Central air conditioning cold and heat source system edge intelligent control method

Abstract

This application relates to the field of edge control technology and discloses an edge intelligent control method for a central air conditioning cold and heat source system. At the start of a control cycle, this method generates feedforward estimated state and timing estimated data by extrapolating a delayed instruction queue; calculates the dynamic distribution delay time based on the opening degree of the terminal regulating valve and the actual flow rate; corrects the target load by combining cross-cycle energy compensation, generates control instructions, and groups them according to equipment response characteristics; after deduplicating historical instructions with the same equipment and parameters, it immediately or delays the issuance of each instruction set based on load trends and dynamic distribution delay time; at the end of the cycle, it compares the estimated and actual states, calculates the energy deviation increment, and updates the cross-cycle energy compensation amount through limiting and periodic release. This invention decouples electromechanical response from pipeline distribution delay, eliminates system operation oscillations, and solves the control stagnation problem caused by integral saturation under constrained operating conditions.

Description

Technical Field

[0001] This invention relates to the field of edge control technology, specifically to an edge intelligent control method for a central air conditioning cold and heat source system. Background Technology

[0002] Central air conditioning systems comprise multiple components, including chillers, pumps, cooling towers, and piping networks. The mechanical response speeds of these components vary, and the water piping network exhibits thermal inertia and physical transmission delays. Existing systems typically employ conventional control logic that synchronously issues control commands. Under varying load conditions, because the time difference between heat / chill production and distribution within the piping network is not differentiated, synchronous control leads to a misalignment between the heat / chill production cycle and the distribution cycle. This results in fluctuations in the system's water supply temperature and ineffective switching of lower-level mechanical and electrical equipment.

[0003] Meanwhile, under conditions of full system load or limited capacity boundaries, the end-load demand cannot be met, and the conventional control closed loop will continuously accumulate supply and demand energy deviations. When the system returns to normal operation, excessive accumulated deviations will cause integral saturation, resulting in the controller being unable to respond to new reverse adjustment demands in a timely manner, causing the system to stagnate.

[0004] In addition, some current optimization schemes that introduce complex algorithms mainly rely on centralized computing in the cloud. The latency and fluctuations of network communication make it difficult to guarantee the continuity of the underlying control cycle, which weakens the anti-interference capability of real-time closed-loop control of field electromechanical equipment. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an edge intelligent control method for central air conditioning cold and heat source systems, which solves the problems of system operation oscillation caused by the lack of decoupling between electromechanical equipment response and pipeline distribution delay in existing control methods, as well as control stagnation caused by integral saturation under limited operating conditions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an edge intelligent control method for a central air conditioning cold and heat source system, comprising the following steps: At the start of the current control cycle, the delay command queue is simulated in combination with the action change rate limit of the controlled equipment, and feedforward estimated state dataset and time-series estimated data with time stamp are calculated and generated. The opening feedback of the regulating valve at the end of the pipeline network is detected to screen effective flow branches and calculate the equivalent effective water volume of the pipeline network. The dynamic transmission and distribution delay time is calculated in combination with the actual flow of the main pipeline. The corrected target load is calculated based on the actual demand load at the terminal and the cross-cycle energy compensation amount fed back from the previous control cycle. A system control command set is generated based on the corrected target load and the feedforward estimated state dataset. The system control command set is then grouped into cold and heat source command set, cooling / source side command set and load side transmission and distribution command set according to the dynamic response characteristics of the equipment. After overwriting the grouped instruction sets with the historical instructions to be executed in the delayed instruction queue with instructions of the same device and parameters, the grouped instruction sets are immediately issued or written into the delayed instruction queue and issued after a delay, based on the trend of the increase or decrease of the target load and the dynamic transmission and distribution delay time. At the end of the current control cycle, the real-time physical state data and time-series estimated data are compared according to the time identifier. The estimated starting point or estimated calculation parameters of the subsequent control cycle are corrected based on the state difference obtained from the comparison. The single-cycle supply and demand energy deviation increment is calculated based on the actual heat and cold distribution and the target energy reference value. After recursion and accumulation, the cross-cycle energy compensation amount is updated and generated after limiting and periodic release processing. The cross-cycle energy compensation amount is fed back to the next control cycle for load correction.

[0007] This invention introduces dynamic transmission and distribution delay time parameters and command time-sharing scheduling logic to decouple and stagger the production actions of cold and heat sources with the physical transmission and distribution actions of the pipeline network, ensuring that the generation of cold and heat and physical transmission are time-aligned at the end of the pipeline network. Furthermore, a global energy compensation closed loop eliminates steady-state deviations, preventing integral saturation of the system under extreme operating conditions.

[0008] Furthermore, the calculation and generation of feedforward estimated state dataset and time-stamped estimated data specifically includes: collecting real-time physical state data at the start of the current control cycle, scanning the delay command queue, obtaining the trigger time and remaining delay time of each control command to be issued, and distinguishing between the waiting stage before command triggering and the execution stage after command triggering; in the waiting stage, determining the estimated state of the waiting stage based on the control objectives that are currently in effect for the controlled equipment; when the estimated time corresponding to the time-stamped estimated data reaches the predetermined trigger time of the control command to be issued and enters the execution stage, calling the action change rate limit value of the corresponding controlled parameter of the controlled equipment, and deducing the state change process after command triggering according to the target state value in the control command to be issued; The feedforward estimated state values ​​of controlled equipment affected by delayed commands are combined with real-time physical state data collected that are not affected by delayed commands to generate a feedforward estimated state dataset. Flow rate estimates and supply / return water temperature difference estimates related to energy transmission and distribution calculations are extracted and stored in the time-series estimated data in chronological order. Estimation is performed by distinguishing between waiting and execution phases to avoid prematurely including untriggered control commands in the system's current operating state, ensuring the accuracy of input for subsequent load calculations.

[0009] Furthermore, effective flow branches are screened and the equivalent effective water volume and dynamic distribution delay time of the pipeline network are calculated. Specifically, effective flow branches are screened through the effective flow determination rules of the branches with hysteresis. Specifically, opening threshold and closing threshold are set in pairs. When the opening degree of the terminal regulating valve is greater than or equal to the opening threshold, the corresponding branch loop is determined as an effective flow branch. When the opening degree of the terminal regulating valve corresponding to the branch loop that is already in the effective flow state is less than or equal to the closing threshold for a preset number of consecutive control cycles, the branch loop is switched to an invalid flow branch. The equivalent effective water volume of the pipe network is calculated by adding the fixed physical volume of the main pipe to the water volume of the corresponding branch pipes of all effective flow branches. The dynamic distribution delay time is obtained by dividing the equivalent effective water volume of the pipe network by the actual flow rate of the main pipe and multiplying by a dimensionless correction coefficient for pipe friction loss and heat exchange loss. A hysteresis-based judgment rule is adopted to address branch state mis-switching caused by fluctuations in the small opening range of the filter-end regulating valve.

[0010] Further, calculating the corrected target load and generating a system control command set specifically includes: converting the cross-cycle energy compensation amount fed back from the previous control cycle into an equivalent compensation load according to the control cycle step size, and applying upper and lower limit constraints to the equivalent compensation load; performing algebraic summation on the equivalent compensation load after upper and lower limit constraints and the actual demand load at the terminal, and applying upper and lower limit constraints to the summation result according to the current available cold and heat source capacity and transmission and distribution capacity to obtain the corrected target load; generating the system control command set based on the corrected target load and the feedforward estimated state dataset, and identifying each control command in the system control command set as a safe start-stop and interlock control command or a normal capacity adjustment command.

[0011] Furthermore, the same device and same parameter instruction overwrite process is performed, specifically including: determining whether the newly generated control instruction and the historical instructions to be executed in the delay instruction queue are for the same controlled device and the same controlled parameter; if so, when the energy action direction of the newly generated control instruction and the historical instructions to be executed are the same, the target state value of the new instruction is replaced with the target state value of the old instruction, and the original predetermined trigger time of the old instruction is inherited. When a newly generated control command has the opposite energy direction to a historical command to be executed, the old command is cleared, and the triggering time is recalculated based on the current load change trend and the latest calculated dynamic transmission and distribution delay time. This overlay mechanism eliminates invalid reversal commands caused by minor disturbances, reducing the frequency adjustment frequency of underlying mechanical equipment.

[0012] Furthermore, when the load change between two adjacent cycles exceeds the preset load change dead zone threshold, and the system is determined to be in a load increase condition, the specific logic for immediately issuing or writing to the delayed instruction queue for each grouped instruction set is as follows: prioritize immediately issuing the cold and hot source instruction set and the cooling / source side instruction set. The load-side distribution command set is written into the delay command queue, and the calculated dynamic distribution delay time is assigned to the load-side distribution command set as the delay trigger time. After the delay trigger time is reached, the corresponding load-side distribution command set is sent to the chilled water pump. Under the condition of increased load, the cold and heat source equipment is started in advance, and the water pump increases the flow rate later to avoid system oscillation caused by a sudden increase in flow rate but insufficient cold and heat.

[0013] Furthermore, when the load change between two adjacent cycles exceeds the preset load change dead zone threshold and the system is determined to be in a load reduction condition, the specific execution logic of the delayed instruction queue is as follows: the load side distribution instruction set is prioritized to be issued to the field controlled equipment to reduce the operating frequency of the chilled water pump. Based on the operating frequency parameters set in the newly issued load-side distribution instruction set, the steady-state expected flow rate after the water pump executes the instruction is calculated. The equivalent effective water volume of the pipe network is divided by the steady-state expected flow rate and multiplied by a dimensionless correction factor for pipe friction loss and heat exchange losses. The delayed trigger time applicable to the current new flow state is then recalculated. This recalculated delayed trigger time is assigned to the cold / heat source instruction set and the cooling / source side instruction set, and written into the delayed instruction queue for delayed issuance. Under load reduction conditions, priority is given to reducing energy consumption on the distribution side, utilizing the remaining effective cold / heat in the pipe network.

[0014] Furthermore, when issuing system control commands: when safety start-stop and interlock control commands and normal capacity adjustment commands exist simultaneously within the same control cycle, the minimum safe flow establishment, equipment start status confirmation, unloading shutdown, or water pump delayed shutdown process corresponding to the safety start-stop and interlock control commands shall be executed first; after the equipment status meets the safe operation conditions, the remaining normal capacity adjustment commands shall be issued in a delayed time-sharing manner.

[0015] Further, the calculation of the single-cycle supply and demand energy deviation increment includes: extracting the flow rate estimate and the supply and return water temperature difference estimate corresponding to the current control cycle number and sampling time from the time-series forecast data; subtracting the result of multiplying the constant pressure specific heat capacity, density, and flow rate estimate with the supply and return water temperature difference estimate from the result of multiplying the constant pressure specific heat capacity, density, and flow rate estimate with the supply and return water temperature difference estimate; and multiplying the difference obtained after the above subtraction by the time step of the discrete control cycle to obtain the single-cycle supply and demand energy deviation increment.

[0016] Furthermore, the cross-cycle energy compensation amount is updated and generated, specifically including: recursively accumulating the incremental supply and demand energy deviation of a single cycle into the global cumulative energy deviation value; comparing the accumulated global cumulative energy deviation value with the preset upper and lower limits of energy deviation to perform a limit verification; and performing periodic release processing on the global cumulative energy deviation value after the limit verification according to the preset compensation release ratio to generate the cross-cycle energy compensation amount that is currently allowed to be fed back to the next control cycle. Unreleased energy deviation balances remain in the global energy deviation accumulation storage unit and participate in the recursive calculation of subsequent control cycles. The state difference between real-time physical state data and time-series prediction data is only used to correct the prediction starting point or prediction parameters of subsequent cycles; it is not considered as the single-cycle supply and demand energy deviation increment, nor is it directly accumulated to the global energy deviation accumulation value. The accumulated value is limited and released proportionally across cycles to solve the integral saturation problem caused by the system capacity boundary limitation.

[0017] This invention provides an edge intelligent control method for a central air conditioning cold and heat source system. It has the following beneficial effects: 1. This invention overcomes the shortcomings of traditional fixed-delay control, which cannot adapt to changes in water system operating conditions, by detecting the opening degree of the terminal valve and the actual flow rate of the main pipeline to calculate the dynamic distribution delay time in real time. Furthermore, the system groups control commands according to equipment response characteristics and sets independent delay trigger times for different command sets based on load increase / decrease trends. This allows chillers, water pumps, and cooling-side equipment to operate asynchronously and in coordination at different time rhythms, thereby decoupling and aligning the production actions on the heat source side with the distribution actions on the pipeline side, eliminating system oscillations caused by differences in equipment response speeds.

[0018] 2. At the start of the current control cycle, this invention combines the controllable equipment's rate of change limit value with the delay command queue to generate a feedforward estimated state dataset. Before the actual feedback data from the underlying pipeline network arrives, the system can pre-calculate the expected operating trajectory of the controlled equipment and accordingly make control corrections to the target load in advance. Combining the single-cycle supply and demand energy deviation increment calculated from real-time physical state data and estimated data, the system further outputs cross-cycle energy compensation through amplitude limiting and periodic release. This compensates for the control lag problem caused by the large inertia of the central air conditioning water system and solves the integral saturation stagnation phenomenon of the system under constrained operating conditions.

[0019] 3. Before issuing control commands, this invention compares historical commands to be executed, replaces the state values ​​of control commands in the same direction, and directly clears and recalculates the trigger time for control commands in the opposite direction. This deduplication mechanism filters out invalid reverse action signals from underlying equipment caused by minor load disturbances. Furthermore, based on load trends, the system adaptively switches the execution order of electromechanical equipment. When the load increases, it prioritizes the operation of heat and cold sources and delays increasing water pump flow; when the load decreases, it prioritizes reducing water pump frequency and delays reducing heat and cold source output. This reduces ineffective frequency regulation wear on mechanical equipment and utilizes the existing residual heat and cold energy in the pipeline network during load reduction phases, thus lowering the system's energy consumption for transmission and distribution. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system architecture of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention; Figure 3 This is a schematic diagram illustrating the iterative calculation principle of state prediction in this invention. Figure 4 This is a flowchart illustrating the delay analysis process of the present invention; Figure 5 This is the target load correction and instruction decoupling logic diagram of the present invention; Figure 6 This is the timing diagram for the misaligned scheduling control of the present invention; Figure 7 This is the closed-loop logic diagram for energy deviation tracking and compensation of the present invention; Figure 8 The system response comparison curve and energy consumption distribution bar chart of the present invention are shown, wherein (a) is the end water supply temperature tracking curve under sudden load step, (b) is the system load unloading response curve after the extreme full load boundary is removed, and (c) is the comparison chart of the relative power consumption of each subsystem during the test period. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Reference Figure 1 The present invention provides an edge intelligent control system for a central air conditioning cold and heat source system, the system including an edge controller and field controlled devices and sensor components that are communicatively connected thereto.

[0023] The controlled equipment on site includes chiller units, cooling towers, chilled water pumps, cooling water pumps, piping networks, and end regulating valves installed in each branch circuit of the piping network.

[0024] In heating operation, the chiller unit can be a hot and cold water chiller unit with the ability to switch between cooling and heating. The cooling / source side equipment corresponding to the cooling tower and cooling water pump can be switched to the source side heat exchange equipment and source side circulation pump of the hot and cold water chiller unit. The cold and heat source instruction set is used to control the heating start and stop status, water supply temperature set value or heating load rate of the hot and cold water chiller unit. The cooling / source side instruction set correspondingly represents the operation instructions of the source side heat exchange equipment and the source side circulation pump.

[0025] In cooling mode, the supply and return water temperature difference is defined as the return water temperature minus the supply water temperature; in heating mode, the supply and return water temperature difference is defined as the supply water temperature minus the return water temperature, so that the energy transmission and distribution under both modes is calculated as a positive value.

[0026] The edge controller is used to perform state prediction, network delay calculation, control command generation and grouping, time-division multiplexing of commands, and cross-cycle energy compensation within discrete control cycles. To ensure clear data transmission relationships between modules, the edge controller must be configured with at least the following functional modules.

[0027] The state prediction module is used to receive physical state data collected by sensor components when the current discrete control cycle starts, scan the delay command queue inside the edge controller, distinguish between control commands that have been issued and taken effect and control commands that have not yet reached their trigger time, and combine the remaining delay time of each control command to be issued with the action change rate limit value of each controlled device to generate a feedforward prediction state dataset and timing prediction data with time stamp.

[0028] The pipeline delay calculation module is used to obtain the opening feedback signal of the terminal regulating valve, calculate the current equivalent effective water volume of the pipeline network in combination with the preset pipeline branch parameter table, and calculate the dynamic transmission and distribution delay time that characterizes the equivalent transmission and distribution delay of the pipeline network based on the equivalent effective water volume of the pipeline network, the actual flow of the main pipeline, and the flow change process formed during the operation of the controlled equipment.

[0029] The control command generation and grouping module is used to obtain the actual demand load at the end, extract the cross-cycle energy compensation amount fed back from the previous control cycle, convert the cross-cycle energy compensation amount into an equivalent compensation load, and sum it with the actual demand load at the end to obtain the corrected target load; then, based on the corrected target load and the feedforward estimated state dataset, a system control command set is generated, and the system control command set is split into a cold and heat source command set, a cooling / source side command set, and a load side transmission and distribution command set.

[0030] Among them, the cross-cycle energy compensation amount is in the dimension of energy. Before participating in the load correction, it is converted into the equivalent compensation load according to the current control cycle step size. The equivalent compensation load has the same dimension as the actual demand load at the end and is used to correct the target load of the current control cycle.

[0031] The instruction time-sharing scheduling module is used to receive each instruction set after grouping, overwrite the newly generated instructions with the historical instructions in the delayed instruction queue that have not yet been triggered with instructions of the same equipment and parameters, and clear the historical instructions to be executed with opposite energy action direction; then, based on the increase or decrease trend of the target load and combined with the dynamic transmission and distribution delay time, the cold and heat source instruction set, the cooling / source side instruction set and the load side transmission and distribution instruction set are immediately issued, or written into the delayed instruction queue and issued after a delay.

[0032] The energy compensation module is used to extract the actual flow rate and actual supply and return water temperature difference from the sensor components at the end of the current discrete control cycle, retrieve the time-series prediction data generated by the state prediction module, and match the actual operating data and prediction data according to the same time identifier.

[0033] The state difference between actual operating data and estimated data is used to correct the estimated starting point or state estimation parameters for subsequent cycles, and is not directly added to the cumulative energy deviation value. The target energy reference value is determined based on the actual demand load at the end of the current control cycle or the constrained target load. The target energy reference value can be obtained by multiplying the actual demand load at the end of the current control cycle or the constrained target load by the control cycle step size.

[0034] The energy compensation module calculates the incremental supply-demand energy deviation for a single cycle based on the actual heat and cold distribution and the target energy reference value. It then recursively accumulates the incremental supply-demand energy deviation for a single cycle to the global cumulative energy deviation value. After limiting and periodic release processing, it updates the cross-cycle energy compensation amount.

[0035] The cross-cycle energy compensation amount is the compensation amount that can be released in the current or next control cycle, rather than the full one-time release value of the global energy deviation accumulation value; the updated cross-cycle energy compensation amount is fed back to the control command generation and grouping module of the next cycle.

[0036] Reference Figure 2 This invention provides an edge intelligent control method for a central air conditioning cold and heat source system, the method comprising the following steps: Step S10: When the current control cycle starts, collect real-time physical state data, scan the delay instruction queue, obtain the trigger time and remaining delay time of each control instruction to be issued, distinguish between the waiting stage before instruction triggering and the execution stage after instruction triggering, and calculate the feedforward estimated state dataset and time-series estimated data with time stamps for the current cycle by combining the action change rate limit value of the controlled equipment.

[0037] The dynamic transmission and distribution delay time required for the feedforward estimation in the current cycle is the dynamic transmission and distribution delay time that has been calculated and saved in the previous control cycle. The dynamic transmission and distribution delay time recalculated in step S20 of the current cycle is used for the instruction time-sharing scheduling of the current cycle and the state estimation of the next control cycle.

[0038] Step S20: Detect the opening feedback of the regulating valve at the end of the pipeline network, screen the effective flow branch through the effective flow judgment rule of the branch with hysteresis, calculate the equivalent effective water volume of the pipeline network, and calculate the dynamic transmission and distribution delay time in combination with the actual flow or flow change process of the main pipeline.

[0039] Step S30: Calculate the corrected target load based on the actual demand load at the terminal and the cross-cycle energy compensation amount fed back from the previous cycle. The cross-cycle energy compensation amount is first converted into an equivalent compensation load according to the control cycle step size, and then algebraically summed with the actual demand load at the terminal; upper and lower limit constraints are applied to the equivalent compensation load and the corrected target load.

[0040] Subsequently, a set of system control commands is generated based on the modified target load and feedforward estimated state dataset. The set of system control commands is then grouped according to the dynamic response characteristics of the equipment into a set of cold and heat source commands, a set of cooling / source side commands, and a set of load side transmission and distribution commands. At the same time, the control commands involved in equipment start-up, shutdown, or safety interlocks are identified.

[0041] Step S40: After grouping each instruction set, perform instruction overwrite processing on the historical instructions to be executed in the delayed instruction queue, and update the delayed instruction queue according to the rule that instructions with the same device, same parameters and the same energy direction inherit the original trigger time, and instructions with the same device, same parameters and opposite energy direction are cleared and the trigger time is recalculated.

[0042] For instructions involving equipment start-up, shutdown, or safety interlocks, the preset safety start-up and shutdown sequence shall be followed first; for capacity adjustment instructions under normal operating conditions, the delayed time-sharing shall be executed based on the trend of increase or decrease of the target load and the dynamic transmission and distribution delay time.

[0043] Step S50: At the end of the current control cycle, compare the real-time physical status data with the time-series estimated data according to the time identifier, and use the status difference between the two to correct the estimated starting point, equipment action change rate parameter or flow rate and supply and return water temperature difference estimated parameters of the subsequent cycle.

[0044] Meanwhile, the target energy reference value is determined based on the actual demand load at the end of the current cycle or the target load after constraints; the incremental energy deviation between supply and demand in a single cycle is calculated based on the actual heat and cold distribution and the target energy reference value, and after being recursively accumulated, the cross-cycle energy compensation amount is updated after being limited and released in stages, and the cross-cycle energy compensation amount is fed back to step S30 of the next cycle for load correction.

[0045] The difference between real-time physical state data and time-series forecast data is only used to correct the forecast starting point or forecast calculation parameters for subsequent cycles. It is not used as the single-cycle supply and demand energy deviation increment, nor is it directly accumulated to the global energy deviation cumulative value.

[0046] To further clarify the implementation of each technical aspect of the present invention, the following will provide a detailed description of the implementation of each functional module involved above and its internal processing flow.

[0047] Reference Figure 3 The specific execution process of step S10 is completed by the state prediction module within the edge controller. This step is used to distinguish between the waiting stage and the execution stage of the command to be issued when there is a delay in the issuance of control commands and a lag in the device response, and to form the state input required for solving subsequent control commands by combining the rate of change limit of the controlled device's actions. This step may specifically include the following sub-steps.

[0048] Step S101, during the discrete control cycle At the start time, the state prediction module acquires real-time physical state data collected by the sensor components via the communication bus. The collected physical state data includes, but is not limited to, the actual flow rate of the main pipeline of the network. Actual supply and return water temperature difference And the current operating frequency or temperature feedback value of each controlled device on site.

[0049] The state prediction module scans the delayed instruction queue maintained in the edge controller's internal memory. This queue stores control instructions generated in the previous control cycle but whose trigger time has not yet arrived. Each control instruction in the delayed instruction queue contains at least the controlled device identifier, controlled parameter identifier, target state value, parameter change direction or energy application direction, instruction generation time, predetermined trigger time, and instruction version identifier.

[0050] The parameter change direction indicates whether the controlled parameter value increases or decreases, while the energy action direction indicates whether the corresponding control command increases or decreases the system's cooling or heating capacity. For different controlled parameters such as outlet water temperature, load rate, and pump frequency, the edge controller can pre-store the correspondence between the parameter change direction and the energy action direction to determine whether new and old commands belong to the same-direction or opposite-direction adjustment.

[0051] The state prediction module traverses the queue to identify instructions currently in a delayed waiting state, locks the specific controlled devices to which these instructions point, and calculates the remaining delay time between the scheduled trigger time and the issuance of each control instruction. The edge controller also stores the control targets that have been issued and taken effect for each controlled device, enabling the state prediction module to distinguish between currently effective instructions and instructions awaiting issuance.

[0052] Step S102: For control commands to be issued that have not yet reached the trigger time, the state prediction module does not include them as control targets that have already taken effect in the current state calculation, so as to avoid prematurely including control actions that have not yet been executed into the state input, causing subsequent control commands to deviate from the actual operating state.

[0053] For controlled devices that are still in the delay waiting stage and awaiting control commands, the status prediction module continues to adopt the actual feedback data uploaded by the sensor components and, in conjunction with the control objectives that are currently in effect for the controlled device, determines the predicted status of the waiting stage.

[0054] Only after the estimated time corresponding to the timing prediction data reaches the predetermined trigger time for the control command to be issued, does the state prediction module call the action change rate limit value of the corresponding controlled parameter of the controlled device. And based on the target state value in the control command to be issued, the state change process after the command is triggered is deduced.

[0055] Motion change rate limit This represents the maximum physical state change that the equipment can perform per unit time, such as the cooling rate limit of a chiller unit or the variable frequency speed limit of a chilled water pump. The rate of change limit value is set separately for the controlled equipment and the controlled parameter, and separate limits can be set for increasing and decreasing the parameter. The unit of the rate of change limit value must be consistent with the unit of the corresponding controlled parameter.

[0056] Among them, the limit value of the rate of change of motion The value range depends on the specific mechanical and electrical characteristics of the equipment. For controlled parameters expressed as a percentage of the rated value, it can be set to a rated change of 0.1% to 5% per second. Rate of change limit value. It can be determined based on the equipment's factory technical data, or it can be obtained through step response testing and calibration.

[0057] During the preliminary calculation, the controlled equipment gradually approaches the target setpoint according to its mechanical and electrical characteristics. The state prediction module uses a discrete-time step calculation method to deduce the change process of the controlled parameters.

[0058] Furthermore, when a control command for a controlled device is first written into the delayed command queue in the current control cycle, the command time-sharing scheduling module, while writing the command, records the real-time physical state data value of the device fed back by the sensor components at that moment, or records the currently valid estimated state value, and uses it as the initial estimated state corresponding to the control command to be issued. The state prediction module reads this initial predicted state when scanning the delay instruction queue in subsequent control cycles. .

[0059] When the control command is issued and the trigger time is reached, during the process of generating future time-series prediction data in the current cycle, the last predicted state value of the waiting stage is used as the iteration starting point of the execution stage; when the system actually runs to the trigger time and obtains sensor feedback, the prediction starting point of the subsequent cycle is corrected using the actual state value corresponding to the trigger time.

[0060] Subsequently, the state prediction module in the previous control cycle The calculated estimated state value is the starting point for iteration. Based on the relationship between the target state value and the current state value, the corresponding calculation logic is used to calculate the controlled device in the current control cycle. Feedforward predicted state values : If the target state value The predicted state value is greater than that of the previous control cycle. This indicates that the controlled equipment is in a process of increasing parameters. At this time, the following formula is used for calculation: ; If the target state value The estimated state value is less than that of the previous control cycle. This indicates that the controlled equipment is in a parameter decreasing process, and the following formula is used for calculation: ; In the formula, This indicates the final target state value set in the delay waiting instruction; This represents the feedforward estimated state value from the previous control cycle; This represents the time step between two adjacent discrete control cycles; the discrete control cycle is... The value is set between 1 second and 5 seconds.

[0061] During the variable load operation of the control system, if a newly generated control command overwrites a historical command to be executed in the delay command queue, the new estimated starting point and trigger time shall be determined as follows: For new and old commands with the same equipment, parameters, and energy direction, the new target state value replaces the target state value of the old command, and inherits the current estimated state and the original scheduled trigger time corresponding to the old command; For new and old commands with the same equipment, parameters, and energy direction, the old command is cleared, and the actual state value or the current valid estimated state value corresponding to the time the new command is generated is used as the new estimated starting point.

[0062] During the execution phase after the instruction is triggered By using the estimated results from the previous cycle, the actual feedback value of the controlled equipment is not abruptly changed due to a similar adjustment in the target state value, thus maintaining the continuity of the estimation process during the execution phase. This process reduces abrupt changes in the estimated state when control commands are overridden, ensuring that the state input used for solving subsequent control commands matches the actual response of the equipment.

[0063] In step S103, the state prediction module combines the feedforward predicted state value of the controlled device affected by the delay command, which has been calculated above, with the real-time physical state data that is not affected by the delay command collected in step S101.

[0064] The state prediction module generates a feedforward prediction state dataset from the combined data. This feedforward prediction state dataset contains the state inputs used for solving control commands within the current control cycle, serving as the input conditions for generating subsequent control commands.

[0065] The state prediction module extracts parameters related to transmission and distribution energy calculation from the feedforward prediction state dataset, specifically including the flow prediction value of the main pipeline. Estimated supply and return water temperature difference .

[0066] Among them, the estimated flow rate of the main pipeline The operating frequency of the chilled water pumps is determined based on the feedforward estimated state dataset. The correspondence between the operating frequency of the chilled water pumps and the flow rate of the main pipe can be stored in the edge controller in the form of a data table or a piecewise function.

[0067] For the estimated water supply temperature, the status estimation module starts with the current outlet water temperature feedback value of the chiller unit and performs phased extrapolation based on the currently effective outlet water temperature setpoint, the trigger time of the outlet water temperature setpoint to be issued, and the corresponding temperature change rate limit value of the chiller unit.

[0068] For the return water temperature prediction, the state prediction module recursively calculates the return water temperature based on the previous control cycle, the current actual demand load at the terminal, the main pipeline flow prediction, and the dynamic transmission and distribution delay time saved in the previous control cycle. It also selects the corresponding supply water temperature at the delay time from the historical supply water temperature sequence as the recursive reference based on the dynamic transmission and distribution delay time.

[0069] In cooling mode, the estimated return water temperature is obtained based on the theoretical temperature rise corresponding to the terminal load; in heating mode, the estimated return water temperature is obtained based on the theoretical temperature drop corresponding to the terminal load, and thus the estimated supply and return water temperature difference is obtained. In this embodiment, the estimated value of the supply and return water temperature difference is... The value ranges from 0℃ to 12℃.

[0070] When the estimated flow rate in the main pipeline is lower than the preset effective calculated flow rate, the state estimation module uses the previous effective supply and return water temperature difference estimate and marks the current period as a low-flow invalid calculation period. This identifier is simultaneously transmitted to the energy compensation module, which does not update the global energy deviation cumulative value based on the flow-temperature difference data under low-flow conditions during this period. In this embodiment, the preset effective calculated flow rate is taken as 10% to 15% of the system's full-load rated main pipeline flow rate.

[0071] The state prediction module stores the flow rate prediction and the supply and return water temperature difference prediction in time sequence into the time-series prediction data, and adds the corresponding control cycle number and sampling time to each set of flow rate prediction and supply and return water temperature difference prediction.

[0072] The estimated time range of the timing prediction data is not less than the maximum range of the dynamic transmission and distribution delay time. Its sampling interval can be consistent with the discrete control cycle step size or less than the discrete control cycle step size.

[0073] When generating timing prediction data, the state prediction module determines point by point whether the sampling time is earlier than the predetermined trigger time of the control command to be issued. Sampling points earlier than the trigger time are generated according to the waiting phase logic; sampling points that reach or are later than the trigger time are generated according to the execution phase logic.

[0074] The time-series prediction data should at least cover historical data covering the maximum range of dynamic transmission and distribution delay times. This time-series prediction data is used to provide reference data for calculating energy deviations in subsequent control cycles, and is used to match actual operating data according to the same control cycle number and sampling time after the data arrives, avoiding direct comparison between actual data and prediction data at different times.

[0075] Reference Figure 4The specific execution process of step S20 is completed by the pipeline delay calculation module within the edge controller. This step determines the effective flow branch currently participating in the circulation based on the opening feedback of the end regulating valve, and calculates the dynamic distribution delay time by combining the pipeline water volume and the actual flow rate of the main pipe, providing time parameters for the time-sharing scheduling of subsequent control commands. This step may specifically include the following sub-steps.

[0076] Step S201: In the current control cycle, the pipeline delay calculation module extracts the opening feedback signal of each end regulating valve in the pipeline in real time through the sensor component.

[0077] The opening degree of the end control valve is used to determine whether the corresponding branch circuit is in an effective flow state. The pipeline delay calculation module compares the extracted opening degree feedback signals of each end control valve with the preset flow determination threshold.

[0078] The flow determination thresholds include an opening threshold and a closing threshold. These thresholds are set in pairs, and the closing threshold must be less than the opening threshold. The opening threshold can be set to an opening value of 5% to 10%, and the closing threshold can be set to an opening value of 2% to 8%. For example, the opening threshold can be set to 8%, and the closing threshold to 5%.

[0079] When the pipeline delay calculation module determines that the opening degree of a certain end control valve is greater than or equal to the opening threshold in the current control cycle, the branch circuit corresponding to the end control valve is identified as an effective flow branch.

[0080] A branch circuit is switched to an invalid flow branch only when the opening degree of the end control valve corresponding to a branch circuit that is already in an effective flow state is less than or equal to the closing threshold. This flow state determination method with hysteresis reduces frequent switching of effective flow branches caused by small fluctuations in the end control valve near the threshold.

[0081] The pipeline delay calculation module reads the preset pipeline branch parameter table. This pipeline branch parameter table is pre-stored in the edge controller's memory, and its data structure includes the unique identifier of each branch loop, the corresponding end regulating valve identifier, and the water volume of the branch pipe of that branch loop.

[0082] Step S202: The pipeline delay calculation module, based on the selected effective flow branches, matches the above pipeline branch parameter table, extracts the corresponding branch pipeline water volume, and calculates the current equivalent effective water volume of the pipeline network.

[0083] The equivalent effective water volume of the pipeline network represents the water volume circulating in the main pipe and effective flow branches within the current control cycle, and is used to calculate the equivalent transmission and distribution delay time of the entire pipeline network. This water volume is used for global scheduling calculations and is not limited to the actual transmission volume experienced by cold and heat to reach each specific terminal branch.

[0084] The pipeline delay calculation module uses the following formula to calculate the equivalent effective water volume of the pipeline during the current control period Tn. : ; In the formula, This represents the fixed physical volume of the system's main pipeline; This represents the total number of effective circulation branches currently in an effective circulation state; Indicates the first The water volume of the branch pipes of each effective circulation branch.

[0085] Step S203: The pipeline delay calculation module obtains the actual flow rate of the main pipeline of the pipeline network from the sensor components. .

[0086] The pipeline delay calculation module combines the current equivalent effective water volume of the pipeline network with the actual flow rate of the main pipeline. Calculate the dynamic transmission and distribution delay time .

[0087] To avoid division by zero errors during system shutdowns or when the actual flow rate in the main pipeline is extremely low, the pipeline delay calculation module adjusts the actual flow rate of the main pipeline. Perform a lower limit check.

[0088] when When the flow rate is less than the preset minimum threshold for delay calculation, the pipeline delay calculation module will dynamically adjust the transmission and distribution delay time. Assign a preset maximum delay time.

[0089] when When the dynamic transmission and distribution delay time is greater than or equal to the minimum flow threshold for delay calculation, the following formula is used to calculate the dynamic transmission and distribution delay time. : ; In the formula, This indicates the actual flow rate of the main pipeline; The dimensionless correction coefficient representing the attenuation of pipeline friction resistance and heat exchange loss can be set between 1.05 and 1.20; the preset minimum flow threshold for delay calculation is 5% to 10% of the rated main pipeline flow of the system; the preset maximum delay time is 60 to 600 times the discrete control cycle step size, for example, it can be set to 300 to 600 seconds, to prevent control commands in the delay command queue from being delayed for a long time due to excessive triggering time when the system is shut down or under extremely low flow conditions.

[0090] Calculated dynamic transmission and distribution delay time This represents the equivalent delay time for heat and cold to be transported to the load side within the pipe network under current operating conditions. For multiple branch loops with significant differences in length, the branch delay can be calculated separately based on the water volume of each branch pipe and the corresponding branch flow rate, and then weighted according to the proportion of the terminal load. Alternatively, the transmission delay time of the main load branch carrying the main load can be selected as the dynamic transmission delay time for system scheduling. After this dynamic transmission delay time is generated, it is passed to subsequent steps to determine the timing of different control commands.

[0091] Reference Figure 5 The specific execution process of step S30 is completed by the control instruction generation and grouping module within the edge controller. This step introduces the cross-cycle energy compensation amount from the previous control cycle based on the current end-point demand load to obtain the corrected target load for the current control cycle. Then, it combines the feedforward estimated state dataset and equipment operating constraints to generate a set of system control instructions. These instructions are then grouped according to equipment type, controlled object, and equipment dynamic response characteristics, providing input for subsequent time-sharing scheduling of instructions. This step may specifically include the following sub-steps.

[0092] Step S301, in the current control cycle The control command generation and grouping module obtains the current actual load demand at the end of the pipeline network. .

[0093] The sensor assembly includes flow sensors and supply and return water temperature sensors installed on the main pipe or each effective flow branch. The control command generation and grouping module calculates the instantaneous demand load of each effective flow branch based on the actual flow rate and the supply and return water temperature difference, and sums the instantaneous demand loads of all effective flow branches to obtain the actual demand load at the terminal. .

[0094] The control command generation and grouping module retrieves the previous control cycle. The cross-cycle energy compensation generated at the end The cross-cycle energy compensation amount is the energy compensation amount that can be released from the previous control cycle to the current control cycle. It is obtained by the energy compensation module after limiting and periodically releasing the global energy deviation cumulative value, and is not equivalent to the full amount of the global energy deviation cumulative value.

[0095] The control command generation and grouping module algebraically sums the current actual demand load at the terminal with the equivalent compensation load to calculate the corrected target load for the current control cycle. The calculation formula is as follows: ; In the formula, This indicates the target load for the current cycle. This indicates the current actual demand load at the end of the pipeline network; This represents the cross-cycle energy compensation amount from the previous control cycle; This indicates the time step of the discrete control cycle.

[0096] Through the above load correction process, the portion of historical energy deviation that is allowed to be released in the current cycle is converted into equivalent load and participates in the target load calculation for the current control cycle.

[0097] Step S302: The control command generation and grouping module obtains the feedforward estimated state dataset as the state input for generating control commands in the current control cycle.

[0098] Combined with the calculated corrected target load The energy balance constraints and energy consumption constraints of the control command generation and grouping module construction system are determined by using the start-stop status of the chiller unit, the load rate or outlet water temperature setpoint, the operating frequency of the chilled water pump, the operating frequency of the cooling water pump, and the fan frequency of the cooling tower as the control variables to be determined.

[0099] The control command generation and grouping module aims to minimize the total input power of the chiller unit, chilled water pump, cooling water pump, and cooling tower, and sets at least the following constraints: The total output capacity of all operating chiller units shall not be less than the corrected target load; the operating parameters of each device shall be within the allowable upper and lower limits; the flow rates on the evaporator and condenser sides of the chiller units shall not be less than the corresponding minimum safe flow rates; the single-cycle state changes of each device shall not exceed the corresponding action change rate limit; the chiller units shall meet the minimum continuous operating time and minimum downtime; devices that have malfunctioned or are under maintenance shall not participate in instruction solving.

[0100] The control command generation and grouping module generates target operating states and target setting parameters for each field-controlled device based on the above-mentioned energy balance constraints, energy consumption targets and equipment operation constraints, forming a set of system control commands.

[0101] When a control command that simultaneously satisfies the above constraints cannot be obtained under the current modified target load conditions, the control command generation and grouping module will limit the modified target load to the maximum safe output capacity of the currently available equipment and call the preset safe operation control command set. The energy compensation amount that cannot be released in the current cycle is retained to subsequent cycles and will participate in the cycle-by-cycle release calculation again in the next control cycle, without being forced to be released in full within a single control cycle.

[0102] Step S303: The control command generation and grouping module divides the system control command set into cold and heat source command set, cooling / source side command set and load side transmission and distribution command set according to the controlled equipment category, controlled parameters and equipment dynamic response characteristics.

[0103] Commands used to control the start-up and shutdown status of chiller units, load rate, or outlet water temperature setpoint are categorized under the cold / heat source command set. Since chiller unit load regulation and outlet water temperature regulation typically have long thermal response times, these commands are treated as slow-response equipment control commands.

[0104] Commands used to control the frequency of cooling tower fans and cooling water pumps are categorized under the cooling / source-side command set. In cooling mode, this command set controls the operation of the cooling tower and cooling water pumps; in heating mode, it controls the operation of the source-side heat exchange equipment and source-side circulating pumps in the chiller / hot water unit. The response time of this type of equipment falls between that of the chiller unit and the load-side distribution equipment, therefore it is treated as a medium-response equipment control command.

[0105] Commands used to control the operating frequency of chilled water pumps are categorized under the load-side distribution command set. Chilled water pumps change the water flow rate on the load side via frequency conversion regulation, and their regulation response time is typically shorter than that of heat source and cooling / source side equipment; therefore, they are treated as fast-response equipment control commands.

[0106] After the above grouping process, the system control command set is divided into cold / heat source command set, cooling / source side command set, and load side transmission and distribution command set. The control command generation and grouping module further identifies the control commands in each command set as either safe start / stop and interlock control commands or normal capacity adjustment commands, based on the nature of their function.

[0107] The control command generation and grouping module also sets the correspondence between the direction of parameter change and the direction of energy action for each controlled parameter, so that whether the new and old control commands are in the same or opposite direction is determined based on their actual direction of action on the system's cooling or heating capacity, rather than simply based on the increase or decrease of the target parameter value.

[0108] Safety start / stop and interlock control commands include pump pre-start, minimum safe flow establishment, chiller unit start, chiller unit unloading and shutdown, pump delayed shutdown, and fault interlock commands. Normal capacity adjustment commands include commands for adjusting outlet water temperature, load rate, pump operating frequency, and fan operating frequency when the equipment is already in operation.

[0109] Safety start / stop and interlock control commands are executed first according to the preset equipment safety start / stop sequence and are not directly applicable to subsequent ordinary time-sharing scheduling based on dynamic transmission and distribution delay time; normal capacity adjustment commands enter the command time-sharing scheduling module after the equipment safety operation conditions are met.

[0110] When both safety start / stop and interlock control commands and normal capacity adjustment commands exist within the same control cycle, the command time-sharing scheduling module first executes the minimum safe flow establishment, equipment start status confirmation, unloading shutdown or water pump delayed shutdown process corresponding to the safety start / stop and interlock control commands; after the equipment status meets the safe operation conditions, the remaining normal capacity adjustment commands are then issued with a delayed time-sharing.

[0111] Reference Figure 6 The specific execution process of step S40 is completed by the instruction time-sharing scheduling module within the edge controller. Before issuing control instructions, this step first performs a process to overwrite historical instructions with the same equipment and parameters in the delayed instruction queue, eliminating instructions that contradict the current control target. Then, based on the changing trend of the target load and the dynamic distribution delay time, it determines the immediate or delayed issuance sequence of the cold / heat source instruction set, the cooling / source side instruction set, and the load side distribution instruction set. This step may specifically include the following sub-steps.

[0112] Step S401: The instruction time-sharing scheduling module receives the cold and heat source instruction set, cooling / source side instruction set, and load side distribution instruction set output by the control instruction generation and grouping module.

[0113] Before sending the above instruction set to the controlled devices on site, the instruction time-sharing scheduling module traverses the delayed instruction queue inside the edge controller and extracts the historical instructions to be executed that have not yet been triggered in the queue.

[0114] The instruction time-sharing scheduling module compares newly generated control instructions with historical instructions to be executed for the same device and parameters. This comparison determines whether the new and old instructions target the same controlled device and the same controlled parameters.

[0115] If it is determined that there are new and old instructions for the same controlled device and the same controlled parameter, in order to avoid the repeated accumulation of instructions in the same direction in the delayed instruction queue, the instruction time-sharing scheduling module will overwrite the corresponding historical instructions to be executed with the newly generated control instructions.

[0116] When the energy direction of the new and old instructions is the same, the target state value of the old instruction is replaced by the target state value of the new instruction, and the original scheduled trigger time of the old instruction is inherited, so as to avoid repeatedly resetting the timer because a new instruction is generated in each control cycle.

[0117] When the energy action directions of the old and new instructions are opposite, the old instructions are cleared, and the triggering time of the new instructions is regenerated based on the current load change trend and the latest dynamic transmission and distribution delay time.

[0118] Each instruction in the delayed instruction queue is also configured with a maximum allowed waiting time. When the instruction waiting time reaches the maximum allowed waiting time, the instruction time-sharing scheduling module executes the instruction, cancels the instruction, or switches the system to a safe operating state based on the current system state, preventing control instructions from being held up for an extended period.

[0119] Specifically, when the current load change direction is still consistent with the energy action direction of the instruction and the equipment safety conditions are met, the instruction is executed; when the current load change direction is opposite to the energy action direction of the instruction, the instruction is canceled; when the equipment safety conditions are not met, the normal capacity adjustment instruction is canceled and the preset safe operation control instruction set is called.

[0120] When a historical instruction pending execution is detected in the delayed instruction queue that targets the same controlled device and the same controlled parameter as the new instruction but has the opposite energy direction, the instruction time-sharing scheduling module determines that there is a conflict between the two control instructions. For example, a historical instruction pending execution in the queue that has not yet been triggered requires the chilled water pump to reduce its frequency, while the newly generated control instruction requires the chilled water pump to increase its frequency.

[0121] To prevent controlled equipment from frequently receiving control commands in opposite directions within adjacent control cycles, the time-sharing command scheduling module performs command overwriting processing on the same equipment and parameters in the delayed command queue, deleting historical commands with opposite energy directions and retaining only newly generated control commands. This process reduces the repeated triggering of reverse control commands caused by load disturbances.

[0122] Step S402: The instruction time-sharing scheduling module extracts the corrected target load of the current control cycle and compares it with the corrected target load of the previous control cycle to determine the load change trend of the current system.

[0123] To avoid frequent switching of dispatch modes caused by pipeline measurement noise, a preset load change dead zone threshold is introduced into the load change trend determination. This load change dead zone threshold is set to 1% to 3% of the system's rated load.

[0124] When the absolute value of the load change between two adjacent cycles is less than the dead zone threshold of the load change, it is determined to be a stable load condition. At this time, the grouped instruction sets can be directly and synchronously issued, but synchronous issuance is only applicable to capacity adjustment instructions when the equipment is already in normal operating condition.

[0125] Control commands involving equipment start-up and shutdown, safety interlocks, minimum safe flow establishment, or fault protection will still be executed according to the preset safe start-up and shutdown sequence. At the same time, pending historical commands for the same controlled equipment and the same controlled parameters in the delayed command queue will be cleared to prevent them from being triggered again at a later time.

[0126] When the load change exceeds the load change dead zone threshold, the load change trend is classified as either a load increase condition or a load decrease condition. The time-sharing command module, based on the determined condition and the dynamic transmission and distribution delay time calculated by the pipeline delay calculation module, then... Then, execute the corresponding delayed time-sharing distribution.

[0127] Step S403: When the instruction time-sharing scheduling module determines that the system is in a load increase condition, since it takes a certain amount of time for the output changes of the cold and heat sources to be transported to the load side with the water flow, the system prioritizes increasing the output capacity of the cold and heat source side and the cooling / source side.

[0128] When the cold and heat source instruction set only involves the load rate or outlet water temperature adjustment of the already running chiller unit, and the actual flow rates on both the chilled water side and the cooling water side meet the minimum safe flow rate requirements of the corresponding equipment, the instruction time-sharing scheduling module immediately issues the cold and heat source instruction set and the cooling / source side instruction set, and writes the load side transmission and distribution instruction set into the delayed instruction queue.

[0129] When the cold and heat source instruction set involves the startup of a new chiller unit, the instruction time-sharing scheduling module does not directly execute the time-sharing issuance sequence of the above-mentioned normal capacity adjustment instructions, but first issues pre-start instructions for the chilled water pump and cooling water pump corresponding to the chiller unit to be started.

[0130] After the actual flow rates on both the evaporator and condenser sides, as indicated by the flow sensors, have reached the minimum safe flow rate, the chiller start-up command and cooling tower operation command are then issued.

[0131] The minimum safe flow rates on the evaporator side and condenser side are set according to the corresponding chiller unit's equipment technical data, rated flow rate, or manufacturer-permitted minimum flow rate. In this embodiment, the minimum safe flow rate on the evaporator side is 40% to 50% of the chiller unit's rated chilled water flow rate; the minimum safe flow rate on the condenser side is 50% to 60% of the chiller unit's rated cooling water flow rate.

[0132] Once the chiller unit reaches a stable operating state, subsequent capacity adjustment commands will be subject to delayed time-sharing scheduling, with a set delay trigger time. This delay trigger time is set to the dynamic transmission and distribution delay time of the current cycle. .

[0133] The delay instruction queue is timed according to the delay trigger time; after the delay trigger time is reached, the instruction time-sharing scheduling module sends the corresponding load-side distribution instruction set to the chilled water pump, so that the increase in water flow on the load side matches the time when the change in output of cold and heat source is transmitted to the load side.

[0134] Step S404: When the instruction time-sharing scheduling module determines that the system is in a load reduction condition, in order to reduce the energy consumption of the load side transmission and distribution and to utilize the heat and cold already formed in the effective water volume of the pipeline network, the system prioritizes reducing the operating frequency of the water pumps on the load side.

[0135] The time-sharing command module immediately sends load-side distribution command sets to the controlled equipment on site, reducing the operating frequency of the chilled water pumps. At the same time, the time-sharing command module writes the cold / heat source command sets and the cooling / source side command sets into the delayed command queue, preparing for delayed issuance.

[0136] Since the load-side distribution command set has been issued immediately, the actual water flow rate of the main pipeline begins to change gradually according to the limit value of the rate of change of the chilled water pump and the hydraulic characteristics of the pipeline. The original dynamic distribution delay time is no longer applicable to the delayed triggering of subsequent cold and heat source command sets and cooling / source side command sets.

[0137] The time-sharing scheduling module calculates the steady-state expected flow rate of the pump after executing the instruction based on the operating frequency parameters set in the newly issued load-side transmission and distribution instruction set. The relationship between the pump operating frequency and the steady-state expected flow rate can be determined based on the pump similarity law, or it can be pre-stored in the edge controller in the form of a data table or piecewise function.

[0138] After calculating the steady-state expected flow Subsequently, to avoid division-by-zero exceptions, the instruction time-sharing scheduling module will... Compare with the preset minimum safe flow threshold for the device. If If the flow rate is below the minimum safe flow rate threshold of the equipment, then the minimum safe flow rate threshold will be used for subsequent calculations. The preset minimum safe flow rate threshold is 20% to 30% of the rated flow rate of the corresponding chilled water pump.

[0139] The instruction time-sharing scheduling module recalculates the delay trigger time applicable to the current new traffic state using the following formula. : ; In the formula, This indicates the recalculated delay trigger time; This represents the equivalent effective water volume of the pipeline network during the current control cycle. This represents the calculated steady-state expected flow rate; It represents the dimensionless correction factor for the decrease in frictional resistance and heat exchange loss along the pipeline.

[0140] The instruction time-sharing scheduling module will set the delay trigger time. Assign the cold / heat source instruction set and the cooling / source side instruction set to the delay instruction queue. The delay instruction queue executes the timing according to the specified time; after the delay trigger time arrives, the corresponding instructions are sent to the chiller unit and cooling tower to complete the delayed time-sharing scheduling under the load reduction condition.

[0141] Reference Figure 7 The specific execution process of step S50 is completed by the energy compensation module within the edge controller. At the end of the current discrete control cycle, this step calculates the single-cycle supply-demand energy deviation increment based on the actual flow rate, actual supply and return water temperature difference, and time-series prediction data matched by time identifiers. This increment is then recursively accumulated to the global energy deviation cumulative value, and after limiting and periodic release processing, a cross-cycle energy compensation amount is generated for target load correction in the next control cycle. This step may specifically include the following sub-steps.

[0142] Step S501: In the current discrete control cycle In the final stage, the energy compensation module extracts the current actual operating data of the pipeline network through sensor components, specifically including the actual flow rate of the main pipeline. Temperature difference between supply and return water .

[0143] When the current cycle is marked as a low-flow invalid calculation cycle by the state prediction module, or when the actual flow rate or actual supply and return water temperature difference fails the validity verification, the energy compensation module does not update the global energy deviation cumulative value and uses the cross-cycle energy compensation amount of the previous valid cycle.

[0144] Verification of the validity of actual flow rate and actual supply and return water temperature difference can include checking whether sensor data is missing, whether it exceeds the sensor range, whether there are abnormal sudden changes, and whether the actual flow rate is lower than the effective calculated flow rate threshold.

[0145] The energy compensation module retrieves the timing prediction data generated by the state prediction module in step S10, and extracts the flow prediction value corresponding to the current control cycle number and sampling time. Estimated supply and return water temperature difference .

[0146] Due to time delays and rate-of-change limitations in the issuance, triggering, and execution of control commands, a deviation will occur between the actual heat and cold distribution in the pipeline network within the current cycle and the corresponding target energy reference value. The energy compensation module matches actual operating data with time-series predicted data according to the same time marker, and based on the approximate assumption that state parameters remain constant within discrete cycles, it uses thermodynamic parameters to calculate the single-cycle supply and demand energy deviation increment for the current control cycle. .

[0147] Among them, the state difference between the actual operating data and the time-series estimated data is used for subsequent estimation starting point or estimation parameter correction; the incremental energy deviation of a single cycle after thermodynamic parameter conversion is included in the global energy deviation cumulative value.

[0148] The energy compensation module uses the following formula for calculation: ; In the formula, Indicates the current control cycle The single-cycle supply and demand energy deviation increment; This represents the constant-pressure specific heat capacity of the circulating fluid in the pipe network; This indicates the density of the circulating fluid in the pipe network; This indicates the time step of the discrete control cycle.

[0149] Step S502: After calculating the incremental supply-demand energy deviation in a single cycle, the energy compensation module performs cross-cycle recursive accumulation on the energy deviation.

[0150] The edge controller is internally configured with a global energy deviation accumulation storage unit to store the cumulative global energy deviation value obtained through cross-cycle recursion. The energy compensation module calculates the single-cycle supply and demand energy deviation increment for the current cycle. The cumulative value of global energy deviation is recursively added to the cumulative value of global energy deviation to update the cumulative value of global energy deviation for the current period. The recursive formula is: ; In the formula, This represents the updated cumulative global energy deviation value. This represents the cumulative global energy deviation value saved in the previous control cycle. During system initialization, the initial value of the cumulative global energy deviation is set to zero.

[0151] During long-term operation, when the controlled equipment in the field approaches its capacity limit or is unable to continue responding and adjusting due to safety constraints, the cumulative global energy deviation may continue to increase. To prevent this cumulative value from exceeding the controllable range, the energy compensation module adjusts the accumulated value. Perform amplitude limiting verification.

[0152] The energy compensation module will The energy deviation is compared with preset upper and lower limit thresholds. These thresholds are set to 5 to 10 times the system's rated single-cycle operating energy, depending on the specific system capacity. For example, if the rated full-load power of the controlled air conditioning system is 1000 kW and the discrete control cycle step size is set to 5 seconds, then the system's rated single-cycle operating energy is 5000 kJ. In this case, the upper and lower limit thresholds for energy deviation can be set between ±25000 kJ and ±50000 kJ. If the value exceeds the set threshold range, it will be limited to the corresponding boundary threshold.

[0153] After completing the recursive accumulation and limit verification, the energy compensation module does not convert the updated global energy deviation accumulation value into a reverse compensation amount all at once. Instead, it performs periodic release processing on the global energy deviation accumulation value based on the preset compensation release ratio, compensation release time constant, current available equipment capacity, and transmission and distribution capacity, generating the cross-cycle energy compensation amount that is currently allowed to be fed back to the next control cycle. .

[0154] The unreleased energy deviation balance remains in the global energy deviation accumulation storage unit and participates in the recursive calculation of subsequent control cycles.

[0155] Cross-cycle energy compensation It has a compensation direction. When the cumulative global energy deviation is positive, it indicates that the actual distributed cooling and heating is greater than the corresponding target energy reference value, and the target load in the next control cycle needs to be reduced under restrictions. When the cumulative global energy deviation is negative, it indicates that the actual distributed cooling and heating is less than the corresponding target energy reference value, and the target load in the next control cycle needs to be increased under restrictions. Both reduction and increase are executed according to the periodic release rule, rather than releasing the cumulative deviation entirely within a single control cycle.

[0156] The energy compensation module saves the updated cross-cycle energy compensation amount. And feed it back to the next control cycle. The control command generation and grouping module. Simultaneously, the energy compensation module stores the remaining energy deviation balance that has not yet been released, so that it can be recursively released in subsequent control cycles.

[0157] The cross-cycle energy compensation amount participates in the target load correction process in step S30 of the next control cycle, and is used to correct the target load for the next control cycle. Through the above-described cycle-by-cycle cumulative and periodic release process, the steady-state control deviation caused by the response lag of the field controlled equipment and the pipeline transmission and distribution delay can be reduced.

[0158] To enable those skilled in the art to more clearly understand the purpose, technical solution, and advantages of this invention, the present invention will be further described in detail below with reference to specific application embodiments, real experimental test data, and corresponding drawings. It should be noted that the embodiments described herein are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0159] This application example uses a single response process of the central air conditioning cooling and heating system of a medium-sized commercial complex under a sudden increase in load.

[0160] System initial settings: Chiller rated power 1000kW; constant pressure specific heat capacity of the piping network. It has a density of 4.18 kJ / kg·℃. 1000 kg / m 3 Discrete control cycle time step The limit for the rate of change of load rate of the chiller unit is 5 seconds. Set to 1% / s, equivalent to 10kW / s. The preset compensation release ratio for energy deviation is set to 10%.

[0161] In a certain initial period The system is operating stably, and the actual load demand at the terminals is within acceptable limits. The actual flow rate of the main pipeline is 500kW. It is 0.024m 3 / s, which is 24L / s, the actual supply and return water temperature difference. The temperature is 5℃.

[0162] In the subsequent cycle Due to a surge in people within commercial complexes, the actual demand load at the end point calculated by the sensor components has changed abruptly. The power is 600kW. The modules within the edge controller execute the following calculation and scheduling processes: The pipeline delay calculation module extracts end-point regulating valves with an opening degree greater than 8% and filters out 3 effective flow branches. Based on the pipeline branch parameter table, the physical volume of the main pipeline... It is 5.0m 3 The sum of the effective water volumes of all the branch channels is 4.6 m³. 3 .

[0163] Calculate the equivalent effective water volume of the pipe network: m 3 ; Combined with correction factor Calculate the dynamic transmission delay time based on the current traffic: s; Since the previous cycle was in a stable state, The control command generation and grouping module calculates and corrects the target load. kW; Decoupling generates corresponding instruction sets: cold and heat source instruction set, which increases the set power of the chiller unit to 600kW, and load-side distribution instruction set, which increases the operating frequency of the chilled water pump.

[0164] The instruction time-sharing scheduling module determines that the load increase is 100kW, which is greater than the set dead zone threshold, 2% of the system's rated load, or 20kW, and confirms that the system is in a load increase condition.

[0165] The system immediately issues a set of cold and heat source commands, and the chiller unit begins ramping up at a rate of 10kW / s. Simultaneously, the load-side distribution command set is pushed into the delay command queue, with the trigger time set to 440 seconds later. During this period, the water pumps maintain their original frequency to avoid unnecessary early pumping energy consumption.

[0166] During the 440-second physical delay period when the chiller unit is climbing the slope and waiting for the water flow to reach the terminal, the actual cooling capacity received by the terminal remains around 500kW, while the target energy reference value is 600kW.

[0167] In the cycle That is, during a certain week of the delay period, the state prediction module predicts that the current theoretical cooling capacity should reach the expected level, but the sensor component reports that the physical cooling capacity converted from the actual flow rate and temperature difference is still 500kW.

[0168] The energy compensation module calculates the incremental supply-demand energy deviation in a single cycle: kJ; This negative value is accumulated to the global energy deviation cumulative value. Assuming that by the end of the delay period, The cumulative value reached -15000kJ. After the amplitude limit check showed no exceedance, 10% of the preset compensation was released. Cross-cycle energy compensation kJ; In the next cycle The compensation amount is converted into an equivalent compensation load: 1500kJ / 5s = 300kW. After safety constraints, such as a maximum single-cycle additional limit of 50kW, it is actually added to the target load. This mechanism prompts the chiller unit to perform controlled, moderate over-generation after the delay period to compensate for the previous energy deficit and eliminate the integral static error caused by physical response hysteresis.

[0169] To verify the actual operational effect of the edge intelligent control method described in this invention, a corresponding central air conditioning cold and heat source system test platform was set up in this embodiment. The test process employed the controlled variable method, comparing and recording the operational data of the conventional synchronous control method and the control method provided in this embodiment under the same terminal dynamic load boundary conditions. The total test duration included four complete variable load cycles of continuous operation, totaling 240 minutes. The results are as follows: Figure 8 As shown: During the sudden load test phase, i.e., when the test had run for 30 minutes, the commanded load was increased from 40% to 80%. Figure 8Subgraph (a) reflects the dynamic tracking curve of the terminal water supply temperature. Under conventional synchronous control methods, the control system simultaneously increases the frequency of the chiller unit and the chilled water pump. Due to the transmission delay in the physical water volume of the pipe network, the pump flow rate increases, but the newly generated cooling capacity of the chiller unit has not yet reached the end of the pipe network, resulting in drastic fluctuations in the terminal water supply temperature. Its maximum temperature overshoot reaches 1.8℃, and the adjustment stabilization time is 18 minutes. Applying the control method provided in this embodiment, the asynchronous scheduling is performed based on the dynamic distribution delay time by the instruction time-sharing scheduling module, so that the cooling capacity production process and the physical water flow transmission process are synchronized at the terminal. The test curve shows that the maximum temperature overshoot of the system at the terminal is limited to within 0.4℃, and the adjustment stabilization time is shortened to 6 minutes, verifying the improvement of the system's anti-disturbance capability.

[0170] During the continuous small disturbance test phase, specifically from 120 to 180 minutes into the test run, the chilled water pump inverter under the conventional synchronous control method recorded 147 forward and reverse operation switches. Such frequent frequency adjustments can easily cause mechanical vibration and reactive power loss in the equipment. However, in the control architecture of this embodiment, by introducing a mechanism for overriding commands with the same parameters for the same equipment and a load change dead zone determination logic, invalid reverse commands are directly overridden and cleared from the delayed command queue. Test results show that the frequency of chilled water pump operation is significantly reduced to 12 times, reducing the mechanical wear of the underlying equipment.

[0171] In system capacity-constrained operating condition testing, the constraint boundaries are artificially set to force the system into a full-load operating state. Figure 8 Subgraph (b) illustrates the response state after the system boundary deviates. Conventional control methods exhibit significant integral saturation under this condition. When the terminal load demand decreases and the operating condition should return to normal, the system remains fully loaded for an extended period (up to 14 minutes) due to excessive accumulated error, failing to respond to new load reduction commands. In this embodiment, the energy compensation module triggers a limiting verification mechanism for the cumulative global energy deviation. After the operating condition deviates from the restricted boundary, it smoothly removes the remaining deviation term using a periodic release rule. Test data shows that the system's response delay after boundary deviates is only two control cycles, approximately ten seconds, avoiding the system response stagnation problem caused by the infinite accumulation of errors.

[0172] according to Figure 8As shown in sub-graph (c), under reduced load conditions, based on the aforementioned time-sharing scheduling logic, the edge controller prioritizes issuing load-side distribution command sets to immediately reduce the pump operating frequency, thereby fully utilizing the residual heat and cold energy within the equivalent effective water volume of the pipe network. Comparative data shows that, under the premise of providing the same total cooling capacity to the terminals, the chilled water pump distribution energy consumption of this embodiment is reduced by 17.4% compared to the conventional scheme, the additional energy consumption for starting and stopping the chiller unit is reduced by 5.2%, and the overall system performance coefficient is improved by 8.9%. The above test data verifies that the control method of the present invention can further reduce the overall operating energy consumption of the system while meeting the load demand of the pipe network terminals.

[0173] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Those skilled in the art can omit, substitute, and modify the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, combining the above method steps to perform substantially the same function and achieve substantially the same result using substantially the same method falls within the scope of the present invention.

Claims

1. A method for edge intelligent control of a central air conditioning cold and heat source system, characterized in that, Includes the following steps: At the start of the current control cycle, the delay command queue is simulated in combination with the action change rate limit of the controlled equipment, and feedforward estimated state dataset and time-series estimated data with time stamp are calculated and generated. The opening feedback of the regulating valve at the end of the pipeline network is detected to screen effective flow branches and calculate the equivalent effective water volume of the pipeline network. The dynamic transmission and distribution delay time is calculated in combination with the actual flow of the main pipeline. The corrected target load is calculated based on the actual demand load at the terminal and the cross-cycle energy compensation amount fed back from the previous control cycle. A system control command set is generated based on the corrected target load and the feedforward estimated state dataset. The system control command set is then grouped according to the dynamic response characteristics of the equipment into a cold / heat source command set, a cooling / source side command set, and a load side transmission and distribution command set. After overwriting the grouped instruction sets with the historical instructions to be executed in the delayed instruction queue using instructions with the same device and parameters, the grouped instruction sets are immediately issued or written into the delayed instruction queue and issued after a delay, based on the trend of the increase or decrease of the target load and the dynamic transmission and distribution delay time. At the end of the current control cycle, the real-time physical state data is compared with the time-series estimated data according to the time identifier. The estimated starting point or estimated calculation parameters of the subsequent control cycle are corrected according to the state difference obtained from the comparison. The single-cycle supply and demand energy deviation increment is calculated based on the actual heat and cold distribution and the target energy reference value. After recursion and accumulation, the cross-cycle energy compensation amount is updated and generated after limiting and periodic release processing. The cross-cycle energy compensation amount is fed back to the next control cycle for load correction.

2. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, The calculation generates a feedforward predicted state dataset and time-stamped predicted data, specifically including: Real-time physical state data is collected when the current control cycle starts, the delayed instruction queue is scanned, the trigger time and remaining delay time of each control instruction to be issued are obtained, and the waiting stage before the instruction is triggered and the execution stage after the instruction is triggered are distinguished. During the waiting phase, the estimated state of the waiting phase is determined based on the control objectives that are currently in effect for the controlled equipment. When the estimated time corresponding to the timing prediction data reaches the predetermined trigger time of the control command to be issued and enters the execution stage, the action change rate limit value of the corresponding controlled parameter of the controlled device is called, and the state change process after the command is triggered is deduced according to the target state value in the control command to be issued. The feedforward estimated state value of the controlled equipment affected by the delayed command is combined with the real-time physical state data collected that is not affected by the delayed command to generate the feedforward estimated state dataset. The flow rate estimate and the supply and return water temperature difference estimate related to the energy transmission and distribution calculation are extracted and stored in the time series estimated data in chronological order. The target energy reference value is determined based on the actual demand load at the end of the current control cycle or the constrained target load.

3. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, The selection of effective flow branches and the calculation of the equivalent effective water volume and dynamic transport and distribution delay time of the pipe network specifically include: Effective flow branches are selected using a branch effective flow determination rule with hysteresis. Specifically, an opening threshold and a closing threshold are set in pairs, and the closing threshold is less than the opening threshold. When the opening degree of the end control valve is greater than or equal to the opening threshold, the corresponding branch circuit is determined as an effective flow branch. Only when the opening degree of the end control valve corresponding to the branch circuit that is already in an effective flow state is less than or equal to the closing threshold for a preset number of consecutive control cycles, is the branch circuit switched to an invalid flow branch. The equivalent effective water volume of the pipe network is calculated by adding the fixed physical volume of the main pipe of the system to the water volume of the branch pipes corresponding to all effective flow branches. The dynamic distribution delay time is obtained by dividing the equivalent effective water volume of the pipeline network by the actual flow rate of the main pipe and multiplying it by a dimensionless correction coefficient for the friction loss and heat exchange loss along the pipeline.

4. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, Calculate and correct the target load and generate a set of system control commands, specifically including: The cross-cycle energy compensation amount fed back from the previous control cycle is converted into an equivalent compensation load according to the control cycle step size, and upper and lower limit constraints are applied to the equivalent compensation load. The equivalent compensation load after upper and lower limit constraints is algebraically summed with the actual demand load at the end, and the summation result is subject to upper and lower limit constraints based on the current available cold and heat source capacity and transmission and distribution capacity to obtain the corrected target load. The system control command set is generated based on the corrected target load and the feedforward estimated state dataset, and each control command in the system control command set is identified as a safe start / stop and interlock control command or a normal capacity adjustment command.

5. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, Perform instruction overwrite processing with the same device and parameters, specifically including: Determine whether the newly generated control command and the historical commands to be executed in the delayed command queue are for the same controlled device and the same controlled parameter; If so, when the energy direction of the newly generated control instruction is the same as that of the historical instruction to be executed, the target state value of the new instruction is used to replace the target state value of the old instruction, and the original scheduled trigger time of the old instruction is inherited. When the newly generated control command has the opposite energy direction to the historical command to be executed, the old command is cleared, and the triggering time is recalculated based on the changing trend of the modified target load and the latest calculated dynamic transmission and distribution delay time.

6. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, When the target load change in two adjacent control cycles exceeds the preset load change dead zone threshold, and the system is determined to be in a load increase condition, the specific logic for immediately issuing or delaying the issuance of each grouped instruction set after writing it into the delayed instruction queue is as follows: Prioritize and immediately issue the aforementioned cold / heat source command set and cooling / source side command set; The load-side distribution instruction set is written into the delay instruction queue, and the calculated dynamic distribution delay time is assigned to the load-side distribution instruction set as the delay trigger time. After the delay trigger time is reached, the corresponding load-side distribution instruction set is sent to the chilled water pump.

7. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, When the target load change in two adjacent control cycles exceeds the preset load change dead zone threshold, and the system is determined to be in a load reduction condition, the specific logic for immediately issuing or delaying the issuance of each grouped instruction set after writing it into the delayed instruction queue is as follows: Prioritize immediately issuing the load-side distribution command set to the controlled equipment on site to reduce the operating frequency of the chilled water pump; Based on the operating frequency parameters set in the newly issued load-side transmission and distribution instruction set, the steady-state expected flow rate after the water pump executes the instruction is calculated. The equivalent effective water volume of the pipeline network is divided by the steady-state expected flow rate and multiplied by the dimensionless correction coefficient of the pipeline friction resistance attenuation and heat exchange loss. The delay trigger time applicable to the current new flow state is then recalculated. The recalculated delay trigger time is assigned to the cold / hot source instruction set and the cooling / source side instruction set, and written into the delay instruction queue in preparation for delayed delivery.

8. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 4, characterized in that, When issuing a set of system control commands: When both safety start / stop and interlock control commands and normal capacity adjustment commands exist within the same control cycle, the minimum safe flow establishment, equipment start status confirmation, unloading shutdown or water pump delayed shutdown process corresponding to the safety start / stop and interlock control commands shall be executed first. Once the equipment status meets the conditions for safe operation, the remaining normal capacity adjustment instructions will be issued after a delay and time-sharing process.

9. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, The calculation of the incremental supply-demand energy deviation in a single cycle includes: Extract the flow rate estimate and supply / return water temperature difference estimate from the time-series prediction data that correspond to the current control cycle number and sampling time. The result of multiplying the constant pressure specific heat capacity and density of the circulating fluid in the pipeline network, the actual flow rate of the main pipe, and the actual supply and return water temperature difference is subtracted from the result of multiplying the constant pressure specific heat capacity, density, the estimated flow rate, and the estimated supply and return water temperature difference. The difference obtained after subtraction is multiplied by the time step of the discrete control cycle to obtain the single-cycle supply and demand energy deviation increment.

10. The edge intelligent control method for a central air conditioning cold and heat source system according to claim 1, characterized in that, The process of updating and generating the cross-cycle energy compensation amount specifically includes: The incremental supply-demand energy deviation of a single cycle is recursively added to the global cumulative energy deviation value. The accumulated global cumulative energy deviation value is then compared with the preset upper and lower limit thresholds of energy deviation to perform a limit verification. According to the preset compensation release ratio, the cumulative value of the global energy deviation after the limit verification is processed in stages to generate the cross-cycle energy compensation amount that is currently allowed to be fed back to the next control cycle. The unreleased energy deviation balance remains in the global energy deviation accumulation storage unit and participates in the recursive calculation of subsequent control cycles; the state difference between the real-time physical state data and the time-series prediction data is only used to correct the prediction start point or prediction calculation parameters of subsequent cycles, and is not used as the single-cycle supply and demand energy deviation increment, nor is it directly accumulated to the global energy deviation accumulation value.

Images