supply, the output or input mode can be AC ​​or DC current. If the number of phases in the input and output modes is known, the important frequencies and the distribution of the total current at these frequencies are also known. Only measured or known power, current, or input voltage values ​​are needed to scale this distribution.

[0054] Compared to evaluating the frequency of the ripple current in a series of ripple current measurements, using information about the number of phases, modes, and / or waveforms in the input and output modes is more computationally efficient and faster. Furthermore, the sampling frequency in the preferred embodiment can be significantly reduced compared to embodiments that measure the frequency of the ripple current.

[0055] In an even more preferred embodiment of the device according to the invention, the controller estimates the effective ripple current I by using one of a set of lookup tables. a The lookup table to be applied is selected based on the device's output mode and the number of phases, mode, and / or waveform at the device input. The lookup table returns the effective ripple current, and preferably returns the effective ripple current I at a given or measured output power value, a given or measured current value, or a given or measured input voltage value. a The uncertainty.

[0056] Since the number of possible input phases, modes, waveforms, and output modes is finite, storing this information in a lookup table allows for the reliable and rapid determination of the effective ripple current I for a given device in all possible scenarios. a However, this information can also be stored in the controller or in memory accessible to the controller in the form of a formula and / or a suitable correction factor.

[0057] In other embodiments, the ripple current is sampled at at least twice the highest frequency of interest in the current situation, and the measurements are analyzed to determine the ripple current at different frequencies. This analysis is preferably a Fast Fourier Transform (FFT). In this embodiment, the ESR value of the current capacitor at different frequencies, or the frequency multiplier, can be obtained by the controller, for example, by storing it in a table or function in memory that is part of the controller or accessible by the controller. The frequency multiplier for a given frequency is preferably a reference ripple current I. sThe effective ripple current I is determined by the square root of the ratio of the ESR value at frequency f0 to the ESR value at the given frequency. In this embodiment, the effective ripple current I is determined. a The square root of the sum of the squares of the ratios of the ripple current at each frequency to the squares of the frequency multiplier at the corresponding frequency is calculated. By determining the effective ripple current under the load condition of expected high-frequency ripple current, assuming a first value and a second value of the highest frequency of interest, and comparing the results, the highest frequency of interest can be determined iteratively. If the difference is less than an acceptable uncertainty, the smaller of the first and second values ​​of the highest frequency of interest is set as the highest frequency of interest for subsequent measurements.

[0058] In one embodiment, the device is a power supply. The power supply includes a rectifier, a large-capacity capacitor, a transformer, and preferably an output capacitor. The large-capacity capacitor includes at least one electrolytic capacitor. The lifespan of the large-capacity capacitor is estimated by the device according to the invention.

[0059] Once their lifespan is exhausted, the power supply will not fail completely, but will begin to exceed its specifications due to increased ripple and noise. In this case, estimating the lifespan is particularly helpful. Power delivery differing from expectations can cause unexpected behavior in devices using the power supply in question. Identifying the cause of such problems is often challenging. However, the device according to the invention points to the user the problem of degradation in the power supply's electrolytic capacitors, thereby aiding in repair.

[0060] A switching circuit can be used before the transformer to generate a high-frequency AC current that allows for efficient conversion to the desired voltage.

[0061] In one embodiment of the device according to the invention, the controller estimates the lifetime (L) of at least one electrolytic capacitor by calculating the product of a first capacitor type, which depends on a constant (C1), and first and second exponential functions. X The exponent of the first exponential function is the casing temperature (T) measured by the temperature sensor. c Preferably, the base of the first exponential function is a constant independent of the capacitor type (Q).

[0062] The exponent of the second exponential function is the square of the effective ripple current. Preferably, the base of the second exponential function is the second capacitor type-dependent constant (C2).

[0063] Written as a formula, the controller in this embodiment calculates... To estimate lifespan.

[0064] The first and second capacitor types depend on constants C1 and C2, which are specific to electrolytic capacitors, while the constant Q is the same for all capacitors. cThis is the casing temperature, i.e., the temperature measured by the sensor. Casing temperature T c It can be given in °C or K; however, the value of the constant C1 also depends on that choice. a This is the frequency-dependent effective ripple current. The effective current can be measured or estimated. a Preferably, the rated ripple current I is normalized to affect the type-dependent constant C2 of the second capacitor. S The frequency. This normalization can be achieved by multiplying by a correction factor. The correction factor preferably depends on the equivalent series resistance at different ripple current frequencies.

[0065] The constants can be determined experimentally, for example by observing the lifetime of a given capacitor at the same effective current but at different case temperatures to determine Q, and by observing the lifetimes of two different capacitors, preferably each at the same case temperature but at the first and second effective currents, to evaluate the capacitor dependence constants C1 and C2 for the two capacitors.

[0066] However, preferably, the constant is chosen to be equal to the following values:

[0067]

[0068] in , , I S These are parameters of the capacitor's thermal model. K represents the unit "Kelvin" and is explicitly stated to allow for checking the units in the results. Specifically, In the reference ripple current I S The temperature difference between the lower core and the surrounding environment. It is the thermal resistance between the outer casing and the surrounding environment. This refers to the thermal resistance between the core and the outer casing. Preferably, this model describes an idealized situation where the capacitor is mounted on a flat surface in a room with a constant temperature, such as 25°C, and a pressure of 1 bar. In this preferred model, the "surrounding environment" is the atmosphere of the room.

[0069] T0, L0, and ESR are parameters describing a capacitor: T0 is its maximum rated category temperature, and L0 is its basic life. The basic life L0 is the time between the ambient temperature equal to the maximum rated category temperature T0 and the reference ripple current I. S The lifespan of the capacitor. ESR is based on the reference ripple current I. S The equivalent series resistance of the capacitor measured at the specified frequency.

[0070] , , I SThe values ​​of ESR, T0, and L0 are usually provided by the electrolytic capacitor manufacturer.

[0071] Lifespan L x It is estimated using the formula given above in time units for the basic lifespan L0. Generally, the casing temperature T... C and / or effective ripple current I a It changes over time, therefore, at the shell temperature T C and / or effective ripple current I a Within a essentially constant time period Δt, only the estimated lifetime L was actually used. x Part of it.

[0072] The thermal model upon which these formulas are based represents a capacitor using a heat source within its core (i.e., dissipated ripple current) and a second heat source (i.e., the surrounding air). Between these two sources, heat is conducted from the core to the outer shell, and from the outer shell to the surrounding environment. The efficiency of this heat conduction is determined by thermal resistance. , describe.

[0073] Internal power losses and capacitor core heating are described by the equivalent series resistance (ESR). At the reference ripple current I... S Below, the power loss of the heating element is W = ESR The simulation model of Ohm's law can be applied to heat conduction under steady-state conditions, where the ratio of temperature difference to the heat energy conducted through a given structure is equal to its thermal resistance. This law applies to structures between the core and its surroundings, as well as structures between the core and its shell. If the ripple current is the same, the power loss W of the heating capacitor will naturally be the same in both cases. Therefore, for a given ripple current, Under the measured conditions, the temperature difference between the surrounding environment and the core... For a reference ripple current I with ripple frequency f0 S It is known, and the shell temperature T c The ripple frequency f is typically different from the ripple frequency f0. a The measurement is performed under different ripple current conditions. Typically, ESR depends on the ripple frequency. However, the effective ripple current I... a It is defined to include a multiplier that eliminates this frequency dependence of ESR. Therefore, if an effective ripple current I is used... a If we use this to characterize the ripple current, then the ESR is the same for all ripple current frequencies. Therefore, the formula given above can be modified to account for the effective ripple current I. S I a I b Differences in:

[0074]

[0075] This allows the ambient temperature, a thermal model parameter that is difficult to measure with sufficient accuracy, to be expressed as... A function of the given effective ripple current I. a Characteristics of capacitor and case temperature:

[0076]

[0077] In the effective ripple current I a Below, the temperature difference between the surrounding environment and the core. It can be derived from the same formula as follows:

[0078]

[0079] Experiments have shown that the lifespan of an electrolytic capacitor can be expressed as:

[0080]

[0081] Where T0 is the reference temperature, I S It is the reference ripple current. It is the reference ripple current I S The temperature difference between the capacitor core and its surrounding environment when current flows through it. K is the unit Kelvin. L0 is the ambient temperature T. a Equal to reference temperature T0 and effective ripple current I a Equal to the reference ripple current I S Lifespan.

[0082] Using the formula derived above, lifespan can be expressed as a function of casing temperature:

[0083]

[0084]

[0085]

[0086] Using the constants defined above, we obtain the lifetime and measured casing temperature T given above. c and the measured or estimated effective ripple current I a The relationship between them: .

[0087] In a preferred embodiment, the controller estimates that the lifetime decay (LC) of at least one electrolytic capacitor over a time period (Δt) is proportional to the quotient of the lifetime over the time period (Δt). The case temperature (T) is measured or assumed to occur during this time period. c ) and effective ripple current (Ia Determine lifespan.

[0088] Preferably, lifetime consumption is determined so that it can be integrated over time to estimate the remaining lifetime. Lifetime consumption (LC) is preferably expressed as a percentage or fraction of lifetime. Since lifetime may vary over time periods, lifetime consumption may also differ between subsequent time periods.

[0089] Lifetime consumption LC and lifetime L at a specific point in time X Inversely proportional: .

[0090] In a preferred embodiment, a time period Δt is selected such that the outer casing temperature T c and effective ripple current I a It remains substantially constant throughout its duration. In this preferred embodiment, the lifetime consumption during this time period is proportional to the quotient of the time period and the lifetime assessed using conditions within that time period: .

[0091] The subscript 'c' is used to indicate the continuous case, where lifetime and lifetime decay are determined individually for each time point. The subscript 'd' is used to indicate the discrete case, where time is divided into potentially small but measurable time intervals Δt.

[0092] Preferably, the lifetime consumption over a time period is expressed as a percentage, in which case... In another preferred embodiment, the lifetime consumption of a time period is represented by a fraction of 1, such that... .

[0093] Using the lifetime estimation expression given above, the lifetime consumption for the time period Δt is as follows: When lifespan consumption is expressed as a fraction of lifespan, it can be represented as follows: .

[0094] Therefore, lifespan can be estimated without having to determine the lifespan in a separate calculation beforehand.

[0095] In one embodiment of the device according to the invention, the controller considers the uncertainty of the housing temperature, the effective ripple current, and the uncertainty of a preferred time period to determine the uncertainty of lifetime consumption.

[0096] Tracking uncertainty helps in interpreting results. Since the measured parameters, such as the casing temperature and, preferably, at least indirectly measured, the effective ripple current, are all exponential functions of lifetime, the uncertainty in the results is difficult to estimate without calculation.

[0097] In a preferred embodiment, lifetime consumption is estimated over a typical time period. In such a preferred embodiment, the time period is also measured and therefore has a lifetime consumption LC that can be calculated in discrete cases.d The small uncertainty is considered in the total uncertainty of (Δt).

[0098] The propagation of uncertainty is preferably determined by means of a truncated Taylor series around the measured value, which, due to the positive and negative uncertainties of the parameters discussed above, evaluates once to give the most negative deviation and once to give the most positive deviation.

[0099] The lifetime consumption estimate in the discrete case is as follows:

[0100]

[0101] In this case, the truncated Taylor series result is:

[0102]

[0103] thereby

[0104] and

[0105] and

[0106] and

[0107] and

[0108] and

[0109] Therefore, the relative uncertainty of lifetime consumption in the discrete case is:

[0110]

[0111] In the case of a typical electrolytic capacitor, T0 = 105°C, I s =1.170A, at T0 and I s With ESR = 0.125Ω, =5.2°C, R thjc =7.01 K / W, R thca =23.4 K / W and L0=131400h, the constant is C1= C2=0.72, Q= =0.93.

[0112] For example, suppose the effective ripple current I a For 1A, T c If the temperature is 50°C, the relative error is as follows:

[0113]

[0114] This means that, in the discrete case, the uncertainty of temperature measurement generally dominates the uncertainty of lifetime consumption. However, as the effective ripple current increases, the uncertainty in C2 and the effective ripple current becomes more significant. Because it contains parameters describing the capacitor and a generally well-known temperature model, C2 typically has only a small uncertainty.

[0115] In a preferred embodiment, uncertainty propagation is thus simplified by ignoring the uncertainty in the constant, which is typically significantly smaller than the measurement uncertainty.

[0116]

[0117] The constants involved are defined as follows, with the logarithmic terms having a negative sign:

[0118]

[0119]

[0120] After rewriting, the effective ripple current I a Casing temperature T c The difference in lifetime loss caused by a small change in the time period Δt is as follows:

[0121]

[0122] Especially over longer time periods, the relative uncertainty within a given time period can be disregarded compared to the increase in relative uncertainty of the effective ripple current and temperature due to variations in these parameters.

[0123]

[0124] This further simplifies the analysis.

[0125] Since the time period is only a part of the lifetime consumption calculation in the discrete case, the uncertainty of lifetime consumption in the continuous case can be estimated using the formula given above by setting the relative uncertainty in the time period to zero.

[0126] By using a positive estimate difference and a negative estimate difference, the positive or negative difference in estimated lifetime consumption caused by measurement uncertainty can be estimated.

[0127] In the discrete case, it is preferable to determine the uncertainty in the positive or negative direction of lifetime consumption for each time period Δt.

[0128] The total lifespan consumption after n time periods is the sum of the lifespan consumption during those time periods:

[0129]

[0130] The uncertainty of the total lifetime consumption after n time intervals Δt is the sum of the uncertainties of lifetime consumption within each time interval:

[0131]

[0132] In a preferred embodiment, it is assumed that the positive estimate difference is equal to the absolute value of the negative estimate difference. In this case, only one uncertainty of lifetime consumption is calculated for each time period, and there is only one estimate for the total lifetime consumption.

[0133] In a preferred embodiment, the controller of the device according to the invention determines lifetime consumption, and preferably determines its uncertainty over a typical time period. Each time period is represented by average temperature and average effective current. The average temperature and average effective current, or average input voltage, are preferably the average of all suitable measurements obtained during the time interval. In this embodiment, total lifetime consumption is the sum of all previous lifetime consumption. The total lifetime consumption is preferably determined and stored.

[0134] Determining lifetime consumption within a sufficiently small, typical time period allows for the use of the above formulas and methods to determine lifetime consumption under constant case temperature and constant effective ripple current, as well as under conditions where case temperature and effective ripple current vary over time. The choice of time period makes the uncertainty arising from the assumption of constant conditions acceptable.

[0135] In a preferred embodiment, the uncertainty estimation formula discussed above can be used to estimate the shell temperature T by setting a reasonable time period length. C and effective ripple current I a The rate of change and the measurement uncertainty for the measurement time period Δt. Users can specify the acceptable uncertainty in the total lifetime consumption. If typical values ​​of the casing temperature and effective ripple current can be further estimated, the formula for estimating the uncertainty of lifetime consumption for a given time period can be reformulated as a quadratic equation for the time period Δt.

[0136] However, preferably, the length of the time period is set as a fraction of the typical time scale of the current device. Preferably, the time period Δt is set to 1 hour, 0.1 hour, 0.01 hour, or 1 / 100 of the duration of the cycle.

[0137] Using it for a regular period of time helps to control and evaluate the lifespan of one or more electrolytic capacitors that are part of the device according to the invention.

[0138] However, in other embodiments, the length of the time period can vary. In a preferred embodiment, a shorter time period is used in the first mode and a different length of time period is used in the second mode, such that the length of the time period in the second mode is preferably influenced by observations made during the application of the first mode. Preferably, the time period of the second mode is longer than the time period of the first mode if the lifetime consumption of the subsequent time periods of the first mode is substantially constant, and shorter than the time period of the second mode is shorter than the time period of the first mode if the lifetime consumption of the subsequent time periods of the first mode frequently varies by an amount exceeding a predefined amount. In this embodiment, the device defines a reasonable time period length itself. Therefore, this embodiment is preferred when the manufacturer is unaware of the use of the device.

[0139] In one embodiment, only one measurement is performed within each time period, and this measurement is a single case temperature reading. In this case, the average temperature is the result of this measurement, and the average effective ripple current is estimated if necessary. This embodiment requires very few measurements, and therefore requires very little computational power.

[0140] In another embodiment, a single measurement of the output power, current, or input voltage is also performed within each time period, and this measurement is used to estimate the average effective ripple current. This embodiment also requires very few measurements but allows for more accurate lifetime estimation of devices whose output power, current, and / or input voltage are unknown and difficult to estimate in advance.

[0141] In another embodiment, the casing temperature is measured multiple times within each time period. For example, ten casing temperature measurements are obtained within a time period, and the average casing temperature is determined by calculating the arithmetic mean of these values.

[0142] Similarly, the average effective ripple current can be evaluated using more than one measurement. However, multiple measurements can be used in different ways: In one embodiment, the ripple current is directly measured and multiple measurements are analyzed to determine the frequency involved. In this case, the multiple measurements produce only a single effective ripple current value, which is then assumed to be the average effective ripple current value. In another embodiment, each measurement or group of measurements of current, input voltage, and / or output power is used to determine multiple effective ripple current values ​​over a single time period, for example, by using a lookup table and information about the number of input phases and the output mode. In this case, the arithmetic mean of the multiple effective ripple current values ​​determined over a single time period can be used as the average effective ripple current. In yet another embodiment, the arithmetic mean of the measured values ​​(e.g., input voltage, current, and / or power) is determined before determining the effective ripple current based on this mean, for example, by using an appropriate lookup table, and is used as the average effective ripple current.

[0143] Before calculating the arithmetic mean of a set of housing temperature measurements and / or current, voltage, and power measurements or a set of effective ripple currents, filters can be applied to remove obviously defective data points or to apply higher weights to high-quality data points.

[0144] In a preferred embodiment of the device according to the invention, the user can specify a usage period. The usage period is expected to be repeated in the future.

[0145] The controller in this embodiment continuously determines the lifetime consumption within a time period or during the usage cycle, and preferably determines its uncertainty. The controller determines the usage cycle lifetime consumption and preferably its uncertainty by integrating or summing the lifetime consumption over the usage cycle.

[0146] Use of lifespan consumption (LC) uc The controller is used to linearly infer total lifetime consumption (LC). total (t) to determine the initial predicted lifetime (L) pp ).

[0147] The slope of this linear inference is preferably calculated using the cycle life consumption (LC). uc ) and the duration of the usage cycle (Δt) uc The quotient of ). Preferably, the function used for linear inference includes the use of the start of the period (t). b ) and total lifetime consumption at the start of the usage cycle (LC) total (t b )).

[0148] Preliminary lifespan prediction (L pp ) is from the start of the capacitor's lifespan to the end of its total lifespan (LC). total The inference of (t) is the duration between the value indicating the end of the lifespan. If lifespan consumption is defined as a fraction of lifespan, this value is 1; if lifespan consumption is defined as a percentage of lifespan, this value is 100%.

[0149] If lifespan depletion is defined as a fraction of lifespan, then the preliminary lifespan prediction is preferably calculated in the following manner:

[0150]

[0151] Preferably, the controller further includes a clock that counts the years (t) of one or more electrolytic capacitors.

[0152] Remaining lifespan L R (t) is estimated by the controller for the lifespan (t) of one or more electrolytic capacitors and the preliminary predicted lifespan (L). pp The longest lifespan t of electrolytic capacitors maxThe difference between the smaller one.

[0153]

[0154] In a device that does not have a clock to calculate the lifespan of one or more electrolytic capacitors, the lifespan can be determined by summing all previous time periods in the discrete case or by integrating a pair of times in the continuous case.

[0155] Preferably, they are returned together to the remaining lifetime L. R (t) and its uncertainty ΔL R (t). Preferably, the uncertainty ΔL R (t) is a function that uses a truncated Taylor series containing only linear terms to consider the uncertainty in determining the lifespan δt of one or more capacitors, and the longest lifespan Δt. max Uncertainty and preliminary predicted lifetime (ΔL) pp The uncertainty is determined by considering the use of cycle life consumption ΔLC. uc and usage cycle duration (ΔΔt) uc ) and years (Δt) b ) and lifetime consumption at the start of the recorded usage period (ΔLC) total (t b The uncertainty of )) is estimated in the same way as the preliminary predicted lifetime (ΔL) pp The uncertainty of ).

[0156]

[0157]

[0158] If the duration of the period (Δt) is used uc ) and the number of years at the start of the usage period (Δt) b The positive uncertainty of ) and the service life consumption ΔLC uc and the lifespan consumption at the start of the recorded usage period (ΔLC) total (t b The negative uncertainty of the initial lifetime prediction appears together with the negative uncertainty of the period (Δt). uc ) and the number of years at the start of the usage period (Δt) b The negative uncertainty of ) and the service life consumption ΔLC uc and the lifespan consumption at the start of the recorded usage period (ΔLC) total (t b The positive uncertainty of the initial lifetime prediction appears together with the positive uncertainty of the initial lifetime prediction.

[0159] In a preferred embodiment, the remaining lifetime is measured as the duration of the recorded usage cycle (Δt). uc (in units)

[0160]

[0161] If we assume that the cycle used repeats again and again since the beginning of the record, then the expression It simply refers to the number of usage cycles that have occurred since the start of the recorded usage cycle.

[0162] This embodiment allows users to obtain an estimate of remaining lifespan in terms of usage cycles. This figure is typically more comprehensive than hours, as users are likely to schedule maintenance and repairs between usage cycles. Furthermore, in most cases, calculating usage cycles yields results with less uncertainty compared to measuring the current time and determining the duration of the usage cycle.

[0163] Most devices, especially power supplies, are used repeatedly in very similar ways. For example, a machine that uses a power supply may be turned on every morning and turned off every night, or the load on a machine may vary in a recurring pattern every day, week, or hour, or within the time frame required to complete a specific task.

[0164] In the manner of using the device according to the invention, the usage cycle is a repeating pattern.

[0165] Defining and documenting such usage cycles improves the quality of remaining lifetime estimates: lifetime consumption per unit time can vary significantly over a usage cycle. Estimating remaining lifetime by interpolating an average over an arbitrary time span carries a significant risk that this arbitrary time span may contain many periods with either high lifetime consumption or very low lifetime consumption. Both can lead to unrealistic estimates of remaining lifetime.

[0166] By allowing users to define or record specific usage cycles, meaningful predictions of remaining lifespan can be generated using the determined usage lifespan consumption observed during the usage cycle.

[0167] Furthermore, if the usage cycle changes, this embodiment allows the user to obtain updated predictions, for example, when the device is used for other purposes: once the controller registers a new usage cycle, previous lifetime consumption will not affect the inferred slope for the future.

[0168] In a preferred embodiment, subsequent usage cycles are used to refine the estimate of usage cycle lifetime consumption. In this embodiment, the controller continuously determines whether the measurement is well represented by the defined usage cycles. If so, the observed usage cycle lifetime consumption is used to calculate an average usage cycle lifetime consumption, and this average usage cycle lifetime consumption is used to estimate the remaining lifetime.

[0169] In another preferred embodiment, although all usage cycles are well represented by defined usage cycles, the controller determines whether the usage cycle lifetime consumption is progressing systematically from one usage cycle to another. In this case, the controller can estimate the remaining lifetime using the observed lifetime consumption of the last measured usage cycle or the average of the last few measured usage cycles. This embodiment allows for automatic improvement of the remaining lifetime estimate under slowly changing background conditions such as, for example, seasonal variations leading to different ambient temperatures or heating due to aging of surrounding components. To detect system progression, the controller preferably stores the usage cycle lifetime consumption of all detected usage cycles.

[0170] Preferably, if the change in housing temperature over time is similar to one of the recorded or defined usage cycles, while accepting a larger difference in the absolute value of the measurement, then the measurement is well represented by the recorded or defined usage cycle. Preferably, the user can set a first threshold for the change in housing temperature measurement and / or ripple current estimation and a second threshold for the absolute values ​​of housing temperature, effective ripple current, and / or lifetime. In this embodiment, if the difference in change is less than the first threshold and if the difference in absolute value is less than the second threshold, the controller assumes that the observed usage cycle is well represented by the recorded or defined usage cycle.

[0171] Preferably, the time period Δt used in determining the lifetime consumption of the usage cycle can be shorter than the time period used during subsequent use of the device, thereby making the uncertainty of the lifetime consumption used to define the usage cycle lower than the uncertainty of subsequent usage cycles. In this embodiment, the determination of lifetime consumption in subsequent usage cycles can be primarily used to determine that subsequent usage cycles are well represented by the recorded usage cycles.

[0172] Preferably, all time periods are integrated to generate the total lifetime consumption. Specifically, time periods that are not considered part of the usage cycle or belong to a previously defined usage cycle are also included in the total lifetime consumption, so that incomplete usage cycles only affect the total lifetime consumption and not the lifetime consumption of each usage cycle.

[0173] In a preferred embodiment, the controller records the measured or estimated lifetime consumption of the housing temperature, and preferably records the estimated effective ripple current for each usage cycle during its specified period. These recorded data sequences are compared with data sequences generated over a period assumed to be a repeating usage cycle. In this preferred embodiment, if the controller detects a deviation between the recorded and measured data greater than a threshold, the controller signals the user, thereby warning the user of quality loss in the remaining lifetime estimate. Preferably, the controller signals the user to specify a new usage cycle.

[0174] In a further preferred embodiment, the controller compares the stored lifetime consumption and its uncertainty with the current lifetime consumption within the defined lifetime duration. If the difference between the detected lifetime consumption and the stored lifetime consumption is greater than the estimated uncertainty, and if this occurs for more than a preset time, the controller issues a warning and suggests recording and defining a more suitable lifetime.

[0175] In a particularly preferred embodiment, during the period when the controller detects a significant difference between the recorded and generated data sequences, the controller attempts to timely shift the recorded data sequence to find a phase shift that minimizes the difference between the current state and the recorded data sequence. If the deviation between the recorded but shifted data and the measured data is less than a threshold, the preliminary predicted lifetime and / or remaining lifetime are updated to include years and total lifetime consumption at the point in time when the generated data sequence is again well represented by the recorded usage period, and the controller preferably stops issuing warnings.

[0176] In a preferred embodiment, the preliminary predicted lifetime and / or remaining lifetime inferences are periodically updated so that the inferences take into account the latest available data on years and total lifetime consumption.

[0177] In the preferred case of linear inference, the preliminary predicted lifetime and / or remaining lifetime are estimated in this embodiment using the following formula, where t u This is the latest available data regarding years and total lifespan.

[0178]

[0179]

[0180] By periodically updating the remaining lifetime and / or preliminary predicted lifetime, preferably after the duration of a specified usage cycle, the controller can consider the impact of individual events on the remaining lifetime. For example, skipping or terminating a usage cycle will extend the remaining lifetime, and updates according to this embodiment allow this impact to be taken into account.

[0181] In a preferred embodiment, the controller includes an interface and stores the housing temperature, preferably, the effective ripple current, and the lifetime consumption for each time period from which these values ​​are obtained. These values, or statistics derived from them, can be exported via the interface. In one embodiment, the interface is an I / O link.

[0182] By considering not only lifespan consumption but also the values ​​of the casing temperature and effective ripple current used to estimate lifespan consumption, users can detect and analyze long-term variations in the device. For example, long-term variations in casing temperature may be caused by seasonality or by the degradation of surrounding components within the device, resulting in more heat being generated than before. Furthermore, users can use the data to optimize the use of the device to extend its lifespan.

[0183] Instead of storing and exporting each data value, exporting and storing statistics derived from them, or a combination of both, conserves memory and allows for a faster overview of desired characteristics. For example, a fixed number of datasets can be stored, and the early history of a capacitor can be recorded by storing only the type of usage cycle and its deviations, or the average case temperature over the entire usage cycle, for each usage cycle. Data can also be categorized to produce histogram data, such as storing the number of time periods with a given lifetime consumption, or the number of time periods with a specific case temperature or a specific assumed effective ripple current. Another statistic that helps in estimating the impact of time period length selection could be a histogram showing the differences in lifetime consumption between subsequent time periods.

[0184] This interface can also preferably be used to customize assessments of lifetime depletion, remaining lifetime, and / or uncertainty estimates. For example, a user can specify a method for estimating the effective ripple current or updating constants describing the capacitor and heat models. They can select different time periods, define usage cycles and thresholds, set the time interval for updating remaining lifetime, and adjust the formulas to infer the total lifetime.

[0185] In a preferred embodiment, the controller includes, in addition to the interface, a button that allows specifying a usage cycle and a display or light indicator that indicates the remaining lifespan. For example, a usage cycle can be defined as follows: pressing the button once at the start of the usage cycle, pressing the button twice at the end of the usage cycle, and the light indicator emitting a green light as long as the remaining lifespan exceeds 25% of the shorter of the initial predicted lifespan and the maximum service life; turning orange once that fraction decreases to 10%; and turning red once there is no remaining lifespan. A flashing light indicator can signal that the currently used usage cycle does not appear to accurately reflect the current use of the device.

[0186] Other advantageous embodiments and combinations of features are derived from the following detailed description and all claims. Attached Figure Description

[0187] The accompanying drawings, used to explain the embodiments, show:

[0188] Figure 1 This is a schematic diagram of an electrolytic capacitor;

[0189] Figure 2a , 2b 2c is a diagram illustrating the location of the housing temperature measurement and temperature sensor according to the present invention;

[0190] Figure 2d , 2e This is an illustration of the volume used to house the temperature sensor according to the present invention;

[0191] Figure 3 This is a thermal model for capacitors;

[0192] Figure 4 This is an example of the effective ripple current and case temperature of an electrolytic capacitor over a single service cycle;

[0193] Figure 5a It shows that the subject is like Figure 4 The estimated lifespan, lifespan at which the effective ripple current and case temperature affect the electrolytic capacitor are shown.

[0194] Figure 5b Is it like this? Figure 5a A magnified view showing lifetime consumption and total lifetime consumption;

[0195] Figure 6 These are examples of the effective ripple current and case temperature of electrolytic capacitors over multiple service cycles, including variations over the service cycle. Figure 6 The updated total lifetime consumption and estimated remaining lifetime are displayed after the detection and definition of the second usage cycle;

[0196] Figure 7 A power supply circuit including a large-capacity capacitor, which is an electrolytic capacitor, is shown, and its lifespan is estimated by the device according to the invention.

[0197] In the accompanying drawings, the same parts are given the same reference numerals. Detailed Implementation

[0198] Figure 1An electrolytic capacitor 1 is shown. The core 10 is formed from two conductor surfaces separated by an insulator. In one example, the core is formed from a rolled sandwich of two aluminum foils separated by impregnated paper, which serves as the insulator. The paper is impregnated with electrolyte. The core 10 is entirely surrounded by a housing 11 except for one side. The open side is sealed by a seal 13, typically made of rubber. Two terminal wires 14 pass through the seal 13, one connected to one conductor in the core 10 and the other connected to the other conductor in the core 10. The exterior of the housing 11 and a portion of the seal 13 are protected by an insulating sleeve 12. The sleeve 12 is typically just a thin plastic layer printed with information such as capacitor type, capacitance, and polarity. Figure 1 The “surrounding environment” 15 of the thermal model of electrolytic capacitor 1 is also shown.

[0199] The lifespan of an electrolytic capacitor 1 is typically limited by the chemical degradation and evaporation of the electrolyte. Therefore, an electrolytic capacitor has a maximum service life 53 independent of use and a temperature-dependent, and thus predictable, service-related lifespan 43. In most cases, the lifespan 43 of an electrolytic capacitor in use is shorter than the maximum service life 53.

[0200] At the end of its lifespan, electrolytic capacitor 1 typically exhibits a decrease in capacitance and an increase in impedance and equivalent series resistance. The increased equivalent series resistance leads to more heat generation. In a device, aging capacitor 1 deviates increasingly from design specifications. Depending on the device and the function capacitor 1 is used for, the user may notice increased noise and / or periodic power loss. The lifespan of electrolytic capacitor 1 is defined as the time during which its characteristics remain within an acceptable range near its nominal values.

[0201] The lifespan of the electrolytic capacitor 1 is estimated to depend on the temperature of its core 10. However, if the core temperature is to be measured directly, a sensor must be inserted through the seal 13. This weakens the seal 13 and accelerates evaporation, thus reducing the maximum service life 53 of the capacitor 1. Furthermore, the capacitor 1 is typically constructed to be as compact as possible. Therefore, in commercially available electrolytic capacitors 1, there is little or no space to place the temperature sensor 32.

[0202] According to one embodiment of the present invention, such as Figure 2a As shown, the temperature sensor 32 is therefore mounted on the housing 11 of the electrolytic capacitor 1, or directly on the sleeve 12. The sleeve 12 is thin enough to be part of the housing 11 in the temperature model. In this embodiment, the temperature sensor 32 is insulated from the surrounding environment 15 by a thermal insulation layer 320. Therefore, the temperature of the temperature sensor 32, and the resulting measured value, is governed by the temperature of the housing 11.

[0203] The thermal insulation layer 320 covers only the sensor and its closest area, which is relatively small compared to the total area of ​​the housing 11. This allows the capacitor 1 to dissipate heat to the surrounding environment 15, thereby minimizing the effects of temperature measurement.

[0204] Temperature sensor 32 is connected to controller 31, which uses the casing temperature measurement to estimate the capacitor's lifespan. In the case shown below, temperature sensor 32 transmits its data via a cable. In other embodiments, wireless transmission may be used.

[0205] According to one embodiment of the present invention, such as Figure 2b and 2c As shown, temperature sensor 32 is mounted between two electrolytic capacitors 1a and 1b connected in parallel. The two electrolytic capacitors 1a and 1b are of the same type and size, and are subjected to the same internal heating due to their parallel connection, and are subjected to the same ambient conditions due to their proximity to each other. Figure 2b The side view is shown. Figure 2c The top view is shown. From all sides, the temperature sensor 32 is positioned between the two capacitors. Since the two capacitors 1a and 1b generate the same amount of heat, the temperature between them will be equal to the common casing temperature. A second capacitor, rather than an added insulating layer, protects the temperature sensor 32 from environmental influences such as localized heat sources or heat sinks. The temperature sensor 32 can be bonded to the sleeve 12 of one of the capacitors 1a and 1b in the area between the first and second capacitors 1a and 1b in the device.

[0206] exist Figure 2b and 2c In the illustrated embodiment, in addition to the temperature sensor 32, there are ammeters 34 and voltmeters 33 that measure the current and voltage across capacitors 1a and 1b. This data is also sent to the controller 31. The controller 31 uses the current and / or voltage measurements to estimate the effective ripple current flowing through capacitors 1a and 1b.

[0207] According to Figure 2a In this embodiment, the effective ripple current is estimated based on information about the number of source phases, output mode, and power required on the output side. This information is preferably shared between a switch controller for controlling the circuit switches of the device and a lifetime estimation controller 31. Preferably, the switch controller and controller 31 are a single unit that controls the switching of the circuit switches of the device and estimates the lifetime consumption and remaining lifetime of the electrolytic capacitors.

[0208] Figure 2d and Figure 2e This illustrates how, in order to insulate the temperature sensor 32 from the surrounding environment 15, according to... Figure 2b and 2cIn the embodiments, the volume of the temperature sensor 32 is preferably set. Figure 2d Two capacitors 1a and 1b are shown from above. The dashed line represents the radius of one of the capacitors, the connecting line between them, and the direction perpendicular to the connecting line and the longitudinal axis of the capacitor. The cube 321 of the temperature sensor 32, which is set in this projection, is shown as a small rectangle positioned between the two capacitors. Figure 2e The situation is shown from the side. In this perspective view, the two capacitors 1a and 1b are shown as rectangles with the same height h, which in this example is represented along their vertical axis. The distance between them is... Figure 2d The same as in the projection. The cube 321 of the temperature sensor 32, set in this projection, is shown as a thin rectangle between two capacitors. Although in Figure 2d and Figure 2e It's hard to see in the example, but part of the capacitor extends into the cube, and the temperature sensor is attached to the surface of that part.

[0209] Figure 3 A thermal model 2, upon which the lifetime estimation according to the present invention is based, is shown. This model describes three parts: the capacitor core 10, the casing 11, and the surrounding environment 15. The core 10 and casing 11 are simulated as thermal capacitors. The surrounding environment 15 is simulated as an equivalent voltage source providing a constant ambient temperature. In addition to the thermal capacitors, the core 10 also includes an equivalent power source connected in parallel with the thermal capacitors, representing the heat source generated by power losses in the core 10 during use. The interfaces between the three parts are simulated by thermistors. The first of these interfaces is between the core 10 and the casing 11, and has a thermal resistance R between the core and the casing 201. thjc The second of these interfaces is between the housing 11 and the surrounding environment 15, with a thermal resistance R between the housing and the environment 215. thca .

[0210] The parameters of this thermal model 2 can be determined through appropriate experiments using electrolytic capacitors in a controlled environment. For example, the casing temperature can be determined based on different effective ripple currents and different ambient temperatures.

[0211] To estimate lifetime consumption in the current method, a steady-state condition with respect to temperature is assumed, therefore only the thermal resistance R between the core and the outer shell 201 is needed. thjc And the thermal resistance R between the casing and the surrounding environment 215. thca And the heat source of the effective ripple current.

[0212] Figure 4 An example of how the housing temperature 42 and effective ripple current 41 evolve over time during the service life is shown. The housing temperature 42 and effective ripple current 41 are given in arbitrary units and plotted on a common y-axis. Time evolves along a common x-axis.

[0213] When the device using the current electrolytic capacitor is turned off, the casing temperature equals the ambient temperature. Once the device is turned on, a ripple current appears and begins to generate heat, which raises the temperature of the core. This heat is conducted to the casing, and the temperature rise is detected. The ripple current changes during use, eventually decreasing to zero when the device is turned off. Without heating, the conductor cools until the casing temperature equals the ambient temperature.

[0214] In a preferred embodiment, the lifetime estimate and / or consumed lifetime are derived from the average temperature 420 and / or the average effective ripple current 410 to represent the situation within a given time period 55. Preferably, the time period 55 is chosen such that the variations in temperature 420 and effective ripple current 410 are small. Figure 4 The long time period of 55 was chosen primarily for illustrative purposes. However, even these long time periods of 55 provide a rough representation of the development of the case temperature and effective ripple current, and can be used, for example, to ensure that a specified usage cycle is being performed.

[0215] exist Figure 4 In the illustrated case, the average temperature 420 and average effective ripple current 410 values ​​are the arithmetic mean of all measurements taken within time period 55. However, in other embodiments, the average temperature 420 and average effective ripple current 410 are representative values ​​for the time period 55 in question. For example, the housing temperature 42 and effective ripple current 41 are measured only once within each time period 55.

[0216] Figure 5a Displayed based on Figure 4 The estimated lifetime 43 is shown for effective ripple current 41 and case temperature 42. Figure 5b The lifetime consumption is displayed as 44 and the total lifetime consumption as 45. Although Figure 5a The estimated lifetime 43 shown is given in time units, while lifetime consumption 44 and total lifetime consumption 45 are given as fractions or percentages of the capacitor's total lifetime. In the case shown, lifetime consumption 44 is expressed as a fraction of the capacitor's total lifetime per unit time, and total lifetime consumption is expressed as a fraction of the capacitor's total lifetime. Therefore, if total lifetime consumption 45 reaches a value of 1, the capacitor has reached the end of its initial predicted lifetime and may begin to show performance degradation effects.

[0217] The lifespan consumption 44 is proportional to a time unit over the estimated lifespan 43, and the total lifespan consumption 45 is the integral of the lifespan consumption 44 over time. Therefore, the total lifespan consumption 45 at the end of the first service cycle can be used as an estimate of the service cycle lifespan consumption. If the service cycle is defined after the service life of the electrolytic capacitor, then the service cycle lifespan consumption is the difference between the total lifespan consumption 45 at the end of the service cycle and the total lifespan consumption 45 at the beginning of the service cycle.

[0218] Figure 5a and 5b The first service cycle of the electrolytic capacitor is shown: the total lifespan consumption 45 is zero at the start of the shown time interval. If the current electrolytic capacitor were to undergo a previous service cycle, the total lifespan consumption 45 would have a non-zero value at the start of the shown time interval. However, the curves for lifespan estimation 43 and lifespan consumption 44 would be the same as those shown.

[0219] exist Figure 5a and 5b In the diagram, the solid lines represent the lifetime estimate 43, lifetime consumption 44, and total lifetime consumption 45, evaluated based on continuous measurements of the housing temperature 42 and effective ripple current 41. The horizontal lines illustrate the values ​​obtained by using the average temperature 420 and average effective ripple current 410 as assumed constant values ​​for the time period 55 in question. While choosing a long time period 55 may result in some deviation from the accurate value of total lifetime consumption 45 when lifetime consumption 44 varies, this difference is small enough at the end of the usage cycle to make the data usable.

[0220] To allow for comparison between the errors caused by using average values ​​of 420 and 410 and the errors caused by uncertain casing temperature measurements, Figure 5b It also showed the casing temperature ratio Figure 4 The total lifetime consumption is shown as either 5% lower (451b) or 5% higher (451a). Due to... Figure 4 The diagram shows an ambient temperature of 20°C, with a 5% error of + / -1°C when the device is off. Clearly, if this temperature uncertainty is acceptable, the uncertainty in total lifetime consumption is even approximately half its value after a single usage cycle. This illustrates that accurate temperature measurement is crucial for estimating effective lifetime consumption, and even more important for effectively predicting remaining lifetime.

[0221] Figure 6 The lifetime estimates and the determination of remaining lifetimes 52a and 52b over multiple service cycles 51a and 51b are shown.

[0222] Figure 6The figure above shows the changes in the housing temperature measurement 42 and the effective ripple current value 41 over time. At the beginning of its service life, the capacitor is used in a device with a first usage cycle 51a. Subsequently, the device is used differently, thus having a second usage cycle 51b. In the case shown here, the first and second usage cycles 51a and 51b are similar in length, but the amplitudes of the housing temperature and the effective ripple current are different. In other embodiments, the amplitudes may be similar, but the duration of the usage cycle may differ. In further embodiments, the usage cycles differ in the shape and / or amplitude of the effective ripple current curve and / or temperature, as well as their duration. There is also a duration during which the development of the housing temperature and the development of the effective ripple current differ from the first and second usage cycles.

[0223] Figure 6 The figure below shows the total lifespan consumption 45 and the maximum lifespan 53 of the capacitor. The total lifespan consumption 45 is given as a fraction of the total lifespan, with the horizontal dashed line representing "1". The estimated time when the total lifespan consumption 45 crosses this dashed line is the preliminary predicted lifespan 54a, 54b. The preliminary predicted lifespans 54a, 54b can be greater than or less than the maximum lifespan 53 of the capacitor. The maximum lifespan 53 is the number of years the capacitor in question begins to fail due to aging processes unrelated to use (e.g., evaporation of electrolyte through seal 13). Figure 6 The remaining lifespan 52 marked with an asterisk in the example is the time until the maximum lifespan 53 or the preliminary predicted lifespan 54a or 54b is reached, whichever comes first.

[0224] exist Figure 6 In this process, the preliminary predicted lifetimes 54a and 54b are updated during the use of the capacitor: at the start of its use, a first use period 51a is defined and a first preliminary predicted lifetime 54a is generated by assuming that similar use periods will occur successively. A first inference 450a of the total lifetime consumption 45 based on the observation of the first use period 51a is shown by a dashed line.

[0225] However, in the example shown, the first usage cycle 51a is repeated only three times. After these three repetitions, the controller notices a deviation from the expected pattern. In the example shown, the user responds to the controller's notification and sets a second usage cycle 51b. The total lifetime consumption during the second usage cycle 51b is less than the total lifetime consumption during the first usage cycle 51a. A second inference 450b is adjusted accordingly to adjust the total lifetime consumption: starting from the total lifetime consumption at the beginning of the second usage cycle 51b, the second inference 450b assumes that the second usage cycle 51b will be repeated over and over again. The second inference 450b is indicated by a dashed line with a smaller slope than the first inference 450a.

[0226] exist Figure 6In the example, the preliminary predicted lifetime 54a based on the first usage cycle 51a is shorter than the capacitor's maximum lifespan 53. Therefore, the remaining lifetime 52a is determined by the preliminary predicted lifetime 54a as long as the usage cycle is not updated. However, after the update, the preliminary predicted lifetime 54b based on the second usage cycle 51b is longer than the capacitor's maximum lifespan 53. Therefore, the remaining lifetime 52a is determined by the maximum lifespan 53.

[0227] While the first usage period 51a was used for inference, it can still be noted that the capacitor usage deviated from the time period of the first usage period 51a. The inference was made by continuing to use a linear function with a slope defined by the first usage period 51a. Figure 6 This deviation in the example is mitigated by shifting the function along the time axis so that it intersects with the actual determined point of total lifetime consumption at a given time. This updated inference 450a This is shown as another dashed line in the figure and results in an updated preliminary predicted lifetime of 54 years. And therefore the remaining lifetime of 52a. .

[0228] exist Figure 6 In the illustrated case, this inference shift occurs once a deviation from the first usage cycle 51a is detected in the housing temperature or effective ripple current measurement. In other embodiments, the inference shift always occurs at the end of the defined usage cycle duration. In further embodiments or beyond further criteria, the inference shift occurs if the difference between the total lifetime consumption and the inference function is greater than a set threshold level.

[0229] Figure 7 A power supply is shown, which is typically a device comprising a large-capacity capacitor 72 that can be constructed using an electrolytic capacitor 1. Figure 7In the example shown, the large-capacity capacitor 72 is implemented by a set of two identical electrolytic capacitors 1 connected in parallel. The lifespan of this large-capacity capacitor 72 is estimated using a device according to the invention, which includes a temperature sensor 32 disposed between the two electrolytic capacitors to measure their common case temperature 42 with high accuracy. In the illustrated embodiment, the device also includes an ammeter 34 for determining the effective ripple current. The temperature sensor 32 and the ammeter 34 are connected to a controller 31, in this case equipped with an output device in the form of a display and an input device in the form of buttons. The display can be used to return the remaining lifespan as an output and to inform the user of the status of the usage cycle recording. For example, the user can press a button once to start recording usage cycles, and the display reads "rec" while recording usage cycles. By pressing the button a second time, the user can indicate to the controller that the usage cycle has ended. By displaying a flashing remaining lifespan, the user can be informed that the stored usage cycles are insufficient to accurately represent the actual use of the device. In other embodiments, the controller includes only or additionally an interface that can connect to an I / O device or storage device carrying instructions about the usage cycles to be used and / or receiving data from the controller. The interface can be wireless or in the form of a socket / plug system. Data to be transferred to a storage device or display, or displayed on the controller itself, may include remaining lifetime, current lifetime estimate, current lifetime consumption, current total lifetime consumption, a timeline of past lifetime estimates, a timeline of past lifetime consumption, a timeline of past values ​​of total lifetime consumption, a timeline of measured case temperature and / or current case temperature, a timeline of effective ripple current values, a timeline of current effective ripple current and timeline, and current values ​​of variables used to estimate the effective ripple current. The controller may represent the data statistically and provide it to the user, for example, a histogram showing the number of time periods with a given lifetime consumption or the number of time periods considered not to belong to the set of usage cycles at the time of occurrence. Furthermore, the controller may provide the user with information about the uncertainty of determined values ​​and the noise observed on temperature sensor readings, and may also provide information about ammeters and / or voltmeters used to estimate the effective ripple current. In addition to specifying the usage cycle, the user can define or update the mode for estimating the effective ripple current and the required information. For example, in some embodiments, the user can upload a lookup table specifying the number of phases on the input side and the output mode of the power supply. In some embodiments, the user can specify whether the controller should determine the ripple current frequency by analyzing current or voltage measurements, or whether the controller should assume a spectrum based on the number of input phases and the output mode. Alternatively, or as a further embodiment, the user can upload the spectrum of the ripple current present in the current device and in the environment in which the device is used.Unlike uploaded spectrum, this information can simply represent one or more correction factors that will be applied to measured or estimated characteristics, such as input current, output voltage, or output or input power, to produce an estimate of the effective ripple current. Additionally, or as a further embodiment, users can upload or download information about the years and total lifespan of one or more electrolytic capacitors whose lifespan is being monitored. Therefore, users can replace capacitors between devices without losing information about their remaining lifespan.

[0230] Figure 7 In addition to the large-capacity capacitor 72, the power supply shown also includes a rectifier 71, a switching unit 73, a transformer 74, and an output capacitor 75. Figure 7 The power supply shown receives input power by transmitting AC current and voltage through two phases 76. The AC current is rectified by a full-bridge rectifier 71. A large-capacity capacitor 72 is grounded at its first terminal and connected to one of the output lines of the rectifier 71 through its second terminal. A switching unit 73 generates a high-frequency AC current to be converted to the desired voltage by a transformer 74. The AC voltage on the secondary side of the transformer 74 can be rectified by another rectifier bridge or used directly as output 77. Figure 7 A single diode and output capacitor 75 are used to rectify the output of transformer 74 to generate a DC output voltage. Preferably, there is feedback from the output side to the switching unit 73, which allows the switch to be adjusted in a suitable manner to suit the needs of the load connected to the power supply. In a preferred embodiment, this feedback is also used by controller 31 to estimate the effective ripple current.

[0231] In summary, it should be noted that the device according to the invention can be used in all devices that use electrolytic capacitors. If two or more electrolytic capacitors are configured to be connected in parallel and adjacent to each other, only one temperature sensor can be placed between any two adjacent electrolytic capacitors, because in this configuration, their lifespan estimates will be the same.

Claims

1. An apparatus for estimating the lifespan of one or more electrolytic capacitors, comprising: a) A temperature sensor, which is insulated from the surrounding environment by being positioned in the following location: i) Between two equal electrolytic capacitors connected in parallel, or ii) The outer casing of the electrolytic capacitor is covered with a layer of heat-insulating material, and b) A controller configured to estimate the lifetime consumption based on measurement data from the temperature sensor; The controller estimates the lifetime consumption of the one or more electrolytic capacitors over a time period in proportion to the quotient of the lifetime over the time period, and determines the lifetime by measuring or assuming the temperature of the housing and the effective ripple current occurring during the time period.

2. The device according to claim 1, comprising: a) An instrument for measuring output or input data, said output or input data being influenced by or affecting the voltage, power, and / or current characteristics of said one or more electrolytic capacitors, and b) wherein the controller is adapted to include the measured output or input data in its estimate of the lifetime consumption of the one or more electrolytic capacitors.

3. The device according to claim 2, comprising any one of the following: a) A first detector, adapted to detect the characteristics of the device, such as input voltage, input current, or input power. b) A second detector, adapted to detect the characteristics of the device, such as output voltage, output current, or output power. And / or, c) An ammeter suitable for measuring the current through the one or more electrolytic capacitors. The controller is adapted to include measurements from the first detector and the second detector or the ammeter in its estimate of the lifetime consumption.

4. The device of claim 1, wherein the controller estimates the effective ripple current by taking into account the output mode of the device, the number of phases and / or modes and / or waveforms of the inputs of the device, and a given or measured output power value or a given or measured current value or a given or measured input voltage.

5. The device of claim 3, wherein the controller estimates the uncertainty of the effective ripple current by taking into account the output mode of the device, the number of phases and / or modes and / or waveforms of the inputs of the device, and a given or measured output power value or a given or measured current value or a given or measured input voltage.

6. The device of claim 3, wherein the controller estimates the effective ripple current by using one of a set of lookup tables; a) wherein the lookup table to be applied is selected based on the output mode of the device, the number of phases and / or mode and / or waveform of the input of the device, and b) wherein the lookup table returns the uncertainty in the effective ripple current and / or the effective ripple current under the following conditions: i) The given or measured output power value, or ii) The given or measured current value, or iii) The given or measured input voltage value.

7. The device of claim 4, wherein the controller considers the temperature of the housing, the effective ripple current, and the uncertainty of the time period to determine the uncertainty of the lifetime consumption.

8. The device according to claim 4, wherein the controller: a) Determine the lifetime consumption and its uncertainty over a typical time period. b) where each time period is represented by an average temperature and an average effective ripple current, said average temperature and said average effective ripple current being preferred averages of all suitable measurements obtained within said time period, and c) Wherein the total lifetime consumption is the sum of all previously determined lifetime consumptions, and the total lifetime consumption is determined and stored.

9. The device according to claim 8, wherein: a) The device allows a user to specify a usage period expected to be repeated in the future, and wherein the controller determines the lifetime consumption and its uncertainty within a time period of the usage period, and wherein the controller determines the usage period lifetime consumption by integrating the lifetime consumption determined continuously over the usage period or by adding the lifetime consumption over all time periods of the usage period, and determines the uncertainty of the usage period lifetime consumption. b) wherein the controller uses the usage cycle lifetime consumption to determine the initial predicted lifetime by linearly inferring the total lifetime consumption. c) wherein the preliminary predicted lifetime is the duration from the start of the capacitor's lifetime to the inference that the total lifetime has been consumed reaches a value representing the end of the lifetime. d) The controller further includes a clock that counts the years of the one or more electrolytic capacitors. e) wherein the remaining lifetime estimated by the controller is the smaller of the following two figures: i) The preliminary predicted lifespan minus the current age of the one or more electrolytic capacitors, ii) The longest lifespan of the one or more electrolytic capacitors minus the current lifespan of the one or more electrolytic capacitors. f) and whereby the remaining lifetime and its uncertainty are returned.

10. The device according to claim 9, wherein: a) The controller records the temperature measurement of the housing or the lifespan consumption and the estimated effective ripple current of the usage cycle during its specification period, and compares these recorded data sequences with data sequences generated at repetition times assumed to be the usage cycle. (b) Wherein if the controller detects that the deviation between the recorded data and the measured data is greater than a threshold, the controller sends a signal to the user, thereby warning of the quality loss in the remaining life estimate, and requiring the user to specify a new usage period.

11. The device of claim 9, wherein the preliminary predicted lifespan and / or the resulting remaining lifespan inference is periodically updated so that the inference takes into account the latest available data regarding the years and total lifespan consumption.

12. The device of claim 8, wherein the controller includes an interface and stores the temperature, the effective ripple current and the lifetime consumption for each time period, such that they or statistical data derived from them can be exported through the interface.

13. The device of claim 1, wherein the controller estimates the lifetime of the one or more electrolytic capacitors by calculating the product of a first capacitor type dependence constant and first and second exponential functions. a) wherein the exponent of the first exponential function is the temperature of the housing measured by the temperature sensor, and the base of the first exponential function is a constant independent of the capacitor type, and b) The exponent of the second exponential function is the square of the effective ripple current, and the base of the second exponential function is the second capacitor type-dependent constant.

14. The device of claim 1, wherein the lifespan consumption is equal to the quotient of the time period over the lifespan within the time period.

Images