Task acceleration execution simulation method based on multi-cycle priority
By using a multi-cycle priority task acceleration simulation method, the problems of low efficiency, high cost, low accuracy and poor versatility in slow dynamic system simulation are solved, and efficient and low-cost simulation result generation and timely response to critical tasks are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANJING RONGGUANG SOFTWARE TECH CO LTD
- Filing Date
- 2026-01-19
- Publication Date
- 2026-05-01
AI Technical Summary
Existing technologies suffer from low efficiency, high cost, low accuracy, and poor versatility in slow dynamic system simulation. In particular, when dealing with tasks of different priorities, critical tasks are easily blocked by low-priority tasks, leading to response delays.
A simulation method based on multi-cycle priority is adopted to accelerate task execution. By determining the original simulation cycle and priority of the slow dynamic system, the system is divided into proportional acceleration cycles, and task scheduling is performed according to priority to ensure that high-priority tasks are executed first and generate simulation results.
It improves simulation efficiency, reduces computational resource consumption and data generation costs, ensures timely response to critical tasks, and enhances the accuracy and versatility of simulation results, making it suitable for a variety of slow-dynamic systems.
Smart Images
Figure CN121959933A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer simulation technology, and in particular to a simulation method for accelerating task execution based on multi-cycle priority. Background Technology
[0002] In the field of modern engineering technology, using computers to simulate and model various real systems has become a core method for design optimization, performance analysis, algorithm verification, and teaching demonstrations. By constructing mathematical models and leveraging computer simulation technology, the operating state, physical change processes, and dynamic response characteristics of a system can be reproduced in a virtual environment. This avoids the high costs, safety risks, or time constraints of actual operation, providing efficient and reliable technical support for system design improvement, energy-saving strategy research, fault diagnosis, and teaching and training.
[0003] In industrial production, systems such as air conditioning systems, urban centralized heating networks, energy transmission networks, and environmental control systems are considered slow-dynamic systems. Currently, the simulation process for slow-dynamic systems typically includes the following steps: First, establishing a physical model of the system, constructing a set of mathematical equations based on the system's physical laws (such as heat conduction equations and fluid mechanics laws), and clarifying the interaction relationships between variables in the system (such as temperature, pressure, and flow rate); second, setting initial and boundary conditions for the simulation, including initial system state parameters (such as initial temperature distribution) and external environmental constraints (such as outdoor temperature and heat source power); third, using numerical calculation methods to perform time evolution, discretizing the continuous physical process into a time-step sequence, and updating the system state time-by-time by solving the discretized equations; finally, outputting simulation results, including curves showing the change of key system parameters over time and steady-state distributions, providing data support for subsequent analysis or applications. In this process, existing technologies generally use a fixed and uniform time step for calculation to ensure numerical stability and result accuracy.
[0004] However, for slow-dynamic systems, whose physical change cycles are often measured in minutes or even hours, the aforementioned unified time-step simulation task execution method will cause a series of problems: First, slow-dynamic tasks consume too many computing resources, resulting in low overall execution efficiency. Due to the slow changes in system state, calculations within a large number of time steps only lead to minor parameter variations, yet still require the same amount of computing resources as significant change phases, making the simulation time close to the actual physical time, which is difficult to meet the needs of rapid verification of schemes in energy-saving algorithm research and real-time demonstration of processes in teaching demonstrations; Second, generating massive amounts of sample data takes too long. When algorithm optimization requires training models with a large number of simulation samples, or when teaching demonstrations need to cover multiple operating conditions, the redundant calculations caused by the unified time step will significantly increase the data generation cycle, restricting the efficiency of research and teaching; Third, critical tasks (such as abnormal state response) may be blocked by low-priority tasks. When processing normal operation simulation and sudden failure analysis tasks of different priorities simultaneously during the simulation process, the unified time-step scheduling method cannot prioritize the allocation of resources to critical tasks, which may lead to delays in the response to abnormal states, affecting the timeliness and reliability of system simulation.
[0005] In summary, the problems with existing technologies are: low efficiency, high cost, low accuracy, and poor versatility in slow dynamic system simulation. Summary of the Invention
[0006] The purpose of this invention is to provide a simulation method for accelerating task execution based on multi-cycle priority, which is highly efficient, low-cost, highly accurate, and versatile.
[0007] The technical solution to achieve the purpose of this invention is as follows:
[0008] A simulation method for accelerating task execution based on multi-cycle priority includes the following steps:
[0009] S10, Slow Dynamic System Simulation Initialization: Determine the original simulation cycle and priority of the slow dynamic system, divide the simulation acceleration cycle, and set the initial simulation parameters;
[0010] S20, Accelerated Simulation of Slow Dynamic Systems: Based on the initial simulation parameters and priorities, accelerated simulation of slow dynamic systems is performed within the acceleration period to obtain the simulation results of each model instance;
[0011] S30, Simulation Result Output: Outputs the simulation results for each model instance.
[0012] Preferably, step S10, the slow dynamic system simulation initialization step, includes:
[0013] S11, Determination of Original Simulation Cycle and Priority: For each model in the virtual simulation of a slow dynamic system, the original simulation cycle and priority of each model are determined based on the dynamic characteristics of its corresponding entity in actual operation.
[0014] S12, Acceleration cycle division: The original simulation cycle of each model is compressed proportionally by a uniform compression factor to obtain the simulation acceleration cycle of each model.
[0015] S13, Initial Parameter Setting: Set the initial simulation parameters for each model instance.
[0016] Compared with the prior art, the significant advantages of this invention are:
[0017] 1. High efficiency: This invention compresses the original task cycle proportionally to obtain the timing simulation cycle. While ensuring the integrity of the simulation logic, it effectively shortens the virtual simulation time axis, greatly improves the simulation efficiency of slow dynamic systems, reduces the consumption of computing resources, and solves the problems of redundant computing resources and simulation time close to actual physical time caused by the uniform time step in the existing technology for slow dynamic tasks. It can quickly meet the timeliness requirements of energy-saving algorithm research and teaching demonstration.
[0018] 2. Low cost: Based on the accelerated timed simulation cycle, this invention can flexibly modify the environmental parameters of the virtual simulation platform in the initial stage or during the simulation process, complete multiple simulations in a short time and generate a large number of historical sample points, efficiently generate massive sample data, reduce data acquisition costs, and solve the problem of excessively long sample generation cycle caused by uniform time step in the prior art, which restricts the efficiency of research and teaching, and provides sufficient data support for the research project.
[0019] 3. High accuracy: This invention ensures that high-priority (smaller numerical value) tasks are executed first through a priority scheduling mechanism, and their output parameters can be used by low-priority (larger numerical value) tasks in a timely manner, avoiding data conflicts, ensuring the timeliness of critical task response, improving simulation reliability, solving the problem of critical tasks being easily blocked by low-priority tasks and response delays in the prior art, and ensuring the accuracy of simulation results.
[0020] 4. Good versatility: It does not depend on specific application scenarios and can be directly applied to various slow dynamic systems such as air conditioning, heating, and industrial control. It has strong versatility, wide applicability, and effectively solves common problems in the simulation of slow dynamic systems in different fields.
[0021] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description
[0022] Figure 1 This is the main flowchart of the simulation method for accelerating task execution based on multi-cycle priority in this invention.
[0023] Figure 2 yes Figure 1 A flowchart of the initialization steps for simulation of a medium-to-slow dynamic system.
[0024] Figure 3 This is a diagram showing the comparison before and after the task cycle was accelerated.
[0025] Figure 4 yes Figure 1 A flowchart of the accelerated simulation steps for medium- and slow-dynamic systems.
[0026] Figure 5 yes Figure 4 A flowchart of the simulation steps for a single-cycle model example.
[0027] Figure 6 This is a schematic diagram of time-axis driven simulation.
[0028] Figure 7 This is a schematic diagram of the execution of the simulation after time acceleration.
[0029] Figure 8 yes Figure 5 A flowchart of the execution steps of a scheduled task. Detailed Implementation
[0030] To facilitate understanding of the present invention, the embodiments of the present invention use an air conditioning system in a slow dynamic system, including an air conditioning unit, a fan, a manifold, a cooling tower, and a room, as examples, where the period value of the model is merely an example value.
[0031] like Figure 1 As shown, the present invention provides a virtual simulation method for accelerating task execution based on multi-cycle priority, comprising the following steps:
[0032] S10, Slow Dynamic System Simulation Initialization: Determine the original simulation cycle and priority of the slow dynamic system, divide the simulation acceleration cycle, and set the initial simulation parameters.
[0033] This step completes the preparation of the simulation cycle, priority, and parameters, laying the foundation for accelerated execution.
[0034] like Figure 2 As shown, the S10 slow dynamic system simulation initialization step includes:
[0035] S11, Determination of Original Simulation Cycle and Priority: For each model in the virtual simulation of a slow dynamic system, the original simulation cycle and priority of each model are determined based on the dynamic characteristics of its corresponding entity in actual operation.
[0036] The original simulation cycle refers to the operating cycle of the entity corresponding to the model in the actual physical system.
[0037] For models of slow-dynamic systems (such as air conditioning systems and heating networks) in virtual simulations, the original simulation cycle and priority need to be determined based on the dynamic characteristics of the corresponding entities in the actual physical system operation. This can be achieved by using the following methods to accurately match the actual operating patterns of the system and provide a foundation for subsequent multi-cycle priority scheduling:
[0038] The rules for determining the original simulation cycle: The original simulation cycle represents the operating cycle of the entity corresponding to the model in the actual physical system, and its determination is strongly correlated with the rate of state change of the entity and the frequency of control interaction.
[0039] 1. When the physical state changes rapidly (e.g., the air conditioning unit needs to respond to temperature control at a high frequency and dynamically adjust the damper opening and refrigerant flow in seconds), or when control commands need to be issued at a high frequency (the fan speed needs to match the air volume demand of the air conditioning unit in real time), a shorter original simulation cycle needs to be set to ensure that the simulation can capture the dynamic characteristics of the physical entity.
[0040] 2. When the physical state changes slowly (e.g., room temperature is affected by building thermal inertia and reaches steady state in minutes; flow regulation of manifolds in heating pipe networks is constrained by pipe network inertia and has a long change cycle), or when the control strategy is mainly based on long-cycle optimization (e.g., cooling tower group control strategy, which balances system heat load in 10-minute intervals), a longer original simulation cycle can be set to avoid redundant calculations.
[0041] Taking air conditioning system simulation as an example, the original simulation cycle for each entity's corresponding task can be determined as follows (this is just an example; in practice, it needs to be verified in conjunction with system design parameters and operating standards):
[0042] 1) Air conditioning unit: As the core unit of air handling, it needs to respond to changes in indoor and outdoor temperature and humidity at a high frequency. The original simulation cycle is set to 4 seconds.
[0043] 2) Fan: Provides airflow support for the air conditioning unit. The speed adjustment needs to be coordinated with the air conditioning unit. The original simulation cycle is set to 2 seconds.
[0044] 3) Manifold: Regulates the flow distribution of the water system. Affected by the inertia of the pipe network fluid, the state changes relatively slowly. The original simulation cycle is set to 8 seconds.
[0045] 4) Cooling tower: The heat of the system is balanced through heat exchange. The heat exchange process has a large inertia, and the original simulation cycle is set to 16 seconds.
[0046] 5) Room: The temperature is affected by the building envelope and the air conditioning delay, and it takes a long time to reach a steady state. The original simulation period is set to 30 seconds.
[0047] Priority determination rules: Priority reflects the real-time impact of tasks on the system state and their data dependencies, following the principle of "prioritizing high real-time tasks and prioritizing critical input tasks."
[0048] 1. If the task output data is a key input for other tasks (such as fan speed data being a core parameter of the air volume-temperature control model of the air conditioning unit), or if the task needs to respond quickly to abnormal states (such as the air conditioning unit fault diagnosis task, which needs to identify and report abnormalities first to avoid system failure), then set a high priority (the smaller the value, the higher the priority).
[0049] 2. If the task is an auxiliary calculation (such as a system energy consumption statistics task that does not directly participate in the real-time control closed loop), or depends on the calculation results of other tasks (such as room temperature distribution simulation that requires obtaining the air supply parameters of the air conditioning unit and the operating status of the fan), then set it to low priority (the larger the value, the lower the priority).
[0050] Taking air conditioning system simulation as an example, the priority of tasks corresponding to each entity can be set as follows (this is just an example; in practice, it needs to be verified in conjunction with system control logic and fault response requirements):
[0051] 1) Air conditioning unit: As the core control unit of the system, its status directly determines the air handling effect, and its priority is set to 0 (highest priority).
[0052] 2) Fan: Provides critical airflow input to the air conditioning unit, needs to operate in coordination with the air conditioning unit, and has a priority of 1;
[0053] 3) Manifold: Regulates the water system flow rate. It needs to match the coordinated control results of the air conditioning unit and fan. The priority is set to 2.
[0054] 4) Cooling tower: Balances the system's heat load and needs to be adjusted based on the comprehensive heat demand of units such as air conditioning units and manifolds. Priority is set to 3.
[0055] 5) Room: Presents the final control effect of the system, which depends on the calculation results of the preceding tasks. The priority is set to 4 (lowest priority).
[0056] Figure 3 This is a diagram showing the comparison before and after the task cycle was accelerated.
[0057] like Figure 3 As shown, in the air conditioning system simulation, the original simulation cycles for the tasks corresponding to the air conditioning unit, fan, manifold, cooling tower, and room can be set to 4 seconds, 2 seconds, 8 seconds, 16 seconds, and 30 seconds, respectively; the priorities are set to 0, 1, 2, 3, and 4, respectively.
[0058] S12, Acceleration Period Division: The original simulation period of each model is compressed proportionally by a uniform compression factor to obtain the simulation acceleration period of each model.
[0059] The setting of the compression factor needs to take into account both simulation efficiency requirements and numerical stability requirements:
[0060] 1. If you need to quickly generate simulation results (such as rapid algorithm verification or real-time teaching demonstration), you can choose a higher compression ratio (such as 10x or 20x).
[0061] 2. If it is necessary to prioritize simulation accuracy (such as the accuracy of the trend of key parameter changes), a lower compression ratio (such as 2x or 5x) can be selected to avoid distortion of system state updates due to excessively high compression ratio.
[0062] Taking air conditioning system simulation as an example (see Figure 3):
[0063] The original simulation cycle of the air conditioning unit is 4 seconds. After being compressed by 10 times, its simulation acceleration cycle is 4 seconds ÷ 10 = 0.4 seconds.
[0064] The original simulation cycle of the wind turbine is 2 seconds. After being compressed by 10 times, its simulation acceleration cycle is 2 seconds ÷ 10 = 0.2 seconds.
[0065] The other models (water manifold, cooling tower, room, etc.) all followed the same compression ratio, resulting in simulation acceleration cycles of 0.8 seconds, 1.6 seconds, and 3 seconds, respectively.
[0066] By using a uniform compression factor, we can ensure that the time evolution logic between models is consistent with the original physical system (such as the relative periodic relationship remains unchanged), and accelerate the overall simulation process by shortening the duration of a single period.
[0067] like Figure 3 As shown, in the simulation of the air conditioning system, the values of the air conditioning unit, fan, manifold, cooling tower, and room model in example S11 were selected. After being accelerated by 10 times, the simulation cycle became 0.4 seconds, 0.2 seconds, 0.8 seconds, 0.16 seconds, and 3 seconds respectively.
[0068] S13, Initial Parameter Setting: Set the initial simulation parameters for each model instance.
[0069] Set initial parameter values for each task participating in the simulation (i.e., instances of the model);
[0070] During the simulation, the calculations are actually performed on instances of each model, and initial values need to be set for multiple input parameters (if any) of each instance of the model.
[0071] S20, Accelerated Simulation of Slow Dynamic Systems: Based on the initial simulation parameters and priorities, accelerated simulation of slow dynamic systems is performed within the acceleration period to obtain the simulation results of each model instance.
[0072] like Figure 4 As shown, step S20, the slow dynamic system accelerated simulation step, includes:
[0073] S21, Simulation Cycle Setting: Set the simulation cycle termination condition. For example, set the simulation duration to 72 hours.
[0074] S22, Single-cycle model instance simulation: Within a single simulation acceleration cycle, the simulation of each model instance is completed through a closed loop of "triggering → scheduling → verification".
[0075] like Figure 5 As shown, the simulation steps for the 22 single-cycle arbitrary model instance include:
[0076] S221, Task Trigger: Start the virtual simulation process. The simulation system triggers task execution requests according to the timed simulation cycle of each task.
[0077] The aforementioned timing simulation cycle refers to the new simulation cycle after compressing the original simulation cycle of the model.
[0078] Figure 6 This is a schematic diagram of time-axis driven simulation.
[0079] like Figure 6 As shown, this illustrates how virtual simulation logic is driven by the system timeline. The computer executes the task scheduling for the current moment at times t0, t1, t2, etc.
[0080] S222, Scheduled task execution: Collect models according to task cycle; group models according to priority; execute priority groups sequentially and execute tasks in groups of the same priority in parallel;
[0081] Figure 7 This is a schematic diagram of the execution of the simulation after time acceleration.
[0082] like Figure 7 As shown, the left side displays the simulation cycle and priority settings for the model, while the right side shows the accelerated execution process. During the air conditioning system simulation, at t=0.2s, instances of all fan models are executed sequentially; at t=0.4s, instances of all air conditioning unit models are executed sequentially, followed by all fan instances; at t=0.6s, instances of all fan models are executed sequentially; at t=0.8s, instances of all air conditioning unit models are executed sequentially, followed by all fan instances; finally, instances of all manifolds are executed; at t=1.6s, all instances of these models are executed sequentially: air conditioning unit, fan, manifold, and cooling tower; at t=3s, all instances of these models are executed sequentially: air conditioning unit, fan, manifold, cooling tower, and room.
[0083] The Figure 7There is a special case where the simulation cycle for both the fan and the air conditioning unit is 0.2 seconds. Then, when t=0.2s, all instances of the fan model and the air conditioning unit are executed concurrently.
[0084] like Figure 8 As shown, step S222, the task execution step, includes:
[0085] S2221, Model set collection; Based on the model's preset simulation cycle and the timed triggering mechanism of the task (i.e., model instance), the models to be executed in the current cycle are selected to obtain the set of models to be executed in the current cycle;
[0086] S2222, Model Priority Grouping: Based on the numerical attribute of model priority, the models that need to be triggered are classified according to priority, and multiple groups of models with the same priority are generated.
[0087] S2223, Priority grouping serial execution: Based on the group priority value and preset sorting rules (from high to low), sort the groups according to priority and schedule the groups in priority order (from high to low);
[0088] S2224, Parallel execution of tasks in the same group: Take the highest priority group at present, and then collect the instances (i.e. tasks) of the model in the group and the parallel execution mechanism according to the model, and execute these tasks of the same priority concurrently.
[0089] S223, Current Cycle Verification: Based on the output results of the scheduled task execution steps in S222, verify the integrity of the current concurrent task execution. If it is complete, proceed to the next cycle; otherwise, output a fault.
[0090] S23, Simulation cycle determination: If the termination condition is not met (such as the simulation time target), proceed to S22; if the termination condition is met (such as the simulation time target), the simulation result of the model instance is obtained.
[0091] S30, Simulation Result Output: Outputs the simulation results for each model instance.
[0092] The simulation results of each model instance include a large number of simulation samples, system characteristics, and optimization basis.
[0093] Through multiple cycles, a massive number of simulation samples are generated, and the system characteristics and optimization basis are output.
Claims
1. A simulation method for accelerating task execution based on multi-cycle priority, characterized in that, Includes the following steps: S10, Slow Dynamic System Simulation Initialization: Determine the original simulation cycle and priority of the slow dynamic system, divide the simulation acceleration cycle, and set the initial simulation parameters; S20, Accelerated Simulation of Slow Dynamic Systems: Based on the initial simulation parameters and priorities, accelerated simulation of slow dynamic systems is performed within the acceleration period to obtain the simulation results of each model instance; S30, Simulation Result Output: Outputs the simulation results for each model instance.
2. The simulation method according to claim 1, characterized in that, The S10 slow dynamic system simulation initialization step includes: S11, Determination of Original Simulation Cycle and Priority: For each model in the virtual simulation of a slow dynamic system, the original simulation cycle and priority of each model are determined based on the dynamic characteristics of its corresponding entity in actual operation. S12, Acceleration cycle division: The original simulation cycle of each model is compressed proportionally by a uniform compression factor to obtain the simulation acceleration cycle of each model. S13, Initial Parameter Setting: Set the initial simulation parameters for each model instance.
3. The simulation method according to claim 2, characterized in that, S20, the slow dynamic system accelerated simulation step includes: S21, Simulation loop period setting: Set the simulation loop termination condition; S22, Single-cycle model instance simulation: Within a single simulation acceleration cycle, the simulation of each model instance is completed through a closed loop of "triggering → scheduling → verification"; S23, Simulation Cycle Determination: Determine whether the simulation cycle meets the termination condition; if the simulation cycle does not meet the termination condition, proceed to S22; if the termination condition is met, obtain the simulation results of the model instance.
4. The simulation method according to claim 3, characterized in that, The simulation steps for the single-period model instance mentioned in point 22 include: S221, Task Trigger: Start the virtual simulation process. The simulation system triggers task execution requests according to the timed simulation cycle of each task. S222, Scheduled task execution: Collect models according to task cycle; group models according to priority; execute priority groups sequentially and execute tasks in groups of the same priority in parallel; S223, Current Cycle Verification: Based on the output results of the scheduled task execution steps in S222, verify the integrity of the current concurrent task execution. If it is complete, proceed to the next cycle; otherwise, output a fault.
5. The simulation method according to claim 4, characterized in that, S222, the task execution steps include: S2221, Model set collection; Based on the model's preset simulation cycle and the task's timed triggering mechanism, the models to be executed in the current cycle are selected to obtain the set of models to be executed in the current cycle; S2222, Model Priority Grouping: Based on the numerical attribute of model priority, the models that need to be triggered are classified according to priority, and multiple groups of models with the same priority are generated. S2223, Priority Grouping Serial Execution: Based on the group priority value and preset sorting rules, sort the groups according to priority and schedule the groups in priority order; S2224, Parallel execution of tasks in the same group: Take the highest priority group at present, and then collect the instances of the model in the group and the parallel execution mechanism according to the model, and execute these tasks of the same priority concurrently.