Intelligent device fault detection method and device based on internet of things and storage medium

Abstract

This invention relates to the field of fault prediction technology, and more particularly to a method, apparatus, and storage medium for fault detection of smart devices based on the Internet of Things (IoT). The method includes: identifying thermal imbalance states based on the surface temperature of individual battery cells, and then determining a thermal runaway risk metric; determining the power supply stress urgency based on the temperature and output current of the switching power supply rectifier module; extracting equalization charging events based on the battery pack float charge current of the first historical cycle, and determining the battery degradation-sensitive voltage range; determining a health risk metric based on the thermal runaway risk metric, the power supply stress urgency, and the battery degradation-sensitive voltage range, and adjusting the determination process of the health risk metric; and generating a float charge voltage adjustment value based on the adjusted health risk metric to determine the float charge voltage for the next control cycle. This invention suppresses the thermal runaway degradation trend of batteries under high-temperature environments while ensuring the reliability of communication backup.

Description

Technical Field

[0001] This invention relates to the field of fault prediction technology, and in particular to a method, apparatus and storage medium for fault detection of smart devices based on the Internet of Things. Background Technology

[0002] The high temperature and humidity of coastal environments pose severe challenges to backup power systems for communication base stations. In such scenarios, the consistently high ambient temperature, coupled with salt spray corrosion, causes the side reaction rate of battery packs in float charging mode to be much higher than in normal environments. Internal heat accumulation is difficult to dissipate effectively, easily inducing overheating and even thermal runaway in individual cells. Simultaneously, the switching power supply rectifier module operates under thermal stress boundaries for extended periods, resulting in insidious performance degradation and reliability reduction. Existing monitoring methods often focus on alarms for single parameters exceeding limits, such as detecting only excessively high voltage or excessively high temperature, making it difficult to capture early signs of failure under the coupling of multiple factors.

[0003] In addition, traditional float charge voltage management strategies are usually mechanically adjusted based on a fixed temperature compensation curve, which lacks correlation analysis of the battery's actual degradation sensitivity and dynamic load stress. This often leads to insufficient battery protection under harsh operating conditions or limited charging efficiency in low-temperature environments, making it impossible to achieve a delicate balance between ensuring communication reliability and extending asset life. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus and storage medium for fault detection of smart devices based on the Internet of Things, so as to solve at least one of the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for fault detection of smart devices based on the Internet of Things, comprising:

[0007] The thermal imbalance state is identified based on the surface temperature of individual cells, and then the thermal runaway risk metric is determined.

[0008] Determine the stress level of the power supply based on the temperature and output current of the switching power supply rectifier module;

[0009] Based on the first historical cycle of the battery pack float charge current, equalization charging events are extracted and the battery degradation sensitive voltage range is determined.

[0010] The health risk metric is determined based on the thermal runaway risk metric, the power supply stress urgency, and the battery degradation sensitive voltage range. The determination process of the health risk metric is adjusted based on the peak-to-peak ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch.

[0011] The float charge voltage adjustment value is generated based on the adjusted health risk metric to determine the float charge voltage for the next control cycle.

[0012] Furthermore, the maximum value Tbimax and minimum value Tbimin of the surface temperature of individual cells within each sub-time window of the analysis period are extracted, and the standard deviation σTbi and average value Tbiavg of the surface temperature of individual cells within each sub-time window of the analysis period are calculated. Then, the range feature TRi and the discrete feature Cti of the individual cell temperature in the sub-time window are constructed, where i is the sub-time window number, i∈[1,Nz], and Nz is the number of sub-time windows in the analysis period.

[0013] Furthermore, the 96th percentile of the Nz TRi values ​​is taken as the single-cell temperature range characteristic TR for the current analysis period;

[0014] The 96th percentile of the Nz Cti values ​​is taken as the single-unit temperature discrete characteristic Ct for the current analysis period;

[0015] TR is compared with a preset range feature threshold tr0 to determine the range risk component Rr. The thermal runaway risk measure RT is determined by combining the individual temperature discrete feature Ct with the range risk component Rr.

[0016] Furthermore, the maximum temperature Tmmax of the switching power supply rectifier module within the analysis cycle is extracted, and Tmmax is compared with each preset temperature threshold to determine the temperature stress WT.

[0017] Calculate the average output current Fp of the switching power supply rectifier module within the analysis cycle, and then calculate the load factor LR, LR=Fp / Fy; determine the load stress FT based on the load factor LR;

[0018] The combined temperature stress WT and load stress FT determine the power supply stress urgency DT.

[0019] Furthermore, the average value of the total voltage of the battery pack in the 10 minutes before the start of the equalization charging event is taken as the trigger open-circuit voltage OCV of the equalization charging event;

[0020] A voltage sample set is constructed based on the trigger open-circuit voltage (OCV) of all equalization charging events within the first historical period. If the number of samples in the current voltage sample set for the first historical period is less than a preset number, then the battery degradation sensitive voltage range is determined to be the preset sensitive voltage range; otherwise:

[0021] The average value of all samples in the voltage sample set is calculated as the center of the degradation-sensitive voltage Vc, and the standard deviation of all samples is calculated as the radius of the degradation-sensitive voltage distribution ΔVc. The battery degradation-sensitive voltage range is defined as [Vc-0.5×ΔVc, Vc+0.5×ΔVc];

[0022] Wherein, the lower limit of the battery degradation sensitive voltage range is not lower than Vmin and the upper limit is not higher than Vmax, where Vmin is the battery pack discharge termination protection voltage and Vmax is the battery pack rated equalization charge voltage.

[0023] Furthermore, the voltage degradation sensitivity factor fv is determined based on the average battery pack voltage Vn of the current analysis cycle and the battery degradation sensitive voltage range: when Vn belongs to the battery degradation sensitive range, fv=|Vn-Vc| / △Vc; otherwise, fv=0.

[0024] The health risk metric He is determined by weighting and fusing thermal runaway risk measures, power supply stress urgency, and voltage degradation sensitivity factors.

[0025] Furthermore, the average value of the peak-to-peak ripple voltage Vrp of the target RRU power supply branch within the analysis period is calculated, and then the ripple coefficient RF is calculated in combination with the rated output voltage Ve, RF=Vrp / Ve; then the first correction factor Z1 is determined, Z1=1-exp(-RF / RFm);

[0026] The peak value Ip and valley value Iv of the instantaneous current of the target RRU power supply branch are calculated within the analysis period. The peak-valley ratio PCR of the instantaneous current is calculated, PCR=(Ip-Iv) / Iv, and then the second correction factor Z2 is determined, Z2=tanh(PCR / PCRm).

[0027] The dynamic stress coefficient ES of the power supply link is determined by combining the first correction factor Z1 and the second correction factor Z2, where ES = ρ1 × Z1 + ρ2 × Z2.

[0028] The health risk metric is adjusted based on the dynamic stress coefficient ES of the power supply link, and the adjusted health risk metric is set as He1, He1=min(1,He×(1+η×ES));

[0029] Where Ip is the 99th percentile of the instantaneous current sampling value within the analysis period, Iv is the 1st percentile of the instantaneous current sampling value within the period, ρ1 is the ripple stress weight, ρ2 is the dynamic tensile stress weight, ρ1+ρ2=1, RFm is the preset ripple coefficient threshold, PCRm is the preset instantaneous current peak-to-valley ratio abnormal threshold, and η is the preset adjustment coefficient.

[0030] Furthermore, at the end of the current control cycle, the He1 value of each analysis cycle is extracted, and its 90th percentile is taken as the control input risk value Hec, thereby determining the float charge voltage adjustment amount ΔVf, ΔVf=-ΔVmax×Hec, where ΔVmax is the preset adjustment range threshold.

[0031] The float charge voltage Vfn for the next control cycle is determined based on the float charge voltage adjustment ΔVf, where Vfn = Vfe + ΔVf, and Vfe is the rated float charge voltage.

[0032] If Vfn is lower than the preset safety lower limit Vfmin, then let Vfn = Vfmin, where Vfmin is the preset safety lower limit.

[0033] According to another aspect of this application, an Internet of Things-based smart device fault detection device is provided, comprising:

[0034] The thermal risk determination unit is used to identify the thermal imbalance state based on the surface temperature of a single cell, and then determine the thermal runaway risk measure.

[0035] The power supply stress determination unit is used to determine the power supply stress urgency based on the temperature and output current of the switching power supply rectifier module.

[0036] The interval calculation unit is used to extract equalization charging events based on the battery pack float charging current of the first historical cycle and determine the battery degradation sensitive voltage interval.

[0037] The integrated risk determination unit is used to determine the health risk measure based on the thermal runaway risk measure, the power supply stress urgency and the battery degradation sensitive voltage range, and to adjust the determination process of the health risk measure based on the peak-to-peak value of the ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch.

[0038] The adjustment unit is used to generate a float charge voltage adjustment value based on the adjusted health risk metric, so as to determine the float charge voltage for the next control cycle.

[0039] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, wherein the computer program is used to control the electronic device in which the computer-readable storage medium is located to perform the Internet of Things-based smart device fault detection method during runtime.

[0040] The beneficial effects of this invention are as follows: Through statistical feature analysis and a historical self-learning mechanism, early tendencies for thermal runaway and the degradation sensitivity of voltage ranges are identified. The solution incorporates power supply stress, battery chemical state, and the dynamic load-pull effect of the end load into a unified risk quantification model, achieving a comprehensive view of system health. Based on this, the resulting float charge voltage regulation strategy is an adaptive action based on real-time risk simulation, which can proactively reduce charging stress when the risk of thermal runaway increases, thereby suppressing side reactions and temperature rise at the source. This method effectively suppresses the capacity drop of batteries in harsh environments while ensuring that base station backup time is not affected, reducing the probability of communication interruptions caused by premature battery failure or overheating protection of the switching power supply, and providing more accurate and robust remote operation and maintenance support for unattended sites. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart illustrating the IoT-based smart device fault detection method of this embodiment.

[0043] Figure 2 This is a flowchart illustrating the health risk measurement and analysis method in this embodiment.

[0044] Figure 3 This is a schematic diagram of the structure of the IoT-based smart device fault detection device in this embodiment. Detailed Implementation

[0045] To more clearly illustrate the present invention, the following description, in conjunction with preferred embodiments and accompanying drawings, further explains the invention. Similar components in the drawings are indicated by the same reference numerals. Those skilled in the art should understand that the specific description below is illustrative rather than restrictive and should not be construed as limiting the scope of protection of the present invention.

[0046] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0047] Specifically, this embodiment is applied to the backup power system of communication base stations in coastal areas with high temperature and humidity.

[0048] Please see Figure 1 As shown, this is a flowchart illustrating the IoT-based smart device fault detection method of this embodiment. Before the method is executed, the following data is collected synchronously, including:

[0049] Battery pack data: The surface temperature of each individual battery cell is collected by a high-precision temperature sensor installed on the terminal of each individual battery cell; the total voltage and charging / discharging current of the battery pack are collected by a Hall voltage sensor and a shunt in the main circuit of the battery pack.

[0050] Switching power supply data: The switching power supply monitoring module collects the output voltage, output current and temperature of the rectifier module. The temperature of the switching power supply rectifier module refers to the temperature value collected by a negative temperature coefficient thermistor installed on the heat sink substrate of the power device.

[0051] Battery cell voltage data: The terminal voltage of each 12V battery cell is collected through the cell voltage inspection line connected to the switching power supply monitoring module.

[0052] Feeder system data: The peak-to-peak value of the ripple voltage of the target RRU power supply branch is collected by miniature Hall sensors installed at the output terminals of the power supply branches of each radio frequency remote unit in the DC power distribution unit; at the same time, the instantaneous current value of the target RRU power supply branch is collected. The target RRU power supply branch refers to the DC power supply branch corresponding to the radio frequency remote unit with the highest average daily downlink data traffic in the past 7 days, according to the statistics of the base station network management system. If the traffic data cannot be obtained temporarily due to communication interruption, function not being configured, or running time not exceeding 7 days, the DC power supply branch corresponding to the radio frequency remote unit that is currently in operation and configured by the base station main control unit as the primary service bearer channel is selected as the target RRU power supply branch.

[0053] This embodiment does not impose specific limitations on the data collection method described above; those skilled in the art can freely set it according to their needs.

[0054] In this embodiment, the analysis period is 15 minutes, the sub-time window is 5 seconds, and the control period is 6 hours.

[0055] The method includes:

[0056] Step S1: Identify the thermal imbalance state based on the surface temperature of a single cell, and then determine the thermal runaway risk measure.

[0057] Specifically, the maximum value Tbimax and minimum value Tbimin of the surface temperature of a single cell within each sub-time window of the analysis period are extracted, and the standard deviation σTbi and average value Tbiavg of the surface temperature of a single cell within each sub-time window of the analysis period are calculated. Then, the temperature range feature TRi and the temperature discrete feature Cti of the single cell within the sub-time window are constructed, where i is the sub-time window number, i∈[1,Nz], and Nz is the number of sub-time windows within the analysis period.

[0058] The expression for TRi is:

[0059] TRi = (Tbimax - Tbimin) / Tbiavg;

[0060] The expression for Cti is:

[0061] Cti = σTbi / Tbiavg;

[0062] The 96th percentile of the Nz TRi values ​​is taken as the single-cell temperature range characteristic TR for the current analysis period;

[0063] The 96th percentile of the Nz Cti values ​​is taken as the single-unit temperature discrete characteristic Ct for the current analysis period;

[0064] TR is compared with a preset range feature threshold tr0 to determine the range risk component Rr. When TR is less than tr0, Rr is determined to be 0. When TR is greater than or equal to tr0, Rr is determined to be min(1,(TR-tr0) / (tr1-tr0)).

[0065] The thermal runaway risk measure RT is determined by combining the discrete temperature characteristics Ct of the individual unit with the range risk component Rr, where RT = α1 × TR + α2 × Ct;

[0066] Where tr1 is the preset range saturation threshold, α1 is the discrete feature weight, α2 is the range feature weight, and α1+α2=1.

[0067] Preferably, in this embodiment, the preset range feature threshold is 0.05, the preset range saturation threshold is 0.2, the discrete feature weight is 0.4, and the range feature weight is 0.6.

[0068] Specifically, by capturing the fluctuations and dispersion of the surface temperature of individual battery cells over a short period of time, rather than simply observing the absolute temperature, it is possible to keenly identify minute signs of thermal imbalance inside the battery. Statistical methods are used to extract the most representative extreme fluctuation characteristics within the analysis period, effectively eliminating the influence of occasional environmental interference or instantaneous sensor noise.

[0069] Please continue reading. Figure 1 As shown, the IoT-based smart device fault detection method further includes:

[0070] Step S2: Determine the power supply stress level based on the temperature and output current of the switching power supply rectifier module.

[0071] Specifically, the maximum temperature Tmmax of the rectifier module of the switching power supply is extracted during the analysis cycle, and Tmmax is compared with each preset temperature threshold to determine the temperature response WT. If Tmmax is greater than or equal to the first preset temperature threshold Ta1, the temperature response WT is determined to be 1. If Tmmax is less than the first preset temperature threshold Ta1 but greater than the second preset temperature threshold Ta2, the temperature response WT is determined to be (Tmmax-Ta2) / (Ta1-Ta2). If Tmmax is less than or equal to the second preset temperature threshold Ta2, the temperature response WT is determined to be 0.

[0072] Calculate the average output current Fp of the switching power supply rectifier module during the analysis cycle, and then calculate the load factor LR, LR=Fp / Fy; based on the load factor LR, determine the load stress FT, FT=max(0,(LR-Lh) / (1-Lh)). 2 ;

[0073] The stress DT of the power supply is determined by combining the temperature stress WT and the load stress FT, where DT = u1 × WT + u2 × FT.

[0074] Where Lh is the preset overload threshold, u1 is the temperature stress weight, u2 is the load stress weight, and u1+u2=1.

[0075] Preferably, in this embodiment, the first preset temperature threshold is 80°C, the second preset temperature threshold is 60°C, the preset heavy load threshold is 0.8, the temperature stress weight is 0.65, and the load stress weight is 0.35.

[0076] Specifically, by setting multi-level temperature thresholds, the assessment of the nonlinear impact of thermal stress on equipment lifespan is refined. At the same time, by weighting the quadratic terms after smoothing the load rate, the accelerated impact effect of heavy load on device junction temperature is more accurately characterized. This composite assessment method allows managers to identify at a glance whether the system is in a normal "high temperature light load" pressure state or in a dangerous boundary state of "high temperature heavy load".

[0077] Please continue reading. Figure 1 As shown, the IoT-based smart device fault detection method further includes:

[0078] Step S3: Extract equalization charging events based on the battery pack float charging current of the first historical cycle, and determine the battery degradation sensitive voltage range.

[0079] Specifically, the equalization charging event is a complete record of the switching power supply actively raising the output voltage from the float charging voltage value to the equalization charging voltage value and continuously executing the equalization charging process for more than 30 minutes. The float charging voltage refers to the constant lower voltage output by the switching power supply to maintain the charge after the battery pack is fully charged. The equalization charging voltage refers to the higher charging voltage output by the switching power supply in stages to eliminate the voltage imbalance of individual batteries.

[0080] The average value of the total voltage of the battery pack in the 10 minutes before the start of the equalization charging event is taken as the trigger open circuit voltage OCV of the equalization charging event;

[0081] A voltage sample set is constructed based on the trigger open-circuit voltage (OCV) of all equalization charging events within the first historical period. If the number of samples in the current voltage sample set for the first historical period is less than a preset number, then the battery degradation sensitive voltage range is determined to be the preset sensitive voltage range; otherwise:

[0082] The average value of all samples in the voltage sample set is calculated as the center of the degradation-sensitive voltage Vc, and the standard deviation of all samples is calculated as the radius of the degradation-sensitive voltage distribution ΔVc. The battery degradation-sensitive voltage range is defined as [Vc-0.5×ΔVc, Vc+0.5×ΔVc];

[0083] Wherein, the lower limit of the battery degradation sensitive voltage range is not lower than Vmin and the upper limit is not higher than Vmax, where Vmin is the battery pack discharge termination protection voltage and Vmax is the battery pack rated equalization charge voltage.

[0084] Specifically, in this embodiment, the first historical period is 30 days prior to the current analysis period. Taking a nominal 48V system as an example, the preset quantity is 10, the preset sensitive voltage range is [48.5V, 51.0V], the Vmin is 43.2V, and the Vmax is 56.4V. During the initial month of system operation, due to the lack of sufficient historical equalization charge records, the system will directly adopt the preset sensitive voltage range.

[0085] Specifically, by analyzing the equalization charge trigger voltage distribution over a long period of time, the system can learn the degradation-sensitive range of the battery pack under specific service conditions. When the subsequent float charge voltage falls into this range, the system can more accurately predict the risk of water loss, sulfation, or increased consistency deviation in the battery pack, so that the adjustment of the float charge voltage truly matches the current chemical state of the battery, rather than blindly applying the factory calibration value.

[0086] Please continue reading. Figure 1 As shown, the IoT-based smart device fault detection method further includes:

[0087] Step S4 involves determining the health risk metric based on the thermal runaway risk metric, power supply stress urgency, and battery degradation sensitive voltage range, and adjusting the determination process of the health risk metric based on the peak-to-peak ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch.

[0088] Please see Figure 2 As shown, the health risk measurement and analysis method includes:

[0089] Step S41: Determine the health risk metric based on the thermal runaway risk metric, power supply stress urgency, and battery degradation sensitive voltage range.

[0090] Specifically, the voltage degradation sensitivity factor fv is determined based on the average battery pack voltage Vn of the current analysis period and the battery degradation sensitivity voltage range: when Vn belongs to the battery degradation sensitivity range, fv=|Vn-Vc| / △Vc; otherwise, fv=0.

[0091] The health risk metric He is determined by weighting and fusing thermal runaway risk measure, power supply stress urgency and voltage degradation sensitivity factor. The expression of the health risk metric He is: He = w1 × RT + w2 × DT + w3 × fv.

[0092] Where w1 is the thermal runaway weight, w2 is the power supply stress weight, w3 is the degradation sensitivity weight, and w1+w2+w3=1.

[0093] Preferably, in this embodiment, the thermal runaway weight is 0.5, the power supply stress weight is 0.3, and the degradation sensitivity weight is 0.2.

[0094] Specifically, a voltage degradation sensitivity factor is introduced, linking static voltage values ​​with dynamic historical degradation patterns. This allows voltage values ​​in the sensitive range to generate additional health risk deductions. Through weighted fusion, this method assigns corresponding decision weights to thermal safety, power supply reliability, and battery aging status, forming a comprehensive health index that facilitates the monitoring center's overall control of the backup power system's health status.

[0095] Please continue reading. Figure 2 As shown, the health risk measurement and analysis method further includes:

[0096] Step S42: Determine the dynamic stress coefficient of the power supply link based on the peak-to-peak value of the ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch within the analysis period, and then adjust the determination process of the health risk metric.

[0097] Specifically, the average value of the peak-to-peak ripple voltage Vrp of the target RRU power supply branch during the analysis period is calculated, and then the ripple coefficient RF is calculated in combination with the rated output voltage Ve, RF=Vrp / Ve; then the first correction factor Z1 is determined, Z1=1-exp(-RF / RFm);

[0098] The peak value Ip and valley value Iv of the instantaneous current of the target RRU power supply branch are calculated within the analysis period. The peak-valley ratio PCR of the instantaneous current is calculated, PCR=(Ip-Iv) / Iv, and then the second correction factor Z2 is determined, Z2=tanh(PCR / PCRm).

[0099] The dynamic stress coefficient ES of the power supply link is determined by combining the first correction factor Z1 and the second correction factor Z2, where ES = ρ1 × Z1 + ρ2 × Z2.

[0100] The health risk metric is adjusted based on the dynamic stress coefficient ES of the power supply link, and the adjusted health risk metric is set as He1, He1=min(1,He×(1+η×ES));

[0101] The system compares He1 with a preset risk threshold h0, and issues a health risk warning to the user when He1 is greater than h0.

[0102] Where Ip is the 99th percentile of the instantaneous current sampling value within the analysis period, Iv is the 1st percentile of the instantaneous current sampling value within the period, ρ1 is the ripple stress weight, ρ2 is the dynamic tensile stress weight, ρ1+ρ2=1, RFm is the preset ripple coefficient threshold, PCRm is the preset instantaneous current peak-to-valley ratio abnormal threshold, and η is the preset adjustment coefficient.

[0103] Preferably, in this embodiment, the ripple stress weight is 0.45, the dynamic tensile stress weight is 0.55, the preset ripple coefficient threshold is 0.05, the preset instantaneous current peak-to-valley ratio abnormal threshold is 5, the preset adjustment coefficient is 0.25, and the preset risk threshold is 0.6.

[0104] Specifically, the rise in peak-to-peak ripple voltage is often a precursor to aging filter capacitors or increased power supply circuit impedance, while the drastic fluctuation in the instantaneous current peak-to-valley ratio of core business units indicates dynamic load stress. By acquiring these two key dimensions and calculating correction factors, this method can dynamically adjust the initial health risk rating. This means that even if the battery and power supply itself do not show any abnormalities temporarily, the deterioration signal at the end of the link can sound an alarm in advance, prompting the system to take preventive voltage adjustment measures before a service interruption or equipment reset occurs.

[0105] Please continue reading. Figure 1 As shown, the IoT-based smart device fault detection method further includes:

[0106] Step S5: Generate a float charge voltage adjustment value based on the adjusted health risk metric to determine the float charge voltage for the next control cycle.

[0107] Specifically, at the end of the current control cycle, the He1 value of each analysis cycle is extracted, and its 90th percentile is taken as the control input risk value Hec, and then the floating charge voltage adjustment amount ΔVf is determined, ΔVf=-ΔVmax×Hec, where ΔVmax is the preset adjustment range threshold.

[0108] The float charge voltage Vfn for the next control cycle is determined based on the float charge voltage adjustment ΔVf, where Vfn = Vfe + ΔVf, and Vfe is the rated float charge voltage.

[0109] If Vfn is lower than the preset safety lower limit Vfmin, then let Vfn = Vfmin, where Vfmin is the preset safety lower limit.

[0110] Preferably, in this embodiment, the preset adjustment range threshold is 1.5V, the rated float charge voltage is 54V, and the preset safety lower limit is 52.8V.

[0111] Specifically, the working principle of the float charge voltage reduction mechanism is as follows: When the comprehensive risk index increases, it indicates that the risk of thermal runaway of the battery pack is aggravated, the stress of the power system is increased, or the battery has entered the deterioration sensitive voltage zone. At this time, appropriately reducing the float charge voltage can reduce the continuous gas evolution and internal side reaction rate of the battery pack in the float charge state, thereby inhibiting the further deterioration of the risk of thermal runaway and delaying the battery aging process.

[0112] Specifically, based on the level of the comprehensive risk value, the float charge voltage setting is adjusted in reverse with a controlled range. When the overall health risk of the system intensifies, the float charge voltage is appropriately reduced. This measure can significantly reduce the internal pressure accumulation and grid corrosion rate of the battery in standby mode, effectively curb the heat accumulation trend under high temperature environment, and delay electrolyte loss.

[0113] Please see Figure 3 As shown, the IoT-based smart device fault detection device includes:

[0114] The thermal risk determination unit is used to identify the thermal imbalance state based on the surface temperature of a single cell, and then determine the thermal runaway risk measure.

[0115] The power supply stress determination unit is used to determine the power supply stress urgency based on the temperature and output current of the switching power supply rectifier module.

[0116] The interval calculation unit is used to extract equalization charging events based on the battery pack float charging current of the first historical cycle and determine the battery degradation sensitive voltage interval.

[0117] The integrated risk determination unit is used to determine the health risk measure based on the thermal runaway risk measure, the power supply stress urgency and the battery degradation sensitive voltage range, and to adjust the determination process of the health risk measure based on the peak-to-peak value of the ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch.

[0118] The adjustment unit is used to generate a float charge voltage adjustment value based on the adjusted health risk metric, so as to determine the float charge voltage for the next control cycle.

[0119] The IoT-based smart device fault detection device provided in this application can execute the IoT-based smart device fault detection method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects of the execution method.

[0120] Those skilled in the art will understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technique for storing information (such as computer-readable programs, data structures, program modules, or other data). Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable programs, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and can include any information delivery medium.

[0121] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is impossible to exhaustively list all the implementation methods here. All obvious variations or modifications derived from the technical solutions of the present invention are still within the protection scope of the present invention.

Claims

1. A method for detecting faults in an intelligent device based on the Internet of Things, characterized by, include: The thermal imbalance state is identified based on the surface temperature of individual cells, and then the thermal runaway risk metric is determined. Determine the stress level of the power supply based on the temperature and output current of the switching power supply rectifier module; Based on the first historical cycle of the battery pack float charge current, equalization charging events are extracted and the battery degradation sensitive voltage range is determined. The health risk metric is determined based on the thermal runaway risk metric, the power supply stress urgency, and the battery degradation sensitive voltage range. The determination process of the health risk metric is adjusted based on the peak-to-peak ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch. The float charge voltage adjustment value is generated based on the adjusted health risk metric to determine the float charge voltage for the next control cycle. 2.The IoT-based smart device fault detection method of claim 1, wherein, The maximum value Tbimax and minimum value Tbimin of the surface temperature of a single cell within each sub-time window of the analysis period are extracted. The standard deviation σTbi and average value Tbiavg of the surface temperature of a single cell within each sub-time window of the analysis period are calculated. Then, the range feature TRi and the discrete feature Cti of the single cell temperature in the sub-time window are constructed, where i is the sub-time window number, i∈[1,Nz], and Nz is the number of sub-time windows in the analysis period. 3.The IoT-based smart device fault detection method of claim 2, wherein, The 96th percentile of the Nz TRi values ​​is taken as the single-cell temperature range characteristic TR for the current analysis period; The 96th percentile of the Nz Cti values ​​is taken as the single-unit temperature discrete characteristic Ct for the current analysis period; TR is compared with a preset range feature threshold tr0 to determine the range risk component Rr. The thermal runaway risk measure RT is determined by combining the individual temperature discrete feature Ct with the range risk component Rr. 4.The IoT-based smart device fault detection method of claim 3, wherein, Extract the maximum temperature Tmmax of the switching power supply rectifier module within the analysis cycle, and compare Tmmax with each preset temperature threshold to determine the temperature stress WT; Calculate the average output current Fp of the switching power supply rectifier module within the analysis cycle, and then calculate the load factor LR, LR=Fp / Fy; determine the load stress FT based on the load factor LR; The combined temperature stress WT and load stress FT determine the power supply stress urgency DT. 5.The IoT-based smart device fault detection method of claim 4, wherein, The average value of the total voltage of the battery pack in the 10 minutes before the start of the equalization charging event is taken as the trigger open circuit voltage OCV of the equalization charging event; A voltage sample set is constructed based on the trigger open-circuit voltage (OCV) of all equalization charging events within the first historical period. If the number of samples in the current voltage sample set for the first historical period is less than a preset number, then the battery degradation sensitive voltage range is determined to be the preset sensitive voltage range; otherwise: The average value of all samples in the voltage sample set is calculated as the center of the degradation-sensitive voltage Vc, and the standard deviation of all samples is calculated as the radius of the degradation-sensitive voltage distribution ΔVc. The battery degradation-sensitive voltage range is defined as [Vc-0.5×ΔVc, Vc+0.5×ΔVc]; Wherein, the lower limit of the battery degradation sensitive voltage range is not lower than Vmin and the upper limit is not higher than Vmax, where Vmin is the battery pack discharge termination protection voltage and Vmax is the battery pack rated equalization charge voltage. 6.The IoT-based smart device fault detection method of claim 5, wherein, The voltage degradation sensitivity factor fv is determined based on the average battery pack voltage Vn and the battery degradation sensitivity voltage range during the current analysis period: when Vn belongs to the battery degradation sensitivity range, fv=|Vn-Vc| / △Vc; otherwise, fv=0. The health risk metric He is determined by weighting and fusing thermal runaway risk measures, power supply stress urgency, and voltage degradation sensitivity factors.

7. The method for fault detection of smart devices based on the Internet of Things according to claim 6, characterized in that, The average value of the peak-to-peak ripple voltage Vrp of the target RRU power supply branch during the analysis period is calculated, and then the ripple coefficient RF is calculated in combination with the rated output voltage Ve, RF=Vrp / Ve; then the first correction factor Z1 is determined, Z1=1-exp(-RF / RFm); The peak value Ip and valley value Iv of the instantaneous current of the target RRU power supply branch are calculated within the analysis period. The peak-valley ratio PCR of the instantaneous current is calculated, PCR=(Ip-Iv) / Iv, and then the second correction factor Z2 is determined, Z2=tanh(PCR / PCRm). The dynamic stress coefficient ES of the power supply link is determined by combining the first correction factor Z1 and the second correction factor Z2, where ES = ρ1 × Z1 + ρ2 × Z2. The health risk metric is adjusted according to the dynamic stress coefficient ES of the power supply link, and the adjusted health risk metric is set as He1, He1=min(1,He×(1+η×ES)); Where Ip is the 99th percentile of the instantaneous current sampling value within the analysis period, Iv is the 1st percentile of the instantaneous current sampling value within the period, ρ1 is the ripple stress weight, ρ2 is the dynamic tensile stress weight, ρ1+ρ2=1, RFm is the preset ripple coefficient threshold, PCRm is the preset instantaneous current peak-to-valley ratio abnormal threshold, and η is the preset adjustment coefficient. 8.The IoT-based smart device fault detection method of claim 7, wherein, At the end of the current control cycle, extract the He1 value of each analysis cycle, take its 90th percentile as the control input risk value Hec, and then determine the floating charge voltage adjustment amount ΔVf, ΔVf=-ΔVmax×Hec, where ΔVmax is the preset adjustment range threshold. The float charge voltage Vfn for the next control cycle is determined based on the float charge voltage adjustment ΔVf, where Vfn = Vfe + ΔVf, and Vfe is the rated float charge voltage. If Vfn is lower than the preset safety lower limit Vfmin, then let Vfn = Vfmin, where Vfmin is the preset safety lower limit.

9. An Internet of Things based smart device fault detection apparatus applied to the Internet of Things based smart device fault detection method according to any one of claims 1-8, characterized in that, include: The thermal risk determination unit is used to identify the thermal imbalance state based on the surface temperature of a single cell, and then determine the thermal runaway risk measure. The power supply stress determination unit is used to determine the power supply stress urgency based on the temperature and output current of the switching power supply rectifier module. The interval calculation unit is used to extract equalization charging events based on the battery pack float charging current of the first historical cycle and determine the battery degradation sensitive voltage interval. The integrated risk determination unit is used to determine the health risk measure based on the thermal runaway risk measure, the power supply stress urgency and the battery degradation sensitive voltage range, and to adjust the determination process of the health risk measure based on the peak-to-peak value of the ripple voltage of the target RRU power supply branch and the instantaneous current of the RRU power supply branch. An adjusting unit is configured to generate a floating voltage adjusting value according to the adjusted health risk metric, so as to determine the floating voltage of the next control period.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, wherein the computer program is used to control an electronic device where the computer readable storage medium is located to execute the smart device fault detection method based on Internet of Things in any one of claims 1-8 when running.

Images