Control method of methanol engine thermal management unit based on air compression thermal compensation and exhaust gas heat exchange

Abstract

This invention belongs to the field of automatic control technology, specifically relating to a control method for a methanol engine thermal management unit based on compressed air heat compensation and exhaust gas heat exchange. The control method includes the following steps: Step 1, unifying the heat transfer process and fluid flow process in the methanol engine thermal management unit into a multi-physics domain bonded graph model containing both thermal and fluid domains, and setting feature vectors containing both thermal and fluid domain features for each node; Step 2, executing message passing on the graph neural network to read the thermal domain temperature features of the corresponding node at the methanol engine intake as the predicted intake temperature; Step 3, the PLC controller, based on the comparison between the predicted and target intake temperatures, ensures the intake temperature reaches the target intake temperature. This invention overcomes the shortcomings of existing technologies that rely on independent control of each heat exchange component and neglect of the system's thermal-fluid coupling relationship through multi-physics domain coupled modeling and online iterative solution using graph neural networks.

Description

Technical Field

[0001] This invention belongs to the field of automatic control technology, specifically relating to a control method for a methanol engine thermal management unit based on air pressure thermal compensation and exhaust gas heat exchange. Background Technology

[0002] Existing technologies for regulating engine intake air temperature mainly include the following: The first method involves directly heating the air in the intake manifold using an electric heating device. This method is simple in structure and has a fast response time, effectively increasing the intake air temperature during cold starts. However, the electric heater consumes a significant amount of energy, and relying on it to maintain the intake air temperature after prolonged engine operation significantly increases the system's total energy consumption and reduces the overall efficiency of the generator set. The second method utilizes engine coolant or exhaust waste heat to heat the intake air. This method can recover waste heat generated by engine operation, thereby reducing external energy consumption. However, it has an inherent limitation: during the initial or pre-start phases of the engine, the coolant and exhaust temperatures are very low, and the waste heat recovery system cannot provide sufficient heating capacity, thus failing to solve the problem of insufficient intake air temperature during cold starts. The third method involves installing an intercooler or heat exchanger in the intake manifold to cool the high-temperature intake air. This solution is widely used in turbocharged engines, but traditional intercoolers are typically designed for a single cooling function and lack the ability to reverse heat in low-temperature environments, limiting their applicability to various operating conditions.

[0003] Most of the existing solutions mentioned above are designed for a specific temperature range or a specific operating condition, lacking a unified thermal management architecture that can simultaneously cover extreme low and high temperature environments. While some technical solutions attempt to combine electric heating with waste heat recovery, they still fall short in terms of switching control between different operating conditions. Traditional control methods typically employ simple threshold judgment based on feedback from a single temperature sensor: heating is activated when the intake air temperature is below a set value, and cooling is activated when it is above the set value. This control method has two significant drawbacks: First, the temperature feedback from a single sensor only reflects the current intake air temperature state and cannot predict the trend of intake air temperature changes, causing control actions to always lag behind temperature changes. This can easily lead to significant overshoot or undershoot of intake air temperature during rapid switching of operating conditions. Second, simple threshold judgment lacks a hysteresis mechanism and minimum hold time constraint. When the intake air temperature fluctuates slightly around the threshold, the heating and cooling devices will frequently start and stop alternately, increasing not only the mechanical and electrical losses of the equipment but also potentially causing larger oscillations in the intake air temperature, which in turn worsens combustion stability. Summary of the Invention

[0004] Therefore, the main objective of this invention is to provide a control method for the thermal management unit of a methanol engine based on air pressure thermal compensation and exhaust gas heat exchange. This invention overcomes the shortcomings of existing technologies, such as independent control of each heat exchange component and neglect of the system's thermal flux coupling relationship, by using multi-physics domain coupled modeling and online iterative solution of graph neural networks. It achieves adaptive intake temperature control of the methanol engine in the entire ambient temperature range, effectively solving the problems of difficult cold start and unstable combustion and easy knocking of methanol engines in low-temperature environments.

[0005] The technical solution adopted in this invention is as follows:

[0006] A control method for a methanol engine thermal management unit based on compressed air heat compensation and exhaust gas heat exchange is proposed. The methanol engine thermal management unit includes a methanol engine, an air compressor, an electric heater, a first heat exchanger, a second heat exchanger, a three-way pipeline, and a PLC controller. The PLC controller is connected to temperature sensors located at a first temperature monitoring point, a second temperature monitoring point, and a third temperature monitoring point. The first temperature monitoring point is located at the outlet side of the first heat exchanger, the second temperature monitoring point is located at the outlet side of the second heat exchanger, and the third temperature monitoring point is located at the methanol engine inlet. The PLC controller is connected to the pressure signal interface of the air compressor. A cooling medium inlet temperature sensor is installed on the cooling side of the first heat exchanger and is connected to the PLC controller. The control method includes the following steps:

[0007] Step 1: The heat transfer process and fluid flow process in the methanol engine thermal management unit are uniformly modeled as a multi-physics domain bond graph model that includes thermal domain and fluid domain. Each element in the multi-physics domain bond graph model is mapped to a node of a graph neural network, each energy bond is mapped to a directed edge of the graph neural network, and feature vectors containing thermal domain features and fluid domain features are set for each node.

[0008] Step 2: Message passing is performed on the graph neural network. Each node updates its own feature vector according to the update rules of the corresponding element type based on the features passed from the adjacent nodes through the directed edges. After completing one round of message passing, pipeline flow modeling is performed. The difference in fluid domain pressure characteristics of the nodes at both ends of each pipe segment is used as the driving pressure difference to solve for the flow velocity and mass flow rate of each pipe segment. The mass flow rate is written back to the feature vector of the corresponding node in the graph neural network. Message passing and pipeline flow modeling are performed alternately until convergence or the preset maximum number of iterations is reached. The thermal domain temperature characteristics of the corresponding node of the methanol engine intake are read as the predicted intake temperature.

[0009] Step 3: Based on the comparison between the predicted intake air temperature and the target intake air temperature, the PLC controller controls the electric heater or the first heat exchanger to work when the methanol engine is not started, and controls the air compressor to work in conjunction with the second heat exchanger or controls the first heat exchanger to work after the methanol engine is started, so that the intake air temperature reaches the target intake air temperature.

[0010] Furthermore, within the thermal domain, the methanol engine is configured as a thermally capacitive element, the electric heater is configured as a controllable thermal potential source element, the first heat exchanger is configured as a first two-port thermally resistive element, the second heat exchanger is configured as a second two-port thermally resistive element, and the cooling medium inlet on the cooling side of the first heat exchanger is configured as a thermal domain boundary node.

[0011] Furthermore, in the fluid domain, the air compressor is configured as a fluid domain converter element; the pipe segment connecting the air compressor outlet to the first heat exchanger inlet in the intake pipe is configured as a series combination of the first fluid inertial element and the first fluid resistive element; the pipe segment connecting the second heat exchanger outlet to the methanol engine inlet is configured as a series combination of the second fluid inertial element and the second fluid resistive element; the three-way pipe is configured as node 1, with the inlet port of node 1 receiving the exhaust heat generated by the methanol engine, and the two outlet ports of node 1 connected to the second heat exchanger and the external heat dissipation pipe, respectively; the junction of the first heat exchanger outlet and the second heat exchanger outlet in the intake pipe is configured as node 0.

[0012] Furthermore, each node is configured with a 4-dimensional feature vector: the first dimension is the thermal domain temperature feature, the second dimension is the thermal domain heat flow rate feature, the third dimension is the fluid domain pressure feature, and the fourth dimension is the fluid domain mass flow rate feature. The thermal domain and the fluid domain are connected by a convection coupling element, which characterizes the heat transport process carried by the fluid as it flows through the heat exchanger. The PLC controller reads the real-time temperature value of the temperature sensor at the first temperature monitoring point and assigns it to the thermal domain temperature feature of the node corresponding to the first dual-port thermal resistive element; reads the real-time temperature value of the temperature sensor at the second temperature monitoring point and assigns it to the thermal domain temperature feature of the node corresponding to the second dual-port thermal resistive element; reads the real-time temperature value of the temperature sensor at the third temperature monitoring point and assigns it to the thermal domain temperature feature of the node corresponding to the methanol engine air inlet; assigns the air compressor outlet pressure to the fluid domain pressure feature of the node corresponding to the fluid domain converter element; and assigns the reading of the cooling medium inlet temperature sensor to the thermal domain temperature feature of the thermal domain boundary node.

[0013] Furthermore, for the node corresponding to node 1, the heat flux characteristics of all incoming adjacent nodes are summed with a sign after being positive in the direction of incoming edge and negative in the direction of outgoing edge. The summation result is then distributed to each downstream adjacent node of outgoing edge as the update increment of heat flux characteristics according to the pre-calibrated flow splitting ratio coefficient. The flow splitting ratio coefficient is determined by the ratio of the equivalent thermal conductivity of the second heat exchanger branch to that of the external heat dissipation branch.

[0014] Furthermore, for the node corresponding to node 0, the fluid domain mass flow characteristics transmitted from all adjacent nodes with incoming edges are summed with a sign according to the direction of incoming edges and the direction of outgoing edges to ensure continuity. At the same time, the thermal domain temperature characteristics and fluid domain mass flow characteristics of each adjacent node with incoming edges are read. The mass flow characteristics of each incoming branch are multiplied by the corresponding thermal domain temperature characteristics one by one, summed, and then divided by the sum of the mass flow characteristics of all incoming branches to obtain the mixed temperature and update it to the thermal domain temperature characteristics of node 0. This temperature is then transmitted to the downstream adjacent nodes along the outgoing edges.

[0015] Furthermore, for the node corresponding to the first two-port thermally resistive element, the thermal temperature characteristics of the adjacent node on the inlet side and the thermal boundary node of the first two-port thermally resistive element are read respectively. The difference between the two temperature characteristics is multiplied by the equivalent thermal conductivity value pre-calibrated by the first heat exchanger to obtain the heat flow rate characteristics through the first heat exchanger. For the node corresponding to the second two-port thermally resistive element, the thermal temperature characteristics of the adjacent node on the inlet side and the adjacent node on the outlet side of node 1 are read respectively. The difference between the two temperature characteristics is multiplied by the equivalent thermal conductivity value pre-calibrated by the second heat exchanger to obtain the heat flow rate characteristics through the second heat exchanger.

[0016] Furthermore, for the node corresponding to the fluid domain converter element, the fluid domain pressure characteristics of the adjacent nodes on the incoming edge are read, scaled according to the compression ratio pre-calibrated by the air compressor, and then updated to the fluid domain pressure characteristics on the outlet side. At the same time, the thermal domain temperature characteristics of the adjacent nodes on the incoming edge are read, and the outlet temperature is found according to the temperature rise mapping table corresponding to the compression ratio pre-calibrated by the air compressor and then updated to the thermal domain temperature characteristics. At the end of each round of message transmission, the convection coupling element multiplies the fluid domain mass flow rate characteristics connected at both ends of the convection coupling element, the difference in thermal domain temperature characteristics between adjacent nodes, and the pre-calibrated equivalent enthalpy coefficient to obtain the convective heat transfer. The sign of the convective heat transfer is determined by the fluid flow direction.

[0017] Furthermore, the convergence criterion is that the change in the thermal domain temperature characteristic of the node corresponding to the methanol engine intake port is less than the preset convergence threshold in two consecutive iterations.

[0018] Furthermore, in step 3, the PLC controller starts the corresponding heating device when the predicted intake air temperature is lower than the target intake air temperature minus the preset hysteresis deviation value, and shuts down the corresponding heating device when the predicted intake air temperature reaches the target intake air temperature plus the preset hysteresis deviation value. After the methanol engine starts and the ambient temperature is lower than the preset low temperature operating threshold, the air compressor compresses and heats the air. The compressed and heated air enters the second heat exchanger, which uses the heat energy supplied by the exhaust gas generated by the methanol engine through the three-way pipe to continue heating the air before sending it into the methanol engine. After the methanol engine starts and the ambient temperature is higher than the preset high temperature operating threshold, the air is compressed by the air compressor and enters the first heat exchanger for heat exchange and cooling. The cooled air then enters the methanol engine. After each working mode switch, the PLC controller performs a preset minimum holding time lock.

[0019] By adopting the above technical solutions, the present invention has produced the following beneficial effects: The methanol engine thermal management unit control method based on air compression thermal compensation and exhaust gas heat exchange provided by the present invention integrates an electric heater, an air compressor, a first heat exchanger, a second heat exchanger, and a three-way pipeline into a system, and combines an alternating coupling iterative solution mechanism of multi-physics domain bond graph modeling, graph neural network message passing, and pipeline flow modeling, thereby realizing adaptive intake temperature control of the methanol engine in the entire ambient temperature range. Compared with the prior art, it has the following beneficial effects.

[0020] This invention constructs a unified thermal management architecture encompassing four temperature regulation methods: compressed air thermal compensation, exhaust gas heat recovery, electric heating assistance, and heat exchange cooling. This architecture enables the methanol engine thermal management unit to simultaneously cover both extreme low-temperature cold start conditions and extreme high-temperature operating conditions. During the methanol engine's inactive state, the electric heater independently handles the intake air heating, resolving the issue of the waste heat recovery system failing to operate due to a lack of heat source during cold starts. After the methanol engine starts, the air compressor and the second heat exchanger work together, utilizing a two-stage heating method of compression heating and exhaust gas waste heat recovery to replace the electric heater. This reduces system energy consumption and fully utilizes the exhaust heat generated by the engine itself, improving energy efficiency. Under high-temperature operating conditions, compressed air is cooled by the first heat exchanger before entering the methanol engine, effectively suppressing the decrease in charging efficiency and the risk of knocking caused by high-temperature intake air.

[0021] This invention unifies the heat transfer and fluid flow processes into a multi-physics domain bonded graph model encompassing both thermal and fluid domains. This model is then mapped to a graph neural network structure for online iterative solving, overcoming the limitations of existing technologies that treat each heat exchange component as an independent control object and ignore the thermal-fluid coupling relationship between components. The message passing mechanism of the graph neural network allows each node to aggregate feature information from adjacent nodes and update according to the physical laws of the corresponding component type. Pipeline flow modeling, after each round of message passing, calculates the actual flow rate of each pipe segment based on the pressure distribution in the fluid domain and writes it back to the graph neural network. These two processes alternately couple and iterate until convergence, resulting in a predicted intake temperature that comprehensively considers the thermal-fluid coupling effect of the entire system. This predicted intake temperature has better foresight and globality compared to the measured temperature from a single sensor, making the PLC controller's operating mode switching judgment more accurate and timely.

[0022] This invention introduces a hysteresis protection mechanism and a minimum hold time locking mechanism into the operating mode switching control, effectively avoiding the problem of frequent alternating start-stop of heating and cooling equipment when the predicted intake air temperature fluctuates slightly around the target value. This reduces mechanical wear and electrical shocks, and improves the operational stability of the control system and the service life of the equipment. Simultaneously, the graph neural network has a small node size, and the computational load for single-round message passing and pipeline flow modeling is limited, making it suitable for real-time operation on a PLC controller, meeting the constraints of control cycle and computational resources in industrial settings. Attached Figure Description

[0023] Figure 1 The present invention provides a switching characteristic curve of electric heater hysteresis control for a methanol engine thermal management unit under cold start conditions.

[0024] Figure 2 The following are the intake air temperature dynamic response curves of the methanol engine thermal management unit under four typical ambient temperature conditions provided in this embodiment of the invention.

[0025] Figure 3 The pressure distribution, velocity distribution, and relationship curves between mass flow rate and convective heat transfer along the intake pipeline obtained by the pipeline flow modeling solution provided in the embodiments of the present invention are shown. Detailed Implementation

[0026] A control method for a methanol engine thermal management unit based on compressed air heat compensation and exhaust gas heat exchange is proposed. The methanol engine thermal management unit includes a methanol engine, an air compressor, an electric heater, a first heat exchanger, a second heat exchanger, a three-way pipeline, and a PLC controller. The PLC controller is connected to temperature sensors located at a first temperature monitoring point, a second temperature monitoring point, and a third temperature monitoring point. The first temperature monitoring point is located at the outlet side of the first heat exchanger, the second temperature monitoring point is located at the outlet side of the second heat exchanger, and the third temperature monitoring point is located at the methanol engine inlet. The PLC controller is connected to the pressure signal interface of the air compressor. A cooling medium inlet temperature sensor is installed on the cooling side of the first heat exchanger and is connected to the PLC controller. The control method includes the following steps:

[0027] Step 1: The heat transfer process and fluid flow process in the methanol engine thermal management unit are uniformly modeled as a multi-physics domain bond graph model that includes thermal domain and fluid domain. Each element in the multi-physics domain bond graph model is mapped to a node of a graph neural network, each energy bond is mapped to a directed edge of the graph neural network, and feature vectors containing thermal domain features and fluid domain features are set for each node.

[0028] Step 2: Message passing is performed on the graph neural network. Each node updates its own feature vector according to the update rules of the corresponding element type based on the features passed from the adjacent nodes through the directed edges. After completing one round of message passing, pipeline flow modeling is performed. The difference in fluid domain pressure characteristics of the nodes at both ends of each pipe segment is used as the driving pressure difference to solve for the flow velocity and mass flow rate of each pipe segment. The mass flow rate is written back to the feature vector of the corresponding node in the graph neural network. Message passing and pipeline flow modeling are performed alternately until convergence or the preset maximum number of iterations is reached. The thermal domain temperature characteristics of the corresponding node of the methanol engine intake are read as the predicted intake temperature.

[0029] Step 3: Based on the comparison between the predicted intake air temperature and the target intake air temperature, the PLC controller controls the electric heater or the first heat exchanger to work when the methanol engine is not started, and controls the air compressor to work in conjunction with the second heat exchanger or controls the first heat exchanger to work after the methanol engine is started, so that the intake air temperature reaches the target intake air temperature.

[0030] In this embodiment, the first step of the methanol engine thermal management unit control method is to abstractly model the physical structure and energy transfer relationships of the methanol engine thermal management unit, establish a unified topological expression that can simultaneously describe both heat transfer and fluid motion, and transform this topological expression into a graph neural network data structure suitable for online iterative solution. The reason for incorporating two different physical domains into the same model framework is that, in actual operation, the intake air temperature of the methanol engine thermal management unit depends not only on the heat exchange capacity of each heat exchange component but also on the air flow state in the pipeline—flow velocity affects the intensity of convective heat transfer, pressure distribution determines the flow distribution of each branch, and the flow distribution, in turn, changes the actual heat transfer capacity of each heat exchanger. If only the thermal domain is modeled separately while ignoring the coupling effect of the fluid domain, the model will not accurately reflect the dynamic changes in intake air temperature under real operating conditions, leading to inaccurate mode switching criteria of the PLC controller.

[0031] Specifically, the construction process of the multi-physics domain bond graph model is divided into four parts: thermal domain component setting, fluid domain component setting, node setting, and cross-domain coupling component setting.

[0032] For the thermal domain element setting, each component in the methanol engine thermal management unit involved in heat generation, storage, and transfer is abstracted as a standard element in the bond graph. During operation, the methanol engine continuously generates heat and releases it externally through media such as the cylinder block and coolant. This characteristic corresponds to a capacitive element capable of storing energy in bond graph theory; therefore, the methanol engine is set as a thermal capacitive element. The physical meaning of a thermal capacitive element is that when heat is injected or extracted from it externally, its temperature (i.e., thermal domain potential change) will change accordingly, and the magnitude of the change is inversely proportional to the equivalent heat capacity of the thermal capacitive element itself. In this embodiment, the equivalent heat capacity of the methanol engine can be pre-calibrated through the product relationship of parameters such as engine cylinder mass, coolant volume, and overall specific heat capacity. The typical value range is 50,000 to 200,000 joules per Kelvin, with the specific value depending on the engine displacement and cooling system volume.

[0033] The electric heater performs the function of actively heating the intake air during the cold start condition when the methanol engine is not running. Its output heating power can be adjusted by the PLC controller as needed. This characteristic corresponds to a controllable thermal potential source element in the bond diagram that can actively inject energy into the system with a controllable injection amount. After setting the electric heater as a controllable thermal potential source element, its behavior in the bond diagram is as follows: the PLC controller sends the target thermal potential value (i.e., target temperature) to the controllable thermal potential source element, which then outputs the corresponding heat flow rate to the energy bond connected to it, thereby achieving active regulation of the intake air temperature. In this embodiment, the rated heating power of the electric heater is 3 to 15 kilowatts, which can be configured according to different unit specifications.

[0034] The first and second heat exchangers in the unit undertake heat exchange tasks under different operating conditions. The first heat exchanger cools the air after it has been compressed and heated by the air compressor in high-temperature environments, and can also be used as an auxiliary cooling method during cold starts. The second heat exchanger specifically utilizes the waste heat from the exhaust gas generated after the methanol engine starts to heat the intake air. Both share the characteristic that heat flows from the high-temperature side to the low-temperature side, and the heat exchange capacity depends on the temperature difference between the two sides and the equivalent thermal conductivity of the heat exchanger itself. This characteristic corresponds to a two-port thermally resistive element with two ports in the bond diagram, where heat flow is driven by the temperature difference between the ports. Therefore, the first heat exchanger is designated as the first two-port thermally resistive element, and the second heat exchanger is designated as the second two-port thermally resistive element. The equivalent thermal conductivity of each two-port thermally resistive element is obtained through offline calibration experiments: the inlet and outlet temperatures and flow rates on both sides of the heat exchanger are measured under different flow and temperature conditions, the equivalent thermal conductivity under the corresponding conditions is calculated, and stored as a lookup table. In this embodiment, the equivalent thermal conductivity of the first heat exchanger ranges from 100 to 800 W / Kelvin, and the equivalent thermal conductivity of the second heat exchanger ranges from 80 to 600 W / Kelvin, with the specific values ​​varying depending on the flow rate and operating conditions.

[0035] Furthermore, the cooling side of the first heat exchanger requires a cooling medium (such as coolant or ambient air) to provide a cold source, and the temperature at the inlet of the cooling medium is a crucial boundary condition determining the heat exchange capacity of the first heat exchanger. To enable the model to acquire this boundary condition, the cooling medium inlet on the cooling side of the first heat exchanger is set as a thermal domain boundary node. The thermal domain boundary node does not participate in the energy conservation calculations within the model; its temperature characteristics are directly provided by external sensors, serving as the known boundary input for the entire thermal domain calculation.

[0036] For the fluid domain component designation, components related to airflow in the methanol engine thermal management unit are abstracted as fluid domain bond graph components. The function of the air compressor is to compress low-pressure air into high-pressure air, and the air temperature rises accordingly due to the thermodynamic effects of the compression process. In bond graph theory, a component that converts one form of power variable (such as pressure and flow rate on the low-pressure side) into another form of power variable (such as pressure and flow rate on the high-pressure side) at a fixed ratio is called a converter component. The air compressor fits this definition perfectly, therefore it is designated as a fluid domain converter component. The core parameter of the fluid domain converter component is the pre-calibrated compression ratio of the air compressor, which is defined as the ratio of the absolute pressure at the air compressor outlet to the absolute pressure at the inlet. In this embodiment, the compression ratio of the air compressor ranges from 1.5 to 4.0, with the specific value determined by the operating power of the methanol engine. The compression ratio also corresponds to a temperature rise mapping relationship: given the inlet temperature and compression ratio, the outlet temperature can be obtained through a pre-established temperature rise mapping table. The method for establishing the temperature rise mapping table is as follows: During the factory testing or field calibration of the air compressor, 5 to 10 inlet temperature points (e.g., from -40 degrees Celsius to 50 degrees Celsius, with an interval of 10 degrees Celsius) and 5 to 8 compression ratio points (e.g., from 1.5 to 4.0, with an interval of 0.5) are used as grid nodes. The outlet temperature corresponding to each grid node is measured. The outlet temperature under any combination of inlet temperature and compression ratio can be obtained by bilinear interpolation.

[0037] Each pipe segment in the intake duct generates frictional resistance and local resistance as air flows through it. Simultaneously, the air inside the pipe possesses a certain inertia, and flow rate changes are not instantaneous. To describe these two effects, the pipe segment connecting the air compressor outlet to the first heat exchanger inlet is designated as a series combination of a first fluid inertial element and a first fluid resistive element; the pipe segment connecting the second heat exchanger outlet to the methanol engine intake is designated as a series combination of a second fluid inertial element and a second fluid resistive element. The fluid inertial element characterizes the resistance of the air mass inside the pipe to flow rate changes; its equivalent inertia is determined by the product of the pipe segment length, the pipe cross-sectional area, and the air density. The fluid resistive element characterizes the resistance of the pipe segment to airflow; its equivalent flow resistance is determined by parameters such as pipe segment length, pipe inner diameter, pipe wall roughness, and air viscosity through the pipe flow resistance relationship. In this embodiment, the pipe inner diameter ranges from 25 to 80 mm, the pipe wall roughness is from 0.02 to 0.1 mm, and the pipe segment length is determined according to the actual layout of the unit, typically from 0.5 to 3 meters.

[0038] In bond graph theory, nodes are used to describe the convergence or distribution of energy. The T-junction in the methanol engine thermal management unit functions to distribute the exhaust heat generated by the methanol engine to two branches: the second heat exchanger and the external cooling pipe. In bond graph theory, a node 1 is characterized by equal flow variables at all connected ports, and the potential variables at each port satisfy the constraint that their algebraic sum is zero. The flow distribution characteristics of the T-junction conform to this description, therefore it is designated as a node 1. The inlet port of node 1 receives the exhaust heat generated by the methanol engine, and the two outlet ports of node 1 are connected to the second heat exchanger and the external cooling pipe, respectively. The distribution ratio of exhaust heat between the two branches is not equal, but determined by the ratio of the equivalent thermal conductivities of the two branches—the branch with the larger equivalent thermal conductivity will receive more heat flow. This distribution ratio coefficient is determined through calibration experiments after system installation and is fixed in the PLC controller. In an optional implementation, an electric regulating valve can be installed at the diversion node of the three-way pipeline. The valve opening can be adjusted by a PLC controller to dynamically change the diversion ratio coefficient, thereby achieving more precise control of the waste gas heat distribution.

[0039] The point where the outlets of the first and second heat exchangers on the intake duct merge is where two air streams with potentially different temperatures and flow rates converge into one before entering the methanol engine. In bond graph theory, a zero node is characterized by equal potential variables at all connected ports, and the flow variables at each port satisfy the constraint that their algebraic sum is zero. The pressure at the air convergence point tends to be uniform, and the total mass flow rate is conserved, which meets the definition of a zero node; therefore, this convergence point is designated as a zero node. The zero node also plays a role in calculating the mixing temperature during subsequent graph neural network message passing, specifically calculating the mixed air temperature after merging based on the mass flow rate and temperature of each inlet branch. This will be explained in detail in step 2.

[0040] In the cross-domain coupling element configuration, the thermal domain and the fluid domain do not operate independently; there is a close physical coupling between them. When air flows through a heat exchanger at a certain velocity, its heat-carrying capacity directly depends on its mass flow rate—the higher the flow rate, the more air flows through the heat exchanger per unit time, and the more heat is carried away or carried in. To characterize this cross-domain coupling relationship in the bond graph model, a convection coupling element is set between the thermal domain and the fluid domain. The two ends of the convection coupling element are connected to heat transfer-related nodes in the thermal domain and flow-related nodes in the fluid domain, respectively. Its function is to transfer the mass flow rate information of the fluid domain to the thermal domain and participate in the calculation of heat flux. Specifically, the convection coupling element multiplies the fluid domain mass flow rate characteristics, the temperature difference between adjacent thermal domain nodes, and the pre-calibrated equivalent enthalpy flow coefficient to obtain the actual convective heat transfer. The physical meaning of the equivalent enthalpy flow coefficient is the heat transferred by a unit mass flow rate of fluid under a unit temperature difference. Its value is determined by factors such as the specific heat capacity and heat transfer efficiency of the fluid, and is obtained through offline calibration experiments. In this embodiment, the equivalent enthalpy flow coefficient on the air side ranges from 800 to 1200 joules per kilogram per Kelvin. In an optional embodiment, the equivalent enthalpy flow coefficient can be set to a lookup table format with segmented values ​​across temperature ranges to improve calculation accuracy in different temperature zones.

[0041] After constructing the multiphysics bond graph model, it needs to be transformed into graph-structured data that can be processed by a graph neural network. The transformation rule is as follows: each element in the multiphysics bond graph model (including thermal capacitive elements, controllable thermal potential source elements, first two-port thermal resistive elements, second two-port thermal resistive elements, thermal domain boundary nodes, fluid domain converter elements, first fluid inertial elements, first fluid resistive elements, second fluid inertial elements, second fluid resistive elements, 1 nodes, 0 nodes, and convection coupling elements) is mapped to one node in the graph neural network, and each energy bond connecting two elements in the bond graph is mapped to one directed edge in the graph neural network. The direction of the directed edge is consistent with the power variable transmission direction determined by the causal lines in the bond graph, which enables the graph neural network to propagate feature information along the correct physical direction during subsequent message transmission.

[0042] In the methanol engine thermal management unit of this embodiment, the graph neural network generated according to the above mapping rules contains 13 nodes and several directed edges. The graph is relatively small, which provides a computational feasibility guarantee for subsequent real-time iterative solving on the PLC controller.

[0043] Each node has a 4-dimensional feature vector to simultaneously carry information on both thermal and fluid domain physical quantities. The first dimension is the thermal domain temperature feature, which represents the temperature value at the location of the corresponding component, in degrees Celsius or Kelvin. The second dimension is the thermal domain heat flow rate feature, which represents the rate of heat transfer through or acting on the corresponding component, in watts. The third dimension is the fluid domain pressure feature, which represents the absolute air pressure at the location of the corresponding component, in Pascals. The fourth dimension is the fluid domain mass flow rate feature, which represents the mass flow rate of air through the pipe section containing the corresponding component, in kilograms per second. For components belonging only to the thermal domain (such as capacitive components and controllable heat potential source components), their third and fourth dimensions are set to zero during initialization and are not updated during message passing. For components belonging only to the fluid domain (such as fluid inertial components and fluid resistive components), their first and second dimensions are set to ambient temperature during initialization and only receive values ​​transferred from convection-coupled components during cross-domain coupling updates.

[0044] The initial assignment of feature vectors is performed by the PLC controller at the beginning of each sampling cycle. The PLC controller reads the real-time temperature value from the temperature sensor at the first temperature monitoring point and assigns it to the thermal domain temperature feature of the corresponding node of the first two-port thermally resistive element. This temperature value reflects the actual air temperature at the outlet of the first heat exchanger. The PLC controller reads the real-time temperature value from the temperature sensor at the second temperature monitoring point and assigns it to the thermal domain temperature feature of the corresponding node of the second two-port thermally resistive element. This temperature value reflects the actual air temperature at the outlet of the second heat exchanger. The PLC controller reads the real-time temperature value from the temperature sensor at the third temperature monitoring point and assigns it to the thermal domain temperature feature of the corresponding node of the methanol engine intake. This temperature value is a key variable that ultimately needs to be controlled within the target range. The PLC controller obtains the outlet pressure of the air compressor through the pressure signal interface and assigns it to the fluid domain pressure feature of the corresponding node of the fluid domain converter element. The PLC controller reads the temperature value from the cooling medium inlet temperature sensor and assigns it to the thermal domain temperature feature of the thermal domain boundary node. This temperature value serves as the boundary condition input model for the cooling side of the first heat exchanger. For the remaining node features not directly covered by the sensor, the iterative convergence value of the previous sampling period is used as the initial value for this period. When running for the first time, the ambient temperature and atmospheric pressure are used as the default initial values.

[0045] In an optional implementation, the graph structure of the graph neural network can be extended according to the actual configuration of the methanol engine thermal management unit. For example, when an intercooler or multi-stage compression system is added to the unit, only the corresponding components and energy bonds need to be added to the multi-physics domain bonded graph model, and they can be transformed into new nodes and directed edges in the graph neural network according to the same mapping rules, without changing the basic algorithm framework for message passing and pipeline flow modeling. This topology mapping-based modeling method gives the control method good scalability and can adapt to methanol engine thermal management units of different specifications and configurations.

[0046] After completing the multi-physics domain bonded graph model construction and graph neural network structure mapping in step 1, the core task of step 2 is to solve for the predicted intake temperature at the methanol engine intake port through alternating coupling iterations of message passing mechanism and pipeline flow modeling on the graph neural network. The reason for using alternating coupling iterations instead of a one-time solution is the bidirectional dependency between the thermal and fluid domains: the heat exchanger capacity of each heat exchanger in the thermal domain depends on the air mass flow rate through the heat exchanger, while the air mass flow rate depends on the pressure distribution of each pipe section in the fluid domain, and the pressure distribution is affected by changes in density and viscosity caused by changes in air temperature. This bidirectional coupling makes it impossible to determine the states of the two domains independently; therefore, iterative approximation is necessary to make the states of both domains simultaneously tend towards self-consistency.

[0047] refer to Figure 3 , Figure 3 This chart contains three sub-charts: upper, middle, and lower. The upper sub-chart's horizontal axis represents the nodes arranged sequentially along the intake pipe from the air compressor outlet to the methanol engine inlet, while the vertical axis represents absolute pressure in kPa. The upper sub-chart contains three curves corresponding to low compression ratio, medium compression ratio, and high compression ratio operating conditions, respectively. All three curves show a gradual decrease in pressure from the higher pressure at the air compressor outlet on the left to the lower pressure at the methanol engine inlet on the right. Taking the high compression ratio condition as an example, the absolute pressure at the air compressor outlet is approximately 354 kPa. After passing through the pipe sections corresponding to the first fluid inertial element and the first fluid resistive element, the pressure gradually decreases, dropping to approximately 330 kPa at the inlet of the first heat exchanger. After passing through the local resistance of the internal flow channels of the first heat exchanger, it further decreases to approximately 280 kPa, reaching approximately 265 kPa at the confluence point, and finally dropping to approximately 225 kPa at the methanol engine inlet. The pressure distribution along the pipeline is not a uniform linear decrease, but rather exhibits a significant pressure drop abruptly within the flow channels of the heat exchanger. This is due to the geometric structure of the corrugated channels inside the heat exchanger, which results in a significantly higher local resistance coefficient compared to straight pipe sections. The horizontal axis node positions of the neutron diagram are related to the upper neutron... Figure 1The vertical axis represents air velocity, measured in meters per second. All three curves show a trend of gradually increasing velocity from zero velocity at the air compressor outlet (since the compressor outlet node itself does not define a velocity) along the pipeline. The increase in velocity is related to changes in the pipe cross-sectional area and pressure drop: higher air velocities are observed at locations with larger driving pressure differences and smaller pipe cross-sectional areas. For example, under high compression ratio conditions, the velocity at the end of the intake pipe can reach 30 meters per second. The following sub-figure shows the mass flow rate. Heat exchange with convection The relationship between them. The horizontal axis of the subgraph below represents mass flow rate. The unit is kilograms per second, with values ​​ranging from 0.01 to 0.12 kilograms per second; the vertical axis represents convective heat transfer. The unit is watts. The subgraph below contains three straight lines, each corresponding to a temperature difference. Equivalent enthalpy coefficients for 10°C, 25°C, and 40°C. The value is 1005 joules per kilogram per Kelvin. All three straight lines originate near the origin and extend upwards to the right. The slope increases with increasing temperature difference, indicating that under the same mass flow rate conditions, a larger temperature difference results in greater convective heat transfer. For example, with a temperature difference of 40 degrees Celsius and a mass flow rate of 0.1 kilograms per second, the convective heat transfer is approximately 4020 watts. Figure 3 It intuitively demonstrates the spatial distribution characteristics of the pressure field and velocity field along the pipeline obtained by pipeline flow modeling, as well as the quantitative correspondence between mass flow rate and convective heat transfer, providing a physical basis for the calculation of heat transfer of convective coupling elements in the message transmission process of graph neural networks.

[0048] The specific process of a graph neural network performing one round of message passing is as follows. In each round of message passing, the graph neural network performs feature update operations on each node sequentially according to a pre-defined node traversal order. Different types of element nodes follow different update rules, which are determined by the constitutive relations and node laws of various elements in bond graph theory.

[0049] For node 1, its update process reflects the distribution pattern of waste gas heat at the T-junction. Node 1 assigns positive signs to the heat flux characteristics of all incoming adjacent nodes and negative signs to the outgoing directions, then performs a signed summation. The meaning of the signed summation is: the heat flux flowing into node 1 from the inlet is recorded as a positive value, and the heat flux flowing from node 1 to the outlet is recorded as a negative value. The summation result represents the net heat flux flowing into node 1. Subsequently, node 1 distributes the net heat flux to each downstream adjacent node according to a pre-defined splitting ratio coefficient, as the update increment of the heat flux characteristics of each downstream node. The splitting ratio coefficient reflects the competitive attraction of heat flux between the two branches; the branch with the larger equivalent thermal conductivity attracts a larger share of heat flux. In this embodiment, assuming the equivalent thermal conductivity of the second heat exchanger branch is 400 W / Kelvin and the equivalent thermal conductivity of the external heat dissipation branch is 200 W / Kelvin, the proportion of heat flow allocated to the second heat exchanger branch is 400 divided by the sum of 400 and 200, which is approximately 0.67, and the proportion allocated to the external heat dissipation branch is approximately 0.33. This flow splitting ratio is obtained through actual measurement during the system calibration phase and is fixed in the PLC controller. In an optional embodiment, if an electric regulating valve is installed at the flow splitting node of the three-way pipe, the flow splitting ratio can be replaced by a lookup table of valve opening and equivalent thermal conductivity. The PLC controller adjusts the valve opening according to the current operating conditions to dynamically adjust the flow splitting ratio.

[0050] For the node corresponding to node 0, its update process reflects the mass conservation and energy mixing law at the confluence of the two air streams. Node 0 first sums the fluid domain mass flow characteristics from all adjacent incoming nodes, assigning positive signs to the incoming direction and negative signs to the outgoing direction, to satisfy the mass continuity constraint—the total air mass flow rate flowing into the confluence equals the total air mass flow rate flowing out. Regarding temperature updates, node 0 needs to calculate the temperature after mixing the two air streams with different temperatures and flow rates. The calculation of the mixing temperature follows energy conservation: the mass flow characteristics of each incoming branch are multiplied by the corresponding thermal domain temperature characteristics, summed, and then divided by the sum of the mass flow characteristics of all incoming branches to obtain the mixing temperature. The physical basis of this calculation method is that, ignoring heat loss and pressure work during the mixing process, the total enthalpy of the mixed gas is equal to the sum of the enthalpies of the gases in each incoming branch. Since the enthalpy of air is approximately proportional to temperature, the mixing temperature is equal to the temperature average weighted by the flow rates of each branch. For example, if the air temperature flowing in from the outlet of the first heat exchanger is 15 degrees Celsius and the mass flow rate is 0.05 kg / s, and the air temperature flowing in from the outlet of the second heat exchanger is 8 degrees Celsius and the mass flow rate is 0.03 kg / s, then the mixing temperature is calculated as 15 multiplied by 0.05 plus 8 multiplied by 0.03, divided by the sum of 0.05 and 0.03, yielding approximately 12.4 degrees Celsius. Node 0 updates the mixing temperature to its own thermal domain temperature characteristics and propagates it along the outgoing edge to the downstream adjacent node, namely the node corresponding to the methanol engine's air inlet.

[0051] For the node corresponding to the first two-port thermal resistivity element, its update process reflects the driving mechanism of heat transfer in the first heat exchanger. The heat transfer capacity of the first heat exchanger is driven by the temperature difference between its two sides: the greater the temperature difference, the more heat is transferred per unit time. The node corresponding to the first two-port thermal resistivity element reads the thermal domain temperature characteristics of the adjacent nodes on the intake side and the thermal domain boundary node, respectively. The temperature of the thermal domain boundary node is provided in real time by the cooling medium inlet temperature sensor. Multiplying the difference between the intake side temperature and the cooling side temperature by the pre-calibrated equivalent thermal conductivity value of the first heat exchanger, the heat flow rate characteristic through the first heat exchanger is obtained. When the intake side temperature is higher than the cooling side temperature, the heat flow rate is positive, indicating that heat is transferred from the intake side to the cooling side, and the air is cooled; when the intake side temperature is lower than the cooling side temperature, the heat flow rate is negative, indicating that heat is transferred from the cooling side to the intake side, and the air is heated. In this embodiment, assuming the equivalent thermal conductivity of the first heat exchanger is 500 watts per Kelvin, the inlet temperature is 70 degrees Celsius, and the inlet temperature of the cooling medium is 30 degrees Celsius, then the heat flow rate through the first heat exchanger is the difference between 500 multiplied by 70 and 30, which is 20,000 watts.

[0052] For the nodes corresponding to the second two-port thermal resistivity element, the update process is similar to that of the first two-port thermal resistivity element. The difference is that the heat source side temperature of the second heat exchanger comes from the adjacent node on the outlet port side of node 1, that is, the temperature at which the exhaust gas heat reaches the heat source side of the second heat exchanger after being distributed through the three-way pipe. The nodes corresponding to the second two-port thermal resistivity element read the thermal domain temperature characteristics of the adjacent nodes on the inlet side and the adjacent nodes on the outlet port side of node 1, respectively. The difference between the two temperature characteristics is multiplied by the pre-calibrated equivalent thermal conductivity value of the second heat exchanger to obtain the heat flow rate characteristics through the second heat exchanger. Under low-temperature conditions, the temperature at which the exhaust gas heat reaches the heat source side of the second heat exchanger through the three-way pipe is usually much higher than the inlet temperature after being compressed by the air compressor. Therefore, the heat flow rate is positive, and the air is heated in the second heat exchanger.

[0053] For the nodes corresponding to the fluid domain converter element, the update process involves both the fluid domain and the thermal domain. In the fluid domain, the fluid domain converter element reads the fluid domain pressure characteristics of the adjacent inlet node (i.e., the air compressor inlet side), scales it according to the pre-calibrated compression ratio of the air compressor, and obtains the fluid domain pressure characteristics of the outlet side. Specifically, the scaling operation multiplies the inlet pressure by the compression ratio to obtain the outlet pressure. For example, if the inlet absolute pressure is 101325 Pascals and the compression ratio is 2.5, then the outlet absolute pressure is 253312 Pascals. In the thermal domain, the fluid domain converter element reads the thermal domain temperature characteristics (i.e., the air compressor inlet temperature) of the adjacent inlet node, looks up the outlet temperature according to the temperature rise mapping table corresponding to the pre-calibrated compression ratio of the air compressor, and updates it to the thermal domain temperature characteristics of the fluid domain converter element. The temperature rise mapping table has been established offline in step 1, covering the operating range of inlet temperature from -40 degrees Celsius to 50 degrees Celsius and compression ratio from 1.5 to 4.0. When looking up the table, a bilinear interpolation method is used, performing linear interpolation on both the inlet temperature and compression ratio dimensions. For example, when the inlet temperature is -35 degrees Celsius and the compression ratio is 2.5, the temperature rise mapping table gives an outlet temperature of approximately a specific value within the range of -15 to -10 degrees Celsius, which depends on the adiabatic efficiency of the air compressor. In an optional implementation, the temperature rise mapping table can be replaced by an online calculation method based on the adiabatic compression relationship. That is, calculations are performed using the adiabatic index of air and the isentropic efficiency of the air compressor based on the inlet temperature (which needs to be converted to an absolute temperature scale) and the compression ratio. However, this method involves slightly more computation than the table lookup method.

[0054] At the end of each round of message passing, the convection coupling element performs a cross-domain feature update. The convection coupling element multiplies the mass flow rate characteristics of the fluid domains connected to its two ends, the temperature difference between adjacent nodes in the thermal domains, and the pre-calibrated equivalent enthalpy coefficient to obtain the convective heat transfer. To illustrate this calculation process more specifically: the convection coupling element reads the mass flow rate characteristic values ​​of the nodes connected to its fluid domain side. ,in This represents the air mass flow rate through the pipe section containing the convection coupling element, in kilograms per second; the temperature characteristic values ​​of the thermal domain nodes at both ends are read and the difference is calculated. ,in This represents the temperature difference between the upstream and downstream thermal nodes of the convection coupling element, expressed in degrees Celsius or Kelvin; the pre-calibrated equivalent enthalpy-current coefficient is read. ,in This represents the heat power transferred per unit mass flow rate of air under a unit temperature difference, measured in joules per kilogram per Kelvin. Its value comprehensively reflects the isobaric specific heat capacity of air and the heat exchanger's efficiency. Convective heat transfer. Determined by the following relationship: ;in The unit is watt. The sign of convective heat transfer is determined by the direction of fluid flow: when air flows from the high-temperature side to the low-temperature side, For positive, A positive value indicates that heat is carried away and transferred along the direction of flow; conversely, a negative value indicates that heat is carried away and transferred along the direction of flow. It is negative. The convection coupling element will The absolute values ​​of the heat flux characteristics are updated to the adjacent thermal domain nodes with corresponding signs—for nodes releasing heat, the heat flux characteristic decreases by the corresponding value; for nodes receiving heat, the heat flux characteristic increases by the corresponding value. In this embodiment, the equivalent enthalpy coefficient... The typical value range is 800 to 1200 joules per kilogram per Kelvin. When the air mass flow rate... 0.06 kg / s, temperature difference At 25 degrees Celsius, the equivalent enthalpy coefficient The convective heat transfer is 1005 joules per kilogram per Kelvin. It is 0.06 multiplied by 1005 multiplied by 25, which is approximately 1508 watts.

[0055] After completing one round of message passing, the next step is to perform pipeline flow modeling. The purpose of pipeline flow modeling is to calculate the actual air velocity and mass flow rate in each pipe segment based on the fluid domain pressure characteristics of each node in the graph neural network, thereby providing a more accurate flow input for the convection coupling calculation in the next round of message passing. Pipeline flow modeling uses the inner diameter, length, and wall roughness of each pipe segment in the intake and coolant pipelines as fixed parameters of the pipeline network. These parameters do not change after system installation. The difference in fluid domain pressure characteristics between the nodes at both ends of each pipe segment after message passing is used as the driving pressure difference of each pipe segment. Starting from the intake pipe inlet segment (i.e., the segment from the air compressor outlet to the inlet of the first heat exchanger), the air velocity and pressure drop in each pipe segment are calculated sequentially.

[0056] The air velocity within the pipe section is calculated using the following method. For each pipe section, the driving pressure difference between its two ends is known. ,in The fluid domain pressure characteristic at the upstream node of the pipe segment is subtracted from the fluid domain pressure characteristic at the downstream node, expressed in Pascals. The pipe segment's inner diameter is... The unit is meters; the length of the pipe section is... The unit is meters; the pipe wall roughness is... The unit is meters; the density of air at the current temperature is... The unit is kilograms per cubic meter; the dynamic viscosity of air at the current temperature is... The unit is Pascal-second. Air density. and dynamic viscosity All are related to the current temperature and are determined by the thermal temperature characteristics of the corresponding pipe segment node in the previous round of message transmission—when the temperature rises, the air density decreases and the viscosity increases; when the temperature falls, the air density increases and the viscosity decreases. Air density Aerodynamic viscosity is determined by the ideal gas state relation from the current temperature and pressure. Retrieved through a pre-stored temperature-viscosity lookup table.

[0057] Average air velocity within the pipe section The relationship between the driving pressure difference and the frictional resistance along the pipe section is determined by the balance between these factors. For fully developed turbulent flow within a pipe, the relationship between the pressure drop and the velocity can be described by the Darcy-Weisbach relation: ;in Darcy's coefficient of friction For the length of the pipe section, The inner diameter of the pipe. air density, This represents the average air velocity within the pipe section. Darcy's coefficient of friction. With Reynolds number relative roughness of pipe wall Relevant. Reynolds number. The definition of is: The meanings of each parameter are as described above. Because... It depends on itself (Using the Reynolds number), the Darcy-Weisbach relation described above constitutes an implicit equation that cannot be solved analytically directly. The PLC controller uses a successive substitution iterative method to solve the implicit equation: first, the flow rate value of the previous sampling period or a reasonable initial guess (such as 5 meters per second) is used as... Given initial values, calculate the corresponding Reynolds number. Then, the Darcy friction coefficient is determined by an explicit approximation of the Körbruck relation based on the Reynolds number and the relative roughness of the pipe wall. Then Substituting the Darcy-Weisbach relation, we can calculate the new... The value is repeated iteratively until the two adjacent iterations are equal. If the difference in values ​​is less than the preset velocity convergence threshold (set to 0.01 m / s in this embodiment), convergence is usually achieved in 2 to 4 iterations. In the special case where the flow inside the pipe is in a laminar state (Reynolds number less than 2300), the Darcy friction coefficient... It can be directly determined by the Reynolds number. At this time, the Darcy-Weisbach relation degenerates into the Hagen-Poiseuille relation, and the flow velocity can be directly obtained analytically.

[0058] The flow velocity of each pipe section can be obtained by solving the problem. Then, the flow rate is converted into mass flow rate. The transformation relationship is as follows: ;in Let be the internal cross-sectional area of ​​the pipe section. From the inner diameter of the pipe Sure, In this embodiment, the inner diameter of the pipe... When it is 50 mm, Approximately 0.00196 square meters; if the flow velocity 8 meters per second, air density If the mass flow rate is 1.2 kg / m³, then the mass flow rate is... It is approximately 0.019 kilograms per second.

[0059] The calculated mass flow rates of each pipe segment are written back to the fluid domain mass flow rate features of the corresponding fluid inertial element nodes in the graph neural network. Specifically, the mass flow rate of the pipe segment connecting the air compressor outlet to the first heat exchanger inlet is written back to the fourth dimension feature of the first fluid inertial element node, and the mass flow rate of the pipe segment connecting the second heat exchanger outlet to the methanol engine inlet is written back to the fourth dimension feature of the second fluid inertial element node. Simultaneously, the mass flow rate of each pipe segment needs to be transmitted to adjacent thermal domain nodes via convection coupling elements, updating the heat flow rate features of adjacent thermal domain nodes according to the aforementioned convection heat transfer calculation method, thereby providing the thermal domain input corrected by the pipeline flow model for the next round of message transmission.

[0060] The graph neural network uses the written-back node features as input to execute the next round of message passing. After message passing is completed, it performs pipeline flow modeling and writes back again, and so on, alternating iteratively. There are two termination conditions for alternating iteration: First, iteration stops when the change in thermal domain temperature feature of the node corresponding to the methanol engine intake is less than a preset convergence threshold in two consecutive iterations. In this embodiment, the convergence threshold is set to 0.5 degrees Celsius, that is, convergence is considered to be achieved when the absolute value of the difference between the predicted intake temperatures in two consecutive iterations is less than 0.5 degrees Celsius; Second, iteration is forcibly stopped when the number of iterations reaches a preset maximum number of iterations. In this embodiment, the maximum number of iterations is set to 5. The consideration for setting the maximum number of iteration rounds is that the PLC controller has limited computing resources. The time window available for iterative calculation within each sampling period is typically 50% to 70% of the sampling period. In this embodiment, the sampling period is 500 milliseconds to 2000 milliseconds. The calculation time for a single round of message passing plus pipeline flow modeling is approximately 50 to 200 milliseconds (depending on the processing power of the PLC controller). Therefore, the total calculation time for 5 iterations is within the range of 250 to 1000 milliseconds, which will not exceed the time budget of the sampling period. In actual operation, since the initial value of each sampling period is inherited from the convergence result of the previous period, the changes in operating conditions are usually gradual. Therefore, convergence can be achieved in 2 to 3 iterations within most sampling periods.

[0061] After the iteration terminates, the thermal domain temperature characteristics of the corresponding node at the methanol engine intake are read as the predicted intake temperature. This predicted intake temperature comprehensively considers the measured temperatures of each sensor, the compression and heating effect of the air compressor, the heat exchange capacity of each heat exchanger, the flow distribution in the pipeline, and the mixing effect after the two air streams merge. It is a temperature estimate obtained through self-consistent calculation coupled with the thermal and fluid domains, which has better predictive power than simply using the measured temperature of the third temperature monitoring point. This is because the measured temperature reflects the result at the current moment, while the predicted intake temperature also includes the influence of the current operating status of each device on the future trend of the intake temperature.

[0062] In an optional implementation, the iteration termination condition can also include a rate-of-change protection: if the change in the thermal temperature characteristic of the node corresponding to the methanol engine intake port in a certain iteration exceeds a preset single-cycle change limit (e.g., 10 degrees Celsius), then the change in this iteration is truncated to this limit value. This rate-of-change protection can prevent the iteration process from diverging due to sudden changes in a sensor signal or noise interference, thus improving the robustness of the algorithm in harsh electromagnetic environments.

[0063] Step 3, based on the predicted intake air temperature output in Step 2, involves the PLC controller performing a working mode switching control with hysteresis protection. The goal of this working mode switching control is to stabilize the air temperature at the methanol engine's intake port near the target intake air temperature, thereby addressing the difficulties of cold starting and the problems of poor combustion and knocking in low-temperature environments. In this embodiment, the target intake air temperature is set to 10 degrees Celsius, a reference value for the methanol engine to achieve good atomization and stable combustion under most operating conditions. In optional embodiments, the target intake air temperature can be adjusted to other values ​​within the range of 8 to 15 degrees Celsius, depending on the specific model and calibration data of the methanol engine.

[0064] The working mode switching control is divided into two stages based on whether the methanol engine has been started.

[0065] When the methanol engine is not started, the unit is in the cold start preparation stage. At this time, the air compressor is not yet working, and the second heat exchanger cannot perform its heat exchange function due to the lack of exhaust gas heat supply. The only available means of regulating the intake air temperature are the electric heater and the first heat exchanger. The PLC controller compares the predicted intake air temperature output in step 2 with the target intake air temperature of 10 degrees Celsius, and determines the start-up and shutdown timing of the electric heater or the first heat exchanger based on the preset hysteresis deviation value.

[0066] Hysteresis control is introduced to avoid frequent switching. Without hysteresis, if the predicted intake temperature fluctuates slightly around the target intake temperature (e.g., repeatedly changing between 9.8°C and 10.2°C), the PLC controller will continuously start and stop the electric heater or the first heat exchanger. This frequent switching not only increases mechanical wear and electrical shocks but may also cause larger oscillations in the intake temperature. Hysteresis control sets a deviation band around the target intake temperature, making the temperature thresholds for starting and stopping the equipment different—the starting threshold is lower than (or higher than) the stopping threshold, forming a temperature dead zone and avoiding unnecessary switching within the dead zone. In this embodiment, the preset hysteresis deviation value is set to 2 degrees Celsius.

[0067] refer to Figure 1 , Figure 1The system comprises two parts. The upper part uses time (in seconds) on the horizontal axis and the predicted intake air temperature (in degrees Celsius) on the vertical axis. The lower part also uses time on the horizontal axis and the electric heater's operating status (start and stop) on the vertical axis. The upper and lower parts share the same time axis and are strictly aligned in the time direction. This allows the system to read the current predicted intake air temperature value from the upper part at any given moment, while simultaneously reading the electric heater's operating status from the lower part. This provides a clear visual representation of the causal relationship between the predicted intake air temperature trend and the electric heater's start-up and shutdown actions. Whenever the temperature curve in the upper part reaches the start-up or stop-up threshold, the state curve in the lower part changes simultaneously. By reading both parts together, the dynamic execution process of the hysteresis control logic can be fully presented. Figure 1 The demonstration uses a cold start scenario with an ambient temperature of -15 degrees Celsius as a typical example. In the upper part, the target intake air temperature is 10 degrees Celsius, marked by a horizontal line; the start threshold is the target intake air temperature minus a preset hysteresis deviation value of 2 degrees Celsius, i.e., 8 degrees Celsius, marked by a horizontal dashed line; the stop threshold is the target intake air temperature plus the preset hysteresis deviation value of 2 degrees Celsius, i.e., 12 degrees Celsius, marked by another horizontal dashed line. The area between the start threshold and the stop threshold is the hysteresis deviation band. At zero time, the predicted intake air temperature starts from -15 degrees Celsius. At this point, the predicted intake air temperature is much lower than the start threshold of 8 degrees Celsius. The PLC controller determines that heating is needed and starts the electric heater, corresponding to the electric heater's state changing from off to on in the lower part. As the electric heater continues to work, the predicted intake air temperature gradually increases along an approximately exponential curve. After about 77 seconds, the predicted intake air temperature reaches the stop threshold of 12 degrees Celsius. The PLC controller then shuts down the electric heater, corresponding to the state changing from on to off in the lower part. After the electric heater is turned off, since the ambient temperature is still -15 degrees Celsius, the air temperature in the intake pipe gradually drops due to natural heat dissipation, and the predicted intake temperature decreases slowly along a downward curve. When the predicted intake temperature drops to the start-up threshold of 8 degrees Celsius, the PLC controller restarts the electric heater, and the temperature begins to rise again. This cycle repeats, with the predicted intake temperature fluctuating periodically within a hysteresis deviation range of 8 to 12 degrees Celsius, and the electric heater exhibiting a periodic start-stop operating mode. Figure 1 It can be clearly seen that the hysteresis control mechanism ensures a clear temperature interval between each start and stop of the electric heater, avoiding the problem of frequent on / off switching of the electric heater when the predicted intake temperature fluctuates slightly around the target value. This protects the electrical components of the electric heater and reduces the switching frequency of the control system.

[0068] When the predicted intake air temperature is lower than the target intake air temperature minus the preset hysteresis deviation value (i.e., lower than 8 degrees Celsius), the PLC controller determines that the current intake air temperature is too low and starts the electric heater to heat the intake air. The electric heater continues to work until the predicted intake air temperature reaches the target intake air temperature plus the preset hysteresis deviation value (i.e., reaches 12 degrees Celsius), at which point the electric heater is turned off. In this way, the electric heater's start-up temperature is 8 degrees Celsius and its shut-off temperature is 12 degrees Celsius, forming a 4-degree Celsius hysteresis band, effectively avoiding frequent start-stop cycles around 10 degrees Celsius. For example, when the ambient temperature is -15 degrees Celsius and the methanol engine is not started, the air temperature in the intake pipe is close to the ambient temperature, and the predicted intake air temperature is much lower than 8 degrees Celsius. The PLC controller starts the electric heater to continuously heat the air, and the intake air temperature gradually rises. After a period of time, it reaches 12 degrees Celsius, at which point the PLC controller turns off the electric heater. At this point, the methanol engine can complete a cold start at an intake air temperature of approximately 10 to 12 degrees Celsius.

[0069] When the predicted intake air temperature is higher than the target intake air temperature plus a preset hysteresis deviation value (i.e., higher than 12 degrees Celsius), the PLC controller determines that the current intake air temperature is too high and starts the first heat exchanger to cool the intake air. The first heat exchanger uses the cooling medium on the cooling side to absorb heat from the intake air, thus lowering the air temperature. The first heat exchanger continues to operate until the predicted intake air temperature reaches the target intake air temperature minus the preset hysteresis deviation value (i.e., reaches 8 degrees Celsius), at which point it shuts down. For example, when the ambient temperature is 65 degrees Celsius, the methanol engine is not started, the air temperature in the intake pipe is close to the ambient temperature, and the predicted intake air temperature is much higher than 12 degrees Celsius. The PLC controller starts the first heat exchanger to cool the intake air, gradually lowering the intake air temperature to 8 degrees Celsius before shutting it down. At this point, the methanol engine can start at an intake air temperature of approximately 8 to 10 degrees Celsius.

[0070] After the methanol engine starts, the air compressor begins operation, compressing and heating the air according to the engine's operating power. Simultaneously, the methanol engine generates exhaust heat, which is supplied to the second heat exchanger via a three-way pipe, enabling the second heat exchanger to perform heat exchange and heating. At this point, the unit has more available temperature control methods, with the PLC controller selecting different operating modes based on the ambient temperature range.

[0071] refer to Figure 2 , Figure 2 The horizontal axis represents time, in seconds, ranging from 0 to 600 seconds; the vertical axis represents intake air temperature, in degrees Celsius, ranging from -35 degrees Celsius to 65 degrees Celsius. Figure 2 The system contains four curves, corresponding to ambient temperatures of -35 degrees Celsius, -15 degrees Celsius, 25 degrees Celsius, and 65 degrees Celsius, respectively. The four curves are distinguished by four different line types: solid line, long dashed line, dotted line, and dotted line. Figure 2 The starting time of the methanol engine is marked by a vertical auxiliary line, approximately at 80 seconds. For extreme low-temperature conditions with an ambient temperature of -35 degrees Celsius, in the 0-80 second interval before engine start, the intake air temperature rises rapidly from -35 degrees Celsius as the electric heater operates, causing the temperature to increase exponentially. Near engine start, as the control mode switches from the electric heater operating alone to the air compressor and second heat exchanger working in tandem, a brief temperature drop occurs, with the intake air temperature falling to around -15 degrees Celsius. This corresponds to the intermediate temperature after the air compressor compresses and heats the ambient air. Subsequently, between 100 and 250 seconds, the second heat exchanger uses the exhaust heat from the methanol engine to continue heating the compressed air, and the intake air temperature gradually approaches the target intake air temperature of 10 degrees Celsius along another exponentially increasing curve. After 250 seconds, the system enters a steady-state operation phase, with the intake air temperature exhibiting a small sinusoidal fluctuation around 10 degrees Celsius, with a fluctuation range not exceeding ±0.5 degrees Celsius. For operating conditions with an ambient temperature of -15 degrees Celsius, the electric heater can heat the intake air temperature from -15 degrees Celsius to near the target value within 0 to 60 seconds. The transition process after the methanol engine starts is even shorter, reaching steady state in about 80 seconds. For operating conditions with an ambient temperature of 25 degrees Celsius, the initial intake air temperature is higher than the target value. The first heat exchanger provides moderate cooling, reducing the intake air temperature to around 10 degrees Celsius within about 30 seconds, and then maintaining a steady state. For operating conditions with an extreme high temperature of 65 degrees Celsius, the initial intake air temperature is 65 degrees Celsius. The first heat exchanger continuously cools the air, and the intake air temperature drops to near the target value within about 50 seconds along an exponentially decreasing curve. After the methanol engine starts, the air compressor operates, further increasing the compressed air temperature, but after cooling by the first heat exchanger, it can still be maintained near the target intake air temperature. Figure 2 It can be seen that, regardless of the ambient temperature range of -35 degrees Celsius to 65 degrees Celsius, the methanol engine thermal management unit can adjust the intake air temperature to around the target intake air temperature of 10 degrees Celsius within a limited time after startup and maintain it stably, which verifies the adaptive intake air temperature control capability of the control method in the entire ambient temperature range.

[0072] When the ambient temperature is below the preset low-temperature operating threshold, the intake air needs to be heated. The PLC controller shuts off the electric heater because the heat from the exhaust gas generated after the methanol engine starts can be recovered and reused by the second heat exchanger, eliminating the need for further electrical energy consumption for heating. The air compressor operates at the corresponding compression ratio based on the current operating power of the methanol engine, compressing and heating the incoming air. Taking an extreme low-temperature condition with an ambient temperature of -35 degrees Celsius as an example, the air temperature rises to approximately -15 degrees Celsius after compression by the air compressor. The compressed and heated air enters the second heat exchanger, which uses the heat energy supplied through the three-way pipeline to further heat the air to 10 degrees Celsius before sending it into the methanol engine. This process achieves synergy between air compression heat compensation and exhaust gas heat exchange: the air compressor is responsible for raising the air from an extremely low temperature to an intermediate temperature, while the second heat exchanger uses the engine's own exhaust heat to further heat the air from the intermediate temperature to the target temperature. The two-stage heating works in synergy, avoiding the high energy consumption of the electric heater during engine operation and fully recovering and utilizing the exhaust gas heat.

[0073] When the ambient temperature exceeds the preset high-temperature operating threshold, the air temperature rises further after compression by the air compressor, requiring cooling to meet the intake temperature requirements of the methanol engine. Under this condition, the compressed air enters the first heat exchanger for cooling. The first heat exchanger uses a cooling medium on its cooling side to absorb the heat from the compressed air, lowering its temperature. The cooled air then enters the methanol engine. Taking an extreme high-temperature condition with an ambient temperature of 65 degrees Celsius as an example, the air temperature after compression by the air compressor may rise above 100 degrees Celsius. After cooling by the first heat exchanger, the temperature drops to near the target intake temperature before being sent to the methanol engine.

[0074] After each operating mode switch, the PLC controller performs a preset minimum hold time lock, maintaining the current operating mode unchanged during the lock period. The purpose of setting the minimum hold time is to prevent frequent switching between heating and cooling modes due to small fluctuations in the predicted intake air temperature at operating boundary conditions (e.g., when the ambient temperature is just near the low-temperature operating threshold). In this embodiment, the preset minimum hold time is set to 30 seconds. After the PLC controller performs an operating mode switch (e.g., switching from electric heater heating mode to air compressor and second heat exchanger co-heating mode), for the subsequent 30 seconds, even if the predicted intake air temperature calculated in step 2 meets the conditions for switching back to the previous mode, the PLC controller still maintains the current co-heating mode unchanged. After the lock time expires, the PLC controller resumes normal mode determination logic. In an optional embodiment, the preset minimum hold time can be adjusted to other values ​​within the range of 15 to 120 seconds according to the thermal inertia characteristics of the unit; units with higher thermal inertia are more suitable for using a longer minimum hold time.

[0075] The PLC controller continuously executes step 2 (alternating coupling iterative solution) and step 3 (operating mode determination) in a fixed sampling cycle. At the beginning of each sampling cycle, the PLC controller first reads the latest real-time data from the temperature sensors and the air compressor's pressure signal interface, assigns the data to the feature vector of the corresponding node in the graph neural network, and then executes step 2 (alternating coupling iteration) until convergence or the maximum number of iterations is reached. It then reads the predicted intake air temperature, and finally executes step 3 (operating mode determination) and outputs control commands. This cycle continues continuously, enabling adaptive intake air temperature control of the methanol engine thermal management unit across the entire ambient temperature range from -35 degrees Celsius to 65 degrees Celsius.

[0076] In an optional implementation, the PLC controller can also incorporate an ambient temperature trend prediction function in the operating mode switching determination step 3. Specifically, the PLC controller saves ambient temperature readings for the most recent sampling periods (e.g., the last 20 periods), calculates the ambient temperature trend (rising or falling), and if the trend indicates that the ambient temperature is about to cross a certain operating threshold, the PLC controller can pre-switch the operating mode, ensuring the heat exchange equipment is ready before the ambient temperature actually crosses the threshold, thereby reducing transient deviations in the intake air temperature. This prediction function is particularly advantageous for applications with drastic ambient temperature changes (e.g., mine ventilation environments, desert environments with large diurnal temperature variations).

[0077] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Those skilled in the art can omit, substitute, and modify the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, combining the above method steps to perform substantially the same function and achieve substantially the same result according to substantially the same method falls within the scope of the present invention. Therefore, the scope of the present invention is defined only by the appended claims.

Claims

1. A control method for a methanol engine thermal management unit based on compressed air heat compensation and exhaust gas heat exchange, wherein the methanol engine thermal management unit includes a methanol engine, an air compressor, an electric heater, a first heat exchanger, a second heat exchanger, a three-way pipeline, and a PLC controller. The PLC controller is connected to temperature sensors located at a first temperature monitoring point, a second temperature monitoring point, and a third temperature monitoring point. The first temperature monitoring point is located at the outlet side of the first heat exchanger, the second temperature monitoring point is located at the outlet side of the second heat exchanger, and the third temperature monitoring point is located at the inlet of the methanol engine. The PLC controller is connected to the pressure signal interface of the air compressor. A cooling medium inlet temperature sensor is provided on the cooling side of the first heat exchanger, and the cooling medium inlet temperature sensor is connected to the PLC controller. The method is characterized in that... The control method includes the following steps: Step 1: The heat transfer process and fluid flow process in the methanol engine thermal management unit are uniformly modeled as a multi-physics domain bond graph model that includes thermal domain and fluid domain. Each element in the multi-physics domain bond graph model is mapped to a node of a graph neural network, each energy bond is mapped to a directed edge of the graph neural network, and feature vectors containing thermal domain features and fluid domain features are set for each node. Step 2: Message passing is performed on the graph neural network. Each node updates its own feature vector according to the update rules of the corresponding element type based on the features passed from the adjacent nodes through the directed edges. After completing one round of message passing, pipeline flow modeling is performed. The difference in fluid domain pressure characteristics of the nodes at both ends of each pipe segment is used as the driving pressure difference to solve for the flow velocity and mass flow rate of each pipe segment. The mass flow rate is written back to the feature vector of the corresponding node in the graph neural network. Message passing and pipeline flow modeling are performed alternately until convergence or the preset maximum number of iterations is reached. The thermal domain temperature characteristics of the corresponding node of the methanol engine intake are read as the predicted intake temperature. Step 3: Based on the comparison between the predicted intake air temperature and the target intake air temperature, the PLC controller controls the electric heater or the first heat exchanger to work when the methanol engine is not started, and controls the air compressor to work in conjunction with the second heat exchanger or controls the first heat exchanger to work after the methanol engine is started, so that the intake air temperature reaches the target intake air temperature. In the thermal domain, the methanol engine is set as a thermal capacitive element, the electric heater is set as a controllable thermal potential source element, the first heat exchanger is set as a first two-port thermal resistive element, the second heat exchanger is set as a second two-port thermal resistive element, and the cooling medium inlet on the cooling side of the first heat exchanger is set as the thermal domain boundary node. In the fluid domain, the air compressor is set as a fluid domain converter element. The pipe section in the intake pipeline connecting the air compressor outlet to the inlet of the first heat exchanger is set as a series combination of the first fluid inertial element and the first fluid resistive element. The pipe section connecting the outlet of the second heat exchanger to the intake port of the methanol engine is set as a series combination of the second fluid inertial element and the second fluid resistive element. The three-way pipe is set as node 1. The inlet port of node 1 receives the exhaust heat generated by the methanol engine, and the two outlet ports of node 1 are connected to the second heat exchanger and the external heat dissipation pipe, respectively. The junction of the outlet of the first heat exchanger and the outlet of the second heat exchanger in the intake pipeline is set as node 0.

2. The control method according to claim 1, characterized in that, Each node is assigned a 4-dimensional feature vector: the first dimension is the thermal domain temperature feature, the second dimension is the thermal domain heat flow rate feature, the third dimension is the fluid domain pressure feature, and the fourth dimension is the fluid domain mass flow rate feature. The thermal domain and the fluid domain are connected by a convection coupling element, which characterizes the heat transport process as the fluid flows through the heat exchanger. The PLC controller reads the real-time temperature value of the temperature sensor at the first temperature monitoring point and assigns it to the thermal domain temperature characteristic of the corresponding node of the first dual-port thermal resistive element. It reads the real-time temperature value of the temperature sensor at the second temperature monitoring point and assigns it to the thermal domain temperature characteristic of the corresponding node of the second dual-port thermal resistive element. It reads the real-time temperature value of the temperature sensor at the third temperature monitoring point and assigns it to the thermal domain temperature characteristic of the corresponding node of the methanol engine air inlet. It assigns the air compressor outlet pressure to the fluid domain pressure characteristic of the corresponding node of the fluid domain converter element and assigns the reading of the cooling medium inlet temperature sensor to the thermal domain temperature characteristic of the thermal domain boundary node.

3. The control method according to claim 2, characterized in that, For the node corresponding to node 1, the heat flux characteristics of the heat domain transmitted from all adjacent nodes with incoming edges are summed with a sign after being assigned a positive sign in the direction of incoming edges and a negative sign in the direction of outgoing edges. The summation result is then distributed to each downstream adjacent node with outgoing edges as the update increment of the heat flux characteristics according to the pre-calibrated flow splitting ratio coefficient. The flow splitting ratio coefficient is determined by the ratio of the equivalent thermal conductivity of the second heat exchanger branch to that of the external heat dissipation pipe.

4. The control method according to claim 3, characterized in that, For the node corresponding to node 0, the fluid domain mass flow characteristics transmitted from all adjacent nodes with incoming edges are assigned positive signs in the direction of incoming edges and negative signs in the direction of outgoing edges, and then summed with a sign to ensure continuity. At the same time, the thermal domain temperature characteristics and fluid domain mass flow characteristics of each adjacent node with incoming edges are read. The mass flow characteristics of each incoming branch are multiplied by the corresponding thermal domain temperature characteristics one by one, summed, and then divided by the sum of the mass flow characteristics of all incoming branches to obtain the mixed temperature, which is then updated to the thermal domain temperature characteristics of node 0 and transmitted to the downstream adjacent nodes along the outgoing edges.

5. The control method according to claim 4, characterized in that, For the node corresponding to the first two-port thermally resistive element, the thermal temperature characteristics of the adjacent node on the inlet side and the thermal boundary node of the first two-port thermally resistive element are read respectively. The difference between the two temperature characteristics is multiplied by the equivalent thermal conductivity value pre-calibrated by the first heat exchanger to obtain the heat flow rate characteristics through the first heat exchanger. For the node corresponding to the second two-port thermally resistive element, the thermal temperature characteristics of the adjacent node on the inlet side and the adjacent node on the outlet side of node 1 are read respectively. The difference between the two temperature characteristics is multiplied by the equivalent thermal conductivity value pre-calibrated by the second heat exchanger to obtain the heat flow rate characteristics through the second heat exchanger.

6. The control method according to claim 5, characterized in that, For the node corresponding to the fluid domain converter element, read the fluid domain pressure characteristics of the adjacent nodes on the inbound side, scale them according to the compression ratio pre-calibrated by the air compressor, and update them to the fluid domain pressure characteristics on the outlet side. At the same time, read the thermal domain temperature characteristics of the adjacent nodes on the inbound side, look up the outlet temperature according to the temperature rise mapping table corresponding to the compression ratio pre-calibrated by the air compressor, and update it to the thermal domain temperature characteristics. At the end of each round of message transmission, the convection coupling element multiplies the mass flow rate characteristic of the fluid domain connected to both ends of the convection coupling element, the difference in thermal domain temperature characteristics between adjacent nodes, and the pre-calibrated equivalent enthalpy coefficient to obtain the convective heat transfer. The sign of the convective heat transfer is determined by the direction of fluid flow.

7. The control method according to claim 1, characterized in that, The convergence criterion is that the change in the thermal temperature characteristic of the node corresponding to the methanol engine intake port is less than the preset convergence threshold in two consecutive iterations.

8. The control method according to claim 1, characterized in that, In step 3, the PLC controller starts the corresponding heating device when the predicted intake air temperature is lower than the target intake air temperature minus the preset hysteresis deviation value, and shuts down the corresponding heating device when the predicted intake air temperature reaches the target intake air temperature plus the preset hysteresis deviation value. After the methanol engine starts and the ambient temperature is lower than the preset low temperature operating threshold, the air compressor compresses and heats the air. The compressed and heated air enters the second heat exchanger, which uses the heat energy supplied by the exhaust gas generated by the methanol engine through the three-way pipe to continue heating the air before sending it into the methanol engine. After the methanol engine starts and the ambient temperature is higher than the preset high temperature operating threshold, the air is compressed by the air compressor and enters the first heat exchanger for heat exchange and cooling. The cooled air then enters the methanol engine. After each working mode switch, the PLC controller performs a preset minimum hold time lock.

Images