Temperature monitoring and liquid cooling control system of energy storage battery

Abstract

The invention discloses a temperature monitoring and liquid cooling control system of an energy storage battery, which relates to the technical field of battery thermal management, and is characterized in that forward and reverse hydraulic flow fields are constructed, the temperature data of a hot area is converted into a numerous number group representing the flow channel topology, and the heat distribution solidification characteristic of a branch inactivation area caused by an air plug when the flow direction is changed is utilized to realize the temperature monitoring and liquid cooling control of the energy storage battery. And accurately locking the fault branch by calculating the sorting consistency and executing the set intersection operation. On the basis, the system automatically matches the optimal exhaust flow direction with a target exhaust valve according to a fault positioning result, controls the liquid cooling pump to execute a non-constant flow excitation exhaust action based on bubble dynamics response time, and verifies a repair effect by quantifying topological characteristic changes before and after exhaust. According to the method, on the premise that branch flow sensors are not additionally arranged, accurate diagnosis and repair of parallel flow channel air plug faults are achieved.

Description

Technical Field

[0001] This invention relates to the field of battery thermal management technology, and more specifically to a temperature monitoring and liquid cooling control system for energy storage batteries. Background Technology

[0002] With the rapid iteration of new energy vehicles and energy storage technologies, battery pack assembly efficiency is constantly improving, and modular, highly integrated architectures are gradually becoming mainstream. To address the thermal management challenges brought about by high-power charging and discharging, such battery systems typically employ a liquid-cooled heat exchange structure with a shunt manifold and multiple parallel flow channel cooling plates. This structure utilizes a main pipe to distribute coolant to dozens or even hundreds of parallel branches, aiming to achieve uniform heat dissipation from the densely arranged cell area through a parallel network, maintaining the temperature consistency of the battery system under different operating conditions.

[0003] However, in the actual operation of liquid cooling systems, due to limitations in pipe assembly processes, coolant gas evolution, or minor leaks after long-term operation, free gas inevitably exists in the circulation loop. This gas tends to accumulate at high points in the flow channel or in local cavities, forming gas pockets. When these gas pockets occupy critical flow sections in parallel branches, the hydraulic resistance of that branch increases dramatically, forcing the coolant to divert to other low-resistance branches, causing the affected area to lose effective convective heat transfer capacity. Due to strict cost and space constraints in engineering applications, it is difficult to individually configure flow or differential pressure monitoring devices for each tiny parallel branch. Battery management systems often only obtain sparsely distributed thermal zone temperature data. Without branch-level flow status feedback, existing monitoring methods cannot accurately distinguish between the temperature rise difference caused by localized flow channel gas blockage and high-power battery heat generation, relying solely on absolute temperature values. This results in the system's inability to promptly identify and eliminate faulty branches in a quiescent inactivation state, posing a hidden risk of thermal runaway. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a temperature monitoring and liquid cooling control system for energy storage batteries. This system solves the problem that, in the absence of branch-level flow monitoring devices, it is impossible to effectively identify and eliminate branch hydraulic deactivation faults caused by air pocket accumulation in parallel flow channels, thus making it difficult to eliminate the potential for localized overheating of the battery pack.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a temperature monitoring and liquid cooling control system for an energy storage battery, comprising: The baseline fingerprint module is used to control the liquid cooling system to maintain the positive liquid supply mode and collect hot zone temperature data, and convert the hot zone temperature data into a positive ranking array according to the preset sorting rules. The topology diagnostic module is used to control the liquid cooling system to switch to reverse liquid supply mode and collect hot zone temperature data to convert it into a reverse ranking array. Based on the forward and reverse ranking arrays, it calculates the ranking consistency, which characterizes the degree of heat distribution solidification. The fault location module is used to filter out the abnormal hot zone set that is within the preset ranking threshold range in both the forward and reverse ranking arrays, and convert the abnormal hot zone set into a set of suspected inactive branches at the flow channel level according to the preset mapping relationship. The directional exhaust module is used to determine the target exhaust flow direction and target exhaust valve based on the set of suspected deactivated branches, and drive the actuator to perform coordinated exhaust actions including flow direction switching and valve opening and closing; The effect verification module is used to reacquire the system's sorting consistency and abnormal hot zone set after the coordinated exhaust action is executed, and compare them with the sorting consistency and abnormal hot zone set before the coordinated exhaust action is executed to determine the conduction recovery status of the deactivated branch.

[0006] Compared with existing technologies, it has the following advantages: This solution proposes a temperature monitoring and liquid cooling control system for energy storage batteries. By constructing two different hydraulic flow states, forward and reverse, the absolute temperature values, which are easily affected by the environment and power, are transformed into a rank array reflecting the flow channel topology. By utilizing the high-temperature location locking behavior of the hydraulic deactivation area caused by airlock when the flow direction changes, the parallel branches in a silent fault state can be accurately identified and screened without adding branch flow monitoring devices. This effectively solves the technical problem that traditional monitoring methods cannot distinguish the temperature rise difference caused by flow channel air blockage and high-power heat generation of the battery, and eliminates the safety blind spot caused by the invisible flow channel state in the thermal management system of energy storage batteries. This solution establishes a spatial mapping relationship from the hot zone to the flow channel actuator based on the fault location results. It automatically matches the optimal exhaust flow direction with the target high-point exhaust valve and controls the liquid-cooled pump to output a non-constant flow excitation waveform at a specific frequency. It effectively overcomes the adhesion force and capillary resistance of the gas plug by utilizing the shear force and resonance effect generated by fluid pulsation, and achieves targeted unblocking of the faulty branch. At the same time, it introduces a convergence comparison mechanism based on the decrease in sorting consistency and the change in the base number of the abnormal set to quantitatively evaluate the recovery of the branch's conductivity after the exhaust operation. It constructs a closed-loop control logic that integrates fault diagnosis, targeted repair and efficacy verification, which significantly improves the self-healing capability and operational safety of the highly integrated battery system. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the system framework of the present invention.

[0008] Figure 2 This is a schematic diagram of the system flow of the present invention. Detailed Implementation

[0009] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] Please see Figures 1 to 2 This application provides a temperature monitoring and liquid cooling control system for energy storage batteries, including a reference fingerprint module, a topology diagnostic module, a fault location module, a directional venting module, and an effect verification module; The reference fingerprint module sends a drive signal to the actuator of the liquid cooling system, locking the manifold reversing valve in the forward liquid supply mode (Mode A) and confirming that the liquid cooling pump is in effective operation. After the flow field stabilizes, the controller synchronously collects real-time raw temperature data of Nz hot zones through a temperature sensor array, and performs noise reduction processing on the sampled values ​​of each hot zone to generate a forward temperature array Ta.

[0011] Specifically, denoising is performed to filter out transient noise caused by electromagnetic interference or unstable sensor contact. In this preferred embodiment, a three-point median filtering algorithm is used. The calculation process for the i-th element Ta[i] in the positive temperature array Ta is as follows:

[0012] In the formula, i is the hot zone number index (the value ranges from 1 to Nz); This represents the original temperature sample value of the i-th hot zone at the current sampling time k; This indicates the median operation. That is, the value of Ta[i] is equal to the median of the values ​​of the three consecutive sampling points of the hot zone at the current time, the previous time, and the time before that. This formula eliminates the interference of fluid dynamic fluctuations on the establishment of the reference by locking the flow direction, and effectively suppresses non-physical spike noise by utilizing the median filtering characteristics, thus establishing reliable positive hydraulic and thermodynamic reference data.

[0013] The controller performs a sorting transformation operation on the positive temperature array Ta to generate a positive ranking array Ra. To ensure the determinism of the algorithm output under thermal equilibrium conditions, this module adopts a cascade sorting rule that prioritizes numerical values ​​and then indexes.

[0014] Specifically, the forward ranking array Ra is an integer sequence of length Nz, where the i-th element Ra[i] represents the relative high temperature ranking of the i-th heat zone in the whole system under forward liquid supply conditions. Its sorting rules are as follows: Primary criterion (numerical comparison): Compare the Ta values ​​of each thermal zone. The zone with the higher temperature value has a higher Ra ranking (for example, the ranking value is set to 1). Secondary criterion (index anchoring): When the Ta values ​​of two or more hot zones are completely equal, compare their hot zone index i. The one with the smaller index i has a higher Ra ranking.

[0015] In detail, the analog temperature value is converted into a ranking that only represents the relative strength relationship. This is because the branch inactivation caused by the gas lock will keep its coverage area in a relatively high temperature state. By ranking the data, the influence of the ambient temperature base value and the charging and discharging power can be separated, and the topological information reflecting the flow distribution characteristics of the flow channel can be accurately extracted. This avoids the false alarm or missed alarm problems that are easy to occur in different seasons or operating conditions by the traditional temperature difference threshold method.

[0016] It should be noted that in this embodiment, the positive ranking array Ra is a process variable that is updated in real time with the sampling period. The index-anchored secondary sorting rule is introduced because when the battery pack is in thermal equilibrium or under low load, it is very easy for large areas of hot zones to have completely equal temperature values ​​within the sensor's digital accuracy range (e.g., all at 25°C). If sorting is based solely on numerical values, the randomness of conventional computer sorting algorithms may cause the positive ranking array Ra to undergo high-frequency jumps in the same state, thus interfering with the stability of the sorting consistency calculation in subsequent modules. This module introduces a structural index as a deterministic anchor point to ensure the uniqueness and reproducibility of the generated sorting fingerprint.

[0017] It should be further explained that although the binary sorting rule of numerical priority followed by indexing used in this embodiment effectively eliminates the ranking jump noise caused by the randomness of the computer algorithm, when the battery pack is in a state of long-term static or absolute thermal equilibrium (i.e., after the liquid supply mode is changed, the temperature values ​​of all hot zones are still completely equal within the sensor's accuracy range), this rule will inevitably cause the forward ranking array Ra to be completely consistent with the reverse ranking array Rb obtained by the subsequent topology diagnostic module, thus making the sorting consistency degree Sscore calculated by the topology diagnostic module equal to 1. In order to avoid misjudging such a normal static thermal equilibrium state as an airlock fault (airlock faults also manifest as a high Sscore), the topology diagnostic module of this system presets a diagnostic start threshold. Specifically, before performing the sorting transformation, the controller will first calculate the statistical characteristics (e.g., maximum temperature difference or temperature variance) of the currently collected Nz hot zone temperature data. Only when the maximum temperature difference exceeds a preset active threshold (e.g., 2°C), indicating that the liquid cooling system is under effective load and the flow field is substantially affecting the heat distribution, will the controller continue with the subsequent sorting consistency Sscore calculation and fault determination process. If the maximum temperature difference is less than the threshold, the controller determines that the current system is in a thermal equilibrium static state, directly suspends the current diagnostic task, and maintains regular monitoring. This design, by introducing the temperature difference dimension as a pre-filter condition, combines the deterministic advantages of cascaded sorting rules with the complexity of actual operating conditions, ensuring high confidence in fault diagnosis.

[0018] Specifically, this module establishes a baseline thermal distribution model of the liquid cooling system under forward hydraulic conditions, and transforms absolute temperature data, which is susceptible to environmental and operating condition drift, into a structured fingerprint characterizing the flow channel topology through signal transformation technology. Before executing this module, the controller has pre-set and read the battery pack structural parameters, including the total number of hot zones Nz, the hot zone-to-branch mapping table, and the branch-to-high point mapping table.

[0019] The topology diagnostic module involves the controller sending a switching command to the manifold reversing valve of the liquid cooling system according to a preset diagnostic sequence. This drives the reversing valve to switch to the reverse liquid supply mode (Mode B) and maintains the operation of the liquid cooling pump. To ensure the validity of subsequent comparison data, this module should be executed immediately after the baseline fingerprint module is completed, or it should be confirmed that there is no abrupt change in the overall heat generation power of the battery pack during the interval between the baseline fingerprint module and the topology diagnostic module, to ensure relatively stable heat load.

[0020] After the reversing valve is in position, the controller maintains the operation of the liquid cooling pump and waits for the preset flow field stabilization time to ensure that the reverse hydraulic flow field is established and stable. Subsequently, the controller again synchronously collects real-time raw temperature data of Nz hot zones through the temperature sensor array and performs noise reduction processing consistent with the algorithm in the benchmark fingerprint module to generate the reverse temperature array Tb. Specifically, the preset flow field stabilization time is usually set to 10 to 60 seconds.

[0021] Specifically, the reverse coolant supply mode (ModeB) is a hydraulic operating state where coolant flows into the cooling plate from its designed outlet and out from its designed inlet. In this state, the fluid flow direction in each parallel branch is reversed compared to the forward coolant supply mode (ModeA), causing the original upstream hot zone to become the downstream hot zone. The reverse temperature array Tb is a vector containing Nz denoised temperature values.

[0022] In detail, a control group was constructed by changing the hydraulic boundary conditions through the actuator. If the flow channel is unobstructed, the cooling priority of the heat source will reverse with the change of flow direction, resulting in a significant migration of heat distribution. If there is an airlock causing the branch to become inactive, the thermal state of the branch will be completely dominated by the self-heating of the battery cell due to the lack of effective flow, exhibiting a solidified characteristic that does not change with the change of flow direction.

[0023] The controller performs a sorting transformation operation on the inverse temperature array Tb to generate an inverse ranking array Rb. To ensure the comparability of the forward and inverse data, the sorting rule for generating Rb must be consistent with the rule for generating the forward ranking array Ra in the baseline fingerprint module, i.e., following a cascaded sorting logic of numerical priority followed by index priority. Subsequently, the controller calls the temporarily stored Ra and the currently generated Rb in memory to calculate the sorting consistency score Sscore. The formula for calculating the sorting consistency score Sscore is as follows:

[0024] In the formula, Nz represents the total number of hot zones, and i represents the hot zone index. Here, is an indicator function, where 1 indicates that the condition is met and 0 indicates otherwise. Specifically, this formula means that the controller traverses all hot zones from 1 to Nz, comparing the ranking Ra[i] of the hot zone in the forward state with the ranking Rb[i] in the reverse state to see if they meet the consistency condition. If they do, the count is incremented by 1; otherwise, no count is added. Finally, the accumulated count is divided by the total number of hot zones Nz to obtain the normalized ratio, which is the sorting consistency score Sscore.

[0025] It should be noted that this module calculates the ranking consistency based on statistical principles; therefore, the total number of hot zones, Nz, has a valid lower limit. To ensure statistical significance of the Sscore, the preferred value of Nz is not less than 10, and the preferred range is 20 to 100. If Nz is too small (e.g., less than 6), random fluctuations in a single hot zone will cause drastic jumps in the Sscore. In this case, it is recommended to use a weighted calculation method instead of the basic normalization formula, thus reducing the module's sensitivity to large-granularity data.

[0026] In detail, this formula uses the change in fluid flow direction to identify the formation mechanism of heat distribution. In normal liquid cooling cycle, the cell temperature field is mainly controlled by the forced convection heat transfer of the coolant. As the coolant continuously absorbs heat along the flow path, the temperature in the upstream area of ​​the channel is lower and the temperature in the downstream area is higher. When the main pipe supply direction is reversed, the upstream and downstream positions inside the cooling plate are interchanged, which will inevitably cause a significant rearrangement and migration of the temperature ranking of each hot zone. However, when a specific branch is in a state of hydraulic deactivation due to airlock, the area loses the heat-carrying effect of the coolant flow, and its heat transfer mode degenerates into the accumulation of heat generated by the cell itself. At this time, no matter how the main pipe supply direction is changed, the area is in a state of fluid stagnation, which causes the area to always be in a relatively high temperature and high ranking in the whole system, showing a thermal solidification characteristic that does not change with the flow direction. Therefore, the ranking consistency degree, as a statistical quantitative indicator, indicates that the heat distribution is more independent of the flow direction, thus directly reflecting the objective fact that there is airlock blocking the flow inside the channel.

[0027] Specifically, by calculating the sorting consistency score (Sscore), the complex morphology of the overall temperature distribution is reduced to a single topological feature scalar. When the Sscore approaches 0, it indicates that the thermal distribution is reconstructed with changes in flow direction, suggesting normal convection function in the flow channel. When the Sscore approaches 1, it indicates that the thermal distribution exhibits solidification characteristics decoupled from the flow direction, suggesting the presence of airlocks causing the affected branches to remain in a relatively high-temperature state. This method effectively avoids the dependence of the absolute temperature difference method on ambient temperature and power conditions, achieving highly robust fault feature extraction.

[0028] It should be noted that, in this embodiment, the criterion condition in the indicator function can preferably be configured as tolerance matching logic in practical engineering applications. Considering the sensor measurement error and the influence of transient thermal capacity, ranking shifts may be accompanied by small-range neighbor jitter. Therefore, a non-negative integer tolerance value is set. ,For example The value is 1, when the absolute value of the difference between Ra[i] and Rb[i] is less than or equal to 1. When the ranking of a hot zone is determined to be substantially the same in both states, the indicator function outputs 1. Furthermore, the ranking consistency score (Sscore) is essentially a statistical indicator that measures the correlation or overlap between two ranking sequences. Besides the normalized counting formula mentioned above, another embodiment can use unnormalized count values, or a weighted consistency calculation that assigns different weights to different hot zones, such as giving higher weights to high-temperature hot zones. As long as the core logic is based on diagnosing the flow channel state by whether the ranking changes before and after the changeover, these should be considered equivalent variations of the technical concept of this invention. Regarding tolerance values... The selection of [the appropriate sensor] should be related to the measurement accuracy of the temperature sensor array and the average temperature gradient within the battery pack. Typically, The value should not exceed 5% of the total number of hot zones Nz. In specific engineering embodiments, if the temperature acquisition error is ±0.5℃ and the average temperature gradient between adjacent hot zones is less than 1℃, it is recommended to use [the appropriate value]. Set to 1 or 2. The core of this range is to filter out non-substantial ranking fluctuations caused by the sensor's own noise, preventing measurement errors from being mistakenly identified as topology changes.

[0029] Specifically, this module closely follows the benchmark fingerprint module. Its core lies in using the change of flow direction as an actively applied hydraulic excitation. By constructing a reverse flow field and comparing the differences in the thermal zone ranking distribution of the system under the two hydraulic states of forward and reverse, the ranking consistency Sscore, which characterizes the degree of solidification of the flow channel topology, is calculated.

[0030] The fault location module involves the controller reading the forward ranking array Ra and the reverse ranking array Rb from the memory and calling the preset extraction quantity parameter Nk. The controller extracts the hot zone numbers that rank in the top Nk positions under both forward and reverse liquid supply states. Through set intersection operations, it filters out the abnormal hot zone set Zset that exhibits relatively high temperatures in both flow states.

[0031] It should be noted that the interception quantity parameter Nk, which is the preset ranking threshold, is a positive integer used to define the range of relatively high temperatures to be examined. Its value setting mainly depends on the flow channel topology of the battery pack liquid cooling system and the granularity of the hot zone division. The value of parameter Nk is positively correlated with the number of hot zones covered by a single parallel flow channel branch. If the average number of hot zones continuously covered by each parallel flow channel branch in the design is m, then the preset value of Nk is usually set to an integer greater than or equal to m. This ensures that when a single branch is completely deactivated due to airlock, all affected hot zones covered by it can be effectively included in the top Nk high-temperature examination sequence, thereby preventing missed detections due to an overly narrow interception range.

[0032] In a preferred embodiment, considering the optimization of computing resources and the accuracy of fault location, the typical value of Nk is usually an integer between 2 and 6, or set to 10% to 20% of the total number of parallel flow channel branches Nb. The core of this value is to achieve a balance between fault capture sensitivity and anti-interference capability. If the value of Nk is set too small, for example, only 1, it may miss real regional faults due to sensor random noise or transient fluctuations in a single hot zone. If the value of Nk is set too large, it may introduce unstructured random high-temperature hot spots into the intersection calculation, leading to a decrease in the focus of fault location. Therefore, by reasonably presetting the value of Nk according to the flow channel structure, the effective coverage and accurate locking of the abnormal hot zone set Zset of gas plug fault characteristics can be guaranteed to the greatest extent.

[0033] Specifically, the Zset of abnormal hot zones is constructed based on the following logical formula:

[0034] In the formula, i is the hot zone index. This formula means that the controller traverses all hot zones and determines whether a hot zone i simultaneously meets the following two conditions: first, its value in the forward ranking array Ra is less than or equal to Nk; second, its value in the reverse ranking array Rb is less than or equal to Nk. If both conditions are met, the hot zone is determined to be a persistent overheated zone across the flow direction, and its index i is added to the abnormal hot zone set Zset.

[0035] In detail, a rigorous fault screening mechanism is constructed by using intersection operations. In a normal liquid cooling system, it is extremely difficult for a single hot zone to maintain an extremely high temperature ranking after the flow direction is reversed. Only when the flow channel branch corresponding to the hot zone is blocked and deactivated will the area exhibit a stubborn high temperature characteristic that is independent of the flow direction. Therefore, Zset can effectively filter out occasional high temperature points caused by load fluctuations, sensor drift or fluid turbulence, and accurately locate areas with structural heat dissipation obstacles.

[0036] The controller calls the hot zone to branch mapping table MapZB stored in non-volatile memory to perform spatial mapping transformation on the abnormal hot zone set Zset, converting the fault information at the hot zone level into control objects at the flow channel level, and generating a suspected inactive branch set Bset.

[0037] Specifically, the hot zone to branch mapping table MapZB is offline calibration data based on the battery pack cold plate flow channel design. It defines which one or more parallel flow channel branches j are mainly used to cool and cover any hot zone number i. The generation process of the suspected deactivated branch set Bset is as follows:

[0038] In the formula, Bset is the set of suspected deactivated branches. Specifically, the controller traverses each hot zone number i in the abnormal hot zone set Zset, queries the hot zone to branch mapping table MapZB to obtain the cooling branch number j corresponding to the hot zone, and adds j to the suspected deactivated branch set Bset. During this process, if multiple abnormal hot zones correspond to the same parallel flow channel branch, or if the branches mapped by different hot zones overlap, the controller will perform a deduplication operation to ensure that the branch numbers in the suspected deactivated branch set Bset are unique. Specifically, the deduplication operation is a necessary result of the union operation at the data processing level, rather than a specific single algorithm step. Regardless of whether the automatic filtering characteristics of hash sets, bitmap marking method, or explicit sorting and deduplication algorithm are used to implement the above logic in the specific implementation, as long as the purpose is to summarize the branches corresponding to multiple hot zones into a set of non-repeating control objects, they all fall within the scope of the union mapping technology described in this embodiment.

[0039] Specifically, by generating a set of suspected deactivated branches Bset, the transformation from the phenomenon space to the entity space is completed. Temperature sensors can only reflect abnormal heat distribution and cannot directly indicate the location of flow channel blockage. Through the structural mapping of this module, the controller transforms the set of abnormal temperatures Zset into a specific set of control objects Bset that can be operated by the actuator, providing a clear target guide for the subsequent directional exhaust recovery action.

[0040] It should be noted that this module implicitly contains necessary process control logic. If the calculated abnormal hot zone set Zset is empty, it indicates that there are no persistent hot spots across the flow direction in the system. The controller will determine that there is no airlock fault, directly terminate the current diagnostic process and return to the normal thermal management strategy, and no longer execute subsequent mapping and exhaust operations. This judgment logic avoids invalid actions of the system and reduces the wear of the actuator.

[0041] Specifically, this module is a crucial link connecting the signal characteristic space and the physical pipeline space. Following the topology diagnostic module's calculation of the sorting consistency score (Sscore), which characterizes the degree of solidification of the overall heat distribution, this module focuses on solving the problem of fault location. Although the sorting consistency score (Sscore) can reflect the trend of flow channel convection failure in the system from a statistical perspective, it cannot directly indicate the specific location of the failure. Therefore, based on the high-temperature overlap characteristics of the bidirectional ranking data, this module uses set operation logic to eliminate occasional hot spot interference, extracts the target area with structural heat dissipation obstacles from the overall hot zone, and locks down the specific faulty flow channel branch through spatial mapping.

[0042] In the directional exhaust module, the controller calls the suspected deactivated branch set Bset, as well as the preset branch-to-high point mapping table MapBH and the branch-to-preferred flow direction mapping table MapDir, to parse out the target exhaust valve object to be operated and the corresponding optimal flushing flow direction.

[0043] Specifically, the controller iterates through each branch number j in the suspected deactivated branch set Bset, retrieves the manifold high-point exhaust valve number h associated with that branch by querying the branch-to-high-point mapping table MapBH, and stores all related exhaust valve numbers after deduplication into the target high-point set Hset. Then, for each high-point position in the target high-point set Hset, the controller queries the branch-to-preferred flow direction mapping table MapDir to determine the liquid supply direction more conducive to pushing gas to the exhaust port of that high-point. If the preferred flow direction corresponding to all high points involved in the target high-point set Hset is consistent, then that direction is confirmed as the target exhaust flow direction Dir. If the target high-point set Hset contains multiple exhaust valves and their corresponding preferred flow directions are inconsistent, the controller calculates the target exhaust flow direction Dir for this operation according to a preset conflict handling strategy. The conflict handling strategy preferably adopts a time-sharing execution logic; that is, if there are both forward and reverse exhaust demands, the action is split into two sequences and executed sequentially, or the one with the higher weight is selected as the single target exhaust flow direction Dir based on the number of branches involved. Specifically, the exhaust strategy is directional based on the flow channel geometry. By using a lookup table to make decisions, the fluid dynamics are always directed in a direction that is conducive to gas exhaust, avoiding backflow or stagnation of bubbles due to improper flow direction, thereby improving the success rate of exhaust operation.

[0044] Based on the determined target exhaust flow direction Dir and the target high point set Hset, the controller sends timing control commands to the actuators of the liquid cooling system to execute a coordinated action sequence including pump speed adjustment, flow direction switching and valve opening and closing, and generates an action log.

[0045] Specifically, the execution process of the coordinated action sequence is as follows: First, the controller sends a command to reduce the speed of the liquid cooling pump to below a safe threshold to reduce the static pressure in the pipeline and prevent water hammer damage to components during flow direction switching; Second, the controller drives the manifold reversing valve to the target exhaust flow direction Dir and confirms that the valve is in place; Third, the controller sends a command to open all high-point exhaust valves included in the target high-point set Hset, connecting them to the gas-liquid separation chamber; Fourth, the controller modulates the operating parameters of the liquid cooling pump to execute a specific exhaust waveform, using the fluid disturbance generated by the non-constant flow to break through the surface tension of the gas plug and carry it to the exhaust port for discharge; Fifth, after the exhaust continues for a preset time, the controller closes the exhaust valves and restores the pump to its normal operating state. The ActionLog stores the target exhaust flow direction Dir used in this operation, the valve number set Hset that was opened, and the pump speed waveform parameters for subsequent verification. A powerful removal mechanism for airlocks is constructed through multi-dimensional collaboration between pumps and valves. In particular, the timing control of first reducing speed and then reversing direction protects pipeline components, while a specific pump speed waveform combined with an exhaust valve can effectively disrupt the stable stagnation state of airlocks at low flow rates, forcing the gas to migrate out of the working flow channel with the fluid, thereby relieving the hydraulic deactivation state of the branch.

[0046] It is important to note the safety threshold speed for preventing water hammer. This parameter aims to limit the pressure surge in the pipeline during flow direction switching. This threshold is typically set to 10% to 30% of the rated maximum speed of the liquid-cooled pump, or to a baseline flow rate that maintains a full pipeline but with extremely low dynamic pressure (e.g., the speed corresponding to 5 L / min to 10 L / min). Operating the directional valve at this speed avoids instantaneous localized overheating caused by fluid stagnation and keeps pressure surges within 20% of the pipeline's pressure limit.

[0047] It should be noted that the specific exhaust waveform is designed to create a sudden, instantaneous change in pressure gradient to overcome the capillary pressure of the gas plug. A configurable liquid-cooled pump can execute sinusoidal flow oscillations, square wave pressure pulses, or short-duration step-like high-speed scavenging. This type of unsteady flow control can generate shear forces different from conventional laminar flow, which helps to detach bubbles adhering to the flow channel walls. To effectively utilize fluid dynamics to overcome the capillary resistance of the gas plug, the unsteady flow waveform executed by the liquid-cooled pump should possess specific frequency characteristics. The preferred waveform frequency range is 0.5 Hz to 2 Hz (i.e., a period of 0.5 seconds to 2 seconds). This frequency range is set based on the dynamic response time of bubbles in the pipeline (i.e., the minimum time required for bubbles to overcome wall adhesion and generate displacement under the action of fluid pressure waves). If the frequency is too high (greater than 2 Hz), the duration of the fluid pressure pulse is shorter than the dynamic response time of the bubbles, and the bubbles only undergo in-situ deformation and cannot detach. If the frequency is too low (less than 0.5 Hz), an effective excitation effect cannot be formed. Therefore, this application defines this characteristic as the bubble dynamic response time. This frequency band matches the time constant of the dynamic response of bubbles in fluid, and can induce resonance or instability at the gas-liquid interface. Simultaneously, the waveform's flow amplitude fluctuation range should reach 50% to 150% of the conventional flow rate to generate shear force peaks sufficient to overcome the surface tension of the gas plug.

[0048] It should be noted that the preset venting time should be sufficient for bubbles to migrate from the farthest end of the flow channel to the venting valve. Based on the flow channel length and average flow rate, this time window is typically set between 10 and 45 seconds. Too short a time may cause bubbles to only move to the middle of the flow channel without being vented, potentially leading to backflow after resetting. Too long a time will consume normal thermal management resources, resulting in deterioration of battery temperature uniformity.

[0049] Furthermore, if the system detects that the input set of suspected deactivated branches, Bset, is empty, indicating that there are no faulty branches requiring processing, the controller will skip this module directly, without performing any valve actions or pump speed adjustments, and mark the ActionLog as empty or invalid, directly proceeding to subsequent processes to avoid unnecessary disturbances to the normally operating system. This process ensures the targeting of control actions and avoids energy consumption and wear caused by blind operation.

[0050] Specifically, this module is the execution and repair link of the present invention. Based on the locked set of suspected deactivated branches Bset, this module actively adjusts the combination state of hydraulic flow direction and exhaust passage to construct a targeted hydrodynamic field to break the mechanical balance established by the gas plug at the high point of the flow channel, and forces the accumulated gas out of the circuit, thereby restoring the liquid conduction and convection heat transfer capacity of the deactivated branch.

[0051] In the effect verification module, after the controller reads the ActionLog to confirm that the exhaust combination action has been completed, it drives the liquid cooling system to enter the retest process. To avoid the hydraulic disturbance during the exhaust process affecting the measurement accuracy, the controller first maintains the operation of the liquid cooling pump and waits for the preset flow field stabilization time. Then, based on the data acquisition and calculation logic from the benchmark fingerprint module to the fault location module, it re-acquires the thermal topology characteristic parameters of the current system and marks them as the post-exhaust state data.

[0052] Specifically, the controller first marks the existing abnormal hot zone set Zset and sorting consistency score Sscore in the memory and transfers them as pre-exhaust baseline data ZsetPre and SscorePre. Then, the controller controls the reversing valve to establish a stable flow field sequentially in forward liquid supply mode ModeA and reverse liquid supply mode ModeB, simultaneously collecting temperature data and generating new forward ranking array Ra and reverse ranking array Rb. Based on the updated bidirectional ranking data, the controller re-executes the consistency calculation formula of the topology diagnostic module to obtain the post-exhaust sorting consistency score SscorePost, and re-executes the set intersection operation of the fault location module to obtain the post-exhaust abnormal hot zone set ZsetPost. By adding a stabilization waiting period after the action and performing standardized retesting, the latest state data with time correlation is obtained, establishing an objective control group for evaluating the exhaust treatment effect and eliminating the interference of ambient temperature drift caused by excessive time span on the verification results.

[0053] It should be noted that the preset flow field stabilization time is used to eliminate the interference of fluid turbulence remnants and transient pressure fluctuations caused by pump speed excitation and valve switching in the directional exhaust module on temperature measurement. Its value should be greater than the theoretical time for the coolant to complete one full circulation cycle inside the liquid cooling plate. In engineering practice, this parameter is typically set between 10 and 60 seconds. Too short a waiting time may result in spurious fluctuations in the acquired temperature data caused by fluid inertia, while too long a waiting time reduces the timeliness of fault diagnosis.

[0054] The controller calculates the changes in parameters after exhaust relative to the baseline data before exhaust, evaluates the exhaust effect based on preset convergence criteria, and executes a hierarchical control strategy based on the evaluation results.

[0055] Specifically, the convergence criteria include set convergence criteria and consistency convergence criteria. The set convergence criterion states that the cardinality of the abnormal hot zone set ZsetPost after exhaust is less than the cardinality of the baseline data ZsetPre before exhaust, or ZsetPost becomes an empty set. The consistency convergence criterion states that the sorting consistency SscorePost after exhaust is less than the baseline data SscorePre before exhaust, and the difference exceeds a preset improvement threshold. If any or all of the above convergence criteria are met, the controller determines that the exhaust is successful, sends a command to terminate the diagnostic process, clears the retry count variable RetryCount to zero, drives the reversing valve to reset to the default flow direction, and restores the pump's normal thermal management control strategy. If the convergence criteria are not met, the controller reads the currently accumulated retry count variable RetryCount. If RetryCount has not reached the preset upper limit, the counter is incremented and the controller returns to the directional exhaust module to replan the flow direction and execute the exhaust. If RetryCount has reached the preset upper limit, it is determined to be an irreversible blockage fault. The controller maintains the current state and sends a power limit request and a thermal management fault alarm to the vehicle system. By quantitatively comparing the changes in topological characteristics before and after exhaust, a closed-loop confirmation of the control effect is achieved. The reduction of set elements directly proves that some deactivated branches have regained conduction, while the decrease in consistency indicates that the system's thermal distribution has regained its sensitivity to the flow direction. Combined with the retry counting mechanism and power limit protection, it not only ensures the removal of stubborn airlocks but also avoids the system falling into an ineffective operation loop and ensures battery safety.

[0056] It should be noted that the ranking consistency score (Sscore) in this embodiment represents the degree of solidification of abnormal heat distribution. When air resistance exists in the flow channel, the corresponding hot zone exhibits high-temperature characteristics in both forward and reverse liquid supply modes, resulting in minimal difference in their ranking and thus maintaining a high Sscore value. When the coordinated venting action successfully clears the air resistance, the temperature of the original abnormal hot zone returns to normal, and its position in the ranking array randomly shifts (i.e., it is no longer fixed at the top), leading to a decrease in the number of hot zones meeting the tolerance conditions, thereby causing a significant drop in the Sscore. The preset improvement threshold... This threshold is used to measure the magnitude of this decrease. It is set based on the proportion of hot zones covered by a single flow channel branch to the total number of hot zones in the system.

[0057] It should be noted that regarding the improvement threshold The threshold value is a dimensionless scalar used to distinguish between random noise in system measurements and substantial changes in the flow channel topology. Its setting primarily depends on the total number of hot zones, Nz. Since the minimum resolution of the Sscore is determined by 1 / Nz, the threshold value is improved to ensure the statistical significance of the judgment results. The value is typically set between 0.5 / Nz and 1.5 / Nz. In a typical power battery liquid cooling system embodiment, it is recommended to improve the threshold. The default value is a fixed value between 0.05 and 0.15. When the decrease in Sscore exceeds this threshold, it indicates that at least 5% to 15% of the hot zone has shown a ranking shift after the flow direction has changed, which is sufficient to prove that the air resistance in the flow channel has been effectively cleared or significantly alleviated.

[0058] It should be noted that the core of the set convergence criterion lies in comparing the potentials of the two sets. If the potential of ZsetPost is less than that of ZsetPre, it indicates that at least one hot zone that was originally in a state of sustained high temperature has exited the bidirectional high-temperature intersection after venting. That is, the thermal state of this region has begun to migrate with the change of flow direction, which in fluid dynamics directly corresponds to the restoration of effective flow in its branch. In addition, when judging whether the set of abnormal hot zones has improved, in addition to comparing the cardinality of the set elements, the logic of set inclusion relationship can also be used. That is, if the set of abnormal hot zones after venting, ZsetPost, is a proper subset of the set of abnormal hot zones before venting, ZsetPre, or the intersection of ZsetPost and ZsetPre is empty, it can be considered that the set convergence criterion is satisfied. This indicates that the original stubborn fault point has disappeared or has been transferred, proving the effectiveness of the fluid dynamic unblocking operation.

[0059] Furthermore, the retry count variable `RetryCount` is set to zero during the initialization of the baseline fingerprint module. The preset upper limit is typically set to 3 to 5 attempts. This mechanism is designed because some stubborn gas plugs may require multiple hydrodynamic impacts to completely expel them. However, if the characteristic parameters do not improve after multiple attempts, it indicates that the fault may be caused by solid foreign object blockage or pipe collapse / deformation, which cannot be resolved by venting. Terminating the attempt and triggering an alarm and power-limiting operation at this point is for safety considerations of protecting the liquid cooling pump and preventing thermal runaway of the battery due to prolonged localized overheating, meeting the redundancy requirements of functional safety design.

[0060] It should be noted that the power limiting request strategy after fault confirmation is as follows: When the retry count reaches the upper limit and is determined to be an irreversible blockage, the power limiting request issued by the controller aims to control the battery heat generation rate within a safe range that can be dissipated by heat conduction alone. Specific execution strategies include, but are not limited to: disabling DC fast charging, limiting regenerative braking power to zero or extremely low levels, and limiting the drive discharge rate to below 0.5C or limiting the motor output torque to 30% to 50% of the peak torque. This strategy actively reduces the thermal load, preventing the cells corresponding to the deactivated branch from inducing thermal runaway due to localized heat accumulation, thereby achieving safe operation of the system during degraded faults.

[0061] Specifically, this module is the verification and decision-making stage of the present invention, which aims to confirm the effect of the directional exhaust action of the directional exhaust module. By quantitatively comparing the thermal topology characteristic parameters before and after exhaust, it determines whether the hydraulic deactivation state of the branch is relieved, and executes a hierarchical strategy of reverting to normal control, cyclic retry, or fault reporting accordingly, thereby constructing a complete control system.

[0062] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A temperature monitoring and liquid cooling control system for an energy storage battery, characterized in that, include: The baseline fingerprint module is used to control the liquid cooling system to maintain the positive liquid supply mode and collect hot zone temperature data, and convert the hot zone temperature data into a positive ranking array according to the preset sorting rules. The topology diagnostic module is used to control the liquid cooling system to switch to reverse liquid supply mode and collect hot zone temperature data to convert it into a reverse ranking array. Based on the forward and reverse ranking arrays, it calculates the ranking consistency, which characterizes the degree of heat distribution solidification. The fault location module is used to filter out the abnormal hot zone set that is within the preset ranking threshold range in both the forward and reverse ranking arrays, and convert the abnormal hot zone set into a set of suspected inactive branches at the flow channel level according to the preset mapping relationship. The directional exhaust module is used to determine the target exhaust flow direction and target exhaust valve based on the set of suspected deactivated branches, and drive the actuator to perform coordinated exhaust actions including flow direction switching and valve opening and closing; The effect verification module is used to reacquire the system's sorting consistency and abnormal hot zone set after the coordinated exhaust action is executed, and compare them with the sorting consistency and abnormal hot zone set before the coordinated exhaust action is executed to determine the conduction recovery status of the deactivated branch.

2. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, The default sorting rule is a cascading sorting rule that prioritizes numerical values ​​and then indexes, including: Hot zone temperature data includes temperature values ​​and hot zone number index; The preset sorting rules include primary criteria and secondary criteria; The primary criterion is: compare the temperature values ​​of each thermal zone, and the thermal zone with the higher temperature value ranks higher in the forward or reverse ranking array. The secondary criterion is: when the temperature values ​​of two or more hot zones are equal, compare the hot zone number indices corresponding to each hot zone. The hot zone with the smaller hot zone number index ranks higher in the forward or reverse ranking array.

3. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, Before generating the forward and reverse ranking arrays, a three-point median filtering algorithm is used to denoise the collected hot zone temperature data. The denoising process involves obtaining three consecutive temperature sample values ​​at the current sampling time, the previous sampling time, and the time before that, and taking the median of these three sample values ​​as the denoised hot zone temperature data.

4. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, The specific methods for calculating sort consistency include: For each hot zone in the forward and reverse ranking arrays, calculate the absolute value of the difference between the ranking value of the hot zone in the forward ranking array and the ranking value in the reverse ranking array. The number of hot zones whose absolute value is less than or equal to the preset tolerance value is counted. Calculate the ratio of the number of hot zones to the total number of hot zones, and use this ratio as the sorting consistency. The preset tolerance value is a positive integer not exceeding five percent of the total number of hot zones.

5. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, The specific methods for filtering sets of abnormal hot zones include: Perform a set intersection operation to include hotspot numbers that simultaneously meet the following two conditions into the set of abnormal hotspots: Condition 1: The rank value of this hot zone in the positive rank array is less than or equal to the preset rank threshold; Condition 2: The rank value of the hot zone in the reverse rank array is less than or equal to the preset rank threshold; The preset ranking threshold is a positive integer, which is an integer between 2 and 6, or is between 10% and 20% of the total number of parallel flow channel branches included in the system.

6. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 5, characterized in that, Generate a set of suspected inactive branches, including: Iterate through each hot zone number in the abnormal hot zone set, and query the preset hot zone number to flow channel branch number mapping relationship to obtain the flow channel branch number corresponding to the hot zone number; Perform a set union operation on all the obtained flow channel branch numbers to generate a set of suspected inactive branches containing unique flow channel branch numbers.

7. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, Perform coordinated exhaust actions, including: Traverse each flow channel branch number in the suspected inactive branch set, query the preset branch-to-high-point exhaust valve mapping relationship, and determine the target exhaust valve corresponding to the flow channel branch number; Query the preset branch-to-preferred flow direction mapping relationship to determine the target exhaust flow direction corresponding to the branch number of the flow channel; The liquid cooling pump in the control actuator is reduced to a speed below a preset safe speed threshold. The manifold directional valve in the drive actuator switches to the target exhaust flow direction; Open the target exhaust valve to connect to the preset gas-liquid separation chamber; The liquid cooling pump is controlled to execute a non-constant flow exhaust waveform for excitation and exhaust; After the preset venting time, the target venting valve is closed and the liquid cooling pump is controlled to return to normal operation.

8. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 7, characterized in that, The non-constant flow exhaust waveform is one of the following: sinusoidal flow oscillation waveform, square wave pressure pulse waveform, or step-type high-speed scavenging waveform. The frequency of the non-constant flow exhaust waveform is set according to the bubble dynamics response time, and the frequency range is 0.5Hz to 2Hz.

9. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, When determining the target exhaust flow direction, the directional exhaust module employs the following conflict resolution strategies, including: When the set of suspected deactivated branches contains multiple flow channel branch numbers, and the target exhaust flow directions corresponding to each flow channel branch number determined according to the branch-to-preferred flow direction mapping relationship are inconsistent, execute any of the following strategies: The coordinated exhaust action is broken down into two independent actions corresponding to different target exhaust flow directions, and executed sequentially according to a preset order; The number of flow channel branch numbers corresponding to each target exhaust flow direction is counted, and the one with the most numbers is selected as the single target exhaust flow direction. The actuator is then driven to perform only the coordinated exhaust action corresponding to that single target exhaust flow direction.

10. The temperature monitoring and liquid cooling control system for an energy storage battery according to claim 1, characterized in that, The effect verification module determines the convergence comparison condition for the restoration of conduction of the deactivated branch by satisfying any one of the following conditions or both of the following conditions: Condition 1: The number of elements in the abnormal hot zone set re-acquired by the effect verification module is less than the number of elements in the abnormal hot zone set generated by the fault location module before the coordinated exhaust action is executed. Condition 2: The sorting consistency value re-acquired by the effect verification module is less than the sorting consistency value calculated by the topology diagnosis module before the execution of the coordinated exhaust action, and the difference between the two is greater than the preset improvement threshold. The improvement threshold is set based on the proportion of the number of hot zones covered by a single flow channel branch to the total number of hot zones.

Images