System and method for stepwise recycling of workpiece after-forging residual heat

Abstract

The present application belongs to the field of waste heat recovery, and particularly relates to a system and method for stepwise recovery and utilization of post-forging workpiece waste heat. The system is provided with a heat-insulating conveying channel at the discharge port of a press, and the inner wall of the channel is provided with radiation heat exchange pipes which are divided into a high-temperature radiation section and a low-temperature convection section. A heating loop is connected between the outlet of the high-temperature radiation section and a workshop heating terminal, and a preheating loop uses heating return water to preheat workpieces in the low-temperature convection section. A sensing module collects product variety codes, stamping cycles and outlet water temperatures. A control module uses a lower limit of a cooling rate and a water temperature range of 70-90 DEG C as constraints, and a stamping cycle as a boundary, and solves conveying speed and circulation flow rate by using a particle swarm algorithm. An execution module synchronously adjusts motor frequency and pump rotating speed. When the proportion of feasible solutions in iteration is lower than a threshold value, a boundary relaxation unit corrects an upper limit threshold value of the conveying speed and initializes a particle swarm, so that the heating water temperature meets the standard and the quality of forgings is not damaged.

Description

Technical Field

[0001] This invention belongs to the field of waste heat recovery, and particularly relates to a system and method for the cascade recovery and utilization of waste heat from forged workpieces. Background Technology

[0002] In hot forging, the temperature of the workpiece after final forging is usually as high as 700-850℃, carrying a large amount of high-grade sensible heat. During natural cooling, almost all of it is dissipated into the workshop environment, which not only causes huge energy waste but also exacerbates thermal pollution in the workshop. At the same time, forging workshops in northern regions need to consume a lot of extra energy for heating in winter. The waste of residual heat from high-temperature forgings and the high energy consumption for heating create a prominent contradiction. In existing technologies, a common practice is to directly immerse the hot forgings after final forging into a water tank or spray device for forced cooling to obtain hot water. However, the rapid cooling process severely affects the microstructure and mechanical properties of the forgings, and the resulting hot water temperature is generally below 40°C, far from the 70-90°C required for floor radiant heating or radiator heating. Another approach integrates water jackets or air-cooled heat exchangers on the conveyor line. Heat transfer mainly relies on local contact heat conduction, which is insufficient for capturing the large amount of high-temperature heat lost from the forging surface in the form of radiation. The outlet water temperature is difficult to stably raise to the heating requirements, and the system cannot automatically adjust the conveyor speed and heat exchange according to the type and size of the workpiece and the forging cycle, often resulting in insufficient recovery or disruption to the normal production cycle. There are also methods using radiant heating... While the heat exchange tube scheme for recovering radiant heat from forgings improves the radiant heat capture capability, it does not construct a cascaded heat energy utilization architecture, nor does it involve using the heating return water circulation for preheating to achieve deeper efficiency improvement. The overall energy efficiency of the system remains limited. Therefore, the technical problem that the existing technology urgently needs to solve is: how to efficiently and cascade the recovery of residual heat from forging workpieces at 700-850℃ without interfering with the forging cycle, directly produce high-temperature hot water at 70-90℃ for heating, and use the heating return water circulation for preheating to achieve deep cascaded energy utilization, while adaptively matching different workpiece types and forging cycles, overcoming the shortcomings of the existing technology such as low outlet water temperature, impact on forging microstructure, insufficient radiant heat recovery, and inability to dynamically adjust the system. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention proposes a system and method for the tiered recovery and utilization of waste heat from forged workpieces. The system features an insulated conveying channel at the press outlet, with radiant heat exchange tubes lining the inner wall, divided into a high-temperature radiant section and a low-temperature convection section. A heating circuit connects the outlet of the high-temperature radiant section to the workshop heating terminal inlet, while a preheating circuit utilizes waste heat from the heating return water to preheat the workpieces in the low-temperature convection section. A sensing module collects workpiece type codes, press stamping cycles, and outlet water temperatures. A control module uses the lower limit of the allowable cooling rate and the target water temperature range as constraints, with the stamping cycle as the boundary for adjusting the conveying speed, and solves for the optimal conveying speed and circulation flow rate using a particle swarm optimization algorithm. An execution module synchronously adjusts the drive motor frequency and the circulation pump speed. When multiple constraints cause the proportion of feasible solutions in the algorithm to be too low, a boundary relaxation unit dynamically corrects the upper limit threshold of the conveying speed and reinitializes the particle swarm optimization, ensuring that the system stably outputs conveying and heat exchange parameters that meet the heating water temperature requirements without compromising the quality of the forgings.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A cascaded heat recovery and utilization system for forged workpieces includes:

[0006] A heat-insulated conveying channel is located at the outlet of the press, and its inner wall is lined with radiant heat exchange tubes. The radiant heat exchange tubes are divided into a high-temperature radiation section and a low-temperature convection section along the workpiece conveying direction.

[0007] The heating circuit connects the outlet of the high-temperature radiant section to the inlet of the workshop heating terminal;

[0008] The preheating circuit connects the return water outlet of the workshop heating terminal with the inlet of the low-temperature convection section, and uses the waste heat of the heating return water to preheat the workpiece flowing through the low-temperature convection section.

[0009] The sensing module is configured to: collect the current workpiece's type code, the press's stamping cycle, and the outlet water temperature of the radiant heat exchanger tube;

[0010] The control module is configured as follows:

[0011] The lower limit of the allowable cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating loop circulation flow rate that satisfy the constraints are solved iteratively by the particle swarm algorithm.

[0012] The execution module is configured to synchronously adjust the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulating pump of the heating circuit according to the conveying speed and the circulation flow rate.

[0013] Specifically, the control module is also equipped with a feasible solution ratio monitoring unit and a boundary relaxation unit: the feasible solution ratio monitoring unit, in each iteration of the particle swarm algorithm, counts the proportion of particles whose outlet water temperature falls within the target range and whose actual workpiece cooling rate is greater than or equal to the lower limit of the allowable cooling rate, as the proportion of the total number of particles, and uses this as the feasible solution ratio; the boundary relaxation unit, when the number of iterations reaches a first preset value and the feasible solution ratio is lower than a preset feasible solution ratio threshold, uses the product of the upper limit threshold of the conveying speed and the preset relaxation coefficient as the corrected adjustment boundary, and triggers the particle swarm to re-initialize, until the feasible solution ratio is greater than or equal to the preset feasible solution ratio threshold, and then outputs the target value of the conveying speed and the target value of the circulating flow rate; the target range corresponds to [70℃, 90℃];

[0014] The upper limit threshold of the conveying speed is determined by the ratio of the stamping cycle value to the minimum allowable dwell time of the workpiece in the heat-insulated conveying channel.

[0015] Specifically, the sensing module is further configured to: collect the surface temperature of the workpiece when it enters the inlet of the heat preservation conveying channel as an initial temperature value; the control module is further configured to: use the initial temperature value and the variety code as joint inputs, query a preset initial temperature-cooling rate correction mapping table, output the corrected allowable cooling rate lower limit value corresponding to the initial temperature value, and use the corrected allowable cooling rate lower limit value to replace the allowable cooling rate lower limit corresponding to the variety code as the constraint condition to participate in the particle swarm algorithm iterative solution.

[0016] Specifically, the sensing module is further configured to: collect the water temperature at the inlet of the high-temperature radiation section as the inlet water temperature; the control module is further configured to: calculate the difference between the outlet water temperature and the inlet water temperature as the current heat exchange temperature difference, use the ratio of the pre-stored internal volume parameter of the radiation heat exchange tube to the current circulating flow rate of the heating loop as the transport delay time, and during the iterative solution of the particle swarm optimization algorithm, use the predicted change in outlet water temperature corresponding to the transport delay time as the feedforward compensation amount, add it to the outlet water temperature, and then determine the constraint conditions.

[0017] Specifically, the preheating circuit is also equipped with a preheating circuit proportional regulating valve; the sensing module is also configured to: collect the surface temperature of the workpiece when it leaves the low-temperature convection section, as the preheated temperature; the control module is also configured to: when the particle swarm algorithm iteratively solves the problem, take the preheated temperature being greater than or equal to a preset lower limit of the preheating temperature as a new constraint condition, which together with the constraint condition constitutes a joint constraint condition; the particle swarm algorithm simultaneously outputs the target value of the conveying speed, the target value of the circulating flow rate of the heating circuit, and the target value of the circulating flow rate of the preheating circuit; the execution module is also configured to: adjust the opening of the preheating circuit proportional regulating valve according to the target value of the circulating flow rate of the preheating circuit.

[0018] Specifically, the sensing module is further configured to: collect the water temperature at the return water outlet of the workshop heating terminal as the heating return water temperature; the control module is further configured to: calculate the real-time heat load of the heating terminal by multiplying the temperature difference between the heating return water temperature and the outlet water temperature and the circulating flow rate of the heating loop; and when the particle swarm algorithm iteratively solves the problem, replace the fitness function with: when the outlet water temperature is within the target range, minimize the absolute value of the difference between the waste heat recovery power and the real-time heat load.

[0019] Specifically, the stamping cycle value and the outlet water temperature are continuously monitored. When the k consecutively collected stamping cycle values ​​all exceed the preset normal range of the stamping cycle, or when the rate of change of the outlet water temperature within the preset time window exceeds the preset water temperature change rate threshold, it is determined to be an abnormal working condition.

[0020] When an abnormal operating condition is determined, the particle swarm algorithm iterative solution is stopped, and a preset safe value for the conveying speed and a preset safe value for the circulating flow rate of the heating circuit are output.

[0021] The execution module is further configured to adjust the frequency of the drive motor of the insulation conveying channel and the speed of the circulation pump of the heating circuit according to the preset safe value of the conveying speed and the preset safe value of the circulating flow rate of the heating circuit.

[0022] Specifically, the control module is also configured to: calculate the variance of the fitness values ​​of all particles in the current particle swarm after each iteration of the particle swarm algorithm, as an indicator of population diversity;

[0023] When the population diversity index is lower than the preset diversity threshold and the current iteration number has not reached the preset maximum iteration number, the inertial weight of the particle swarm optimization algorithm is multiplied by a preset perturbation factor and the iteration continues until the population diversity index recovers to a level greater than or equal to the preset diversity threshold. Then, the inertial weight is restored to its original set value.

[0024] Specifically, the control module is also configured to: when the outlet water temperature falls into the target range for m consecutive times and the target values ​​of the conveying speed and the circulating flow rate output by the particle swarm algorithm are both within the preset steady-state fluctuation range, record the initial temperature value and the current lower limit of the corrected allowable cooling rate participating in the iteration to form a correction sample point;

[0025] When the number of accumulated correction sample points reaches a preset correction sample number threshold, the updated initial temperature-cooling rate correction mapping table is refitted with all correction sample points and replaced with the preset initial temperature-cooling rate correction mapping table.

[0026] Methods for the graded recovery and utilization of residual heat from forged workpieces include:

[0027] Collect the current workpiece's type code, press stamping cycle, and radiant heat exchanger tube outlet water temperature;

[0028] The lower limit of the allowable cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating loop circulation flow rate that satisfy the constraints are solved iteratively by the particle swarm algorithm.

[0029] Based on the conveying speed and circulation flow rate, the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulating pump of the heating circuit are adjusted synchronously.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] This invention constructs a tiered heat exchange architecture consisting of a high-temperature radiation section and a low-temperature convection section. This architecture directly utilizes the high-grade radiant heat from the forged workpiece (700–850°C) to produce high-temperature heating water (70–90°C). Simultaneously, it leverages the low-temperature waste heat from the heating return water to preheat the workpiece, achieving tiered and in-depth utilization of thermal energy and significantly improving the overall system energy efficiency. At the control level, by introducing the lower limit of the allowable cooling rate corresponding to the workpiece type code and the target range of the outlet water temperature as constraints into the particle swarm optimization algorithm, the waste heat recovery process is always strictly limited within a process window that does not damage the metallographic structure of the forging. This fundamentally solves the problem of traditional waste heat recovery methods affecting forging quality. In particular, by setting up a feasible solution ratio monitoring unit and a boundary relaxation unit, when the ratio of feasible solutions in the algorithm is too low within a preset number of iterations due to multiple constraint coupling, the system automatically multiplies the upper limit threshold of the conveying speed by the relaxation coefficient to broaden the solution space and reinitialize the particle swarm. This effectively avoids the dilemma of the particle swarm algorithm getting stuck in local optima or having no solution due to excessive constraints. It ensures that the system can always output conveying speed and circulation flow parameters that meet the heating water temperature requirements without compromising the cooling quality of the forgings under complex and variable working conditions. It has the triple advantages of process compatibility, heat recovery efficiency and operational robustness. Attached Figure Description

[0032] Figure 1 This is a system architecture diagram of the waste heat recovery and utilization system for forged workpieces in Embodiment 1 of the present invention;

[0033] Figure 2 This is a flowchart of the method for graded recovery and utilization of residual heat from forged workpieces in Embodiment 3 of the present invention; Detailed Implementation

[0034] Example 1

[0035] Please see Figure 1One embodiment of the present invention provides a cascade recovery and utilization system for waste heat from forged workpieces, comprising:

[0036] A heat-insulated conveying channel is located at the outlet of the press, and its inner wall is lined with radiant heat exchange tubes. The radiant heat exchange tubes are divided into a high-temperature radiation section and a low-temperature convection section along the workpiece conveying direction.

[0037] The heating circuit connects the outlet of the high-temperature radiant section to the inlet of the workshop heating terminal;

[0038] The preheating circuit connects the return water outlet of the workshop heating terminal with the inlet of the low-temperature convection section, and uses the waste heat of the heating return water to preheat the workpiece flowing through the low-temperature convection section.

[0039] The sensing module is configured to: collect the current workpiece's type code, the press's stamping cycle, and the outlet water temperature of the radiant heat exchanger tube;

[0040] The control module is configured as follows:

[0041] The allowable lower limit of the cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating loop circulation flow rate that satisfy the constraints are iteratively solved using the particle swarm optimization algorithm. In this embodiment, the formulaic expression of the constraints, adjustment boundary and fitness function is as follows:

[0042] Constraint 1 is that the actual cooling rate of the workpiece is not less than the lower limit of the allowable cooling rate corresponding to the product code, expressed as v_cool≥v_cool_min, where v_cool is the actual cooling rate of the workpiece in the heat preservation conveying channel, in ℃ / s, and v_cool_min is the lower limit of the allowable cooling rate obtained by querying the product code. If the system adopts the initial temperature correction mechanism, v_cool_min is replaced by the corrected lower limit of the allowable cooling rate.

[0043] The second constraint is that the outlet water temperature of the radiant heat exchanger tube falls within the target range, which is expressed as 70℃≤T_out≤90℃, where T_out is the outlet water temperature of the radiant heat exchanger tube collected by the sensor module in ℃. If the system adopts a feedforward compensation mechanism, then T_out is replaced by the predicted outlet water temperature after superimposing the transport delay compensation amount.

[0044] In this embodiment, both constraint one and constraint two must be satisfied simultaneously for the particle to be marked as a feasible solution.

[0045] The adjustment boundary is manifested as an upper limit constraint on the conveying speed, expressed as V≤V_max, where V is the conveying speed variable in m / s, and V_max is the upper limit threshold of the conveying speed, determined by V_max=T_cycle / t_min, where T_cycle is the stamping cycle value of the press in seconds, and t_min is the minimum allowable dwell time of the workpiece in the heat-insulated conveying channel in seconds.

[0046] The fitness function represents the maximization of waste heat recovery power, expressed as F = c × ρ × Q × T_out, where F is the fitness value in kW, c is the specific heat capacity of the circulating medium water (taken as 4.18 kJ / (kg·℃), ρ is the density of the circulating medium water (taken as 1000 kg / m³), Q is the circulating flow rate of the heating loop in m³ / s, and T_out is the outlet water temperature of the radiant heat exchanger tube in ℃.

[0047] The execution module is configured to synchronously adjust the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulating pump of the heating circuit according to the conveying speed and the circulation flow rate.

[0048] In this embodiment, the product code is issued by the production execution system based on the material and specifications of the currently processed workpiece. The corresponding lower limit of the allowable cooling rate is determined according to the critical cooling rate in the continuous cooling transformation curve of the material to avoid the formation of non-target structures. For example, to avoid the formation of bainite in 42CrMo steel forgings, the cooling rate needs to be controlled above a certain threshold. The target range for the outlet water temperature is set to [70℃, 90℃]. The lower limit of 70℃ is determined according to the water supply temperature requirements for floor radiant heating and radiators in the "Code for Design of Heating, Ventilation and Air Conditioning of Civil Buildings" (GB50736-2012), and the upper limit of 90℃ is determined by atmospheric pressure heating. The upper limit of the safe operating temperature of the water system and the thermodynamic critical temperature for preventing scaling on the surface of the radiant heat exchange tubes are determined; the stamping cycle serves as the adjustment boundary for the conveying speed, and its value is determined by the time interval between two adjacent stamping actions collected in real time by the cycle sensor installed on the press slider or crankshaft end, ensuring that the conveying cycle of the workpiece in the insulated conveying channel is strictly matched with the press discharge cycle; the waste heat recovery power serves as a fitness function, and its value is determined by the heat power value after converting the product of the outlet water temperature and the circulating flow rate of the heating circuit into specific heat capacity and density, directly reflecting the effective heat recovered by the system per unit time that can be used by the heating terminal.

[0049] In this embodiment, the high-temperature forging (700-850℃) first undergoes high-intensity radiant heat exchange with the radiant heat exchange tubes laid on the inner wall within the insulated conveying channel. The high-temperature radiant section absorbs the highest-grade sensible heat released by the workpiece, and this high-grade heat energy is directly transferred to the workshop heating terminal through the heating circuit, producing 70-90℃ high-temperature hot water that meets the heating requirements, completing the first stage of high-grade heat recovery. Although the temperature of the return water after the heat is released by the heating terminal has decreased, it still carries a considerable amount of low-grade waste heat. This return water is not directly discharged or returned to the cooling tower, but is preheated and then returned to the cooling tower. The low-temperature convection section of the radiant heat exchanger tube is introduced into the pipeline. The low-grade waste heat of the heating return water is used to preheat the forged workpieces that flow through it. On the one hand, this increases the initial temperature of the workpieces before they enter the high-temperature radiant section, reducing the consumption of high-grade heat. On the other hand, it allows the heating return water to be recirculated or discharged after completing the two-stage heat exchange. This achieves a closed-loop cycle in which the temperature of the heat medium decreases step by step and the heat is utilized in a tiered manner, from "high-temperature radiant heat of the workpiece → high-temperature hot water of the heating system → low-temperature waste heat of the return water → preheating of the workpiece". Thus, two-stage heat energy recovery of high-grade heating and low-grade preheating is completed within a single transport channel.

[0050] Example 2

[0051] In actual forging production, frequent changes in workpiece types and fluctuations in the stamping cycle of the press cause strong coupling or even conflict between the constraints faced by the particle swarm optimization algorithm (lower limit of cooling rate, target water temperature range) and the adjustment boundary (upper limit of conveying speed determined by the stamping cycle). This can lead to iterative solutions getting trapped in local optima or generating infeasible solutions, resulting in the system's inability to output effective control parameters in real time, thereby affecting the continuity of heating and the quality of forgings. To address this, this embodiment further introduces feasible solution proportion monitoring and boundary relaxation mechanisms into the control module.

[0052] It should be further explained that, in this embodiment, the control module is also configured with a feasible solution ratio monitoring unit and a boundary relaxation unit: the feasible solution ratio monitoring unit, in each iteration of the particle swarm algorithm, counts the proportion of particles whose outlet water temperature falls within the target range and whose actual workpiece cooling rate is greater than or equal to the lower limit of the allowable cooling rate, as the proportion of the total number of particles, and uses this as the feasible solution ratio; the boundary relaxation unit, when the number of iterations reaches a first preset value and the feasible solution ratio is lower than a preset feasible solution ratio threshold, uses the product of the upper limit threshold of the conveying speed and the preset relaxation coefficient as the corrected adjustment boundary, and triggers the particle swarm to re-initialize, until the feasible solution ratio is greater than or equal to the preset feasible solution ratio threshold, and then outputs the target value of the conveying speed and the target value of the circulating flow rate; the target range corresponds to [70℃, 90℃]; wherein, the upper limit threshold of the conveying speed is determined by the ratio of the stamping cycle value to the minimum allowable residence time of the workpiece in the heat preservation conveying channel.

[0053] In this embodiment, the first preset value is selected based on 60% to 80% of the total number of iterations of the particle swarm optimization algorithm, in order to determine if the constraints are too tight in the later stages of the algorithm. Those skilled in the art will understand that this value range is set based on the following: when the number of iterations is below 60%, the algorithm is still in the global exploration stage, and the particle swarm has not yet fully converged. At this time, the statistical proportion of feasible solutions cannot reliably reflect the true tightness of the constraints. However, if boundary relaxation is triggered after exceeding 80%, the algorithm may reach the maximum number of iterations before reaching convergence, thus outputting a suboptimal solution. Therefore, those skilled in the art can also use other percentage values ​​based on the convergence speed of the actual system. For example, when the particle swarm size is large or the search space dimension is high, the first preset value can be set to 85% of the total number of iterations to reserve more convergence margin.

[0054] The second preset value is set to 20%–30%. A value below this indicates a severe shortage of solutions satisfying the process constraints in the current search space. Those skilled in the art will understand that this range is set based on the following: when the proportion of feasible solutions is below 20%, most particles in the particle swarm violate the constraints, making it difficult for the algorithm to effectively discover feasible solution regions in subsequent iterations through its own evolutionary mechanism. A boundary relaxation mechanism is needed to expand the search space. Conversely, if the threshold is set too high (e.g., above 40%), it may trigger the relaxation mechanism too frequently due to excessive sensitivity, leading to unnecessary and repeated adjustments to the upper limit of the conveying speed, which in turn causes system fluctuations. Therefore, those skilled in the art can adjust this value according to the sensitivity of the forging material to the cooling rate constraint: for alloy steel forgings with high cooling rate sensitivity, the second preset value can be set to 25% to maintain stricter constraints; for ordinary carbon steel forgings with low cooling rate sensitivity, the second preset value can be set to 30% to obtain more solution opportunities. The preset relaxation coefficient is set to 1.1–1.5 to gradually relax the upper limit of the conveying speed, avoiding sudden and significant expansion of the boundary that could cause system oscillations. The minimum allowable residence time is calculated based on the length of the high-temperature radiation section and the minimum radiative heat transfer required by the workpiece, ensuring that the workpiece can release sufficient heat to the heating circuit within the shortest conveying time to raise the outlet water temperature to above 70°C. When reinitializing the particle swarm, the initial position of the particles is randomly generated between the corrected upper and lower limits of the conveying speed, while the initial speed remains within the original range to preserve the search experience from the early stages of the algorithm. In this embodiment, when reinitializing the particle swarm, the initial position of the heating circuit circulation flow rate is randomly generated between a preset lower limit and a preset upper limit. The preset lower limit is determined based on the minimum flow rate required for the heating circuit to maintain a minimum antifreeze cycle, and the preset upper limit is determined based on the rated maximum flow rate of the circulation pump.

[0055] As a specific example, the total number of particle swarm iterations is set to 100 generations, the first preset value is 70 generations, the preset feasible solution ratio threshold is 25%, the preset relaxation coefficient is 1.2, the press stamping cycle is 10 seconds, and the minimum allowable residence time in the heat preservation conveying channel is 30 seconds. Then, the initial upper limit threshold of the conveying speed is the ratio of the effective channel length to the minimum residence time. For example, when the effective channel length is 9 meters, the upper limit is 0.3 meters / second. When iterating to the 70th generation, the feasible solution ratio is only 15%. The boundary relaxation unit corrects the upper limit threshold of the conveying speed to 0.36 meters / second and reinitializes the particle swarm to continue optimization until the feasible solution ratio recovers to more than 25%. The final output is such as the target value of conveying speed of 0.32 meters / second and the target value of circulation flow of 1.5 cubic meters / hour. The frequency of the drive motor and the speed of the circulation pump are adjusted synchronously to ensure that the system can stably provide heating hot water of 70-90℃ while ensuring the cooling quality of the forging.

[0056] In actual forging production, even for workpieces with the same product code, the actual temperature after final forging can vary within the range of 700–850°C due to fluctuations in furnace temperature, die temperature drop, and differences in operating cycle time. If the single allowable lower limit of cooling rate corresponding to the product code is always used as the constraint, when the actual initial temperature of the workpiece is high, the allowable lower limit of cooling rate will remain unchanged, which may lead to the workpiece cooling too quickly and producing a non-target microstructure; when the initial temperature is low, the constraint is too lenient, resulting in insufficient residual heat recovery. Therefore, this embodiment introduces a dynamic correction mechanism for the lower limit of cooling rate based on the initial temperature of the workpiece in the sensing and control modules.

[0057] It should be further explained that, in this embodiment, the sensing module is also configured to: collect the surface temperature of the workpiece when it enters the inlet of the heat preservation conveying channel, as the initial temperature value; the control module is also configured to: use the initial temperature value and the variety code as joint input, query a preset initial temperature-cooling rate correction mapping table, output the corrected allowable cooling rate lower limit value corresponding to the initial temperature value, and use the corrected allowable cooling rate lower limit value to replace the allowable cooling rate lower limit corresponding to the variety code as the constraint condition to participate in the particle swarm algorithm iterative solution.

[0058] In this embodiment, the sensor for collecting the surface temperature of the workpiece when it enters the insulated conveyor channel is an infrared thermometer, installed at a preset height above the workpiece's passage plane above the inlet of the insulated conveyor channel. Its temperature measurement response time is less than 1 / 5 of the time the workpiece stays at the inlet, ensuring that the initial temperature value is collected before the workpiece enters the high-temperature radiation section. The preset initial temperature-cooling rate correction mapping table is based on the continuous cooling transformation curve of the workpiece material corresponding to the product code at different initial temperatures. An initial temperature point is selected at preset temperature intervals, and the maximum allowable cooling rate corresponding to the material avoiding the formation of non-target structures is determined at each initial temperature point. The one-to-one correspondence between the initial temperature value and the maximum allowable cooling rate is pre-stored in the control module as the initial temperature-cooling rate correction mapping table. The corrected lower limit of the allowable cooling rate is replaced before the constraint violation judgment step of each iteration of the particle swarm optimization algorithm. That is, in each iteration, the correction value corresponding to the current actual initial temperature value of the workpiece is used as the lower limit of the allowable cooling rate to participate in the constraint condition judgment. In this embodiment, the preset temperature interval is 50°C. Four initial temperature points are selected within the range of 700°C to 850°C, namely 700°C, 750°C, 800°C, and 850°C. The mapping table uses the initial temperature value as the index key and the corresponding maximum allowable cooling rate as the value. It is stored in the control module using a one-dimensional array structure. When querying, the value corresponding to the index key that is closest to the current initial temperature value is used as the query result.

[0059] As a specific example, the product code corresponds to the material 42CrMo steel, whose static allowable cooling rate lower limit is 5℃ / s. When the infrared thermometer collects the initial temperature value of 820℃ when the current workpiece enters the insulated conveyor channel, the control module queries the initial temperature-cooling rate correction mapping table and outputs the corrected allowable cooling rate lower limit corresponding to 820℃ as 4.2℃ / s. This corrected value replaces the static value of 5℃ / s and is used as a constraint condition in the particle swarm algorithm iterative solution to participate in the constraint violation judgment. This allows the system to appropriately relax the cooling rate limit when the initial temperature of the workpiece is high, maximizing the waste heat recovery power while ensuring the metallographic structure is qualified.

[0060] In actual forging production, the water flow from the inlet to the outlet of the high-temperature radiation section in the radiant heat exchanger tube has an inherent transport delay determined by the tube's internal volume and circulation flow rate. This causes the water temperature collected by the sensing module at the outlet to reflect the heat exchange state several seconds or even tens of seconds prior. When the particle swarm optimization algorithm uses this lagging water temperature to make constraint judgments and output adjustment commands, the adjustment action of the execution module will inevitably lag behind the changes in the actual heat exchange conditions. This can easily lead to low-frequency oscillations of the outlet water temperature outside the target range due to adjustment overshoot, affecting both the stability of the heating water temperature and potentially causing excessive fluctuations in the workpiece cooling rate. Therefore, this embodiment introduces a feedforward compensation mechanism based on the inlet water temperature into the control module.

[0061] It should be further explained that, in this embodiment, the sensing module is also configured to: collect the water temperature at the inlet of the high-temperature radiation section as the inlet water temperature; the control module is also configured to: calculate the difference between the outlet water temperature and the inlet water temperature as the current heat exchange temperature difference, use the ratio of the pre-stored internal volume parameter of the radiation heat exchange tube to the current circulating flow rate of the heating loop as the transport delay time, and when the particle swarm algorithm iteratively solves the problem, use the predicted change in outlet water temperature corresponding to the transport delay time as the feedforward compensation amount, add it to the outlet water temperature, and then determine the constraint conditions.

[0062] In this embodiment, the water temperature at the inlet of the high-temperature radiation section is collected by an immersion temperature sensor installed on the inlet pipe of the high-temperature radiation section. The response time constant of this sensor is less than 0.5 seconds, ensuring that the collection time of the inlet water temperature and the collection time of the outlet water temperature strictly correspond to the same control cycle in the time series. The pre-stored internal volume parameter of the radiation heat exchanger tube is the total internal volume of all parallel radiation heat exchanger tubes in the high-temperature radiation section, calculated from the inner diameter, single tube length, and number of parallel tubes of the radiation heat exchanger tube. It is measured and stored in the control module during the initial system debugging. The transport delay time is calculated by dividing the internal volume parameter by the current... The circulating flow rate of the pre-heating circuit is obtained, in seconds; the predicted outlet water temperature change is calculated using a first-order inertial plus pure lag model based on the current heat exchange temperature difference, transport delay time, and convective heat transfer coefficient between the radiant heat exchanger tube wall and the water flow. This model's time constant and gain coefficient are identified through a step response test during initial system debugging and stored in the control module for online calculation. The feedforward compensation is superimposed before each iteration of the particle swarm optimization algorithm performs the constraint violation judgment, ensuring that the outlet water temperature used for constraint judgment is the predicted value after transport delay compensation, rather than the original lag sampled value. In this embodiment, the formula for calculating the predicted outlet water temperature change ΔT_out is ΔT_out=K×(T_out-T_in)×(1- In this equation, K is the gain coefficient, τ is the time constant, Δt is the transport delay time, T_in is the inlet water temperature, and T_out is the outlet water temperature. The gain coefficient K and time constant τ are identified during the initial system commissioning through the following step response test: When the circulating flow rate of the heating circuit is stable, the speed of the circulating pump is instantaneously increased by a preset step amplitude. The entire response curve of the outlet water temperature from the start of the step to reaching the new steady-state value is recorded. The ratio of the steady-state change of the response curve to the step amplitude is used as the gain coefficient K, and the time required for the response curve to reach 63.2% of the steady-state change is used as the time constant τ. Those skilled in the art will understand that this step response test is usually performed offline after the system's initial commissioning or after replacing the radiant heat exchange tubes. The identified K and τ values ​​are stored in the control module for online calculation. After long-term system operation, if the heat exchange characteristics change significantly due to factors such as pipe wall scaling, those skilled in the art can re-execute the step response test according to the actual operating conditions to update the K and τ values, ensuring the accuracy of the feedforward compensation calculation. The preset step amplitude is typically taken as 10% to 20% of the rated speed of the circulating pump, so as to generate a sufficiently identifiable outlet water temperature response without affecting normal heating.

[0063] As a specific example, the pre-stored internal volume parameter of the radiant heat exchanger tube is 0.08 cubic meters, and the current circulation flow rate of the heating circuit is 1.6 cubic meters per hour. Therefore, the transport delay time is 0.08 ÷ (1.6 ÷ 3600) = 180 seconds. When the sensor module detects that the inlet water temperature of the high-temperature radiant section is 55℃ and the outlet water temperature is 78℃, the current heat exchange temperature difference is 23℃. According to the first-order inertial plus pure time delay model, under a 180-second transport delay, the predicted change in outlet water temperature is +3.5℃, meaning the predicted actual outlet water temperature at the current moment is 78℃ + 3.5℃ = 81.5℃. The particle swarm optimization algorithm uses 81.5℃ instead of 78℃ as the outlet water temperature to determine the aforementioned constraint conditions. This allows the algorithm to detect the rising trend of the water temperature in advance and output corresponding transport speed and circulation flow rate adjustment commands before the actual outlet water temperature reaches the upper limit of the target range, thus avoiding adjustment lag and outlet water temperature overshoot oscillation caused by transport delay.

[0064] This process constructs a tiered heat exchange architecture consisting of a high-temperature radiation section and a low-temperature convection section. The high-grade radiant heat from the forged workpiece is directly used to produce high-temperature hot water for heating, while the low-grade waste heat from the heating return water is used to preheat the workpiece. This achieves progressive utilization of heat energy and a significant improvement in overall system energy efficiency. At the control level, the particle swarm optimization algorithm uses the lower limit of the allowable cooling rate, the target water temperature range, and the lower limit of the preheated temperature as multiple constraints to ensure that the waste heat recovery process always operates within a process window that does not damage the metallographic structure of the forging. In particular, the introduction of feasible solution ratio monitoring and boundary relaxation mechanisms ensures that when multiple constraints cause the iterative solution to become unsolvable or locally optimal, the system automatically widens the upper limit of the conveying speed with a relaxation coefficient and initializes the particle swarm, effectively guaranteeing the solvability of control parameters and output stability under complex operating conditions. Furthermore, by dynamically correcting the lower limit of the cooling rate using an initial temperature-cooling rate correction mapping table, the problem of constraint deviation caused by the final forging temperature fluctuation of the same type of workpiece is solved, ensuring that the constraint conditions always accurately match the actual initial temperature. Through a feedforward compensation mechanism based on the inlet water temperature, the predicted outlet water temperature change corresponding to the transport delay time is superimposed onto the measured outlet water temperature before constraint determination, fundamentally eliminating feedback lag and outlet water temperature overshoot oscillation caused by water flow transport delay, significantly improving the stability and control accuracy of the heating water temperature. The synergistic effect of these technologies enables the system to adaptively output transport speed and circulation flow parameters that balance meeting heating water temperature standards, ensuring qualified forging cooling quality, and optimizing cascade recovery energy efficiency, even under complex production conditions such as frequent workpiece changes, fluctuating stamping cycle time, and gradual changes in heat transfer characteristics. This combination offers the triple advantages of process compatibility, heat recovery efficiency, and control robustness.

[0065] In actual forging production, when the execution module synchronously adjusts the circulating pump speed according to the target value of the heating loop circulation flow rate output by the particle swarm optimization algorithm, the change in the water flow velocity in the radiant heat exchanger tube will simultaneously affect the high-temperature radiant section and the low-temperature convection section. Although increasing the flow rate is beneficial to improving the heat exchange power and outlet water temperature of the high-temperature radiant section, it will shorten the residence time of the heating return water in the low-temperature convection section. This will cause the workpiece temperature at the outlet of the low-temperature convection section to drop after preheating. Insufficient preheating will reduce the initial temperature of the workpiece before entering the high-temperature radiant section, which in turn will increase the consumption of high-grade heat and weaken the overall energy efficiency of the cascade recovery. To address this, this embodiment adds an independent regulating valve to the preheating loop and incorporates the preheated temperature into the constraint system of the particle swarm optimization algorithm to achieve decoupled control of the heat exchange power of the two sections.

[0066] It should be further explained that, in this embodiment, a preheating circuit proportional regulating valve is also provided on the preheating circuit; the sensing module is also configured to: collect the surface temperature of the workpiece when it leaves the low-temperature convection section, as the preheated temperature; the control module is also configured to: when the particle swarm algorithm iteratively solves the problem, take the preheated temperature being greater than or equal to a preset lower limit of the preheating temperature as a new constraint condition, which together with the constraint condition constitutes a joint constraint condition; the particle swarm algorithm simultaneously outputs the target value of the conveying speed, the target value of the circulating flow rate of the heating circuit, and the target value of the circulating flow rate of the preheating circuit; the execution module is also configured to: adjust the opening of the preheating circuit proportional regulating valve according to the target value of the circulating flow rate of the preheating circuit. In this embodiment, the search range boundary of the heating circuit circulation flow rate is consistent with the adjustment boundary of the heating circuit circulation flow rate. The lower limit of the search range of the preheating circuit circulation flow rate is zero, and the upper limit is the maximum circulation flow rate corresponding to the full opening of the preheating circuit proportional regulating valve. The joint constraint condition is an AND relationship of three constraint conditions, that is, the particle must simultaneously meet the three conditions of the outlet water temperature falling into the range of [70℃, 90℃], the actual cooling rate of the workpiece being greater than or equal to the lower limit of the allowable cooling rate, and the temperature after preheating being greater than or equal to the lower limit of the preset preheating temperature, in order to be marked as a feasible solution. The fitness function of the three-dimensional particle follows the definition of waste heat recovery power. When the outlet water temperature meets the target range, the fitness value is c×ρ×Q_heating×T_out, where Q_heating is the heating circuit circulation flow rate.

[0067] In this embodiment, the surface temperature of the workpiece when it leaves the low-temperature convection section is collected by an infrared thermometer installed at the outlet of the low-temperature convection section. The infrared thermometer is installed at a distance no more than a preset distance from the end outlet of the low-temperature convection section to ensure that the collected preheated temperature accurately reflects the actual surface temperature of the workpiece after preheating and heat exchange within the low-temperature convection section. The preset lower limit of the preheating temperature is determined according to the process requirements of avoiding the tempering brittle temperature zone of the workpiece material. For example, for 42CrMo steel forgings, the preheated temperature should not be lower than a certain safety threshold below the Ms point of the material to avoid excessive residual stress during subsequent air cooling. The lower limit value is entered into the control module according to the material process specification corresponding to the workpiece type code during the initial system debugging. The joint constraint condition is executed in the constraint violation judgment step of each iteration of the particle swarm algorithm. That is, the particle must simultaneously meet three conditions: the outlet water temperature value falls within the range of [70℃, 90℃], the actual cooling rate of the workpiece is greater than or equal to the allowable lower limit value of the cooling rate, and the temperature after preheating is greater than or equal to the preset lower limit value of the preheating temperature, in order to be marked as a feasible solution. The particle swarm algorithm encodes the conveying speed, the heating circuit circulation flow rate, and the preheating circuit circulation flow rate into the three-dimensional position vector of the particle, and solves the target values ​​of the three variables simultaneously in each iteration. Those skilled in the art will understand that the general principle for determining the preset lower limit value of the preheating temperature is: first, obtain the martensitic transformation start temperature (Ms point) and temper brittleness temperature range of the material, and then select a temperature value that is higher than the upper limit of the temper brittleness temperature range and lower than the Ms point as the safe lower limit value of the preheating temperature. When the material does not have an obvious temper brittleness range, the room temperature or workshop ambient temperature can be used as the lower limit value of the preheating temperature, and participate in the constraint construction of the particle swarm algorithm only from the perspective of maximizing the waste heat recovery efficiency. Those skilled in the art can also refer to the preheating temperature range recommended in publicly available literature such as the "Forging Heat Treatment Process Handbook" for setting the temperature.

[0068] As a specific example, the product code corresponds to the material 42CrMo steel, with a preset preheating temperature lower limit of 200℃; in a certain iteration of the particle swarm optimization algorithm, the target value of the heating loop circulation flow rate corresponding to the current particle position is 1.6 cubic meters / hour, the target value of the preheating loop circulation flow rate is 1.2 cubic meters / hour, and the target value of the conveying speed is 0.32 meters / second; the sensor module collects the outlet water temperature as 78℃, the actual workpiece cooling rate as 5.2℃ / s, and the preheated temperature as 215℃; it is verified that the outlet water temperature falls within [70℃, 90℃]. If the following three conditions are met simultaneously: the actual cooling rate of the workpiece is 5.2℃ / s, which is greater than or equal to the lower limit of the allowable cooling rate of 5℃ / s, and the temperature after preheating is 215℃, which is greater than or equal to the preset lower limit of the preheating temperature of 200℃, then the particle is marked as a feasible solution and participates in the fitness value sorting. Finally, the target value of the preheating loop circulation flow rate output by the particle swarm algorithm is converted by the execution module into the opening control signal of the preheating loop proportional regulating valve, so that the preheating loop operates at a flow rate independent of the heating loop, ensuring that the preheating effect of the low-temperature convection section is not disturbed by the flow rate regulation of the heating loop.

[0069] In actual forging production, the real-time heat load of the workshop heating terminals dynamically changes with factors such as outdoor temperature, workshop door opening and closing, and the number of heating terminals turned on. Using the waste heat recovery power within the target range of outlet water temperature as the fitness function, the optimization direction is to maximize the recovery power. However, when the heating terminal heat load significantly decreases due to a rise in outdoor temperature, maximizing the waste heat recovery power leads to the system supplying heat continuously exceeding the terminal consumption capacity. Excess heat accumulates in the return water, pushing up the heating return water temperature, which in turn causes an abnormally high return water temperature entering the low-temperature convection section, disrupting the cascade utilization effect of the preheating loop, and even causing the preheated temperature to exceed the process allowable upper limit. Therefore, this embodiment replaces the fitness function with minimizing the supply-demand power deviation, enabling dynamic matching between system heat generation and terminal heat consumption.

[0070] It should be further explained that, in this embodiment, the sensing module is also configured to: collect the water temperature at the return water outlet of the workshop heating terminal as the heating return water temperature; the control module is also configured to: calculate the real-time heat load of the heating terminal by multiplying the temperature difference between the heating return water temperature and the outlet water temperature and the circulating flow rate of the heating loop; and when the particle swarm algorithm iteratively solves the problem, the fitness function is replaced with: when the outlet water temperature is within the target range, minimizing the absolute value of the difference between the waste heat recovery power and the real-time heat load. In this embodiment, when the outlet water temperature corresponding to a particle does not fall into the [70℃, 90℃] range, its fitness value is set to the maximum value of the fitness values ​​of all particles that satisfy the target range constraint in the current iteration plus a preset offset, where the preset offset is 10% of the maximum value; the update logic of the individual optimal position and the global optimal position is synchronously changed to select the particle with the smaller fitness value as the better solution, while the other particle swarm velocity update and position update formulas remain unchanged.

[0071] In this embodiment, the water temperature at the return water outlet of the workshop heating terminal is collected by an immersion temperature sensor installed on the return water main pipe of the workshop heating terminal. The installation position is no more than a preset distance from the pipe length of the return water collection point of the heating terminal to reduce the impact of pipe heat dissipation on the measured return water temperature. When calculating the real-time heat load of the heating terminal, the circulating medium in the heating circuit is water, and its specific heat capacity and density are taken as preset constants. The product of the water temperature difference and the circulating flow rate determines the kilowatt value of the real-time heat load. The calculation method of the waste heat recovery power is the same as the calculation method of the fitness function, which is determined by the outlet water temperature and the circulating flow rate of the heating circuit. The product of quantities is used to determine the thermal power value after conversion by specific heat capacity and density. The replacement of the fitness function is performed in the fitness value calculation step of each iteration of the particle swarm algorithm. That is, when calculating the fitness value of each particle, it is first determined whether the outlet water temperature corresponding to the particle falls into the range of [70℃, 90℃]. If it falls into the range, the fitness value of the particle is calculated as the absolute value of the difference between the waste heat recovery power and the real-time heat load. The smaller the fitness value, the better the particle is. If the outlet water temperature does not fall into the range, a penalty is imposed on the particle, and its fitness value is set to a preset maximum value, so that the particle is naturally eliminated in the sorting.

[0072] As a concrete example, if the sensor module collects data showing an outlet water temperature of 78℃, a heating loop circulation flow rate of 1.5 cubic meters per hour, and a heating return water temperature of 52℃, then the real-time heat load of the heating terminal is approximately (78℃ - 52℃) × 1.5 cubic meters per hour × water specific heat capacity 4.18 kJ / (kg·℃) × water density 1000 kg / m³ ÷ 3600 seconds ≈ 45.3 kW. Simultaneously, the waste heat recovery power is approximately 48.5 kW corresponding to the temperature of 78℃. In a certain iteration of the particle swarm optimization algorithm, a certain particle... The outlet water temperature is 78℃ and within the target range, with a fitness value of |48.5-45.3|=3.2; the outlet water temperature of another particle is 83℃, with a waste heat recovery power of approximately 52.0 kW, and a fitness value of |52.0-45.3|=6.7; after sorting by fitness value, the particle with a fitness value of 3.2 is better. The particle swarm algorithm converges towards a direction closer to the real-time heat load, and the final output conveying speed and circulation flow rate make the system's heat generation and terminal heat consumption tend to be balanced, avoiding abnormal rise in return water temperature that would damage the cascade recovery effect.

[0073] This process achieves decoupled control of heat exchange power between the heating and preheating circuits by adding an independent proportional regulating valve to the preheating circuit and incorporating the preheated temperature as a new constraint into the joint constraint system of the particle swarm optimization algorithm. The particle swarm optimization algorithm encodes the delivery speed, the circulation flow rate of the heating circuit, and the circulation flow rate of the preheating circuit into three-dimensional position vectors for synchronous solution. This ensures that the preheating effect of the low-temperature convection section is no longer passively affected by the disturbance of the flow rate regulation of the heating circuit, fundamentally solving the problem of insufficient or excessive preheating caused by flow coupling, and ensuring the independence and stability of the two-stage heat exchange in the cascade recovery architecture. Meanwhile, the fitness function is replaced by minimizing the absolute value of the difference between the waste heat recovery power and the real-time heat load of the heating terminal. This changes the optimization direction of the particle swarm optimization algorithm from simply pursuing the maximization of recovered heat to a dynamic supply-demand match between heat generation and terminal heat consumption. When the heat load of the heating terminal decreases due to factors such as changes in outdoor temperature, the algorithm automatically converges to a heat generation level closer to the real-time heat load, avoiding excessive heat generation that pushes up the return water temperature and thus undermines the cascade utilization effect of the preheating loop. Conversely, when the heat load increases, it can fully release the recovery potential. The synergistic effect of these two technical means enables the system to consistently output optimal control parameters that balance the requirements of heating water temperature, preheating temperature, and real-time supply and demand power under complex operating conditions of dynamic fluctuations in heating load and strong coupling of two-stage heat exchange flow. This significantly improves the operational stability and overall energy efficiency of the cascade heat recovery system.

[0074] In actual forging production, presses may experience prolonged shutdowns due to equipment failure, mold replacement, or unexpected power outages. During these shutdowns, the stamping cycle value becomes infinitely large or remains zero, rendering the data collected by the sensing module inaccurate and failing to reflect normal production rhythm. Simultaneously, radiant heat exchange tubes may leak due to weld cracks or pipe wall corrosion, causing a rapid drop or abnormal fluctuation in outlet water temperature within a short period. The outlet water temperature data collected by the sensing module then fails to accurately reflect the normal heat exchange state. If the particle swarm optimization algorithm continues to iterate and output adjustment commands based on this distorted data, the execution module will drive the motor and pump according to incorrect target values, potentially leading to workpiece stagnation and overheating in the conveyor channel, idling of the circulating pump, or overpressure in the heating circuit, resulting in safety accidents. Therefore, this embodiment establishes an abnormal operating condition detection and protection mechanism in the control module, immediately switching to safe preset values ​​upon detecting an abnormal event.

[0075] It should be further explained that, in this embodiment, the stamping cycle value and the outlet water temperature are continuously monitored. When the k consecutively collected stamping cycle values ​​all exceed the preset normal range of the stamping cycle, or when the rate of change of the outlet water temperature within the preset time window exceeds the preset water temperature change rate threshold, it is determined to be an abnormal working condition. When an abnormal working condition is determined, the particle swarm algorithm iterative solution is stopped, and a preset safe value for the conveying speed and a preset safe value for the heating circuit circulation flow rate are output. The execution module is also configured to adjust the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulation pump of the heating circuit according to the preset safe value for the conveying speed and the preset safe value for the heating circuit circulation flow rate.

[0076] In this embodiment, the number of continuously collected samples, k, is set to a positive integer greater than or equal to 3 to avoid accidental triggering of abnormal operating condition judgment due to instantaneous abnormal values ​​caused by single sensor signal jumps or electromagnetic interference. The lower limit of the preset normal range of the stamping cycle corresponds to the stamping cycle value corresponding to the maximum normal operating cycle of the press, and the upper limit corresponds to the longest standby interval allowed by the press. When the stamping cycle value exceeds the upper limit, it indicates that the press is in a shutdown or fault state. The preset time window is three to five times the rise time of the radiant heat exchanger tube outlet water temperature response to flow regulation, to ensure that the time window is sufficient to cover the dynamic changes in water temperature during normal regulation, while also being able to capture abnormal change rates caused by faults such as leaks in a timely manner. Those skilled in the art will understand that the "rise time" refers to the transition process time of the outlet water temperature response to the step change in circulating flow, and its value can be obtained through the aforementioned step response test, that is, the time from the start of the step to the outlet water temperature reaching 90% of the new steady-state value.

[0077] The preset water temperature change rate threshold is determined based on the maximum normal change rate of the outlet water temperature under the maximum allowable flow rate adjustment range of the heating circuit, and is multiplied by a preset safety factor to avoid misjudging normal flow rate adjustment response as abnormal. Those skilled in the art will understand that the "maximum normal change rate" can be obtained by applying the maximum flow rate adjustment range (e.g., from the minimum safe flow rate to the rated maximum flow rate) during the system commissioning phase, recording the change curve of the outlet water temperature, and taking the maximum value of the absolute value of the slope of the curve as the maximum normal change rate. The preset safety factor is usually taken as 1.2 to 1.5 to leave a reasonable judgment margin between normal adjustment and abnormal leakage.

[0078] The preset safe conveying speed value is set as the safe conveying speed corresponding to the minimum allowable residence time of the workpiece in the insulated conveying channel, ensuring that the workpiece can still pass through the conveying channel at a safe cooling rate without overheating and accumulation under abnormal working conditions; the preset safe value of the heating circuit circulation flow rate is set to maintain the minimum safe circulation flow rate of the heating circuit to prevent the circulation pump from running dry and being damaged at zero flow or extremely low flow rate, while maintaining the minimum antifreeze flow rate of the heating terminal.

[0079] As a specific example, k is set to 3, the preset normal range of the pressing cycle is set to 4 to 60 seconds, the preset time window is 30 seconds, the preset water temperature change rate threshold is set to 5℃ / minute, the preset safe value of the conveying speed is set to 0.15 m / s, and the preset safe value of the heating circuit circulation flow rate is set to 0.5 cubic meters / hour. When the pressing cycle values ​​collected by the sensor module for three consecutive times are 65 seconds, 68 seconds, and 70 seconds, respectively, all exceeding the upper limit of the normal range of 60 seconds, the control module determines that the press has stopped and enters an abnormal working condition. At this time, the control module stops the particle swarm algorithm iterative solution, outputs the safe value of the conveying speed of 0.15 m / s and the safe value of the heating circuit circulation flow rate of 0.5 cubic meters / hour, and the execution module adjusts the frequency of the drive motor of the insulation conveying channel and the speed of the heating circuit circulation pump accordingly to maintain the safe operation of the system during the fault. After the pressing cycle value returns to the normal range, the particle swarm algorithm iterative solution is restarted and normal optimization control is restored. In this embodiment, when the k consecutively collected stamping cycle values ​​all return to the preset normal range of the stamping cycle and the rate of change of the outlet water temperature within the preset time window returns to less than or equal to the preset water temperature change rate threshold, the control module restarts the particle swarm algorithm iterative solution. When restarting, the initial position of the particle swarm is randomly generated between the upper and lower limits of the conveying speed and the upper and lower limits of the heating circuit circulation flow rate, and the initial speed is randomly initialized. The particle swarm state before the abnormal working condition is not inherited.

[0080] In actual forging production, when the particle swarm optimization (PSO) algorithm solves for conveying speed and circulation flow rate that satisfy multiple constraints, the particle swarm may prematurely cluster in a local region as iterations progress. The positions and fitness values ​​of all particles in the population tend to be consistent, at which point the algorithm loses its ability to explore better solutions in a wider search space. Even if the current optimal solution is not the global optimum, the algorithm cannot autonomously escape this local region, resulting in the final output of the target conveying speed and circulation flow rate as locally suboptimal solutions. This causes the waste heat recovery power to be lower than the system's actual achievable optimal level. Therefore, this embodiment introduces a population diversity monitoring and inertial weight adaptive perturbation mechanism in the control module.

[0081] It should be further explained that, in this embodiment, the control module is also configured to: calculate the variance of the fitness values ​​of all particles in the current particle swarm after each iteration of the particle swarm algorithm, as a population diversity index; when the population diversity index is lower than a preset diversity threshold and the current iteration number has not reached the preset maximum iteration number, multiply the inertia weight of the particle swarm algorithm by a preset perturbation factor and continue iterating until the population diversity index recovers to a value greater than or equal to the preset diversity threshold, and then restore the inertia weight to its original set value.

[0082] In this embodiment, the variance of the fitness values ​​of all particles in the current particle swarm is calculated by dividing the sum of squared deviations of the fitness value of each particle in the swarm from the mean of the fitness values ​​of all particles in the current iteration by the total number of particles. The calculation method of the fitness value is consistent with the definition of the fitness function. The preset diversity threshold is determined based on the lower limit of the population diversity index recorded during the convergence process when the particle swarm algorithm is run multiple times during the initial debugging phase. 1.2 to 1.5 times this lower limit is taken as the preset diversity threshold to avoid triggering perturbation when the algorithm converges normally to the global optimum, and only to trigger it when the population loses diversity prematurely. Those skilled in the art will understand that the " "Multiple runs during the initial debugging phase" refers to running the particle swarm optimization algorithm at least 10 times offline, using typical workpiece types and typical stamping cycles as input. The lowest value of the fitness variance is recorded during each run as it gradually decreases from the initial value to a convergent stable value. The average of all the lowest values ​​recorded is taken as the lower limit. If offline testing conditions are limited, those skilled in the art can directly use 1% of the theoretical maximum value of the fitness function as the initial diversity threshold, and adjust it according to the triggering frequency during online system operation: if disturbance triggers are too frequent, the threshold can be appropriately lowered; if they are almost never triggered, the threshold can be appropriately increased to enhance the global search capability.

[0083] The preset maximum number of iterations is consistent with the total number of iterations of the particle swarm optimization algorithm; the preset perturbation factor is greater than 1.0 and less than or equal to 2.0, increasing the inertia weight makes the particles gain greater motion inertia in the current velocity direction, thereby expanding the search range and jumping out of the local optimum region; the operation of restoring the inertia weight to the original set value is executed immediately after the population diversity index is restored to greater than or equal to the preset diversity threshold in the current iteration, so as to avoid the algorithm failing to converge near the global optimum due to continuously excessive inertia weight.

[0084] As a specific example, the preset diversity threshold is set to 50, the preset maximum number of iterations is 100 generations, the original setting of the inertia weight is 0.7, and the preset perturbation factor is 1.5. When the particle swarm algorithm iterates to the 30th generation, the control module calculates that the variance of the fitness values ​​of all particles in the current particle swarm is 28, which is lower than the preset diversity threshold of 50, and the current iteration number of 30 has not reached the preset maximum number of iterations of 100. Therefore, the inertia weight is corrected from 0.7 multiplied by the perturbation factor 1.5 to 1.05. The iteration continues with the corrected inertia weight of 1.05. When the population diversity index recovers to 62 in the 45th generation, which is greater than or equal to the preset diversity threshold of 50, the inertia weight is restored to the original setting of 0.7 and the subsequent iterations continue until the maximum number of iterations is reached, at which point the target values ​​of the conveying speed and the circulating flow rate are output.

[0085] This process calculates the variance of the population fitness value as a population diversity index after each iteration of the particle swarm optimization algorithm and compares this index with a preset diversity threshold. This enables the system to self-awareness of the particle swarm search state; that is, when the population prematurely clusters in a local optimum due to multiple constraints and loses its ability to continue exploring, the system can identify this premature convergence trend in real time. Based on this, the inertia weight is temporarily increased by a preset perturbation factor, giving the particles greater inertia in the current velocity direction, thereby breaking through the constraints of the local optimum and expanding the search range. Once the population diversity recovers to above the threshold, the inertia weight is immediately restored to its original value, allowing the algorithm to continue its fine-grained search near the global optimum even after escaping local optima, maintaining its original convergence characteristics. This mechanism, at the cost of temporarily increasing and promptly restoring the inertial weight, effectively balances global exploration and local development capabilities without increasing the overall number of algorithm iterations or changing the original convergence architecture of the particle swarm. This allows the particle swarm algorithm to maintain sufficient population diversity in complex feasible domains with multiple constraints and narrow boundaries. The final output of the target values ​​for transport speed and circulation flow rate is closer to the global optimum, significantly improving the system's waste heat recovery power and the reliability of optimization under all operating conditions.

[0086] In actual forging production, oxide layers and scale gradually form on the inner wall of the radiant heat exchanger tubes over time, and the performance of the insulation material in the heat-insulating conveying channel also slowly deteriorates. These gradually changing factors cause a progressive drift in the correspondence between the actual cooling characteristics of the workpiece and the initial temperature-cooling rate correction mapping table established during the initial system commissioning. If the mapping table is not updated for a long period, the corrected allowable lower limit of cooling rate output by the control module based on the mapping table, which has deviated from the actual heat exchange characteristics, will generate cumulative deviations. This makes the constraints unable to accurately reflect the safe boundary of the cooling rate of the workpiece under the current actual heat exchange conditions, thereby affecting the consistency of the forging's metallographic structure and the efficiency of waste heat recovery. Therefore, this embodiment introduces a mapping table self-correction mechanism based on automatic acquisition of steady-state operating condition data into the control module.

[0087] It should be further explained that, in this embodiment, the control module is also configured to: when the outlet water temperature falls into the target range for m consecutive times and the target values ​​of the delivery speed and circulation flow rate output by the particle swarm optimization algorithm are both within the preset steady-state fluctuation range, record the initial temperature value and the currently participating modified lower limit value of the allowable cooling rate to form a correction sample point; when the number of accumulated correction sample points reaches the preset correction sample number threshold, refit the updated initial temperature-cooling rate correction mapping table with all correction sample points and replace the preset initial temperature-cooling rate correction mapping table.

[0088] In this embodiment, the m consecutive values ​​are positive integers greater than or equal to 5 to ensure that the system has entered a true steady-state operation rather than a transient coincidence when collecting calibration sample points, thus avoiding the misidentification of non-steady-state data points that accidentally meet the conditions as calibration samples. The preset steady-state fluctuation range is centered on the average of the target values ​​of the conveying speed and the circulating flow rate in the most recent p outputs, with a preset percentage above and below the average value as the fluctuation range. When the target values ​​output for m consecutive times do not exceed this range, the system is determined to be in steady-state operation. Those skilled in the art will understand that the preset percentage is usually set to 10% to 20%, and its setting is based on the fact that when the target value fluctuation range is within this range, the adjustment accuracy of the execution module (frequency converter and circulating pump) can respond sufficiently, and the outlet water temperature will not deviate from the target range due to small fluctuations. If the percentage is set too small (e.g., below 5%), the steady-state conditions may be difficult to meet due to sensor noise or normal fine-tuning, affecting the collection rate of calibration samples. If the percentage is set too large (e.g., exceeding 25%), the operating condition that has not yet been fully stabilized may be misjudged as steady state, resulting in the collection of non-representative calibration sample points. Those skilled in the art can select an appropriate percentage value within this range based on the adjustment precision of the execution module and the magnitude of the system's inertia.

[0089] The preset threshold for the number of calibration samples is determined based on the number of preset temperature intervals in the initial temperature-cooling rate correction mapping table, and is taken as an integer multiple of the number of preset temperature intervals to ensure that there are a sufficient number of calibration sample points covering the heat transfer characteristic drift in each preset temperature interval. The refitting uses the least squares method to perform polynomial fitting on all accumulated calibration sample points, and the order of the fitted curve is consistent with the curve type of the preset initial temperature-cooling rate correction mapping table to generate a smooth and continuous updated mapping table. The replacement operation takes effect immediately after the mapping table is updated, and the next time the particle swarm algorithm iterates, the updated initial temperature-cooling rate correction mapping table is used to query the corrected lower limit of the allowable cooling rate.

[0090] As a specific example, if the value is 5 for m consecutive times, the steady-state fluctuation range of the target conveying speed is centered on the average of the last 10 outputs (0.32 m / s) with a fluctuation range of 15% above and below it, i.e., [0.272, 0.368] m / s. The steady-state fluctuation range of the target circulating flow rate is centered on the average of the last 10 outputs (1.5 cubic meters / hour) with a fluctuation range of 15% above and below it, i.e., [1.275, 1.725] cubic meters / hour. During a certain operating period, the outlet water temperature is collected 5 times consecutively at 76℃, 78℃, 77℃, 79℃, and 78℃, all falling within the target range of [70℃, 90℃]. The corresponding target conveying speed and circulating flow rate are... The target values ​​are all within their respective steady-state fluctuation ranges. At this time, the control module records the initial temperature value of the current workpiece (810℃) and the current corrected lower limit of the allowable cooling rate (4.0℃ / s) participating in the iteration, forming a calibration sample point (810, 4.0). When the number of calibration sample points accumulated by different batches of workpieces at different initial temperatures reaches the preset calibration sample number threshold of 50, the control module performs polynomial fitting on all 50 calibration sample points using the least squares method to generate an updated initial temperature-cooling rate correction mapping table, replacing the old mapping table that has been run for a preset duration or a preset number of workpieces, so that the correction benchmark of the lower limit of the cooling rate automatically adapts to the actual drift state of the heat transfer characteristics.

[0091] This process accurately identifies the system as having entered a truly stable operating state by monitoring the dual conditions of the outlet water temperature continuously falling within the target range and the delivery speed and circulation flow rate both being within the steady-state fluctuation range. In this state, it automatically collects the initial temperature value and the currently corrected lower limit of the allowable cooling rate to form calibration sample points. This ensures that the data source for the self-calibration of the mapping table comes from the actual operating data of the system under optimal conditions, rather than unreliable data from transient disturbances or unstable conditions. When the accumulated calibration sample points reach a preset threshold, a polynomial fit is performed on all calibration sample points using the least squares method to generate an updated initial temperature-cooling rate correction mapping table and replace the old table. This allows the correction benchmark for the lower limit of the cooling rate to automatically follow the drift of actual heat transfer characteristics caused by slowly changing factors such as oxidation and scaling on the inner wall of the radiant heat exchanger tubes and the attenuation of insulation materials. This self-calibration mechanism requires no manual intervention or shutdown for calibration. It can autonomously complete the incremental update of the mapping table during continuous system operation, fundamentally eliminating the cumulative deviation of constraint conditions caused by the long-term lack of updating of the mapping table. This ensures that the particle swarm algorithm can perform optimization solutions at the cooling rate safety boundary that accurately reflects the current actual heat transfer characteristics in all operating stages, thereby maintaining the consistency of the forging metallographic structure and the optimal level of waste heat recovery efficiency in the long term.

[0092] Example 3

[0093] Please see Figure 2 Another embodiment of the present invention provides a method for the graded recovery and utilization of residual heat from forged workpieces, comprising:

[0094] S1. Collect the current workpiece's type code, press stamping cycle, and radiant heat exchanger tube outlet water temperature;

[0095] S2. The lower limit of the allowable cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating circuit circulation flow rate that satisfy the constraints are solved iteratively by the particle swarm algorithm.

[0096] S3. Based on the conveying speed and circulation flow rate, synchronously adjust the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulation pump of the heating circuit.

[0097] Example 4

[0098] An electronic device includes a memory and a control module, wherein the memory stores a computer program, and the control module executes the computer program to realize a method for the cascaded recovery and utilization of residual heat from forged workpieces.

[0099] A computer-readable storage medium storing computer instructions that, when executed, perform a method for the cascaded recovery and utilization of residual heat from forged workpieces.

[0100] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the claims. All of these variations are within the protection scope of the present invention.

Claims

1. A system for the cascaded recovery and utilization of residual heat from forged workpieces, characterized in that, include: A heat-insulated conveying channel is located at the outlet of the press, and its inner wall is lined with radiant heat exchange tubes. The radiant heat exchange tubes are divided into a high-temperature radiation section and a low-temperature convection section along the workpiece conveying direction. The heating circuit connects the outlet of the high-temperature radiant section to the inlet of the workshop heating terminal; The preheating circuit connects the return water outlet of the workshop heating terminal with the inlet of the low-temperature convection section, and uses the waste heat of the heating return water to preheat the workpiece flowing through the low-temperature convection section. The sensing module is configured to: collect the current workpiece's type code, the press's stamping cycle, and the outlet water temperature of the radiant heat exchanger tube; The control module is configured as follows: The lower limit of the allowable cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating loop circulation flow rate that satisfy the constraints are solved iteratively by the particle swarm algorithm. The execution module is configured to synchronously adjust the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulating pump of the heating circuit according to the conveying speed and the circulation flow rate.

2. The waste heat recovery and utilization system for forged workpieces as described in claim 1, characterized in that, The control module is also equipped with a feasible solution ratio monitoring unit and a boundary relaxation unit: the feasible solution ratio monitoring unit, in each iteration of the particle swarm algorithm, counts the proportion of particles whose outlet water temperature falls within the target range and whose actual workpiece cooling rate is greater than or equal to the lower limit of the allowable cooling rate, as the proportion of the total number of particles, and uses this proportion as the feasible solution ratio; the boundary relaxation unit, when the number of iterations reaches a first preset value and the feasible solution ratio is lower than a preset feasible solution ratio threshold, uses the product of the upper limit threshold of the conveying speed and the preset relaxation coefficient as the corrected adjustment boundary, and triggers the particle swarm to reinitialize, until the feasible solution ratio is greater than or equal to the preset feasible solution ratio threshold, and then outputs the target value of the conveying speed and the target value of the circulating flow rate; wherein, the upper limit threshold of the conveying speed is determined by the ratio of the stamping cycle value to the minimum allowable residence time of the workpiece in the heat preservation conveying channel.

3. The waste heat recovery and utilization system for forged workpieces as described in claim 2, characterized in that, The sensing module is further configured to: collect the surface temperature of the workpiece when it enters the inlet of the heat preservation conveying channel as an initial temperature value; the control module is further configured to: use the initial temperature value and the variety code as joint inputs, query a preset initial temperature-cooling rate correction mapping table, output the corrected allowable cooling rate lower limit value corresponding to the initial temperature value, and use the corrected allowable cooling rate lower limit value to replace the allowable cooling rate lower limit corresponding to the variety code as the constraint condition to participate in the particle swarm algorithm iterative solution.

4. The waste heat recovery and utilization system for forged workpieces as described in claim 3, characterized in that, The sensing module is further configured to: collect the water temperature at the inlet of the high-temperature radiation section as the inlet water temperature; the control module is further configured to: calculate the difference between the outlet water temperature and the inlet water temperature as the current heat exchange temperature difference, use the ratio of the pre-stored internal volume parameter of the radiation heat exchange tube to the current circulating flow rate of the heating loop as the transport delay time, and during the iterative solution of the particle swarm optimization algorithm, use the predicted change in outlet water temperature corresponding to the transport delay time as the feedforward compensation amount, add it to the outlet water temperature, and then determine the constraint conditions.

5. The waste heat recovery and utilization system for forged workpieces as described in claim 4, characterized in that, The preheating circuit is also equipped with a preheating circuit proportional regulating valve; the sensing module is also configured to: collect the surface temperature of the workpiece when it leaves the low-temperature convection section, as the preheated temperature; the control module is also configured to: when the particle swarm algorithm iteratively solves the problem, take the preheated temperature being greater than or equal to a preset lower limit of the preheating temperature as a new constraint condition, which together with the constraint condition constitutes a joint constraint condition; the particle swarm algorithm simultaneously outputs the target value of the conveying speed, the target value of the circulating flow rate of the heating circuit, and the target value of the circulating flow rate of the preheating circuit; the execution module is also configured to: adjust the opening of the preheating circuit proportional regulating valve according to the target value of the circulating flow rate of the preheating circuit.

6. The waste heat recovery and utilization system for forged workpieces as described in claim 5, characterized in that, The sensing module is further configured to: collect the water temperature at the return water outlet of the workshop heating terminal as the heating return water temperature; the control module is further configured to: calculate the real-time heat load of the heating terminal by multiplying the temperature difference between the heating return water temperature and the outlet water temperature and the circulating flow rate of the heating loop; and when the particle swarm algorithm iteratively solves the problem, replace the fitness function with: when the outlet water temperature is within the target range, minimize the absolute value of the difference between the waste heat recovery power and the real-time heat load.

7. The waste heat recovery and utilization system for forged workpieces as described in claim 6, characterized in that, The stamping cycle value and the outlet water temperature are continuously monitored. When the k consecutively collected stamping cycle values ​​all exceed the preset normal range of the stamping cycle, or when the rate of change of the outlet water temperature within the preset time window exceeds the preset water temperature change rate threshold, it is determined to be an abnormal working condition. When an abnormal operating condition is determined, the particle swarm algorithm iterative solution is stopped, and a preset safe value for the conveying speed and a preset safe value for the circulating flow rate of the heating circuit are output. The execution module is further configured to adjust the frequency of the drive motor of the insulation conveying channel and the speed of the circulation pump of the heating circuit according to the preset safe value of the conveying speed and the preset safe value of the circulating flow rate of the heating circuit.

8. The waste heat recovery and utilization system for forged workpieces as described in claim 7, characterized in that, The control module is also configured to: calculate the variance of the fitness values ​​of all particles in the current particle swarm after each iteration of the particle swarm algorithm, as an indicator of population diversity. When the population diversity index is lower than the preset diversity threshold and the current iteration number has not reached the preset maximum iteration number, the inertial weight of the particle swarm optimization algorithm is multiplied by a preset perturbation factor and the iteration continues until the population diversity index recovers to a level greater than or equal to the preset diversity threshold. Then, the inertial weight is restored to its original set value.

9. The waste heat recovery and utilization system for forged workpieces as described in claim 8, characterized in that, The control module is further configured to: when the outlet water temperature falls into the target range for m consecutive times and the target values ​​of the conveying speed and the circulating flow rate output by the particle swarm algorithm are both within the preset steady-state fluctuation range, record the initial temperature value and the current lower limit of the corrected allowable cooling rate participating in the iteration to form a correction sample point. When the number of accumulated correction sample points reaches a preset correction sample number threshold, the updated initial temperature-cooling rate correction mapping table is refitted with all correction sample points and replaced with the preset initial temperature-cooling rate correction mapping table.

10. A method for the cascaded recovery and utilization of residual heat from forged workpieces, implemented based on the cascaded recovery and utilization system for residual heat from forged workpieces according to any one of claims 1-9, characterized in that, include: Collect the current workpiece's type code, press stamping cycle, and radiant heat exchanger tube outlet water temperature; The lower limit of the allowable cooling rate corresponding to the variety code and the target range of the outlet water temperature of the radiant heat exchanger tube are used as constraints. The stamping cycle is used as the adjustment boundary of the conveying speed. The waste heat recovery power of the outlet water temperature of the radiant heat exchanger tube within the target range is used as the fitness function. The conveying speed and heating loop circulation flow rate that satisfy the constraints are solved iteratively by the particle swarm algorithm. Based on the conveying speed and circulation flow rate, the frequency of the drive motor of the heat preservation conveying channel and the speed of the circulating pump of the heating circuit are adjusted synchronously.

Images