A method and system for cooperative operation control of a forest fire pump group

Abstract

The present application relates to the technical field of industrial control system, specifically to a kind of forest fire pump group's collaborative operation control method and system, comprising the following steps: decomposition forest fire pump group collaborative operation process, aggregate all local pump controller's discretization control task list, establish whole system real-time task set.The present application is by forest fire pump group collaborative operation process discretization into task level control unit, and constructs complete whole system real-time task set, enhances the structural clarity and operation accuracy of task timing scheduling.With quantitative extraction and contrast execution to task resource allocation parameters, a clear priority order is assigned to the tasks, and accurate scheduling boundaries are provided by the deterministic execution time window table, allowing each pump controller to respond collaboratively within a strict time limit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial control system, and particularly relates to a cooperative operation control method and system of forest fire pump group. BACKGROUND

[0002] The cooperative operation control method of forest fire pump group is a real-time control method for cooperative operation of multiple fire pump devices in the process of forest fire extinguishing, aiming to realize efficient, safe and continuous operation of the fire pump group in the fire environment by coordinating the operating states of the fire pumps and uniformly scheduling parameters such as engine speed, water outlet pressure and resource allocation of the devices.

[0003] Although the prior art has realized the uniform scheduling of the operating states of multiple fire pump devices, it only sets the operating elements such as engine speed and water outlet pressure based on static parameters, lacks a dynamic adjustment mechanism for the operation task rhythm and control priority, causes the problem of non-uniform scheduling rhythm and waste of control resources, and easily causes execution conflicts or response lag in high-frequency task or multi-pump linkage scenarios. Therefore, improvement is needed. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art and provide a cooperative operation control method and system of forest fire pump group.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme, a cooperative operation control method of forest fire pump group, comprising the following steps:

[0006] Decompose the cooperative operation process of forest fire pump group, aggregate the discretized control task list of all local pump controllers, and establish a full-system real-time task set;

[0007] Based on the full-system real-time task set, obtain a task resource allocation original parameter set, compare the task resource allocation original parameter set, assign an execution order to each task, and generate a deterministic execution time window table;

[0008] According to the time point set by the deterministic execution time window table, obtain the measurement values of fuel level, water pump outlet pressure, water pump engine speed and water source level from each local pump controller, calculate the fuel consumption rate through the fuel level change amount and time interval, calculate the device load through the water pump outlet pressure, obtain a multi-dimensional resource depletion time sequence, select the minimum value from the multi-dimensional resource depletion time sequence, and calculate the combat sustainability quantization score;

[0009] The received minimum sustainability threshold is continuously compared with the current combat sustainability quantification score, while evaluating the expected impact of adjusting the water pump engine speed or water pump outlet pressure of individual local pump controllers on the combat sustainability quantification score, forming a control strategy adjustment candidate set, when the combat sustainability quantification score is lower than the minimum sustainability threshold, selecting the adjustment scheme that makes the score rebound and has the least impact on the total water output from the control strategy adjustment candidate set, generating the cooperative operation instruction set under logistics constraints.

[0010] Preferably, the obtaining step of the full-system real-time task set is:

[0011] The cooperative operation process of the forest fire pump group is analyzed, and data acquisition actions, state estimation actions, control law calculation actions, network communication actions, and actuator driving actions contained in the process are identified one by one, the process is decomposed into independent basic action units, and the operation content of each action unit is determined one by one to obtain a set of independent actions of the forest fire pump group;

[0012] Based on the set of independent actions of the forest fire pump group, a fixed execution period is set for each independent action unit, the worst-case execution time of each independent action unit is estimated and confirmed, the execution priority of each action unit is determined according to the importance of real-time control, and a strict execution deadline is set for each action unit to generate a discretized control task list for a single pump controller;

[0013] Based on the discretized control task list of the single pump controller, the discretized control task lists of all local pump controllers are integrated and merged, sorted according to the task execution period, worst-case execution time, and priority, to obtain a full-system real-time task set.

[0014] Preferably, the obtaining step of the task resource allocation original parameter set is:

[0015] Based on the full-system real-time task set, all control tasks and network communication tasks are extracted according to local pump controller nodes, the worst-case execution time and execution period corresponding to each control task are recorded, the data packet size and sending period corresponding to each communication task are recorded, and the maximum processing capacity and maximum network bandwidth preset for each node are read to generate a task resource allocation original parameter set for each local pump controller node.

[0016] Preferably, according to the task resource allocation original parameter set, the system comprehensive load of the local pump controller node is calculated;

[0017] The system comprehensive load value calculated by each local pump controller node is compared with the preset resource capacity threshold of each node, the task set whose load value exceeds the upper limit of the threshold is filtered out, and the priority is sorted according to the task deadline and the cycle stability, so as to generate a deterministic execution time window table meeting the real-time constraint.

[0018] Preferably, the acquisition step of the multi-dimensional resource depletion time sequence is:

[0019] Based on the time point set by the deterministic execution time window table, the fuel level change value, the water pump outlet pressure, the water pump engine speed and the water source level of each local pump controller node are collected, and the fuel consumption rate, the outlet pressure change rate, the speed reduction rate and the water level drop rate in the current period are calculated respectively by combining the historical average sampling period of each type of measurement data, so as to obtain the multi-dimensional resource depletion time sequence.

[0020] Preferably, the acquisition step of the operation sustainability quantitative score is:

[0021] According to the multi-dimensional resource depletion time sequence, the independent depletion time of the four types of resources at the node is calculated by using the current resource remaining amount divided by the change rate for the fuel level, the water pump outlet pressure, the water pump engine speed and the water source level respectively, the independent depletion time of the four types of resources is arranged in parallel to form the resource depletion time sequence of the current node, the minimum value is selected as the shortest resource depletion time, and the corresponding resource type and consumption rate are recorded to obtain the shortest depletion time and the normalized consumption rate.

[0022] Based on the shortest depletion time and the normalized consumption rate, the operation sustainability quantitative score is calculated.

[0023] Preferably, the acquisition step of the control strategy adjustment candidate set is:

[0024] Based on the lowest sustainability threshold and the operation sustainability quantitative score, the numerical comparison is continuously carried out at a fixed time interval to determine whether the operation sustainability quantitative score is less than the lowest sustainability threshold, and the comparison time and the numerical difference are recorded to generate a sustainability threshold real-time comparison sequence;

[0025] Based on the sustainability threshold real-time comparison sequence, a single node is selected from all local pump controller nodes one by one, the water pump engine speed and the water pump outlet pressure are adjusted slightly respectively, the change trend of the operation sustainability quantitative score before and after the adjustment is simulated and calculated, the expected influence of each adjustment scheme on the operation sustainability quantitative score and the change of the total water output are evaluated, and a control strategy adjustment candidate set is formed.

[0026] Preferably, the acquisition step of the cooperative operation instruction set under the logistics constraint is:

[0027] adjustment candidate set, when the latest operational sustainability quantification score in the real-time comparison sequence is lower than the minimum sustainability threshold, compare the recovery amplitude of the operational sustainability quantification score and the change of the total water output of each scheme after adjustment, select the adjustment scheme that meets the threshold requirement in the recovery amplitude of the operational sustainability quantification score and has the minimum change of the total water output, and generate the cooperative operation instruction set under the logistics constraint in combination with the deterministic execution time window table.

[0028] The application also provides a cooperative operation control system, comprising:

[0029] A task set construction module, which decomposes the cooperative operation process of the forest fire pump group, aggregates the discretized control task list of all local pump controllers, and establishes a full-system real-time task set;

[0030] A task scheduling module, which obtains a task resource allocation original parameter set based on the full-system real-time task set, compares the task resource allocation original parameter set, allocates an execution sequence for each task, and generates a deterministic execution time window table;

[0031] A resource evaluation module, which obtains the measurement values of the fuel level, the water pump outlet pressure, the water pump engine speed and the water source level from each local pump controller at the time points set by the deterministic execution time window table, calculates the fuel consumption rate through the fuel level change and the time interval, calculates the equipment load through the water pump outlet pressure, obtains a multi-dimensional resource depletion time sequence, selects the minimum value from the multi-dimensional resource depletion time sequence, and calculates the operational sustainability quantification score;

[0032] A control optimization module, which continuously compares the received minimum sustainability threshold with the current operational sustainability quantification score, simultaneously evaluates the expected influence of adjusting the water pump engine speed or the water pump outlet pressure of a single local pump controller on the operational sustainability quantification score, forms a control strategy adjustment candidate set, and when the operational sustainability quantification score is lower than the minimum sustainability threshold, selects the adjustment scheme that has the minimum influence on the total water output and makes the score recover from the control strategy adjustment candidate set, and generates the cooperative operation instruction set under the logistics constraint.

[0033] Compared with the prior art, the application has the advantages and positive effects that:

[0034] The present application enhances the structural clarity and operational accuracy of task timing scheduling by discretizing the forest fire pump group cooperative operation process into task-level control units and constructing a complete full-system real-time task set. With quantitative extraction and comparative execution of task resource allocation parameters, a clear priority order is assigned to the tasks, and the precise scheduling boundaries provided by the deterministic execution time window table enable each pump controller to respond cooperatively within a strict time limit. In the operation monitoring phase, four types of core resource data, including fuel level, water pump outlet pressure, water pump engine speed and water source level, are collected in real time and converted into calculable physical indicators such as fuel consumption rate and equipment load, forming a multi-dimensional depletion time sequence. The shortest depletion time is selected to construct a sustainability quantification score with comprehensive indicators such as predictability, continuity and compressibility. The sustainability score result is continuously compared with the minimum threshold value. When there is a risk trend, the speed or outlet pressure of a single controller is immediately adjusted through simulation, and the optimal intervention strategy is selected based on the score change trend and total water output influence analysis, ensuring that the control result takes into account the performance recovery and stability of the water output capacity. The whole process establishes a close link from the task layer to the scheduling layer, from real-time collection to strategy adjustment, from resource estimation to response optimization, which enhances the system cooperation, improves the predictability of resource scheduling, enhances the efficiency of risk response and extends the guarantee time. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 The present application provides a technical solution, a cooperative operation control method of a forest fire pump group, comprising the following steps: DETAILED DESCRIPTION

[0036] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0037] Please refer to Figure 1 The present application provides a technical solution, a cooperative operation control method of a forest fire pump group, comprising the following steps:

[0038] Decompose the cooperative operation process of the forest fire pump group, aggregate the discretized control task list of all local pump controllers, and establish a full-system real-time task set;

[0039] Based on the full-system real-time task set, obtain the original parameter set of task resource allocation, compare the original parameter set of task resource allocation, assign the execution order to each task, and generate a deterministic execution time window table;

[0040] According to the time point set by the deterministic execution time window table, the measurement values of fuel level, water pump outlet pressure, water pump engine speed and water source level are obtained from each local pump controller, the fuel consumption rate is calculated by the fuel level change amount and time interval, the equipment load is calculated by the water pump outlet pressure, the multi-dimensional resource depletion time sequence is obtained, the minimum value is selected from the multi-dimensional resource depletion time sequence, and the combat sustainability quantitative score is calculated;

[0041] The received minimum sustainability threshold is continuously compared with the current combat sustainability quantitative score, and the expected impact of adjusting the water pump engine speed or water pump outlet pressure of a single local pump controller on the combat sustainability quantitative score is evaluated to form a control strategy adjustment candidate set, when the combat sustainability quantitative score is lower than the minimum sustainability threshold, the adjustment scheme that makes the score rebound and has the least impact on the total water output is selected from the control strategy adjustment candidate set, and the cooperative operation instruction set under logistics constraints is generated.

[0042] The acquisition step of the whole system real-time task set is:

[0043] The cooperative operation process of the forest fire pump group is analyzed, the data acquisition action, state estimation action, control law calculation action, network communication action and actuator driving action contained in the process are identified one by one, the process is decomposed into independent basic action units, and the operation content of each action unit is determined one by one to obtain the independent action set of the forest fire pump group;

[0044] Based on the independent action set of the forest fire pump group, a fixed execution cycle is set for each independent action unit, the worst-case execution time of each independent action unit is estimated and confirmed, the execution priority of each action unit is determined according to the importance of real-time control, and a strict execution deadline is set for each action unit to generate the discretized control task list of a single pump controller;

[0045] Based on the discretized control task list of a single pump controller, the discretized control task lists of all local pump controllers are integrated and merged, and are sorted according to the task execution cycle, worst-case execution time and priority to obtain the whole system real-time task set.

[0046] Specifically, first, the pre-prepared forest fire pump group collaborative operation process document is called, which records in detail the whole process from equipment start-up, pipeline pressurization, stable water supply to collaborative pressure regulation and final shutdown of the pump in the form of a flowchart. By functionally disassembling each link in the flowchart, five different types of basic actions are identified one by one, namely data acquisition action, state estimation action, control law calculation action, network communication action and actuator driving action. For example, in the "pipeline pressurization link", "reading the water pump outlet pressure sensor" is identified as the data acquisition action, "calculating the difference between the current pipeline pressure and the target pressure" is identified as the state estimation action, "calculating the engine throttle adjustment amount according to the pressure difference" is identified as the control law calculation action, "broadcasting the local pressure value to the adjacent pump" is identified as the network communication action, and "sending a PWM signal to the engine throttle controller" is identified as the actuator driving action. Subsequently, these identified actions are further refined into independent basic action units that cannot be further divided, and are assigned unique identifiers and explicit operation contents. For example, "reading the water pump outlet pressure sensor" is refined into "action unit 001: sending a read instruction to the pressure sensor (model MSP300) through the SPI bus and receiving 16-bit pressure data returned", and "sending a PWM signal to the engine throttle controller" is refined into "action unit 002: outputting a pulse width modulation signal with a period of 20 milliseconds and adjustable duty cycle to the microcontroller GPIO pin (e.g. PB4)". The identification and refinement process is repeated for all links in the flowchart until the entire operation process is completely decomposed. Finally, all defined independent basic action units are collected to form a list containing the unique identifier of each action unit, the action category it belongs to, the specific hardware interface, the communication protocol and the operation instruction details, obtaining the forest fire pump group independent action set.

[0047] Based on the independent action set of the forest fire pump group, the time characteristic parameters of each independent action unit in the list are configured. First, a fixed execution period is set for each independent action unit. The setting of the period is based on the dynamic characteristics of its physical process and the control requirements. For example, for "action unit 001: read water pump outlet pressure sensor", since high real-time performance is required for pressure control, the execution period is set to 50 milliseconds. For "read fuel level" and other slow-changing state monitoring, the execution period can be set to 2000 milliseconds. Then, the worst-case execution time (WCET) of each independent action unit is estimated and confirmed by repeatedly executing tests on the target hardware platform (for example, a microcontroller with ARM Cortex-M7 core). The specific method is to execute each action unit continuously not less than 10000 times, and use the high-precision timer of the hardware to record the time consumption of each execution to form a time consumption data set. Then, the maximum observed execution time and the standard deviation of the data set are calculated. The final confirmation value of the worst-case execution time (WCET) is determined by a calculation containing a safety margin. The calculation method is to multiply the maximum observed execution time by a safety factor, and then add the standard deviation multiplied by a risk factor. For example, if the maximum observed execution time of a certain action unit is 85 microseconds, the standard deviation is 5 microseconds, the safety factor is set to 1.2, and the risk factor is set to 2.0, then the worst-case execution time is confirmed to be 1.2*85+2.0*5=112 microseconds. Subsequently, the execution priority of the action unit is determined according to its influence on the safety and stability of the system. The priority is divided into three levels: high, medium and low, and is mapped to numerical values 1 to 255 (the smaller the value, the higher the priority). For example, the "actuator driving action" directly controlling the water pump operation and the "control law calculation action" of the core are assigned high priority of 1 to 50, the actions for state estimation and regular communication are assigned medium priority of 51 to 150, and the auxiliary actions such as data recording and display updating are assigned low priority of 151 or more. Finally, a strict execution deadline is set for each action unit. For hard real-time tasks such as pressure control related tasks, the execution deadline is strictly set to equal to the execution period to ensure the predictability of the system. Each independent action unit and its corresponding execution period, worst-case execution time, execution priority and execution deadline are combined into a task record, and all task records together constitute the discrete control task list of a single pump controller.

[0048] Based on the discretization control task list generated by the local pump controller deployed on each water pump, the integration merging process is started, which first collects and summarizes the discretization control task list of all local pump controllers (for example, there are N pumps in the system, that is, N controllers). Since the basic task set of each pump controller is the same, the core of the integration process is to distinguish between local execution tasks and shared resource (such as wireless communication channel) competition tasks in the entire system, and add a source node identifier (such as "pump ID_01", "pump ID_02") to each task record, so as to distinguish the "read water pump outlet pressure sensor" task of "pump ID_01" from the task of the same name of "pump ID_02", and form a preliminary global task list containing all tasks of all nodes. Then, the preliminary global task list is reordered according to the key attributes of the tasks to build the final system real-time task set. The sorting adopts multiple criteria, the first sorting criterion is to arrange the tasks in order from high to low (that is, from small to large) according to the execution priority value in the task record, to ensure that the tasks that are crucial to the safety and stability of the system are given the highest priority. In the case of the same execution priority, the second sorting criterion is to compare the execution period of the task, and the task with shorter execution period is given higher actual scheduling priority, which is consistent with the principle of rate monotonic scheduling strategy. Therefore, the tasks with shorter execution period are arranged in front. If the execution priority and execution period are the same, the third sorting criterion is used, that is, the worst-case execution time of the task is compared, and the task with shorter worst-case execution time is arranged in front, so as to facilitate the scheduler to complete more tasks in a limited time and reduce system load. Through the application of the three sorting rules in turn, all tasks in the preliminary global task list are uniquely and deterministically sorted, and a structured and globally ordered list containing all tasks to be processed and their scheduling attributes in the system is finally generated, and the system real-time task set is obtained.

[0049] The task resource allocation original parameter set acquisition step is:

[0050] Based on the system real-time task set, all control tasks and network communication tasks are extracted according to the local pump controller node classification, the worst-case execution time and execution period of each control task are recorded, the data packet size and sending period of each communication task are recorded, and the maximum processing capacity and maximum network bandwidth preset by each node are read to generate the task resource allocation original parameter set of each local pump controller node.

[0051] Specifically, based on the real-time task set of the whole system, each task record in the set is traversed, and the source node identifier (such as "pump ID_01") contained in the task record is classified and aggregated. Tasks belonging to the same local pump controller are collected together. Next, in the task aggregation of each node, the tasks are classified again according to their properties. Specifically, by consulting the "action category" field in the task record, tasks with categories such as "data collection action", "state estimation action", "control law calculation action", or "actuator driving action" are classified as control tasks, and tasks with the category "network communication action" are classified as network communication tasks. For all entries classified as control tasks, the system extracts and records the "worst-case execution time" and "execution period" parameters. For all entries classified as network communication tasks, the system extracts the "packet size" and "sending period" parameters. For example, for the "pump ID_01" node, the "read water pump outlet pressure sensor" task is classified as a control task with a worst-case execution time of 112 microseconds and an execution period of 50 milliseconds, and the "broadcast local state to adjacent pump machines" task is classified as a network communication task with a packet size of 64 bytes and a sending period of 100 milliseconds. After extracting all task parameters, the system reads two key preset performance indicators, maximum processing capacity and maximum network bandwidth, from the hardware configuration file of each local pump controller node. The maximum processing capacity is quantified in millions of instructions per second (MIPS) according to the core processing performance of the microprocessor (such as the STM32H7 series) used by the controller, and a conservative value is set considering the overhead of the operating system and background services in actual applications, for example, the rated value is 120 MIPS, and the preset value is 100 MIPS. The maximum network bandwidth is determined according to the theoretical maximum data transmission rate of the wireless communication module (such as the SX1262 chip based on LoRa technology) under the current configuration (spreading factor, coding rate, bandwidth), for example, the theoretical rate is 293 kbps under a specific configuration, and the maximum network bandwidth is preset to 250 kbps considering channel interference and protocol overhead. The extracted control task parameter set, network communication task parameter set, and read maximum processing capacity and maximum network bandwidth are integrated together to form a structured data set for each node. Finally, after all the structured data sets of the nodes are aggregated, the task resource allocation original parameter set for each local pump controller node is generated.

[0052] The acquisition step of the deterministic execution time window table is:

[0053] According to the task resource allocation original parameter set, the system comprehensive load of the local pump controller node is calculated, and the calculation formula is:

[0054]

[0055] wherein, represents the system comprehensive load value of the kth local pump controller node, represents the computing resource occupancy of the kth local pump controller node, the worst-case execution time of the jth task is T kj , the corresponding task period is P kj , and the total number of tasks is m k , represents the communication resource occupancy, the packet size of the ith communication task is S ki , the sending period is I ki , and the preset maximum network bandwidth is the number of communication tasks is r k , represents the nonlinear interaction load effect between computing and communication resources, w calc , w comm , and w int are dimensionless normalized weighting coefficients corresponding to the computing resource, the communication resource, and the coupling effect, respectively, and satisfy the constraint condition w calc +w comm +w int =1.

[0056] Based on the system comprehensive load value calculated by each local pump controller node, the preset resource capacity threshold of each node is compared item by item, the task set whose load value exceeds the upper limit of the threshold is screened out, and the priority is sorted according to the task deadline and the period stability, to generate a deterministic execution time window table that meets the real-time constraint.

[0057] Specifically, the formula is: The beneficial effect of the formula is that a nonlinear interaction term The interaction term can quantify the additional system overhead caused by the mutual interference of concurrent execution of computing tasks and communication tasks, for example, frequent interrupts triggered by high-intensity network communication will preempt CPU time, thereby affecting the execution efficiency of computing tasks, and vice versa, intensive computing tasks may also cause delay in the processing of network packets by the operating system. This coupling effect is often ignored in traditional linear load models, resulting in underestimation of the real load of the system. By introducing the interaction term, the present method can more accurately depict the complex load characteristics of the embedded real-time system.

[0058] w calc , w comm , w intThese are dimensionless, normalized weighting coefficients. The coefficients are set based on offline calibration experiments of the system's performance under different load combinations. During the experiments, the inflection points of system performance were tested by gradually increasing pure computational and pure communication tasks, and the changes in system throughput and task execution latency under mixed loads were recorded. Based on the collected experimental data, a multiple linear regression analysis method was used to fit the relationship between load components and the overall system performance indicators, thereby determining the optimal weight allocation. For example, in a specific calibration test, a performance evaluation function P is set. perf =a·(1-U calc )+b·(1-U comm )+c·(1-U calc (1-U) comm The coefficients a, b, and c are obtained by fitting a large number of test data points using the least squares method, and then normalized to obtain the weights. A specific calculation example is as follows: The original coefficients obtained through regression analysis are a = 3.0, b = 1.5, and c = 0.5, with a total sum of 3.0 + 1.5 + 0.5 = 5.0. Therefore, the normalized weights are w. calc =3.0 / 5.0 = 0.6, w comm =1.5 / 5.0=0.3, w int =0.5 / 5.0=0.1. This set of weights indicates that in this system, the computational load has the greatest impact on the overall performance, followed by the communication load. The coupling effect between the two also accounts for a certain proportion, and the constraint condition 0.6+0.3+0.1=1 is satisfied.

[0059] This represents the computational resource utilization of the k-th local pump controller node, and its calculation depends on the worst-case execution time T of all control tasks on that node. kj and task cycle P kj These parameters are extracted directly from the original parameter set of task resource allocation generated in the previous step, without the need for remeasurement. For example, for node k=1, two control tasks are extracted from its parameter set: Task 1 (stress monitoring), T 11 = 0.000112 seconds, P 11 = 0.05 seconds; Task 2 (Fuel Calculation), T 12 = 0.000250 seconds, P 12 =2.0 seconds, the computing resource utilization of this node is the sum of the utilization of these two tasks, and the calculation process is as follows:

[0060] This represents the communication resource utilization rate of the k-th local pump controller node, and its calculation requires the data packet size S of all communication tasks of that node. ki Sending period Iki and preset maximum network bandwidth These parameters are also directly obtained from the original parameter set of task resource allocation, and the units should be unified during calculation. Usually, the packet size is converted from byte (Byte) to bit (bit), and the network bandwidth is converted from kbps to bps. For example, for node k = 1, the maximum network bandwidth is 250 kbps, i.e. 250000 bps, and this node has a communication task: task 1 (state broadcast), the packet size S 11 = 64 bytes, and the sending period I 11 = 0.1 seconds. First, the total data sending rate is calculated: ∑(S ki / I ki ) = (64 bytes / 0.1 seconds) = 640 bytes / second, then the rate is converted to bit rate: 640*8 = 5120 bps, and finally the communication resource occupation rate is calculated:

[0061] Calculation process:

[0062] Taking the first local pump controller node (k = 1) as an example, the specific values obtained according to the above parameter acquisition steps are brought into the formula for calculation.

[0063] The known parameters are as follows:

[0064] Calculation of resource occupation rate

[0065] Calculation of communication resource occupation rate

[0066] Normalization weighting coefficient w calc = 0.6, w comm = 0.3, w int = 0.1;

[0067] The calculation process is as follows:

[0068]

[0069]

[0070]

[0071]

[0072] The result shows that the system comprehensive load value of the first local pump controller node is about 0.0076, which is a dimensionless value, intuitively reflecting the resource pressure level of the node under the current task allocation. This value is much smaller than 1, indicating that the computing and communication resources of the node are very abundant, and the system is in an extremely low load state. In the subsequent steps, this value will be compared with the preset resource capacity threshold (for example, 0.85). Since 0.0076 is much smaller than 0.85, it is determined that the resource load of the node is within the safe range.

[0073] Based on the system comprehensive load value calculated by each local pump controller node, the system will read these load values one by one and compare them with the resource capacity threshold preset by each node, which is not fixed but dynamically set according to the hardware performance of the controller and the criticality level of the task, and its setting benchmark comes from a large number of stress tests on the controller under full load operating conditions, by observing at what comprehensive load level the task response delay starts to exceed its deadline, and multiplying this critical load value by a safety factor (e.g. 0.9) to determine it finally, for example, if the test shows that a certain type of controller starts to frequently appear task timeout when the system comprehensive load value reaches 0.95, then its resource capacity threshold is set to 0.95 x 0.9 = 0.855, when comparing, if the system comprehensive load value of a certain node, for example 0.91, is greater than its set threshold 0.855, then the task set of the node is marked as "overloaded", the system will start the task screening logic, which will operate according to the priority of the task, first select the task with the lowest priority (such as log recording or status display update task with priority value between 200 and 255) from the task list of the node, and temporarily move it into a pending queue, then recalculate the system comprehensive load value after removing these tasks, if the new load value is lower than the threshold, the screening process ends, otherwise, continue to select the task with the second lowest priority from the remaining tasks for removal, until the load meets the requirements, for all nodes whose load values are within the threshold range and their adjusted nodes after screening, their task sets are considered as valid task sets, next, the system will uniformly prioritize these valid task sets to build the final schedule, the first criterion for sorting is the deadline of the task, the more urgent the deadline, the higher the ranking, following the "earliest deadline first" (EDF) scheduling principle, if the deadlines of multiple tasks are the same, the second sorting criterion is enabled, which is the cycle stability requirement of the task, for tasks with high cycle stability requirements (such as PWM signal generation tasks that need to accurately control the speed of the water pump, any jitter in the execution time of which may cause output pressure fluctuations), they are given higher sorting positions, through such comparison, screening and multi-criteria sorting process, finally a detailed task execution sequence is generated for each pump group node, which clearly shows the execution order of each task, and generates a deterministic execution time window table that meets the real-time constraints.

[0074] The acquisition step of the multi-dimensional resource depletion time sequence is:

[0075] Based on the time point set by the deterministic execution time window table, the fuel level change value, the water pump outlet pressure, the water pump engine speed and the water source level of each local pump controller node are collected, and the average sampling period of each type of measurement data is combined to calculate the fuel consumption rate, the outlet pressure change rate, the speed reduction rate and the water level drop rate in the current period, respectively, to obtain a multi-dimensional resource depletion time sequence.

[0076] Specifically, according to the execution time point of each task item in the deterministic execution time window table, the system scheduler will accurately trigger the corresponding data collection action at the specified time. Specifically, when the time stamp of the scheduler matches the next execution time point of the "collect fuel level" task, the microcontroller will send a read instruction to the ultrasonic liquid level sensor installed in the fuel tank through the I2C or SPI bus, the sensor returns an original data representing the current liquid level, the data is converted to get the fuel level value (unit: liter) after calibration, and is stored in a ring buffer in the memory together with the time stamp at the collection time. Similarly, the system will trigger the reading of the water pump outlet pressure sensor at a higher frequency (for example, every 50 milliseconds) set by the window table, the voltage output of the pressure sensor is obtained through the analog-to-digital converter (ADC), and the current water pump outlet pressure value (unit: MPa) is calculated according to the linear relationship of the sensor (for example, 4-20mA corresponds to 0-1.6MPa). The speed of the water pump engine is monitored by the Hall effect sensor or optical encoder installed on the flywheel, and the controller obtains the real-time speed (unit: RPM) by calculating the number of pulses received per unit time. For the water source level, a similar ultrasonic or float type sensor is used for remote or local measurement (unit: meters). In each collection period, when a new measurement value is obtained, the system will immediately retrieve the last valid reading and its time stamp from the ring buffer, and calculate the average change rate by dividing the difference between the current value and the last value by the time interval between the two measurements (directly obtained by subtracting the time stamps, unit: seconds). For example, the fuel level is 45.5 liters at time t1, and the fuel level is 45.3 liters at time t2, the time interval is 120 seconds, then the fuel consumption rate is calculated as (45.5-45.3) / 120=0.00167 liters / second. This process is parallel processed for all four types of measurement data, and the fuel consumption rate, outlet pressure change rate (may be positive or negative), speed reduction rate (usually non-positive) and water level drop rate are calculated respectively. The four real-time calculated rate values are used as a data vector, associated with the current time stamp, continuously generated and updated to obtain a multi-dimensional resource depletion time sequence.

[0077] The acquisition step of the combat sustainability quantification score is:

[0078] According to the multi-dimensional resource depletion time sequence, the current resource remaining amount divided by the change rate is used to calculate the independent depletion time of the four types of resources under the node, respectively, including the fuel level, the water pump outlet pressure, the water pump engine speed and the water source level. The independent depletion time of the four types of resources is arranged in parallel to form the resource depletion time sequence of the current node, and the minimum value is selected as the shortest resource depletion time. The corresponding resource type and consumption rate are recorded to obtain the shortest depletion time and normalized consumption rate.

[0079] Based on the shortest depletion time and the normalized consumption rate, the quantitative score of the operational sustainability is calculated, and the calculation formula is:

[0080]

[0081] Wherein, C s is the quantitative score of the operational sustainability (dimensionless, ranging from 0 to 100), D min is the shortest depletion time (unit: second), ρi min is the normalized dynamic consumption rate of the resource type causing the shortest depletion time, which is equal to the current instantaneous consumption rate divided by the historical average consumption rate (dimensionless), α is the dynamic risk sensitivity coefficient (dimensionless, indicating the degree of punishment), τ is the reference task response period (unit: second), D eff is the risk-adjusted effective depletion time.

[0082] Specifically, according to the latest set of change rate data in the real-time updated multi-dimensional resource depletion time series, the system predicts the continuous supply capability of four key resources (fuel, water pressure, engine speed, water source) of each local pump controller node. First, the system obtains the current remaining amount of each resource. For fuel, the current fuel level value is directly read as the remaining amount (e.g., 45.3 liters). For the water level of the water source, the current reading of the water level sensor is subtracted by a preset minimum available water level reference value (e.g., current water level 3.5 meters, minimum reference 1.0 meter, then remaining amount is 2.5 meters). For performance indicators such as water pump outlet pressure and engine speed, the "remaining amount" is defined as the difference between the current value and the minimum threshold required to maintain operation. For example, if the set minimum outlet pressure is 0.8 MPa and the current pressure is 1.2 MPa, the pressure "remaining amount" is 0.4 MPa. If the set minimum engine speed is 1500 RPM and the current speed is 2200 RPM, the speed "remaining amount" is 700 RPM. Then, the system divides the remaining amount of each resource by the current change rate of the corresponding resource obtained from the multi-dimensional resource depletion time series to calculate the independent depletion time of each resource. If the change rate is zero or causes the resource to increase (e.g., pressure rises), its depletion time is considered infinite. For example, the fuel remaining amount is 45.3 liters and the consumption rate is 0.00167 liters / second, so the independent depletion time of the fuel is 45.3 / 0.00167 ≈ 27125 seconds. The calculated independent depletion times of the four types of resources (fuel depletion time, pressure failure time, speed failure time, water source depletion time) form a vector containing four values, which is the resource depletion time series of the current node. The system traverses this sequence and selects the smallest positive value in it. This value is determined as the shortest resource depletion time of the current node. At the same time, the system records the resource type (e.g., "fuel") that leads to this shortest time and its corresponding consumption rate (0.00167 liters / second). Finally, the system calculates the normalized value of the consumption rate by dividing the current instantaneous consumption rate by the average consumption rate of the resource under similar historical conditions (e.g., the past hour) (e.g., the historical average is 0.00150 liters / second), obtaining a dimensionless ratio (1.11). The shortest depletion time and the normalized consumption rate are obtained.

[0083] Formula: The advantage of the formula is that it evaluates the operational sustainability by introducing the risk-adjusted effective depletion time D eff , instead of simply using the shortest depletion time D min linearly extrapolated. The calculation of D eff takes into account the abnormality of resource consumption, which is quantified by the normalized dynamic consumption rate . When the resource consumption rate suddenly far exceeds the historical average (

[0084] ), D eff will be significantly smaller than D min , thus leading to a sustainability score C s that drops sharply, which is equivalent to a built-in "risk amplifier" that provides early warning of potential failures (such as fuel leaks, pipe breakage) or unexpected large load changes, much earlier than the resource stock itself drops to a dangerous level, and the dynamic risk sensitivity coefficient a provides a configurable means of tuning the sensitivity of this early warning, the final score formula adopts an exponential decay form, so that the score stays at a high level when D eff is large, but drops rapidly when D eff is close to the reference period τ, which is consistent with the non-linear nature of risk perception.

[0085] D min is the minimum depletion time, which represents the time that the most scarce resource in the system is expected to sustain operation at the current consumption rate, it is directly obtained from the calculation result of the previous step, and is the minimum value in the resource depletion time series, the acquisition of this parameter does not involve new measurements, but is a result of processing existing data, for example, in the previous step, through the calculation and comparison of the independent depletion times of the four resources of fuel, water pressure, speed, and water source, the four time values obtained are 27125 seconds, infinity, infinity, and 36000 seconds, respectively, then the system selects the minimum value among them, and obtains D min = 27125 seconds.

[0086] pi min is the normalized dynamic consumption rate of the resource type that causes the minimum depletion time, this parameter quantifies the deviation of the consumption speed of the most critical resource from its normal level, it is calculated by dividing the current instantaneous consumption rate by the historical average consumption rate, both of these rate values come from the previous steps, the instantaneous consumption rate is calculated in real time, while the historical average consumption rate is calculated by taking a long-term (for example, the last 1 hour continuously or 10 hours cumulatively) moving average of the consumption rate of the specific resource under similar working conditions during the operation of the system, and is stored in the non-volatile memory of the controller, for example, if the resource that causes the minimum depletion time is fuel, its current instantaneous consumption rate is 0.00167 liters / second, and the historical average consumption rate of this pump under the current load is calculated to be 0.00150 liters / second according to historical data, then pi min =

[0087] 0.00167 / 0.00150 ≈ 1.113, this value greater than 1 indicates that the current fuel consumption is about 11.3% faster than usual.

[0088] α is the dynamic risk sensitivity coefficient, which is a dimensionless adjustment parameter used to control the degree of punishment of the system to abnormal consumption rate. Its value is set according to the risk strategy of fire fighting operation and the requirement of early warning sensitivity. The setting process usually combines expert experience and simulation test. Fire fighting command experts will define different risk tolerance for different scenarios (such as initial control, general attack, and cleaning up afterfire), and quantify them as specific α values. For example, in the general attack stage which requires the rapid establishment of fireproof isolation belt, any resource anomaly is extremely sensitive, and α can be set to a higher value (such as 2.0). In the relatively moderate cleaning up afterfire stage, α can be appropriately reduced (such as 1.0). The specific value is optimized through playback simulation of historical fire rescue data, and the goal is to find an optimal α value that can provide sufficient warning time (such as 30 minutes in advance) before the real depletion of resources and a false positive rate of less than 5%. In this example, the value of α is set to 1.5 under a standard operation scenario.

[0089] τ is the reference task response period, which is a time scale reference with a unit of seconds, used to adjust the effective depletion time D eff after risk adjustment. The setting of τ reflects the response ability of the system's logistics support or the expected duration of a typical tactical action. For example, it can be set as the average time required from issuing a supply request to the actual arrival of supplies (such as fuel, water) at the front pump site, or the cycle of a combat team rotation. Through statistical analysis of historical fire operations, if the average logistics response time is found to be 45 minutes, τ can be set to 45 × 60 = 2700 seconds. The setting of this value makes the sustainability score C eff drop to a significant warning interval when D s drops to a level comparable to the logistics response time, triggering the corresponding decision-making process.

[0090] Calculation process:

[0091] First, calculate the risk-adjusted effective depletion time D eff , and bring in the parameter values obtained in the previous example: D min = 27125 seconds, α = 1.5.

[0092]

[0093] D eff ≈ 23193.67 seconds;

[0094] Then, based on the calculated D eff and the set reference task response period τ = 2700 seconds, calculate the final combat sustainability quantification score C s .

[0095]

[0096] C s = 100 · (1 - e -2.9309 );

[0097] C s = 100 · (1 - 0.05335);

[0098] C s ≈ 94.67;

[0099] The result shows that the current node's operational sustainability quantification score is 94.67, which is a dimensionless score between 0 and 100. This score is very high, indicating that the operational sustainability of the fire pump node is very strong, and there is still a long time to resource depletion. The current resource consumption pattern is relatively stable, although there is a slight over-speed consumption (pi min > 1), but the effective time D eff after risk adjustment is still much larger than the reference task response period τ, so the score is close to 100.

[0100] The acquisition step of the control strategy adjustment candidate set is:

[0101] Based on the minimum sustainability threshold and the operational sustainability quantification score, numerical comparison is continuously carried out at fixed time intervals to determine whether the operational sustainability quantification score is less than the minimum sustainability threshold, and the comparison time and numerical difference are recorded to generate a sustainability threshold real-time comparison sequence.

[0102] Based on the sustainability threshold real-time comparison sequence, a single node is selected from all local pump controller nodes one by one, and the water pump engine speed and water pump outlet pressure are adjusted slightly. The change trend of the operational sustainability quantification score before and after the adjustment is simulated and calculated, the expected influence of each adjustment scheme on the operational sustainability quantification score and the change of the total water output are evaluated, and the control strategy adjustment candidate set is formed.

[0103] Specifically, based on the received minimum sustainability threshold and the real-time calculated updated operational sustainability quantification score, the system initiates a periodic monitoring and recording procedure, the execution period of which is set to be consistent with the update period of the operational sustainability quantification score, for example, once every 10 seconds, the minimum sustainability threshold is set by the command center according to different operational stages and risk preferences, and is issued to each local pump controller through a wireless network, for example, in the intense total attack stage of the fire situation, the threshold value may be set to a higher 75 points to reserve sufficient redundancy, while in the smooth continuous water supply stage, it may be set to 60 points, the basis of this setting is to ensure that the risk-adjusted effective depletion time is at least twice the logistics response period, the comparison process is executed immediately at the beginning of each period, the system directly compares the numerical size of the latest operational sustainability quantification score (for example, 68.5 points) and the currently effective minimum sustainability threshold (for example, 75 points), and judges whether the Boolean condition “score < threshold” is true, in this example, 68.5 is less than 75, the condition is true, the comparison result is true, at the same time, the system calculates the numerical difference between the two, that is, 75-68.5=6.5, and appends the current system timestamp, operational sustainability quantification score, minimum sustainability threshold, numerical difference and Boolean comparison result as a structured data entry to a first-in-first-out queue stored in memory, which is the sustainability threshold real-time comparison sequence, the length of this sequence is fixed, for example, only the last 60 minutes of records are retained, that is, 360 data entries, by continuously executing this comparison and recording operation, the sustainability threshold real-time comparison sequence is generated.

[0104] Based on the real-time comparison of the latest entries recorded in the sequence, once the combat sustainability quantification score is found to be below the minimum sustainability threshold, the system immediately triggers the simulation evaluation process of the control strategy, which first selects the local pump controller nodes as adjustment objects in numerical order (for example, starting from "Pump ID_01") from the pump group, and for each selected node, the system generates a series of virtual instructions for micro-adjustment around its current two key operating parameters, namely the water pump engine speed and the water pump outlet pressure. The so-called micro-adjustment refers to small step increments and decrements based on the current operating value, with the step size being pre-set, for example, the adjustment step size of the engine speed is 50 RPM, and the adjustment range is ±200 RPM of the current speed, i.e. eight adjustment schemes are generated: -200, -150, -100, -50, +50, +100, +150, +200 RPM, and the adjustment step size of the water pump outlet pressure is 0.05 MPa, and the adjustment range is ±0.2 MPa. For each virtual adjustment instruction (for example, reducing the engine speed of "Pump ID_01" by 50 RPM), the system will call an internal simulation based on a physical model to predict the adjusted system state, which includes an engine fuel consumption model (for example, a two-dimensional lookup table mapping speed and load to fuel consumption rate) and a water pump performance model (based on the affinity law of water pumps). Through simulation, the system can calculate the changes in parameters such as fuel consumption rate and water output of the adjusted node, and then recalculate the new shortest resource depletion time and the new combat sustainability quantification score. At the same time, the system also aggregates the water output of all nodes (including adjusted and unadjusted nodes) to calculate the change in total water output. Each adjustment scheme (for example, "Node ID_01, speed, -50 RPM"), the pre-adjustment score, the simulation-predicted post-adjustment score, the expected change in score, and the expected impact on total water output (for example, -3.5 cubic meters / hour) are added to a temporary list as a complete record. After traversing all the micro-adjustment schemes for all nodes, the temporary list constitutes the control strategy adjustment candidate set.

[0105] The acquisition step of the collaborative operation instruction set under logistics constraints is:

[0106] Based on the control strategy adjustment candidate set, when the latest combat sustainability quantification score in the real-time comparison sequence of the sustainability threshold is below the minimum sustainability threshold, the adjustment scheme that meets the threshold requirement in terms of the combat sustainability quantification score rebound amplitude and has the smallest total water output change is selected, and the deterministic execution time window table is combined to generate the collaborative operation instruction set under logistics constraints.

[0107] Specifically, based on the control strategy adjustment candidate set, when the system detects that the latest combat sustainability quantitative score in the real-time comparison sequence is lower than the set minimum sustainability threshold (for example, the current score 68.5 is lower than the threshold 75), the decision and execution process of the optimal adjustment scheme is started. The process first screens all the alternative schemes in the control strategy adjustment candidate set, and the screening basis is whether the scheme can make the combat sustainability quantitative score rise above a safe level, which is defined as a "score recovery target value". The target value is set in the form of "minimum sustainability threshold + a fixed safety margin", for example, the safety margin is set to 5 points, and the score recovery target value is 75+5=80 points. The system will go through the candidate set and eliminate all schemes with "simulated predicted adjusted score" less than 80 points. After the first round of screening, the system will face a subset containing all adjustment schemes that can restore the system to a safe state. Then, the system sorts the schemes in this subset in the second round, and the only standard for sorting is "the absolute change in total water output". The system will select the scheme that causes the least decrease in total water output. If there are multiple schemes that can increase water output, the scheme with the smallest increase will be selected, that is, the adjustment that has the least impact on the current fire fighting operation is sought. For example, scheme A can make the score rise to 82 points, but the total water output decreases by 5 cubic meters / hour, and scheme B can make the score rise to 81 points, and the total water output decreases by only 2 cubic meters / hour. The system will select scheme B. After determining the only optimal adjustment scheme (for example, reducing the engine speed of "pump ID_01" by 50 RPM), the system will convert the high-level instruction into specific actuator instructions and query the deterministic execution time window table to find the next "actuator driving action" time window reserved for the engine controller of "pump ID_01". The instruction is inserted into the time window for execution, and finally one or more tuples containing the exact execution time, target node, target actuator and target value are generated to form the coordinated operation instruction set under logistical constraints.

[0108] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any skilled person in the art can modify or change the above disclosed technical content to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made according to the technical essence of the present application to the above embodiments without departing from the technical solution content of the present application still falls within the protection scope of the present application.

Claims

1. A method for coordinated operation control of a forest fire pump group, characterized in that, Includes the following steps: Decompose the collaborative operation process of forest fire pump groups, aggregate the discrete control task list of all local pump controllers, and establish a real-time task set for the entire system; Based on the real-time task set of the entire system, the original parameter set for task resource allocation is obtained, the original parameter set for task resource allocation is compared, the execution order of each task is allocated, and a deterministic execution time window table is generated. Based on the time points set in the deterministic execution time window table, the measured values ​​of fuel level, water pump outlet pressure, water pump engine speed and water source level are obtained from each local pump controller. The fuel consumption rate is calculated by the change in fuel level and the time interval, and the equipment load is calculated by the water pump outlet pressure to obtain a multi-dimensional resource depletion time series. The minimum value is selected from the multi-dimensional resource depletion time series to calculate the combat sustainability quantitative score. The received minimum sustainability threshold is continuously compared with the current operational sustainability quantification score. At the same time, the expected impact of adjusting the pump engine speed or pump outlet pressure of a single local pump controller on the operational sustainability quantification score is evaluated to form a control strategy adjustment candidate set. When the operational sustainability quantification score is lower than the minimum sustainability threshold, the adjustment scheme that makes the score rise and has the least impact on the total water output is selected from the control strategy adjustment candidate set to generate a collaborative operation instruction set under logistics constraints. The steps for obtaining the real-time task set for the entire system are as follows: The collaborative operation process of forest fire pump groups is analyzed, and the data acquisition action, state estimation action, control law calculation action, network communication action and actuator driving action included in the process are identified one by one. The process is decomposed into independent basic action units, and the operation content of each action unit is determined one by one, so as to obtain the independent action set of forest fire pump groups. Based on the independent action set of the forest fire pump group, a fixed execution cycle is set for each independent action unit, the worst-case execution time of each independent action unit is estimated and confirmed, the execution priority of each action unit is determined according to the importance of real-time control, and a strict execution deadline is set for each action unit to generate a discrete control task list for a single pump controller. Based on the discrete control task list of the individual pump controller, the discrete control task lists of all local pump controllers are integrated and merged, and sorted according to the task execution cycle, worst case execution time and priority to obtain the real-time task set of the entire system.

2. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, The steps for obtaining the original parameter set for task resource allocation are as follows: Based on the real-time task set of the entire system, all control tasks and network communication tasks are extracted according to the local pump controller node. The worst-case execution time and execution cycle of each control task are recorded, and the data packet size and sending cycle of each communication task are recorded. At the same time, the preset maximum processing capacity and maximum network bandwidth of each node are read to generate the original parameter set of task resource allocation for each local pump controller node.

3. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, Calculate the overall system load of the local pump controller node based on the original parameter set of the task resource allocation. The overall system load value calculated by each local pump controller node is compared with the preset resource capacity threshold of each node. The set of tasks whose load value exceeds the upper limit of the threshold is filtered out. Based on the task deadline and cycle stability, the tasks are prioritized and a deterministic execution time window table that meets the real-time constraints is generated.

4. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, The steps for obtaining the multidimensional resource depletion time series are as follows: Based on the time points set in the deterministic execution time window table, the fuel level change value, water pump outlet pressure, water pump engine speed and water source level of each local pump controller node are collected. Combined with the historical average sampling period of each type of measurement data, the fuel consumption rate, outlet pressure change rate, speed deceleration rate and water level drop rate in the current time period are calculated to obtain a multidimensional resource depletion time series.

5. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, The steps for obtaining the quantitative score of operational sustainability are as follows: Based on the multidimensional resource depletion time series, the independent depletion time of the four types of resources under the node is calculated by dividing the current remaining amount of resources by the rate of change for fuel level, water pump outlet pressure, water pump engine speed and water source level. The independent depletion times of the four types of resources are listed side by side to form the resource depletion time series of the current node. The minimum value is selected as the shortest resource depletion time. The corresponding resource type and consumption rate are recorded to obtain the shortest depletion time and normalized consumption rate. Based on the shortest exhaustion time and normalized exhaustion rate, a quantitative score for operational sustainability is calculated.

6. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, The steps for obtaining the control strategy adjustment candidate set are as follows: Based on the minimum sustainability threshold and the operational sustainability quantitative score, numerical comparisons are continuously performed at fixed time intervals to determine whether the operational sustainability quantitative score is less than the minimum sustainability threshold. At the same time, the comparison time and numerical difference are recorded to generate a real-time sustainability threshold comparison sequence. Based on the real-time comparison sequence of the sustainability threshold, individual nodes are selected one by one from all local pump controller nodes, and the pump engine speed and pump outlet pressure are slightly adjusted respectively. The change trend of the operational sustainability quantitative score before and after the adjustment is simulated and calculated. The expected impact of each adjustment scheme on the operational sustainability quantitative score and the change in the total water output are evaluated to form a control strategy adjustment candidate set.

7. The method for coordinated operation control of forest fire pump groups according to claim 1, characterized in that, The steps for obtaining the collaborative operation instruction set under the logistics constraints are as follows: Based on the control strategy, the candidate set is adjusted. When the latest operational sustainability quantitative score in the real-time comparison sequence of sustainability thresholds is lower than the minimum sustainability threshold, the recovery rate of the operational sustainability quantitative score after each adjustment and the change in total water output are compared item by item. The adjustment scheme that meets the threshold requirement for the recovery rate of the operational sustainability quantitative score and has the smallest change in total water output is selected. Combined with the deterministic execution time window table, a set of collaborative operation instructions under logistics constraints is generated.

8. The collaborative operation control system of the collaborative operation control method for forest fire pump groups according to any one of claims 1-7, characterized in that, include: The task set construction module decomposes the collaborative operation process of the forest fire pump group, aggregates the discrete control task list of all local pump controllers, and establishes a real-time task set for the entire system. The task scheduling module obtains the original parameter set of task resource allocation based on the real-time task set of the entire system, compares the original parameter set of task resource allocation, assigns an execution order to each task, and generates a deterministic execution time window table. The resource assessment module, based on the time points set in the deterministic execution time window table, obtains the measured values ​​of fuel level, water pump outlet pressure, water pump engine speed, and water source level from each local pump controller. It calculates the fuel consumption rate by the change in fuel level and the time interval, and calculates the equipment load by the water pump outlet pressure, thus obtaining a multi-dimensional resource depletion time series. The minimum value is selected from the multi-dimensional resource depletion time series to calculate the operational sustainability quantitative score. The control optimization module continuously compares the received minimum sustainability threshold with the current operational sustainability quantification score, and simultaneously evaluates the expected impact of adjusting the pump engine speed or pump outlet pressure of a single local pump controller on the operational sustainability quantification score, forming a control strategy adjustment candidate set. When the operational sustainability quantification score is lower than the minimum sustainability threshold, the module selects the adjustment scheme from the control strategy adjustment candidate set that will raise the score and have the least impact on the total water output, and generates a collaborative operation instruction set under logistics constraints.

Images