A method and system for oil temperature regulation of a gearbox lubricant

Abstract

The application relates to the technical field of oil temperature regulation, in particular to an oil temperature regulation method and system for gear box lubricating oil, and the application proposes the following scheme: obtaining oil temperature data, establishing a graph model in combination with a gear box structure, and constructing an oil temperature thermal diagram; calculating node heat potential based on the thermal diagram, positioning a node to be processed, and generating a basic loop connected with a cold end; determining the sequence of the node to be processed through entropy production sensitivity; presetting an oil temperature regulation matrix according to the node heat potential, calculating a temperature response kernel in combination with the basic loop, iteratively updating the matrix until convergence, and implementing oil temperature regulation accordingly. The method realizes accurate identification and distribution of hot spots, avoids global overcooling, reduces energy consumption while improving local cooling precision, and improves the operation reliability and service life of the gear box.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of oil temperature regulation, and in particular to an oil temperature regulation method and system for gear box lubricating oil. BACKGROUND

[0002] The existing gear box oil temperature regulation method mainly depends on the overall oil temperature signal as the regulation basis. When the overall oil temperature is in the normal range, the local area may still have overheating phenomenon, which cannot be effectively reflected and timely processed. Due to the lack of pertinence of the regulation means to the local hot spots, the global cooling is often started when the temperature rise is abnormal, resulting in a lag in response and difficulty in timely reducing the local temperature. On the other hand, the overall driving cooling device is easy to cause the overall oil temperature to drop excessively, causing an increase in energy consumption and abnormal lubricating oil viscosity, which adversely affects the transmission efficiency and operating life. In order to solve the above problems, the present application designs an oil temperature regulation method and system for gear box lubricating oil. SUMMARY

[0003] The technical problem to be solved by the present application is to provide an oil temperature regulation method and system for gear box lubricating oil, which acquires oil temperature data, establishes a graph model in combination with the gear box structure, and constructs an oil temperature thermodynamic graph. The node heat potential is calculated based on the thermodynamic graph, the nodes to be processed are located, and the basic loop connected with the cold end is generated. The order sequence of the nodes to be processed is determined by the entropy production sensitivity. The oil temperature regulation matrix is preset according to the node heat potential, the temperature response kernel is calculated in combination with the basic loop, the matrix is iteratively updated until convergence, and the oil temperature regulation is implemented accordingly. The present method realizes accurate identification and distribution of hot spots, avoids global overcooling, improves the local cooling precision while reducing energy consumption, and improves the operation reliability and life of the gear box.

[0004] To achieve the above purpose, the present application provides the following technical scheme:

[0005] An oil temperature regulation method for gear box lubricating oil, the method comprising:

[0006] Acquiring the oil temperature data of the lubricating oil, and constructing an oil temperature thermodynamic graph in combination with the graph model corresponding to the gear box structure;

[0007] Calculating the node heat potential according to the oil temperature thermodynamic graph and locating a plurality of nodes to be processed;

[0008] Generating a corresponding number of basic loops according to the nodes to be processed, calculating the entropy production sensitivity of the basic loops, and determining the order sequence of the nodes to be processed according to the entropy production sensitivity;

[0009] According to the node thermal potential, a current oil temperature regulation matrix is preset, an iteration operation is performed, a node to be updated is selected according to the sequence, a temperature response kernel of the node to be updated is calculated in combination with the basic loop, the temperature response kernel is updated and replaces the current oil temperature regulation matrix, until the sequence is empty, the current oil temperature regulation matrix is output, and oil temperature regulation is performed according to the current oil temperature regulation matrix.

[0010] In combination with the graph model corresponding to the gearbox structure, an oil temperature thermal diagram is constructed, including:

[0011] According to the thermal coupling relationship of the structure in the graph model, the gearbox is divided into nodes and edges, wherein the nodes represent local thermal capacity and oil quantity, and the edges represent thermal transfer connectivity and oil flow directional attributes;

[0012] According to the thermal transfer connectivity and the oil flow directional attributes, an edge function for representing heat contribution attenuation is calculated, wherein the edge function is calculated according to path length, residence time of lubricating oil on the edge, shunt ratio and cold end absorption capacity;

[0013] According to the local thermal capacity and the oil quantity, a node function for representing heat accumulation is calculated, wherein the node function is calculated according to a heat conservation constraint, in combination with temperatures of adjacent nodes and a node temperature;

[0014] The oil temperature data is propagated along the edges by the edge function, and incident contributions are accumulated at the nodes according to the node function to obtain temperature estimates of the nodes;

[0015] According to the temperature estimates, an oil temperature thermal diagram is constructed.

[0016] According to the oil temperature thermal diagram, a node thermal potential is calculated and a plurality of nodes to be processed are located, including:

[0017] The oil temperature thermal diagram is subjected to Gaussian filtering denoising processing, and a two-dimensional temperature gradient vector field is constructed, wherein the two-dimensional temperature gradient vector field represents the node thermal potential;

[0018] According to the two-dimensional temperature gradient vector field, temperature gradient values of each node in X-axis and Y-axis directions of a rectangular coordinate system are calculated to generate a temperature gradient vector;

[0019] If a modulus of the temperature gradient vector is greater than or equal to a preset temperature gradient threshold value, a corresponding node is marked as a candidate node;

[0020] An iteration operation is performed on the candidate nodes to obtain a temperature associated region of the candidate nodes, and a node to be processed is determined according to a number of nodes in the temperature associated region.

[0021] An iteration operation is performed on the candidate nodes to obtain a temperature associated region of the candidate nodes, including:

[0022] create a separate set for each candidate node, and calculate an average principal gradient direction of the set, wherein the average principal gradient direction is determined based on a local density parameter and a temperature gradient value;

[0023] perform an iteration operation to obtain a current principal gradient direction of a neighboring node of any node in the set, and if an included angle between the current principal gradient direction and the average principal gradient direction is less than a preset same direction threshold, add the corresponding neighboring node to the set, until the included angle between the current principal gradient direction of the neighboring node and the average principal gradient direction is greater than or equal to the same direction threshold, and output the set as a temperature correlation region.

[0024] generate a corresponding number of basic loops according to the to-be-processed node, calculate an entropy production sensitivity of the basic loop, and determine a sequence of the to-be-processed node according to the entropy production sensitivity, including:

[0025] extract a basic loop set passing through the corresponding temperature correlation region and connected to a cold end node according to the connectivity of the graph model, wherein the cold end node represents a boundary node for heat exchange with the outside world;

[0026] for each basic loop, calculate a loop entropy generation rate in the basic loop through a discrete thermodynamic relationship, and calculate the entropy production sensitivity according to the loop entropy generation rate, wherein the loop entropy generation rate is calculated by a heat transfer flux and a temperature potential gradient;

[0027] weight and aggregate the entropy production sensitivities of multiple basic loops passing through the same to-be-processed node to obtain an entropy production sensitivity indicator of each to-be-processed node, wherein the weight is calculated according to a residence time and path accessibility of the basic loop passing through the corresponding to-be-processed node;

[0028] determine the sequence of the to-be-processed node according to a descending order of the entropy production sensitivity indicator.

[0029] a row of the current oil temperature regulation matrix corresponds to a cold end node, a column of the current oil temperature regulation matrix corresponds to a to-be-processed node, and a matrix element of the current oil temperature regulation matrix is calculated according to a node thermal potential.

[0030] preset a current oil temperature regulation matrix according to the node thermal potential, including:

[0031] for each cold end node, obtain a cold end capability weight and a time delay registration parameter according to the corresponding instantaneous capacity upper limit and the response time delay;

[0032] According to the corresponding node thermal potential, temperature gradient strength and the number of nodes in the temperature correlation region, a node cooling demand weight and a confidence weight are obtained for each to-be-processed node;

[0033] According to the basic loop of the to-be-processed node, the reachability weight between each cold end node and each to-be-processed node is calculated in combination with the time delay registration parameter;

[0034] The matrix elements are calculated according to the cold end capacity weight, the node cooling demand weight, the confidence weight and the reachability weight, and the current oil temperature regulation matrix is obtained.

[0035] The temperature response kernel of the to-be-updated node is calculated in combination with the basic loop, including:

[0036] According to the current oil temperature regulation matrix and the to-be-updated node, a node oil temperature sub-matrix is determined;

[0037] According to each cold end node in the node oil temperature sub-matrix, the residual transmission coefficient from the cold end node to the to-be-updated node is obtained by multiplying and accumulating the edge function in the basic loop corresponding to the to-be-updated node;

[0038] According to the node thermal potential, the admission weight of the to-be-updated node is calculated, wherein the admission weight is used to represent the cooling conversion capability of the to-be-updated node;

[0039] The effective share vector of the current column of the node oil temperature sub-matrix is calculated according to the matrix elements of the current column of the node oil temperature sub-matrix, and the temperature response kernel is calculated in combination with the residual transmission coefficient and the admission weight.

[0040] The current oil temperature regulation matrix is updated and replaced by the temperature response kernel, including:

[0041] The matrix elements of the current column of the node oil temperature sub-matrix are updated by weighting according to the temperature response kernel;

[0042] The matrix elements of the current column of the node oil temperature sub-matrix after the weighted update are replaced by the matrix elements in the corresponding column of the current oil temperature regulation matrix.

[0043] An oil temperature regulation system for gear box lubricating oil, the system comprising:

[0044] A data acquisition module is used to acquire oil temperature data of the gear box lubricating oil, and process in combination with the graph model corresponding to the gear box structure;

[0045] A thermodynamic diagram construction module is used to construct an oil temperature thermodynamic diagram according to the thermal coupling relationship of the graph model, and estimate the temperature distribution of each node;

[0046] A node identification module is configured to calculate node thermal potential according to the oil temperature thermodynamic diagram, locate a plurality of to-be-processed nodes, and determine a sequence of the to-be-processed nodes by combining the entropy production sensitivity calculated based on the basic loop.

[0047] A matrix generation module is configured to preset a current oil temperature regulation matrix according to the node thermal potential, the cold end node capacity and the reachability relationship.

[0048] A matrix updating module is configured to calculate a temperature response kernel of a to-be-updated node by combining the basic loop in an iteration process, update the oil temperature regulation matrix until the sequence is empty, and output an updated oil temperature regulation matrix.

[0049] An oil temperature regulation module is configured to regulate the oil temperature according to the output of the matrix updating module.

[0050] Compared with the prior art, the application has the following beneficial effects:

[0051] The application can accurately reveal the local temperature rise distribution characteristics by mapping the oil temperature data into an oil temperature thermodynamic diagram corresponding to the gear box structure, realize the priority ordering of the to-be-processed nodes by combining the calculation of the basic loop and the entropy production sensitivity, ensure that the limited cooling resource acts on the hotspot area that needs to be processed most, and dynamically match the cold end contribution and the node receiving capacity by introducing the temperature response kernel and the oil temperature regulation matrix in the iteration and updating, thereby avoiding global overcooling and improving the local regulation precision. The method of the application not only reduces the energy consumption, improves the lubricating oil viscosity stability, but also prolongs the service life of the gear and bearing components, and improves the reliability and efficiency of the overall operation of the gear box. BRIEF DESCRIPTION OF DRAWINGS

[0052] Other features, objects and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0053] Figure 1 FIG. 1 is a schematic diagram of a gear box oil temperature regulation problem principle according to an embodiment of the application;

[0054] Figure 2 FIG. 2 is a schematic diagram of an exemplary application scenario according to an embodiment of the application;

[0055] Figure 3 FIG. 3 is a flowchart of an oil temperature regulation method for gear box lubricating oil according to an embodiment of the application;

[0056] Figure 4 FIG. 4 is a schematic diagram of candidate node screening according to an embodiment of the application;

[0057] Figure 5 FIG. 5 is a schematic diagram of temperature correlation area according to an embodiment of the application;

[0058] Figure 6An oil temperature regulation matrix schematic diagram for an embodiment of the present application;

[0059] Figure 7 A principle schematic diagram for updating the oil temperature regulation matrix by a node oil temperature sub-matrix of an embodiment of the present application. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application.

[0061] In this document, the term “embodiment” means that a specific feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily mutually exclusive of other embodiments. It will be explicitly and implicitly appreciated by those skilled in the art that the embodiments described herein can be combined with each other.

[0062] In the actual operation of the sealed gear transmission device, the lubricating oil not only bears the function of reducing friction and forming a film, but also bears the function of carrying heat and exchanging heat. Engineering experience shows that when the overall oil tank temperature is in a conventional range, local areas such as meshing pairs and bearing seats can still form short-time heat peaks due to load pulsation, oil flow loop time delay, and uneven heat diffusion; if only the overall oil temperature threshold is relied on to trigger cooling, response lag and excessive cooling of the whole system often occur, resulting in rising energy consumption and viscosity penalty.

[0063] On the other hand, the available sensor configuration in the industrial field is usually sparse, and it is difficult to directly obtain the temperature field details consistent with the structure topography; the effect of the cooling execution side on the target area also has propagation attenuation and arrival time difference. The present application proposes a way to abstract the internal heat flow of the gear box as a graph model and carry out regulation and control on it, realizing on-demand suppression of local heat peaks without adding hardware.

[0064] It can be understood that the present application does not rely on fixed partitioning of the specific box or preset cavity division of the oil circuit, but maps the box, meshing pairs, bearing cavities, oil return channels, and cooling boundaries and other elements into a topology composed of nodes and directed edges; on the topology, the attenuation function of the edge and the aggregation function of the node are defined on the basis of heat carrying and heat conduction behavior, so that a limited number of temperature and working condition signals can be projected as temperature distribution estimates consistent with the structure, i.e., oil temperature thermographs. On this basis, by calculating the node heat potential reflecting the strength of local overheating and the risk of overflow, and determining the temperature correlation area by neighborhood expansion combined with gradient consistency, the target set that needs to be intervened preferentially is identified as a constraint of structure and transmission characteristics.

[0065] To avoid rough sorting by temperature level, the application further calculates the entropy generation rate and its sensitivity to node disturbance and control quantity change based on the minimum cycle covering the aforementioned temperature correlation region and communicating with the cold end boundary, converges the sensitivities of multiple cycles in the node dimension, and obtains a sorting basis that can balance the cooling benefit and path cost. The processing order determined accordingly is not limited to local temperature peaks, but reflects the comprehensive trade-off of heat reachable path, arrival delay, and energy consumption budget.

[0066] In the core concept, to accurately deliver cooling capacity to the target region, the application constructs a current oil temperature regulation matrix indexed by the cold end node and the node to be processed. The rows correspond to boundary units that can exchange heat with the outside world (such as oil cooler branches, oil tank heat sinks, etc.), and the columns correspond to the nodes to be processed identified by the foregoing. The matrix elements are comprehensively calculated based on node thermal potential, cold end capacity, path reachability, and time delay matching parameters, and are used as an initial allocation framework for control resources in space and time. The regulation process adopts a rolling iteration method: along the basic cycle, the propagation from the cold end to the target node is time-sequenced and attenuated, forming a temperature response kernel that characterizes the effectiveness of unit control quantity arrival; the response kernel is used to perform constrained weighted update and replacement on the corresponding columns in the matrix, so that each iteration produces a current solution that can be issued, and gradually converges to an allocation result that meets the temperature constraints within the energy consumption and execution boundary. Since the update of the graph model and the modification of the matrix are synchronized, the application can maintain proactive suppression of local thermal peaks under load steps and environmental changes, while avoiding overcooling of the entire system.

[0067] It should be noted that the method of the application is applicable to gear transmission equipment with a closed cavity, complex oil return and external heat exchange path, sparse sensing, and propagation delay on the execution side, including wind power gearboxes, vehicle electric drive gearboxes, heavy industry reducers, etc. In different devices, only the device structure and calibration data are needed to update the graph model parameters and cold end capacity, and the thermal diagram construction, node thermal potential evaluation, basic cycle and entropy production sensitivity analysis, and matrix iteration strategy based on the temperature response kernel of the application can be reused. Through the integrated design of modeling and control, the precise constraint of local oil temperature peaks and the optimization of global energy efficiency can be achieved under the premise of not changing the hardware configuration, guided by structural constraints and targeting energy minimization.

[0068] Reference Figure 1 , Figure 1 The schematic diagram of the gear box oil temperature regulation problem provided by the embodiment of the application.

[0069] Figure 1The gearbox can only collect overall oil temperature data during operation. When the overall oil temperature is within the normal range, the local area may still form a hot spot due to the concentration of friction power. The prior art usually uses overall oil temperature data as the basis for regulation and control. When the overall oil temperature rises, the cooling device is driven, and there is often a problem of response lag. When a local thermal peak is formed, the cooling strategy can only adopt global overcooling because the overall oil temperature cannot accurately reflect the local thermal peak, resulting in increased energy consumption and abnormal lubricating oil viscosity, making it difficult to achieve precise regulation and control of the local hot spot.

[0070] It can be understood that, Figure 1 The gearbox oil temperature regulation problem shown is not due to insufficient cooling device capacity, but due to the limitations of the single and timeliness of the regulation basis. When only relying on overall oil temperature as the input signal, the transient heat aggregation in the local area cannot be captured in real time, resulting in a lag or even failure of the cooling action. Further, to avoid the risk of hot spots, the system often has to adopt a conservative strategy of global overcooling, but this brings additional energy consumption and lubricating oil viscosity deviation, causing the gear meshing area and bearing area to operate in a non-optimal viscosity state, affecting transmission efficiency and service life.

[0071] Reference Figure 2 , Figure 2 An example application scenario graph is provided for the embodiments of the present application.

[0072] Figure 2 An application scenario is shown, which includes an oil pool 100, an oil pump 101, a cooler 102, a gearbox body 103, an oil temperature sensor 104, and a processor 105, wherein:

[0073] The oil pool 100 is used to store lubricating oil and serves as the starting and return end of the system circulation;

[0074] The oil pump 101 is used to extract the lubricating oil in the oil pool 100 and provide circulation pressure;

[0075] The cooler 102 is arranged in the oil circuit and is used to perform heat exchange treatment on the circulating lubricating oil, so that the oil temperature is maintained within a target range;

[0076] The gearbox body 103 is the action area of the lubricating oil and contains components such as gear pairs and bearings inside, and is the main object of heat generation and oil temperature regulation and control;

[0077] The oil temperature sensor 104 is used to collect oil temperature data at different positions of the gearbox body 103 to represent the overall temperature level and local temperature rise;

[0078] The processor 105 is electrically connected with the oil temperature sensor 104, used for processing the collected oil temperature data, and performing the steps of heat map construction, node heat potential calculation, basic loop identification, and oil temperature regulation matrix iterative updating described in the present application, so as to generate the regulation instructions for the oil pump 101 and the cooler 102.

[0079] It should be noted that the gear box body 103 is usually provided with a plurality of oil passage channels inside, which are used for guiding the lubricating oil to different gear meshing areas and bearing areas. The figure is only used for exemplary illustration, and does not constitute a limitation on the actual number and layout of the oil passages. In addition, the cooler 102 can also be provided with multiple coolers in parallel or series to form a cooling unit in actual application, so as to meet the heat exchange requirements under different working conditions. In the abstract modeling of the present application, the cooler 102 can be understood as a cold end node, which forms a heat transfer transmission path with the oil passages and hot spot nodes inside the gear box through edges and loops, so as to realize accurate cooling of local hot spots and energy consumption optimization and regulation of global oil temperature.

[0080] Next, a method for regulating the oil temperature of the lubricating oil of the gear box provided by the present application is further described in combination with the drawings, Figure 3 The method shown is applied as follows S1-S5:

[0081] S1: Obtain the oil temperature data of the lubricating oil, and construct an oil temperature heat map in combination with the graph model corresponding to the structure of the gear box;

[0082] In the present embodiment, the oil temperature data can be collected by the oil temperature sensor arranged near the gear box oil pool, the cooling circuit or the bearing seat, or can be calculated according to the gear box shell temperature, lubricating oil circulation time and load running data. By mapping the collected oil temperature data to the graph model nodes based on the gear box cavity, bearing cavity and meshing pair, the oil flow direction and attenuation law are described by the edge function, and the node function is used to aggregate the contributions of multiple paths, so as to obtain the heat map reflecting the local temperature distribution inside the gear box. The oil temperature heat map constructed in this way can compensate for the observation blind area and obtain the estimation of the local temperature rise under the condition of only relying on a small number of sensors, so as to solve the problem that the traditional overall oil temperature signal cannot describe the hot spot.

[0083] S2: Calculate the node heat potential and locate a plurality of to-be-processed nodes according to the oil temperature heat map;

[0084] In the present embodiment, the node heat potential is obtained by weighting the temperature gradient of the node and the adjacent node and the centrality index of the node in the heat transfer topology, and is used to represent the overheating risk and diffusion risk of the node under the current state. By filtering and gradient analysis on the oil temperature heat map, candidate nodes that form continuous temperature rise in local areas can be identified, and the temperature correlation area can be obtained by iterative expansion, so as to determine the to-be-processed nodes.

[0085] S3: generating a corresponding number of basic loops according to the to-be-processed node, calculating entropy production sensitivity of the basic loop, and determining a sequence order of the to-be-processed node according to the entropy production sensitivity;

[0086] In the embodiment, the basic loop refers to the minimum cycle covering the to-be-processed node and communicating with the cold end node, and the entropy generation rate in the basic loop can be calculated by accumulating the product of the heat flux and the temperature gradient on the edge. By calculating the entropy production sensitivity, the influence degree of the temperature disturbance of the loop on the overall entropy production can be quantified, and the regulation priority of different to-be-processed nodes is evaluated accordingly. By weighting and aggregating the entropy production sensitivities of multiple loops in the node dimension, the entropy production sensitivity index of each to-be-processed node is obtained, and the cold end reachability and the time delay are combined for correction, and finally the sequence order is formed.

[0087] S4: presetting a current oil temperature regulation matrix according to the node thermal potential, performing an iteration operation, selecting a to-be-updated node according to the sequence order, calculating a temperature response kernel of the to-be-updated node in combination with the basic loop, updating and replacing the current oil temperature regulation matrix according to the temperature response kernel, until the sequence order is empty, and outputting the current oil temperature regulation matrix;

[0088] In the embodiment, the rows of the current oil temperature regulation matrix correspond to the cold end nodes, the columns correspond to the to-be-processed nodes, and the matrix elements are calculated by combining the cold end capacity, the node thermal potential, the reachability and the time delay, and represent the allocation amount of the cooling resource to the target node. In the iteration process, the to-be-updated node is selected one by one according to the sequence order, the temperature response kernel is calculated based on the corresponding loop to characterize the residual effective share of the cooling resource arriving at the node after transmission along the loop, and the acceptance weight is combined for weighting to update the corresponding column element of the matrix. Each iteration obtains a new oil temperature regulation matrix and outputs it for real-time issuance of control instructions. Through this step-by-step updating method, the columns of the matrix can be corrected one by one with low computational complexity without relying on global one-time solution, ensuring the real-time and stability of regulation.

[0089] S5: performing oil temperature regulation according to the current oil temperature regulation matrix;

[0090] In the embodiment, the current oil temperature regulation matrix finally reflects the action allocation of each cold end node to different to-be-processed nodes in the current time window, which can be mapped to the opening of the cooler, the flow regulation of the oil pump or the shunt ratio of the bypass valve. After execution, the limited cooling resource can be more accurately delivered to the local area with the risk of overheating, avoiding global overcooling, reducing energy consumption and keeping the viscosity of the lubricating oil in a reasonable range.

[0091] Before expanding the specific technical content corresponding to the steps, the embodiments of the present application need to be emphasized again:

[0092] In industrial scenarios, the oil temperature characteristics of gearboxes differ significantly from those of conventional thermal management objects. The interior of a typical gearbox is a cavity-enclosed structure, and the gear pairs and bearings continuously bear high-frequency friction and load impact. The heat generated is not uniformly distributed, but is concentrated in the meshing points and rolling contact areas. Because the flow path of the lubricating oil in the gearbox is limited by the oil passage layout and the cavity geometry, the overall oil pool temperature is often within the normal range, while the local hot spots may instantaneously exceed the oil film stability interval. If only the overall oil temperature is used as the control signal, the response will be delayed, and fatigue damage will occur before the local overheating is detected. Conversely, to avoid risks, global overcooling will lead to increased energy consumption and abnormal oil viscosity, affecting transmission efficiency and service life.

[0093] The processing logic used in this embodiment does not involve adding more sensing points or hardware modifications to the oil circuit. Instead, it involves structuring modeling of existing oil temperature data to calculate the behavior trend of local heat transfer and aggregation.

[0094] Specifically, the gearbox is abstracted as an oil temperature thermodynamic graph composed of nodes and edges, the edge function is used to describe the attenuation transmission of the lubricating oil in the channel, and the node function is used to reflect the local heat capacity and aggregation effect, so as to reconstruct the local temperature rise distribution under limited data. Unlike the traditional single threshold judgment based on the average oil temperature, this application further introduces the entropy production sensitivity as an evaluation index to analyze and quantify the diffusion and dissipation trend of thermal disturbance in different channels. When it is found that the thermal potential of a certain regional node is not only higher than that of the neighborhood, but also shows a stronger coupling amplification effect in the entropy production sensitivity of the loop, it can be logically deduced that the node is a hot spot that needs to be processed first.

[0095] Further, to avoid resource waste caused by global one-time cooling, this embodiment constructs an oil temperature control matrix indexed by cold nodes and to-be-processed nodes to allocate limited cooling capacity in space and time. The update of the matrix is not based on empirical rules, but on the temperature response to iteratively correct the transmission residual and reception efficiency of each cooling action, so as to output an executable allocation solution at each moment. Compared with the traditional scheme, this control logic based on the oil temperature thermodynamic graph and the temperature response core can effectively eliminate the misjudgment and lag caused by the overall temperature averaging without the need to increase hardware, and realize precise constraint of local hot spots and simultaneous optimization of global energy efficiency.

[0096] Next, the technical content of the method of the application for constructing the oil temperature thermodynamic graph is further expanded.

[0097] In one example, the oil temperature thermodynamic graph is constructed in combination with the graph model corresponding to the gearbox structure, including:

[0098] S1.1: According to the thermal coupling relationship of the structure in the graph model, the gearbox is divided into nodes and edges, where the nodes represent local thermal capacity and oil volume, and the edges represent thermal transfer connectivity and oil flow direction attributes;

[0099] Specifically, in the actual gearbox interior, the generation and transmission of heat are not uniformly distributed, but are coupled by the meshing pair, bearing arrangement, and oil passage geometry. Therefore, before constructing the thermal diagram, the gearbox structure needs to be abstracted into a topology composed of nodes and edges. The nodes are used to represent the oil volume and thermal capacity characteristics of the local area, such as gear meshing points, bearing seats, oil pool areas, and oil return slots, and the edges are used to represent the lubricating oil flow path and the connectivity of heat transfer.

[0100] In this embodiment, the division of nodes can be determined based on the structural drawing of the gearbox, the design of the fluid passage, and the distribution area of the oil during operation. The parameters carried by each node include the oil volume, specific heat capacity, and contact area with solid components of the region, ensuring that they can reflect the thermal inertia characteristics of the region. The establishment of edges depends on the geometric connectivity relationship between adjacent nodes, and is accompanied by oil flow direction attributes.

[0101] For example, the gear meshing pair and the oil pool can establish an edge through the splashing path, and the oil pump outlet and the bearing inlet can establish an edge through the forced oil supply pipeline, which can truly reflect the movement route of the lubricating oil in the gearbox and provide a topology basis for subsequent temperature propagation and distribution estimation.

[0102] Further, the definition of nodes and edges does not require a one-to-one correspondence with the actual physical parts, but is abstracted according to the laws of heat coupling and fluid flow. For example, multiple adjacent oil passages can be combined into a node or an edge to reduce computational complexity. Even in the case of limited number of sensors, limited data can be mapped to the abstract graph structure, thereby achieving a description of the internal thermal distribution of the entire gearbox.

[0103] S1.2: According to the thermal transfer connectivity and oil flow direction attributes, calculate an edge function for representing heat contribution decay, wherein the edge function is calculated according to path length, lubricating oil residence time on the edge, shunt ratio, and cold end absorption capacity;

[0104] Specifically, when heat is transmitted through oil, there will inevitably be a decay phenomenon along the path.

[0105] For example, if the oil stays in the pipeline for too long, it will exchange heat with the surrounding metal wall, causing the temperature contribution to gradually weaken. For another example, when a flow of oil is shunted to multiple branches, the temperature transfer capacity obtained by each branch will also decrease accordingly.

[0106] In this embodiment, the definition of edge function takes into account the path length, the residence time of oil in the edge, the split ratio and the cold end absorption capacity. The longer the path, the greater the temperature attenuation; the longer the residence time, the more significant the heat loss; the lower the split ratio, the weaker the heat contribution; and when the path end is connected to the cold end node, the attenuation effect will be further amplified.

[0107] Further, the edge function is used to give how much effective contribution a unit temperature (or heat) can leave after propagating along the edge, and the input information includes:

[0108] The effective hydraulic length of the edge, the geometric bending and cross-section change, the transmission oil volume between nodes, the split ratio of the edge, whether the end is connected to the cold end, and the available heat exchange capacity of the cold end at the moment, including wall material, surface roughness and flow working condition as correction terms.

[0109] In some optional specific embodiments, the calculation of the edge function also includes:

[0110] One is the initial calibration of the structure, which records the time and amplitude of the upstream and downstream temperature rise and fall on the bench with fixed pump speed steps and cold end switching steps, and inversely calculates the initial value table of length and time factor.

[0111] The second is the online correction, which makes a small disturbance to the pump speed and cold end opening under the premise of not affecting safety, uses the ratio of node temperature change before and after the disturbance and the time difference to correct the table entries of the split factor and the cold end absorption factor, and limits the correction amount within a certain range to avoid drift caused by short-term noise.

[0112] Thirdly, different default function templates are used for different categories of edges: forced oil supply edges can use high flow and weak energy storage profile templates; splash edges use medium delay and strong diffusion templates; and return oil edges use gravity dominated and easy to stack templates.

[0113] All edge functions are subject to three operating constraints: cold end instantaneous capacity upper limit, actuator switching frequency limit and energy consumption budget upper limit. Any one reaching the boundary will trigger temporary depression of the attenuation coefficient of the edge and widening of the profile. The edge function calculated in this way can not only distinguish the real contribution of different paths to temperature propagation, but also adjust online with the working condition, and finally provide reliable edge propagation weight for the update of the heat map in space and time.

[0114] S1.3: calculating a node function for representing heat accumulation according to the local heat capacity and the oil volume, wherein the node function is calculated according to a heat conservation constraint, in combination with the temperatures of adjacent nodes and the node temperature;

[0115] Specifically, the lubricating oil does not flow along a single path when inside the gearbox, but tends to accumulate in certain areas, with the contributions of multiple paths superimposed together. This accumulation effect has a decisive influence on the local temperature. For example, in a bearing cavity, it may simultaneously receive multiple contributions from gear splashing, oil pump injection and oil return channels, which requires the introduction of a node function in the graph model to describe the accumulation law of these contributions.

[0116] In the present embodiment, the node function is constrained by heat conservation, ensuring that the temperature contributions of the incoming nodes are balanced with the temperature state of the node itself. In specific calculation, the temperature increments transmitted by the adjacent nodes through the edge function are weighted and superimposed, and then combined with the temperature and heat capacity characteristics of the node itself to obtain the updated temperature estimate, which can dynamically balance the temperature contributions from different sources and avoid excessive amplification of a certain source, thereby ensuring the physical rationality of the temperature estimate.

[0117] S1.4: propagate the oil temperature data along the edges through the edge functions, and accumulate the incident contributions at the nodes according to the node functions to obtain temperature estimates of each node;

[0118] Specifically, after the edge function and node function are defined, the collected oil temperature data can be used as initial input, the temperature information is propagated along each edge through the edge function, and then the incident contributions are accumulated at the nodes using the node function to calculate the temperature estimate value of each node. The temperature distribution formed in this way is the oil temperature thermal map, which can intuitively reflect the temperature state of each region of the gearbox.

[0119] In the present embodiment, the oil temperature data can come from different sensor distribution points, such as monitoring points at the inlet of the oil pool, the outlet of the cooler or the vicinity of the bearing. By mapping these observation point data to the corresponding nodes as boundary conditions, the calculation of the thermal map can be carried out on the entire graph model. The final thermal map not only contains the real data of the measurement points, but also includes the temperature estimates of the unmeasured regions calculated, thereby making up for the information loss caused by the sparse distribution of sensors.

[0120] S1.5: constructing an oil temperature thermal map according to the temperature estimates;

[0121] Next, the technical content of the method of the present application for positioning a plurality of nodes to be processed is further expanded.

[0122] In one example, calculating a node thermal potential and positioning a plurality of nodes to be processed according to the oil temperature thermal map, comprising:

[0123] S2.1: performing Gaussian filtering denoising processing on the oil temperature thermal map, and constructing a two-dimensional temperature gradient vector field, wherein the two-dimensional temperature gradient vector field represents the node thermal potential;

[0124] Specifically, the oil temperature heat map often contains a large number of high-frequency components from sensor noise, local disturbances and calculation errors after being propagated through the node and edge functions. If these data are directly used to determine the local temperature rise, false hot spots may be misjudged. Therefore, before constructing the node heat potential, the oil temperature heat map needs to be smoothed and denoised.

[0125] In this embodiment, the Gaussian filter is used to perform weighted average on the temperature values of the local neighborhood through the convolution kernel, which can suppress isolated high-frequency noise points while maintaining the trend of the overall temperature distribution. The parameter setting of the Gaussian filter is not fixed, but is dynamically adjusted according to the operating conditions of the gearbox: when the oil pump flow and speed are low, the oil flow disturbance is small, and the filter radius can be set small to retain details; while in high-speed impact conditions, the temperature distribution fluctuates violently, and the filter radius needs to be increased to obtain a smooth trend field. After filtering, the temperature distribution is converted into a two-dimensional temperature gradient vector field, in which the heat potential of each node not only reflects its own temperature level, but also combines the gradient directionality with adjacent nodes, thereby depicting the temperature change trend rather than a single point value.

[0126] S2.2: Calculate the temperature gradient values of each node in the X-axis direction and the Y-axis direction of the rectangular coordinate system according to the two-dimensional temperature gradient vector field, and generate a temperature gradient vector;

[0127] Specifically, after obtaining the two-dimensional temperature gradient vector field, the temperature gradient values of each node in the two directions of the rectangular coordinate system need to be further calculated to generate a specific temperature gradient vector.

[0128] It can be understood that a local hot spot is not only represented by a high temperature value, but more importantly, its temperature distribution has a mutation or a strong gradient relative to the surrounding nodes. If only the absolute temperature value is used, potential dangerous areas with slightly high temperature but steep change trend may be missed.

[0129] In this embodiment, the gradient calculation is performed by directional differentiation of the temperature difference between the node and the adjacent nodes. The gradient in the X-axis direction is obtained by the temperature difference between the node and the left and right neighborhoods, and the gradient in the Y-axis direction is obtained by the temperature difference between the node and the upper and lower neighborhoods. In the graph model, if there is an irregular geometric structure, the adjacency matrix is used to expand the directionality to ensure that the gradient vector reflects the true spatial relationship. The obtained temperature gradient vector contains both size information and direction information, so that the development trend of the hot spot can be judged to be diffusion to a certain area or limited to a local point.

[0130] S2.3: If the modulus of the temperature gradient vector is greater than or equal to a preset temperature gradient threshold, the corresponding node is marked as a candidate node;

[0131] Specifically, after calculating the temperature gradient vector of each node, a standard is needed to distinguish abnormal gradient from normal temperature fluctuation. The embodiment introduces a preset temperature gradient threshold. If the gradient vector size of a node is greater than or equal to the threshold, it is considered that the node has an abnormal temperature rise trend, and it is marked as a candidate node. The threshold is not fixed, but can be determined by experiment calibration and online adaptive adjustment. For example, under normal temperature light load working condition, the threshold can be set lower to capture subtle abnormal trends in time; while under high temperature heavy load working condition, the overall temperature gradient of the gearbox is larger, and the threshold needs to be increased to avoid misjudgment in a large area. Through this method, the sensitivity of hot spot identification can be combined with the characteristics of working conditions to improve the robustness of detection.

[0132] S2.4: performing an iteration operation on the candidate nodes to obtain a temperature associated region of the candidate nodes, and determining a node to be processed according to the number of nodes in the temperature associated region;

[0133] Specifically, the candidate node is only a potential hot spot, and it is still necessary to determine whether it is a node that needs to be processed through region expansion. The embodiment uses an iterative operation method to expand the candidate node to its adjacent nodes to form a temperature associated region. In the expansion process, not all neighbors are simply added, but the average main gradient direction is combined for judgment: if the gradient direction of a neighboring node is less than the same direction threshold with the average main gradient direction of the set, it means that the node and the candidate node are in the same heat diffusion trend, and the node should be added to the set; otherwise, the expansion is stopped. In this way, an associated region that conforms to the physical heat flow direction can be formed in the graph model, rather than an arbitrary diffusion clustering result.

[0134] In the embodiment, the size of the temperature associated region directly affects whether the candidate node is confirmed as a node to be processed. When the number of nodes in the associated region exceeds the preset scale, it means that there is a large range of sustained heat aggregation in the local region, and the candidate node is confirmed as a node that needs to be processed. The advantage of this processing is that it can identify both single-point intense hot spots and regional hot spots composed of multiple nodes, avoiding misjudgment caused by relying solely on temperature values or gradient values.

[0135] Reference Figure 4 , Figure 4 The candidate node screening schematic diagram provided by the embodiment of the present application.

[0136] As Figure 4 shown, in the two-dimensional temperature gradient vector field calculated based on the oil temperature thermodynamic diagram, some nodes are preliminarily marked as candidate nodes because the temperature gradient exceeds the preset threshold, such as candidate node A, candidate node B and candidate node C. These candidate nodes are distributed in different local regions, indicating that there is an abnormal temperature change trend near them.

[0137] Reference Figure 5 , Figure 5 The temperature correlation region schematic diagram provided by the embodiment of the application.

[0138] Figure 5 It is shown that on the basis of candidate node screening, temperature correlation region A, temperature correlation region B and temperature correlation region C are obtained by iterative expansion with candidate node A, candidate node B and candidate node C as the starting points and in combination with the temperature gradient directions of the nodes in the neighborhood of the candidate nodes. It can be seen that the temperature correlation region not only contains the candidate node itself, but also covers the adjacent node set maintaining the same trend of temperature rise. The formation of these regions reflects the local heat aggregation phenomenon existing in the gearbox. Through the size and distribution characteristics of the region scale, the severity of the hot spot and whether it needs to be preferentially regulated can be judged. When the number of nodes in the temperature correlation region exceeds the set threshold value, the corresponding candidate node can be confirmed as a to-be-processed node, so as to ensure that the subsequent cooling resource allocation focuses on the real hot spot region and avoids misjudgment caused by single-point noise or local abnormality.

[0139] It can be understood that when the number set threshold value is 2, temperature correlation region A and temperature correlation region B meet the set threshold value, that is, candidate node A and candidate node B can be used as to-be-processed nodes.

[0140] In one example, an iterative operation is performed on the candidate node to obtain the temperature correlation region of the candidate node, comprising:

[0141] S2.4.1: An independent set is created for each candidate node, and the average main gradient direction of the set is calculated, wherein the average main gradient direction is determined based on a local density parameter and a temperature gradient value, and the local density parameter is obtained by sum average calculation according to the modulus of one or more temperature gradient vectors in the set;

[0142] S2.4.2: An iterative operation is performed to obtain the current main gradient direction of the adjacent node of any node in the set, and if the included angle between the current main gradient direction and the average main gradient direction is less than a preset same direction threshold value, the corresponding adjacent node is added to the set, until the included angle between the current main gradient direction of the adjacent node and the average main gradient direction is greater than or equal to the same direction threshold value, and the set is output as a temperature correlation region, wherein the same direction threshold value can be determined by experiment, and in principle, it is within the range of acute angle;

[0143] Next, the technical content of the method of the application about determining the order sequence of the to-be-processed node is further expanded.

[0144] It can be understood that the order sequence does not mean that only one cold end can only serially cool the nodes, but is used to determine the calculation and writing order of each column (corresponding to the to-be-processed node) of the oil temperature regulation matrix. In any iteration period, the cold end node participates in parallel as the row dimension, and multiple cold ends can simultaneously act on the same column or different columns. The significance of the order sequence is that:

[0145] When updating at the column level, which node is processed first and which node is processed last, so that each column update is based on the latest matrix of the consumed row capacity, the formed arrival delay distribution and the generated coupling effect, thereby avoiding row capacity overallocation, response cores canceling each other or causing oscillation on the shared loop, and ensuring the convergence and energy utilization efficiency of iteration.

[0146] In one example, a corresponding number of basic loops is generated according to the to-be-processed node, the entropy production sensitivity of the basic loop is calculated, and the order sequence of the to-be-processed node is determined according to the entropy production sensitivity, including:

[0147] S3.1: According to the connectivity relationship of the graph model, a set of basic loops passing through the corresponding temperature correlation area and connected with the cold end node is extracted, wherein the cold end node represents a boundary node for heat exchange with the outside world;

[0148] Specifically, whether a local hot spot is easily cooled depends on whether there is a reachable and closed oil heat transmission channel between the local hot spot and the cold end. Only by judging according to the geometric shortest path or a single oil supply pipeline, the circulating effect brought by the oil return path and splash return will be ignored, and it is difficult to reflect the true cooling feasibility. Therefore, it is necessary to find a closed channel in the graph model which passes through the temperature correlation area and is connected with the cold end, and abstract it as a basic loop to represent the minimum circulation unit of the cooling capacity entering the area, completing heat exchange and returning to the oil pool or the cold end.

[0149] In this embodiment, the original graph is first directionally cropped: only the edges consistent with the oil flow direction and the effective heat conduction direction are retained, and the edges that are not conductive or counter-flow under the current working condition are temporarily shielded; then the temperature correlation area where the to-be-processed node is located is taken as the core to construct a restricted subgraph, so that the candidate loop necessarily passes through the area.

[0150] Further, a set of minimum loop bases is generated on the restricted subgraph: a directed spanning tree is preferentially constructed, and the feasible edges not on the tree are sequentially back-connected to form a simple closed loop; each time a closed loop is formed, it is checked whether at least one cold end node and a returnable path are included in the connectivity relationship, if yes, it is recorded as a basic loop, otherwise it is discarded. In order to avoid false closed loops that detour far, two criteria are introduced:

[0151] One is the upper limit of the loop length, which comes from the combination of the structure size and the allowed propagation delay;

[0152] The second is a path reachability lower limit, a combination threshold of attenuation coefficients from each edge and shunt ratio.

[0153] The basic loop obtained through the foregoing screening not only guarantees physical reachability, but also guarantees executable significance within a time window.

[0154] S3.2: For each basic loop, a loop entropy generation rate is calculated in the basic loop through a discrete thermodynamic relationship, and the entropy production sensitivity is calculated according to the loop entropy generation rate, wherein the loop entropy generation rate is calculated through a heat flux and a temperature potential gradient;

[0155] Specifically, uneven cooling and heating is embodied in a loop in the joint action of a heat flux and a temperature potential difference. If heat transfer along a certain loop is mainly consumed by invalid diffusion, and effective cooling is not formed in a target area, the loop will have a higher entropy generation level; on the contrary, if the cooling capacity can successfully reach the hot spot and maintain a small potential difference loss in the return flow, the entropy generation level is lower. By measuring invalid dissipation, the influence of different loops on the overall thermal balance can be compared in a global sense.

[0156] In the embodiment, for each basic loop, the edges and nodes of the loop are traversed, the propagation attenuation, residence time and arrival time sequence information provided by the edge function are called, and the aggregation and local thermal inertia information provided by the node function is called, the heat transfer and potential difference dissipation of the loop are accumulated in a fixed prediction time window to obtain the time window integral value of the loop entropy generation rate. To obtain the sensitivity, two types of small perturbations are applied respectively:

[0157] One type is a node-side perturbation, that is, a small change is applied to the temperature estimation of the node to be processed, the entropy generation level of the loop is recalculated, and the difference between the two reflects the influence of the temperature change of the node on the entropy production of the loop;

[0158] The other type is a control-side perturbation, that is, a small duty cycle change is applied to the cold end channel related to the loop, the loop entropy production is recalculated, and the difference between the two reflects the response of the loop to the control allocation change.

[0159] Both types of perturbations use the same amplitude as the calibration, and are consistent with the current working condition in the arrival time sequence, so that the sensitivity value is consistent with the running state rather than a static nominal value.

[0160] S3.3: The entropy production sensitivities of multiple basic loops passing through the same node to be processed are weighted and aggregated to obtain an entropy production sensitivity index of each node to be processed, wherein the weight is calculated according to the residence time and path reachability of the basic loop passing through the corresponding node to be processed;

[0161] Specifically, the same to-be-processed node is often covered by multiple basic loops, and the propagation efficiency, delay characteristics, cold end occupation, and coupling strength with the surrounding nodes of each loop are not the same. If the maximum or average is simply taken, the difference between a small amount of coverage of an efficient loop and a large amount of coverage of an inefficient loop will be masked, and therefore it is necessary to weight and aggregate the entropy production sensitivity of the multi-loop in the node dimension to obtain an index that can represent the global influence of the node.

[0162] In the present embodiment, the weight is composed of two types of quantities:

[0163] One type is the reachability weight, which comes from the comprehensive reachability of the edge function along the way and the overlap degree of arrival within the time window, describing whether the cooling effect can arrive and take effect in the current window;

[0164] The other type is the cold end occupation weight, which comes from the residual equivalent capacity of the associated cold end of the loop, describing the allocable space of the loop in the cold end node resource.

[0165] After normalization, the weight is multiplied by the entropy production sensitivity of the corresponding loop and summed to obtain the node-level entropy production sensitivity index. To avoid repeated measurement of overlapping loops, a loop overlap penalty is introduced: if two loops share a key edge or have a high degree of overlap in arrival timing near the node, the weight of the loop added later is reduced by the overlap ratio, so that the aggregation result is closer to the number and quality of independent available cooling paths.

[0166] S3.4: According to the descending order of the entropy production sensitivity index, determine the order sequence of the to-be-processed node;

[0167] Specifically, the order sequence is used to determine the column-level writing order of the oil temperature regulation matrix in the present rolling iteration, so that each column update is completed under the constraint of the latest row capacity and arrival timing, without limiting the parallel action of multiple cold ends. Arranging only in descending order of node index will ignore the fact that the row capacity has been occupied by the previous column, the shared loop will cause coupling conflicts, and the arrival window will be superimposed on each other, and therefore the combined action of immediate feasibility and coupling suppression needs to be introduced in the sorting process.

[0168] In this embodiment, using the node-level entropy production sensitivity index as the initial priority, and combining the remaining row-side capacity of the current matrix, the cold-end switching frequency limit, and the arrival overlap within the time window, the marginal benefit is calculated for each candidate column: that is, the effective cooling increment that the column can contribute if it is written first, without violating existing constraints. Then, the coupling degree with already included nodes is evaluated, where already included nodes refer to nodes in the sequential sequence. This mainly considers the overlap of shared basic loops, the in-phase overlap of arrival times, and resource contention for the same cold-end row, and applies a reduction to the marginal benefit based on the coupling strength. The corrected marginal benefit is used as the immediate priority to select the column for the current round; immediately after writing, the row capacity, arrival profile, and coupling relationship are updated, and the remaining priorities are recalculated recursively until the column-level writing within the current window is completed, where the arrival profile represents the transmission order between column nodes.

[0169] Next, we will further elaborate on the technical content of the method in this application regarding the construction of the oil temperature control matrix.

[0170] refer to Figure 6 , Figure 6 This is a schematic diagram of the oil temperature control matrix provided in an embodiment of this application.

[0171] Figure 6 The oil temperature control matrix shows that the rows correspond to multiple cold-end nodes C1 to Cn, and the columns correspond to multiple nodes to be processed N1 to Nm. Figure 6 The matrix consists of three elements (for illustrative purposes only), where m and n are both integers greater than 1. Each element represents the cooling contribution allocation of the cold-end node to the node under the current operating conditions. This allocation is calculated based on a combination of factors, including node thermal potential, loop entropy production sensitivity, path reachability, and cold-end capacity. During the iterative update of the matrix, a column corresponding to one node to be updated is selected each time, and its temperature response is used to correct the matrix column elements, thus ensuring the allocation result better reflects the actual heat transfer capacity and requirements. Once all nodes to be processed have completed column-level updates, the output matrix represents the oil temperature control scheme within that time window, which can be directly used to adjust the cooler opening, oil pump flow rate, or bypass valve allocation ratio.

[0172] It is easy to understand that in the oil circuit topology of the gearbox, the connectivity between the cold end nodes such as coolers, oil pump bypass or splash oil supply and different hot spot areas is different, and not all cold ends can act on all nodes to be processed through a feasible path. Therefore, in the oil temperature regulation matrix, for those cold end hot spot combinations that do not have an effective loop or the loop accessibility is lower than a threshold, the corresponding matrix elements will be set to zero, indicating that the cold end has no actual cooling effect at the node. This processing not only conforms to the physical reality, but also avoids false distribution in the calculation process of the regulation matrix, ensuring that the matrix only considers feasible paths when updating and executing mapping. Further, the presence of zero elements makes the regulation matrix present a sparse structure, which not only reduces the calculation complexity in engineering implementation, but also highlights the real allocation boundary of limited cold end resources, thereby improving the effectiveness and executability of the regulation strategy.

[0173] In one example, a current oil temperature regulation matrix is preset according to the node thermal potential, including:

[0174] For each cold end node, a cold end capacity weight and a time delay registration parameter are obtained according to the corresponding instantaneous capacity upper limit and response time delay.

[0175] Specifically, the cooling capacity of each cold end node is not unlimited, but is affected by the instantaneous capacity upper limit and the response time delay.

[0176] In this embodiment, the cold end capacity weight can be determined by two parts: one is the maximum heat exchange capacity of the cooler or the oil pump, and the other is the remaining available capacity obtained by folding according to the dynamic occupation ratio of the current working condition. The response time delay parameter is obtained by measuring the delay time from the input of the control signal to the actual generation of the cooling flow. In the calculation, the remaining capacity is normalized to the interval [0, 1] as the cold end capacity weight, and the time delay is inversely mapped as a registration correction coefficient. The shorter the delay, the closer the registration factor to 1.

[0177] For each node to be processed, a node cooling demand weight and a confidence weight are obtained according to the corresponding node thermal potential, temperature gradient intensity and the number of nodes in the temperature correlation area.

[0178] In this embodiment, the node cooling demand weight is determined based on the node thermal potential. The greater the node thermal potential, the more serious the temperature rise of the node relative to the periphery, and the weight is increased in a proportional relationship. Specifically, the difference between the node temperature and the average temperature of the neighborhood can be used, and normalized processing is performed in combination with the temperature gradient intensity to obtain a demand weight in the interval [0, 1]. The confidence weight is used to correct the reliability of the data. When the historical sampling fluctuation of the oil temperature sensor at the node is small and the consistency with the predicted value is high, the confidence weight takes a high value; when the sampling noise is large or the data is lost, the weight decreases. In this way, it is ensured that abnormal distribution will not occur in matrix updating due to unreliable data.

[0179] According to the basic loop of the to-be-processed node, the reachability weight between each cold end node and each to-be-processed node is calculated in combination with the time delay registration parameter;

[0180] In the embodiment, the reachability weight is used to describe the possibility of the cold end capability reaching the node through the basic loop. In the calculation, the basic loop length, the attenuation coefficient of the edge function along the way, and the shunt ratio are multiplied step by step, and the residual contribution ratio of the cold end action to the node is obtained in combination with the cold end capacity limit and the time delay registration correction, that is, as the reachability weight. If there is a break in the path or the attenuation exceeds the threshold, the reachability weight is directly set to zero.

[0181] According to the cold end capability weight, the node temperature drop demand weight, the confidence weight, and the reachability weight, the matrix element is calculated to obtain the current oil temperature regulation matrix;

[0182] Next, the technical content of the method of the application on the iterative update of the oil temperature regulation matrix is further expanded.

[0183] In one example, the specific steps of S4 are as follows:

[0184] S4.1: According to the current oil temperature regulation matrix and the to-be-updated node, a node oil temperature sub-matrix is determined;

[0185] Specifically, the oil temperature regulation matrix often contains the corresponding relationship of multiple cold end nodes and all to-be-processed nodes, but when updating a node, it is not necessary to process the entire matrix, but to extract the part related to the to-be-updated node separately to form a local sub-matrix. That is, the column update operation of the matrix only depends on the cold end action row directly related to the node, so that separate extraction of the sub-matrix can avoid repeated calculation of the full matrix, improve real-time update efficiency, and at the same time avoid numerical oscillation caused by interference of irrelevant columns.

[0186] In the embodiment, the node oil temperature sub-matrix is constructed in the following manner: taking the column where the to-be-updated node is located as the core, the matrix elements of all cold end rows corresponding to the column are extracted. If a cold end is unreachable from the node or the matrix element is zero, the cold end row is retained but set to zero in the sub-matrix as part of the structural constraint. The sub-matrix obtained in this way is a local matrix with smaller dimension, which clearly indicates the distribution capability of all cold ends in the direction of the node.

[0187] S4.2: According to each cold end node in the node oil temperature sub-matrix, the residual transmission coefficient from the cold end node to the to-be-updated node is obtained by multiplying and accumulating according to the edge function in the basic loop corresponding to the to-be-updated node;

[0188] Specifically, the effect of the cold end node on the node to be updated is not directly linear, but needs to be realized through the transmission path of the oil circuit loop. During transmission, the temperature contribution will gradually decay due to factors such as path length, oil residence time, and shunt ratio, so the edge function on the path needs to be multiplied and added to obtain the effective contribution of the cold end to the node to be updated after passing through the loop, that is, the residual transmission coefficient.

[0189] In the present embodiment, the calculation method of the residual transmission coefficient is: starting from the cold end node, traversing along the basic loop to the node to be updated, calling the edge function to obtain the attenuation factor of the path piece by piece, and continuously multiplying; at the same time, the shunt points in the path are added and corrected to ensure that when a loop is shared by multiple branches, the contribution is distributed in proportion. The final residual transmission coefficient is between 0 and 1, indicating the effective reservation ratio of the initial contribution of the cold end when reaching the node to be updated.

[0190] S4.3: According to the node thermal potential, calculate the admission weight of the node to be updated, wherein the admission weight is used to represent the cooling conversion ability of the node to be updated;

[0191] Specifically, whether the ability of the cold end can be effectively converted into the cooling effect of the node depends on the temperature demand and physical characteristics of the node. Only the cold end contribution is not enough, if the node itself is in low thermal potential or has great thermal inertia, even if the cooling resources are allocated, it is difficult to significantly reduce the temperature, therefore, the admission weight is introduced to depict the cooling conversion ability of the node to be updated. The purpose of this is to allocate resources to the nodes that can really produce cooling effect in a short time in matrix updating, rather than wasting the ability in the area with low cooling efficiency.

[0192] In the present embodiment, the admission weight is a quantitative parameter for measuring the absorption and conversion efficiency of the cold end contribution of the node to be updated, and its value depends on the thermal potential intensity of the node, the heat capacity of the node, and the aggregation degree of the temperature related region.

[0193] Specifically, the node thermal potential directly serves as the dominant factor of the admission weight, the higher the thermal potential value, the more prominent the temperature rise of the node, and the stronger the urgency of cooling, so the admission weight is positively adjusted on the basis of the reference value. This processing can ensure that the hot spot area is given priority in matrix updating, and the cooling resources are more allocated to the nodes that urgently need to be cooled.

[0194] In the embodiment, the thermal capacity of the node is introduced as a correction factor. The node with larger thermal capacity has stronger temperature inertia, that is, even if the cooling capacity is put in, it is not easy to produce obvious cooling effect in a short time, so the admission weight is adjusted downward accordingly, so that the cold end distribution amount is reserved; on the contrary, the node with smaller thermal capacity can quickly respond to the cooling effect, and the admission weight is adjusted upward to highlight the efficiency of short-time regulation. In this way, the cooling resource allocation not only considers the temperature level, but also considers the actual conversion efficiency after regulation.

[0195] Further, the stability of the temperature correlation region is used to avoid the spread of hot spots. When a node is in a large-scale high-temperature region, even if the thermal capacity of the node is large, the admission weight of the node needs to be increased to prevent the overall temperature rise of the region from further spreading. Therefore, when calculating the admission weight, a region size coefficient is introduced, and when the number of high-temperature nodes in the region exceeds a threshold, an additional gain is added to the admission weight of the node. In this way, the risk of ignoring the overall risk of the region due to the low priority of a single point can be prevented, and the hot spot group can be inhibited early.

[0196] In the specific implementation process, the calculation method of the admission weight is: taking the node thermal potential as the reference input, combining the inverse proportional correction coefficient of the node thermal capacity and the proportional correction coefficient of the region stability, to obtain the final comprehensive weight value. In other words, the higher the thermal potential, the smaller the thermal capacity, and the more concentrated the region, the larger the admission weight value; when the thermal potential is low or the thermal capacity is extremely large, even if there is local temperature rise, the admission weight will be reduced, thereby avoiding resource waste. The admission weight obtained in this way can effectively reflect the actual cooling conversion capacity of the node under the current working condition, so that the calculation result of the temperature response kernel is consistent with the physical process.

[0197] S4.4: calculating an effective share vector of the current column of the node oil temperature sub-matrix according to the matrix elements of the current column of the node oil temperature sub-matrix, and calculating a temperature response kernel combining the residual transmission coefficient and the admission weight;

[0198] Specifically, the temperature response kernel is a core parameter for updating the matrix column elements, which comprehensively considers the effective share of the cold end transmission to the node and the admission efficiency of the node to the cold end capacity, and can quantify the real cooling response of the cold end node under the current working condition. The reason for calculating the temperature response kernel is that simple matrix element updating cannot reflect the coupling relationship of path loss and node conversion, and the response kernel can mathematically unify the two to form an update quantity that conforms to the principles of thermodynamics and is convenient for iterative calculation.

[0199] In the embodiment, the calculation process is as follows: first, according to the matrix elements of the current column in the sub-matrix, the effective share vector of the node in the cold end dimension is extracted as the nominal allocation proportion of the cold end to the node; then, the share vector is multiplied by the residual transmission coefficient of the corresponding cold end item by item to obtain the cold end effect corrected by path attenuation; and then the result is combined with the admission weight to complete the correction of the node side conversion efficiency. The final output temperature response kernel is a vector, and each component of the vector represents the actual cooling contribution of the cold end at the node.

[0200] S4.5: Weighted update is performed on the matrix elements of the current column of the node oil temperature sub-matrix according to the temperature response kernel;

[0201] S4.6: The matrix elements of the current column of the node oil temperature sub-matrix after the weighted update are replaced by the matrix elements in the corresponding column of the current oil temperature regulation matrix;

[0202] Reference Figure 7 , Figure 7 The principle diagram of the node oil temperature sub-matrix provided by the embodiment of the present application for updating the oil temperature regulation matrix is shown.

[0203] As Figure 7 shown, in the iteration process, when a node to be updated is selected, the corresponding column is extracted to form a node oil temperature sub-matrix, and a temperature response kernel is calculated by combining the residual transmission coefficient of the basic loop and the node admission weight. Subsequently, the temperature response kernel is used to perform weighted correction on the matrix elements of the current column to generate an updated column vector, and the column vector is used to replace the corresponding column in the original oil temperature regulation matrix. In this way, the matrix maintains the overall consistency of the cold end nodes in the row dimension, and converges in the column dimension through iteration, so that the allocation of cooling resources gradually approaches the actual heat transfer capacity of the oil circuit and the cooling demand of the nodes. Finally, with the column-level update of all nodes to be processed in turn, the output matrix is the stable regulation result under the current time window, which can be directly mapped to the adjustment instruction of the cooler opening degree or the oil pump flow.

[0204] It can be understood that, Figure 7 only for the purpose of illustrating the iterative update process of the oil temperature regulation matrix in the embodiment of the present application, the example is used to show how the node oil temperature sub-matrix replaces the corresponding column of the original matrix in the update process to reflect the dynamic relationship between the cold end contribution and the node admission. It should be noted that, Figure 7 the diagram is a schematic diagram and does not limit the specific number of rows, the number of columns, the value of the matrix elements, and the visualization form. In different application scenarios, the size of the matrix can vary with the complexity of the gearbox oil circuit structure, the number of cold ends, and the number of hot nodes, and the specific calculation of the matrix elements can also be adjusted according to the working condition parameters, the temperature measurement accuracy, and the execution end capability. Therefore, Figure 7The role of the above-mentioned embodiments is to assist in explaining the logical process of updating and replacing, and should not be considered as a limitation of the technical solutions of the present application. Those skilled in the art can make various equivalent modifications and extensions under the inspiration of the present application.

[0205] In one example, the present application provides an oil temperature regulation system for gear box lubricating oil, the system comprising:

[0206] A data acquisition module for acquiring oil temperature data of the gear box lubricating oil and processing in combination with a graph model corresponding to the gear box structure;

[0207] A heat map construction module for constructing an oil temperature heat map according to the heat coupling relationship of the graph model and estimating the temperature distribution of each node;

[0208] A node identification module for calculating node heat potential according to the oil temperature heat map and locating a plurality of to-be-processed nodes, and calculating entropy production sensitivity in combination with a basic loop to determine a sequence order of the to-be-processed nodes;

[0209] A matrix generation module for presetting a current oil temperature regulation matrix according to the node heat potential, cold end node capacity and reachability relationship;

[0210] A matrix update module for calculating a temperature response kernel of a to-be-updated node in combination with the basic loop in an iterative process, updating the oil temperature regulation matrix until the sequence order is empty, and outputting an updated oil temperature regulation matrix;

[0211] An oil temperature regulation module for regulating oil temperature according to the output of the matrix update module.

[0212] Although the embodiments of the present application have been shown and described above, it can be understood that the above-mentioned embodiments are exemplary and cannot be understood as a limitation of the present application. Those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method of oil temperature regulation for a gear box lubricating oil, characterized in that, The method comprises: obtaining oil temperature data of the lubricating oil, and constructing an oil temperature thermodynamic diagram in combination with a graph model corresponding to a gearbox structure, wherein the gearbox structure comprises a gearbox cavity, a bearing cavity and a meshing pair, and the construction of the oil temperature thermodynamic diagram comprises: dividing the gearbox into nodes and edges according to a thermal coupling relationship of the structure in the graph model, wherein the nodes represent local thermal capacity and oil quantity, and the edges represent thermal transfer connectivity and oil flow directional attribute; calculating an edge function for representing heat contribution attenuation according to the thermal transfer connectivity and the oil flow directional attribute, wherein the edge function is calculated according to path length, residence time of the lubricating oil on the edge, shunt ratio and cold end absorption capacity; calculating a node function for representing heat accumulation according to the local thermal capacity and the oil quantity, wherein the node function is calculated according to a heat conservation constraint in combination with temperatures of adjacent nodes and a node temperature; propagating the oil temperature data along the edges through the edge function, and accumulating incident contributions at the nodes according to the node function to obtain temperature estimation of each node; and constructing an oil temperature thermodynamic diagram according to the temperature estimation; calculating node thermal potential and locating a plurality of to-be-processed nodes according to the oil temperature thermodynamic diagram; generating a corresponding number of basic loops according to the to-be-processed nodes, and calculating entropy production sensitivity of the basic loops, comprising: extracting a basic loop set that passes through a corresponding temperature correlation region and is connected with a cold end node according to a connectivity relationship of the graph model, with the temperature correlation region being a constraint of each to-be-processed node, wherein the cold end node represents a boundary node for heat exchange with the outside world; for each basic loop, calculating a loop entropy generation rate in the basic loop through a discrete thermodynamic relationship, and calculating the entropy production sensitivity according to the loop entropy generation rate, wherein the loop entropy generation rate is calculated through heat transfer flux and temperature potential gradient; weighting and aggregating the entropy production sensitivity of a plurality of basic loops passing through the same to-be-processed node to obtain an entropy production sensitivity index of each to-be-processed node, wherein the weight is calculated according to residence time and path reachability of the basic loop passing through the corresponding to-be-processed node; and determining a sequence order of the to-be-processed nodes according to a descending order of the entropy production sensitivity index; presetting a current oil temperature regulation matrix according to the node thermal potential, performing an iteration operation, selecting a to-be-updated node according to the sequence order, calculating a temperature response kernel of the to-be-updated node in combination with the basic loop, updating and replacing the current oil temperature regulation matrix according to the temperature response kernel, until the sequence order is empty, outputting the current oil temperature regulation matrix, and performing oil temperature regulation according to the current oil temperature regulation matrix. The temperature response kernel of the to-be-updated node is calculated in combination with the basic loop, including: determining a node oil temperature sub-matrix according to the current oil temperature regulation matrix and the to-be-updated node; multiplying and accumulating according to the edge function in the basic loop corresponding to the to-be-updated node from the cold end node in the node oil temperature sub-matrix to obtain a residual transmission coefficient; calculating the admission weight of the to-be-updated node according to the node thermal potential, wherein the admission weight is used to represent the cooling conversion capability of the to-be-updated node; calculating the effective share vector of the current column of the node oil temperature sub-matrix according to the matrix elements of the current column of the node oil temperature sub-matrix, and calculating the temperature response kernel in combination with the residual transmission coefficient and the admission weight.

2. A method for regulating the temperature of lubricating oil in a gear box as claimed in claim 1, wherein, The node thermal potential is calculated according to the oil temperature thermodynamic diagram, and a plurality of to-be-processed nodes are located, including: The oil temperature thermodynamic diagram is subjected to Gaussian filtering denoising processing, and a two-dimensional temperature gradient vector field is constructed, wherein the two-dimensional temperature gradient vector field represents the node thermal potential; The temperature gradient values of each node in the X-axis direction and the Y-axis direction of the rectangular coordinate system are calculated according to the two-dimensional temperature gradient vector field, and a temperature gradient vector is generated; If the modulus of the temperature gradient vector is greater than or equal to a preset temperature gradient threshold value, the corresponding node is marked as a candidate node; An iteration operation is performed on the candidate node to obtain a temperature associated region of the candidate node, and the number of nodes in the temperature associated region is used to determine the to-be-processed node.

3. A method for regulating the temperature of lubricating oil in a gear box as claimed in claim 2, wherein, The iteration operation is performed on the candidate node to obtain the temperature associated region of the candidate node, including: An independent set is created for each candidate node, and the average main gradient direction of the set is calculated, wherein the average main gradient direction is determined based on the local density parameter and the temperature gradient value; An iteration operation is performed to obtain the current main gradient direction of the adjacent node of any node in the set, and if the included angle between the current main gradient direction and the average main gradient direction is less than a preset same direction threshold value, the corresponding adjacent node is added to the set, until the included angle between the current main gradient direction of the adjacent node and the average main gradient direction is greater than or equal to the same direction threshold value, and the set is output as the temperature associated region.

4. A method for regulating the temperature of lubricating oil in a gear box as claimed in claim 1 wherein, The rows of the current oil temperature regulation matrix correspond to the cold end nodes, the columns of the current oil temperature regulation matrix correspond to the to-be-processed nodes, and the matrix elements of the current oil temperature regulation matrix are calculated according to the node thermal potential.

5. A method for regulating the temperature of lubricating oil in a gear box as claimed in claim 4 wherein, A current oil temperature regulation matrix is preset according to the node thermal potential, including: For each cold end node, the cold end capability weight and the delay registration parameter are obtained according to the corresponding instantaneous capacity upper limit and the response time delay; For each to-be-processed node, the node cooling demand weight and the confidence weight are obtained according to the corresponding node thermal potential, the temperature gradient intensity and the number of nodes in the temperature associated region; The reachability weight between each cold end node and each to-be-processed node is calculated according to the basic loop of the to-be-processed node in combination with the delay registration parameter; The matrix elements are calculated according to the cold end capability weight, the node cooling demand weight, the confidence weight and the reachability weight to obtain the current oil temperature regulation matrix.

6. A method for regulating the temperature of lubricating oil in a gear box as claimed in claim 1 wherein, The temperature response kernel is updated and replaces the current oil temperature regulation matrix, including: According to the temperature response, the matrix elements of the current column of the node oil temperature sub-matrix are updated by weighting; The matrix elements of the current column of the node oil temperature sub-matrix after the weighting update are substituted into the matrix elements in the corresponding column of the current oil temperature regulation matrix.

7. An oil temperature regulating system for gear box lubricating oil for implementing a method of regulating the temperature of gear box lubricating oil as claimed in any one of claims 1 to 6, characterized in that, The system comprises: A data acquisition module is configured to acquire oil temperature data of the gearbox lubricating oil and process the data in combination with a graph model corresponding to the gearbox structure; A thermal map construction module is configured to construct an oil temperature thermal map according to a thermal coupling relationship of the graph model and estimate temperature distribution of each node; A node identification module is configured to calculate node heat potential according to the oil temperature thermal map, locate a plurality of to-be-processed nodes, and calculate entropy production sensitivity in combination with a basic loop to determine a sequence of the to-be-processed nodes; A matrix generation module is configured to preset a current oil temperature regulation matrix according to the node heat potential, cold end node capacity and reachability relationship; A matrix update module is configured to calculate a temperature response kernel of a to-be-updated node in combination with the basic loop in an iteration process, update the oil temperature regulation matrix until the sequence is empty, and output the updated oil temperature regulation matrix; An oil temperature regulation module is configured to regulate oil temperature according to the output of the matrix update module.

Images